Download Testing stabilisation policy limits in a small open policy forum

Testing stabilisation policy limits in a small open
economy: proceedings from a macroeconomic
policy forum
Contents
Preface
3
Testing stabilisation policy limits in a small open economy:
5
Editors’ summary of a macroeconomic policy forum
Bob Buckle, The Treasury
Aaron Drew, Reserve Bank of New Zealand
Macroeconomic policy challenges: monetary policy
13
Stephen Grenville, Lowy Institute
Discussion by Christopher Allsopp, University of Oxford
31
Stabilisation policy in New Zealand: Counting your blessings, one by one
37
Willem Buiter, London School of Economics
Discussion by Pierre Siklos, Wilfrid Laurier University
75
New Zealand’s monetary and exchange-rate policy in international comparison
83
Klaus Schmidt-Hebbel, Central Bank of Chile
Discussion by John Edwards, HSBC
145
External imbalances in New Zealand
149
Sebastian Edwards, University of California, Los Angeles
Discussion by William Cline, Institute for International Economics
181
Panellist comments on the proceedings
Val Koromzay, Organisation for Economic Co-operation and Development
189
Stephen Dunaway, International Monetary Fund
193
John McDermott, Victoria University of Wellington
195
This document is also available at www.rbnz.govt.nz and www.treasury.govt.nz
Copyright © 2006 Reserve Bank of New Zealand and The Treasury
ISBN 0-9582675-2-9
Testing stabilisation policy limits in a small open economy
1
2
Reserve Bank of New Zealand and The Treasury
Preface
Monetary and fiscal policies in New Zealand over the past
This volume presents the proceedings from a conference
decade have achieved considerable success in meeting their
held in Wellington on June 12th 2006 organised to present
objectives of positively contributing to lower inflation, and
the findings of the visiting experts. It contains their papers,
improved economic growth and stability. These policies
discussants comments, thoughts from a panel of long-term
have been undertaken within institutional arrangements
observers of the New Zealand economy, and an overview
introduced in the late 1980s and early 1990s which at the
chapter from the Editors. The overall conclusion that
time attracted international interest. Although there have
emerged is that the essential elements of New Zealand’s
been modifications introduced since then, the fundamental
macroeconomic policy frameworks are still fundamentally
features of the institutional arrangements have remained
sound and remain appropriate. Furthermore, some
broadly the same over the past decade. As part of good
fluctuations in the current account and some volatility in the
economic policy practice the Treasury and the Reserve
exchange-rate and other relative prices are to be expected
Bank considered it an appropriate time to re-appraise the
and are an important part of the process of adjustment to
institutional arrangements and application of monetary and
changing international and domestic events. Nevertheless,
fiscal policy in New Zealand.
several suggestions to improve the way structural, fiscal and
There have been periods of significant external imbalance
and large fluctuations in the New Zealand dollar which
has placed stress on the externally exposed sectors of
the economy. These conditions were apparent during the
mid-1990s and emerged again during the last 3 years. It
remains an open question whether better external balance
and smaller swings in the exchange-rate can be achieved
monetary policies interacted and impacted on the economy
were raised and debated. While it seems that the proverbial
silver bullet remains as elusive as ever, the material in this
volume provides much food for thought for policy makers
in New Zealand, and indeed policy makers in any small open
economy charged with running independent policies in an
increasingly integrated world.
through the application of alternative policies or policy
strategies, while also maintaining or enhancing overall
Grant Spencer
economic performance and prospects.
Assistant Governor
In early 2006, at the request of the Reserve Bank of
New Zealand and New Zealand Treasury, four international
academic experts and practitioners in the macro economic
policy arena visited New Zealand. Their brief was to critically
examine New Zealand’s macro economic policy frameworks
and consider whether alternative, possibly non-conventional,
Reserve Bank of New Zealand
New Zealand
Peter Bushnell
Deputy Secretary
The Treasury
New Zealand
policy tools might be used to provide a smoother ride for
the externally exposed sectors of the economy over the
October 2006
business cycle. As part of the brief, the visitors interviewed
policy makers, academics, bank economists, business
leaders and others to give them a first-hand impression of
New Zealand’s macroeconomic environment.
Testing stabilisation policy limits in a small open economy
3
4
Reserve Bank of New Zealand and The Treasury
Testing stabilisation policy limits in a small open economy:
Editors’ summary of a macroeconomic policy forum
Bob Buckle, The Treasury and Aaron Drew, Reserve Bank of New Zealand1
1
Macroeconomic issues
CAD may have increased the likelihood of a sudden and
motivating the conference
disruptive exchange-rate adjustment to levels uncomfortably
New Zealand has been one of the faster growing OECD
economies since the early 1990s, driven by both strong
employment and labour productivity gains. During the last
five years, however, this growth was accompanied by the
emergence of macroeconomic imbalances. While some
of these imbalances are in common with several other
relatively fast growing economies, including Australia and
the United States, this does not mitigate concerns that the
eventual process of adjustment might be both painful and
prolonged.
The so-called imbalances have manifested themselves in
a number of areas: CPI inflation has increased beyond the
top end of the Reserve Bank’s target band; the balance of
payments Current Account Deficit (CAD) increased to over
9 per cent of GDP (one of the largest amongst developed
economies); asset prices (notably house prices) increased
rapidly; and household debt levels rose to historic highs.
below historical averages. From a monetary policy point of
view, this would be particularly unwelcome at the present
juncture where inflation is already high. Alternatively, the
sharp rise in household debt levels may have increased
the vulnerability of the household sector to a fall in house
prices, particularly if accompanied by a significant increase
in unemployment levels. If large enough, in combination
these shocks could potentially pose a systematic risk to the
banking system, given the banks’ reliance on foreign capital
and their exposure to the household sector.
Another concern is that the imbalances may adversely
impact productivity growth. New Zealand business cycles
are characterised by variations in the relative growth of the
tradable and non-tradable production sectors. The current
cycle has been sustained by strong non-tradables growth,
particularly in residential investment (a pattern that has
been accentuated by large cyclical net migration swings).
In contrast, growth in the tradables sector has been much
While the term “imbalances” is commonly used to describe
weaker, even though this sector tends to have higher trend
these macroeconomic outcomes, the term could in some
productivity growth. Associated with this pattern, the
instances be regarded as pejorative. For example, in an
New Zealand economy also experiences large exchange-rate
inter-temporal context, a CAD can be viewed as the
swings over the cycle. These large swings may have adverse
consequence of a reshuffling of demand across time, which
effects on investment and productivity, again particularly in
results in differences in the levels of contemporaneous
the tradables sector.
domestic demand and supply. Hence, while a CAD is
sometimes described as a situation in which a country is
“living beyond its means,” in an inter-temporal sense a
CAD is not necessarily inconsistent with life-time “means.”
Nevertheless, there are several legitimate reasons why these
recent developments could be a concern.
One explanation for the large increase in New Zealand’s
CAD is that it reflects an unusual international distribution
of savings. High excess savings in East Asia are flowing to
economies where yields are higher, in part financing the
growth in business and housing investment in New Zealand
(and similarly Australia and the US). The redistribution
One concern is that the economy may experience a “hard
of these savings through global financial markets has
landing” if a normal cyclical downturn is amplified by a
caused yields on long-term securities to converge across
significant fall in asset prices. For example, the size of the
countries. This has reduced the level of real interest rates
Bob Buckle is a Principal Advisor at The Treasury and
Adjunct Professor of Economics, Victoria University of
Wellington. Aaron Drew manages the research division in
the Economics Department, Reserve Bank of New Zealand.
The views expressed herein do not necessarily fully represent
those of the Treasury or Reserve Bank of New Zealand.
that New Zealand residents may have otherwise faced
1
Testing stabilisation policy limits in a small open economy
and has therefore contributed to domestic investment and
consumption spending staying higher for longer.
5
This process of internationalisation of financial markets has
a summary of the panellists’ comments. Finally, we offer
given rise to a third concern, namely that the Reserve Bank
our thoughts on policy areas that may warrant further
of New Zealand now has less leverage over longer-term
attention.
interest rates. In order to influence domestic inflation, the
Bank may therefore have to rely more on short-term interest
rate movements and, as a consequence, the exchange-rate
channel. In these circumstances, more of the burden of
adjustment may fall on the tradable goods sector, with the
potential ramifications for productivity growth expressed
2
Summary of papers
Macroeconomic policy challenges:
monetary policy
above.
Authored by Stephen Grenville (Lowy Institute),
discussion by Christopher Allsopp (University of Oxford)
A final and closely related issue is that the imbalances reflect
It is widely accepted that New Zealand’s inflation targeting
excess demand pressures in the economy more generally,
approach has become more ‘flexible’ as low inflation and
and these could indicate that macro economic policy
inflation expectations have become embedded in the
(encompassing both monetary and fiscal policy) settings
economy. Stephen Grenville reflects that the concerns
have not been very effective over recent times in preventing
over the stress placed on the externally exposed sectors of
the build up of these imbalances, therefore contributing to
the economy in the recent cycle could be regarded as an
the risk of a costly adjustment phase.
extension of a flexible approach. However, he is doubtful
In June 2006, The Treasury and the Reserve Bank of
New Zealand co-hosted a Macroeconomic Policy Forum
that brought together international and domestic experts
to examine the policy issues relating to these recent
New Zealand macroeconomic developments.
whether monetary policy alone, with a conventional single
instrument (the OCR), can reasonably be expected to
address cyclical strains caused by sectoral imbalances. This
would be especially so if recent international and financial
developments have significantly shifted the transmission
channel of monetary policy further towards the exchange-
The overall assessment of the invited speakers and
rate. This view is endorsed by Christopher Allsopp,
discussants at the Forum was that the essential elements
who suggests that if, in principle, policy instruments or
of New Zealand’s macroeconomic policy institutions are
interventions can be found that are better targeted to the
sound and remain appropriate. They also emphasised
source of a sectoral shock, then institutional responsibility for
that changes in real exchange-rates and fluctuations in
this in almost all circumstances should lie with The Treasury.
current account balances are often an essential part of the
In his view, monetary policy should remain primarily focused
processes of adjusting to domestic and international shocks.
on its price stability objective.
Further, some expressed the view that recent international
developments and the way they have impacted on
New Zealand may have been unique. Hence, there was a
general tone that there is no reason for New Zealand policy
makers to panic. Participants at the Forum did not go so far
to suggest, however, that policy makers in New Zealand can
be complacent about the economic outlook, or that there
are no policy areas that warrant further scrutiny.
There are several policy areas, however, where Grenville
thinks there may be scope to modestly reduce sectoral
stresses. First, he suggests the RBNZ should be even more
forthcoming in pointing out to the public when it thinks
asset prices (e.g. the exchange-rate and house prices) are
misaligned. Regarding currency misalignments, he proposes
that the RBNZ should use foreign exchange intervention
more systematically, in the spirit of the Reserve Bank of
The following section provides a high-level snapshot of the
Australia’s approach, to “lop the peaks and fill the troughs”
papers in this volume that were presented at the Forum
of movements in the currency. This view is predicated on
in June. This snapshot focuses on some (but by no means
there being systematic arbitrage opportunities over the
all) of the policy suggestions that arose. This is followed by
6
Reserve Bank of New Zealand and The Treasury
currency cycle, an assertion that Grenville makes but many
inflation expectations. He notes these expectations have
other participants of the Forum disagreed with. Indeed,
remained relatively low, albeit with some sign of modest
many considered that the RBNZ should not entertain
upward drift in the current cycle. Nevertheless, he suggests
intervening in currency markets at all.
the effectiveness of monetary policy might be enhanced
In relation to the risks posed by a booming housing market,
Grenville proposes that much more comprehensive data
on housing loan exposures should be collected and given
widespread and critical public coverage. Two specific policy
suggestions are also offered. First, that the mortgage levy
idea raised (and largely dismissed) in the recent SSI report
produced by the RBNZ and New Zealand Treasury deserves
further attention given its potential to curtail a housing led
boom.2 Second, that insurance for loans with loan-to-value
ratios above 80% should be required (which is encouraged in
Australia by the application of a higher capital requirement if
this insurance is not in place). These suggestions on housing
market measures merge into the arena of prudential policy.
He sees these policies as modest measures to improve the
chances that banks remain efficient financial intermediaries
throughout the cycle, and in the face of low-probability
events such as a “sudden stop” of foreign funding sources.
Further, more “speculative” prudential policy options are
also offered.
through several changes to the current framework. First,
he advocates replacing the inflation target band with a
point target. This suggestion, in some guise, found favour
with many of the Forum participants. Second, he favours
replacing the single-decision maker arrangement with a
Monetary Policy Committee, along the lines of the Bank of
England.3 Third, Buiter advocates taking the “flexible’”out
of “flexible inflation targeting” and replacing it with
lexicographic inflation targeting. Finally, Buiter is very critical
of the Bank’s foreign exchange intervention framework. He
is sceptical that intervention can work to reduce exchangerate volatility and suggests it raises the temptation to try
and target both inflation and the nominal exchange-rate, a
policy well known to be infeasible. All these suggestions are
broadly endorsed by Pierre Siklos, who further claims that
intervention decisions linked to whether the exchange-rate
departs excessively from fundamentals is confusing because
of the lack of reliable evidence of the factors that determine
“fundamentals.”
In the fiscal policy area, Buiter thinks there is scope for
policy to make a larger contribution to cyclical stabilisation,
Stabilisation policy in New Zealand:
both by increasing the effectiveness of the automatic fiscal
Counting your blessings, one by one
stabilisers and, somewhat more speculatively, by more active
Authored by Willem Buiter (London School of
Economics), discussion by
Pierre Siklos (Wilfrid Laurier University)
use of discretionary fiscal policy. Regarding the former, Buiter
Willem Buiter’s main conclusions are that New Zealand’s
monetary and fiscal policy frameworks are fundamentally
sound, and top of the international class. He nevertheless
considers that some of New Zealand’s inflation, business
cycle and structural characteristics imply there is scope to
improve the macroeconomic policy framework.
proposes (a) broadening the GST base4, (b) taxing capital
gains at the same rate as other income, and (c) index-linking
income, corporate and capital gains taxes. To enable active
discretionary fiscal policy, Buiter suggests the GST rate
could be adjusted by raising during a boom and lowering
in a downturn. He suggests designing a policy rule to guide
variations in the GST rate and delegating this policy to an
operationally independent GST Committee. Siklos is not in
In the monetary policy area, Buiter is not convinced that
favour of this idea due to the “daunting” technical issues
recent international financial developments have reduced the
effectiveness of monetary policy. His argument is that what
is crucial is how changes in the OCR impact on long-term
3
Reserve Bank of New Zealand and The Treasury (2006),
“Supplementary Stabilisation Instruments,” April. Wellington.
http://www.treasury.govt.nz/ssip/.
4
2
Testing stabilisation policy limits in a small open economy
This recommendation is not based on any particular concern
with historical policy. Rather, he sees a Committee structure
as reducing the risk of getting a “bad draw” for a Governor.
By eliminating the exemptions for financial services and
housing rentals, including the imputed consumption of
housing services by owner-occupiers.
7
associated with implementing an effective counter-cyclical
can impact growth in an asymmetric way (i.e. significant
GST policy. Nor is he in favour of Buiter’s other proposals
over or under valuations retard growth, while mild under-
to enhance automatic fiscal stabilisers on the grounds that,
valuations improve growth). Schmidt-Hebbel finds that
while all these ideas may have some economic merit, they
currency misalignments in New Zealand have rarely entered
are not likely to be politically acceptable.
and do not stay long in the “danger territory,” while mild
pro-growth under-valuations are apparent. He concludes
that real exchange-rate misalignments in New Zealand have
New Zealand’s monetary and exchange-rate
policy in international comparison
While not critical of the Bank’s foreign exchange intervention
Authored by Klaus Schmidt-Hebbel (Central Bank of
Chile), discussion by John Edwards (HSBC)
Klaus Schmidt-Hebbel utilises a range of cross-country
empirical
techniques
to
examine
not negatively impacted growth.
New Zealand’s
macroeconomic outcomes and policy performance. The
empirical work suggests that the strength and nature of
the monetary policy transmission process in New Zealand
is not significantly different to the group of comparable
economies (Australia, Canada, Norway, Sweden and Chile).
He also finds that the RBNZ has, on average, better met its
inflation target objective than most, and performance has
improved between 1990-97 and 1998-2005. Nevertheless,
he advocates a change in the PTA to incorporate a more
precise horizon for the achievement of the inflation
target. While John Edwards thinks policy makers can take
framework, Schmidt-Hebbel is sceptical that foreign
exchange-rate intervention can do much to moderate
the exchange-rate cycle. Moreover, his empirical analysis
weakens the case for intervention even if it were effective.
To assist monetary policy in the case of domestic-sourced
shocks, he instead thinks policy makers should consider
evaluating the scope for fiscal policy and financial policy
instruments to play more prominent short-run stabilisation
roles. His suggestions include: a Chilean-styled structural
balance rule; pro-cyclical tax rates or specific countercyclical spending measures; investing public savings abroad
in assets that are negatively correlated to the New Zealand
cycle; and issuing public debt indexed to commodity prices
as insurance against “sudden stops.”
considerable comfort from these findings, he raises the point
that the results pertain to longer term average outcomes
and do not negate the real difficulties faced by the RBNZ
in the recent period and that the transmission mechanism
External imbalances in New Zealand
Authored by Sebastian Edwards (UCLA), discussion by
William Cline (Institute for International Economics)
did seem to take longer than usual. Edwards also notes
The primary purpose of Sebastian Edward’s paper is to
that the improvement in performance that Schmidt-Hebbel
analyse the potential consequences of New Zealand’s
identifies coincides with changes in the PTAs that permit
balance of payments position. Specifically, he investigates
the Bank more flexibility over the horizon in which it brings
the probability that New Zealand will undergo a costly
inflation back to the target range.
adjustment, the proverbial hard landing, in the form of a
A concern often heard in New Zealand, as discussed in
“sudden stop” in capital flows and an abrupt and large
Section 1, is that its exchange-rate cycles are damaging
reversal in the CAD.
for growth. Schmidt-Hebbel confronts this concern, first
Edwards identifies a number of characteristics of
by reviewing the theoretical and empirical literature. This
New Zealand’s external position that sets it apart from
presents a mixed picture, with conclusions depending on
most other advanced countries. Notably, that the CAD is
modelling techniques and data samples (with the notable
presently one of the largest in the OECD, and the most
exception of “currency crisis” episodes). Second, he
important contribution to the CAD tends to be the deficit
empirically examines whether New Zealand currency cycles
on net investment income. This in turn reflects the fact that
have negatively impacted growth. The methodology is
New Zealand’s net international investment position (NIIP)
based on a recent study that shows currency misalignments
8
Reserve Bank of New Zealand and The Treasury
is one of the most negative amongst advanced countries (at
have increased, but remain modest. He concludes that the
around 90% of GDP). Given New Zealand’s close economic
present external balances should not be a cause for great
relationship with Australia, Edwards investigates how the
concern - the adjustment to the current account, when it
trans-Tasman relationship affects New Zealand’s external
does come, will likely be benign. William Cline is not so
balances. He shows that when trans-Tasman transactions are
sanguine, largely basing his arguments on the implications
excluded, external balances are not as large. However, even
of the simple arithmetic of sustainable external debt.
after making the trans-Tasman adjustment, New Zealand’s
present CAD is still almost double the size of what Edwards
estimates to be sustainable. William Cline uses a similar
approach to support the conclusion that the present
position is unsustainable.
While Edwards proffers some suggestions for improving
the conduct of monetary policy, Cline is sceptical that
monetary policy can effectively change the savings and
current account situation, principally because the impact
of changes in the interest rate on the current account are
Do these facts imply New Zealand is at risk of a sudden stop?
likely to be ambiguous. Rather, Cline considers that policy
One perspective Edwards uses to address this question is the
should focus on maintaining or increasing the fiscal surplus
inter-temporal present value model of the current account,
and on policy-settings that affect incentives to both invest
which posits that fluctuations in the current account are
in and supply residential property, fund residential property
due to rational consumption-smoothing behaviours in
through foreign investment, and incentives that impact
the presence of macroeconomic shocks. Recent research
on the willingness of foreigners to invest in New Zealand
estimating this type of model on the New Zealand data
financial instruments. He also suggests that New Zealand
does not reject consumption-smoothing behaviours.
policy makers agree on a ceiling NIIP and integrate a serious
Edwards notes, however, that the recent deterioration in
intention of staying within that limit.
the external trade account is not consistent with the longterm solvency condition in these models, perhaps still
implying that the external balance will have to go through
a significant correction. Cline is more sceptical that recent
behaviours represent optimal consumption smoothing,
arguing that the rise in New Zealand’s CAD has been
associated with a decline in household saving rather than
with a surge in private investment, or as he puts it, a case of
“absorption roughing.” Cline argues that New Zealanders
3
Panellists comments
Val Koromzay (OECD)
Val Koromzay stresses that New Zealand’s policy frameworks
are sound and stand out favourably in international
comparisons. He warns that in asking whether they could be
adjusted to reduce volatility, such adjustment could come at
the expense of losing what New Zealand presently has.
are transferring absorption from the future to the present.
In doing so, they are imposing an undue welfare burden
on future generations and may therefore be politically
unsustainable. In this respect, the large CAD and NIIP are
a problem.
Koromzay agrees with Buiter‘s and Allsopp’s views that
among the tasks assigned to monetary policy, inflation
control should have priority. He is not in favour of systematic
intervention in currency markets, arguing it poses a serious
risk to being able to sustain a clear, effective and credible
To quantify the risk of a sudden stop, Edwards utilises a
cross-country data set to estimate a probit model of the
determinants of the probability that a country will experience
an abrupt current account reversal (defined as a reduction
in the current account deficit of at least 3% of GDP in a
one year period). The model is then applied to New Zealand
for the early 2000s and then for 2005-06 when the CAD
communication strategy for monetary policy. Koromzay
therefore sees little need to change the present monetary
policy framework, and thinks there is little that monetary
policy can do to moderate exchange-rate swings. Instead,
to minimise the potential costs of volatility he advocates
focusing on maximising the flexibility of product and labour
markets.
was larger. Edwards finds that abrupt reversal probabilities
Testing stabilisation policy limits in a small open economy
9
financial
less explicit, approach is in fact already being practiced in
supervision and regulatory framework as sound, and
New Zealand. Citing the example of the experience of the
considers financial market supervision to be too important
US during the late 1990s, unlike Koromzay, Dunaway thinks
a matter to subject it to secondary, so called “macro-
there is a place for financial regulatory and supervision
financial” considerations. He also sees few opportunities
policy to play a role in preventing excesses in the financial
for fiscal policy to lean more strongly against macro
sector during economic expansions.
Koromzay
regards
New Zealand’s
present
fluctuations. In his view discretionary fiscal policy is not to be
recommended; and stronger “automatic fiscal stabilisers”
can only effectively be achieved by raising tax and spending,
John McDermott (Victoria University of
and these are decisions that should not be made on the
Wellington)
basis of smoothing cycles. Koromzay does think, however,
John McDermott sees no obvious missing instrument that
there is merit in considering an ex ante fiscal rule based on
would improve monetary policy performance. He does,
“normal” terms of trade that specifies how the budget will
however, agree that there may be scope for some marginal
deal with revenue windfalls and shortfalls.
improvements and that is where the focus should be. He
With regard to structural policies that may help lift national
raises several specific issues.
saving and reduce potential external vulnerability, Koromzay
McDermott disagrees with Buiter’s analysis that plays-
rejects suggestions to make foreign credit more expensive.
down the importance of external terms-of-trade shocks
Nor does he think the international evidence supports
to New Zealand. He suggests these are important given
tax incentives as an effective means of raising aggregate
New Zealand’s economic structure, and this structure
savings. Koromzay does agree with Buiter that increasing
means that the business cycle will be sensitive to changes in
property taxes has merit, although the political-economy of
the terms of trade, as research has shown.
such moves is daunting and would require “extra-ordinary
The structure of the New Zealand economy is, in McDermott’s
political leadership.”
view, also relevant to the debate about foreign exchange
intervention. He argues that empirical evidence suggests
Steve Dunaway (IMF)
Steve Dunaway suggests, like Koromzay, that some of the
factors that lead to the large macroeconomic imbalances
in New Zealand over the last five years are likely to have
been one-off factors. Specifically, the strong inflow of
that the large real exchange-rate swings are fundamentally
broadly determined by commodity price swings, in which
case the exchange-rate plays a buffering role. Like others,
he is concerned that direct intervention risks distracting
attention from the core role of monetary policy.
foreign savings and possibly also the degree to which the
McDermott is not in favour of introducing prudential
New Zealand cycle was out of synch with the international
instruments to complement monetary policy. He suggests
cycle. He acknowledges international financial integration
the information required to do this efficiently is not adequate
may mean that the way monetary policy impacts has
and it will introduce distortions to the credit market. Nor
changed, and policy makers will need to be alert to this
does he regard the CAD as a problem. Even though some
change, but it does not mean that monetary policy is
of the recent growth in the deficit is due to consumption
ineffective.
growth, a significant proportion is durables consumption
Dunaway is not persuaded by the arguments for greater
emphasis on discretionary counter-cyclical fiscal policy. He
does, however, think there is merit in some form of counter-
which he suggests is a form of investment in a future stream
of household services. That is, some of Cline’s so-called
“absorption roughing” is a form of investment.
cyclical fiscal rule of the types discussed by Klaus SchmidtHebbel and Val Koromzay. He argues a similar, though
10
Reserve Bank of New Zealand and The Treasury
4
Possible areas of future policy
The stabilisation role of fiscal policy
research
A debate running through the Forum revolved around the
The key policy issues that arose in the Forum can be
summarised under four broad topics: the role and conduct
of monetary policy; the stabilisation role of fiscal policy;
exchange-rate volatility and implications for the economy;
merits of more active fiscal stabilisation to complement the
inflation targeting role of monetary policy. There are three
issues for analysis under this theme:
•
What has been the macroeconomic impact of fiscal
and structural policies (including taxation structures) that
policy over the business cycle? Has fiscal policy
may be impacting on housing demand and household
exacerbated or ameliorated cycles in GDP, interest rates,
savings. In what follows we suggest possible areas for
the exchange-rate and the current account?
future policy research under each of these topics.
•
What are the merits and implications of a more active
stabilisation role for fiscal policy, over and above the
effect of automatic stabilisers? How should these
The role and conduct of monetary policy
short-run objectives be weighed up against long term
A number of speakers at the Forum raised questions about
sustainability and economic growth objectives?
monetary policy. There are five broad areas of investigation
in this sphere:
•
•
business cycle, or if there is a case for more activist fiscal
What has been the recent impact of monetary policy on
policy, what type of institutional arrangements should
the business cycle, and on exchange-rates?
•
be considered?
Inflation expectations have risen over the recent cycle.
Is this simply a cyclical increase or is there evidence that
inflation expectations have adjusted to a higher trend
rate of expected inflation?
•
Exchange-rate volatility
A popular assumption is that exchange-rate volatility is
costly and should be managed although, as Klaus Schmidt-
Does the recent experience suggest there is a need
Hebbel discusses, the economic case for this is far from
to change any of the PTA parameters? Examples to
obvious. Four broad areas of analysis could be undertaken
consider might include:
to shed more clarity in this area:
(i) sharper specification of the medium-term time
•
frame for targeting inflation;
(ii) the attention given, within an hierarchical structure,
rates? What is the contribution of domestic policy
to the exchange-rate, interest rates and GDP
(monetary and fiscal policy) versus other “suspects”
volatility; and
such as commodity prices, growth differentials, and
remains a band or instead is re-specified as a point
target.
Does the decision making structure for monetary policy
matter in practice?
•
What do the stylised facts tell us about New Zealand
exchange-rates and factors influencing the exchange-
swings in international investor sentiment?
(iii) consideration of whether the target inflation rate
•
If there is evidence that fiscal policy has exacerbated the
•
Does the New Zealand exchange-rate buffer or amplify
shocks? Are there any factors influencing the exchangerate that systematically “amplify” movements in the
exchange-rate, and if so, can policy do anything about
this?
Is there scope to more effectively communicate
New Zealand economic conditions and monetary policy
to international financial markets?
Testing stabilisation policy limits in a small open economy
11
•
•
What are the implications of exchange-rate cycles and
5.
volatility for longer-term productivity growth and the
Concluding remarks
The recent business cycle in New Zealand has tested
structure of the New Zealand economy? Are the effects
macroeconomic policy stabilisation limits. In part, this may
asymmetric? Can costs be found at the micro-firm level
have reflected an unusual confluence of shocks. Rising
if not in aggregate?
international commodity prices, including New Zealand’s
Are there gaps in the range of financial market products
agriculturally-based basket, are not normally seen in an
available to New Zealand firms that may help them
environment of weak growth in the G7 economies. Long-
hedge against currency volatility? If so, is there a role
term interest rates do not normally stay at historic lows when
for policy to help create a market for these types of
policy rates are rising. House prices do not usually increase at
products?
rates of plus 15 per cent per annum over a sustained period.
However, the recent experience bears some resemblance to
the cycle in the mid-1990s, where the exchange-rate also
Structural policy issues relating to
rose to levels well beyond its “fundamentals” as monetary
residential investment
policy leaned against similar domestically-sourced inflation
Although the brief of the Forum participants was to focus
pressures. And similar stresses were placed on the externally
on policy options for smoothing the economic cycle, a
exposed sectors of the economy.
number highlighted that the New Zealand cycle may be
affected by underlying structural policies that give rise to
low household savings rates and (possibly) a concentration
of wealth in housing assets. Policy issues that may warrant
further attention here include:
•
•
•
The purpose of the policy forum at which the papers in
this volume were presented was to test the robustness of
New Zealand’s macroeconomic policy frameworks and
to evaluate opportunities to improve those frameworks.
Although the overall conclusion that emerged from this
To what extent are New Zealanders “overweight” in
forum was that the essential elements of New Zealand’s
housing? Can taxation structures be altered to reduce
macroeconomic policy frameworks are fundamentally
any biases?
sound, there were also many questions asked and ideas
Do the potential distortions in the structure of household
raised that may warrant deeper investigation.
balance sheets increase the vulnerability of households
The complex issues involved imply many of these questions
and the financial sector to adverse shifts in investor
have no straight forward answers. Trade-offs that are difficult
sentiment, interest rates and the exchange-rate? If so,
to quantify with any degree of precision are inherent, for
are there any implications for prudential policies?
example, in policy suggestions to modify savings incentives
Do flows in net migration exacerbate the residential
investment cycle? Is it possible to improve the stability
and/or predictability of migration flows? Can regulatory
structures be improved to reduce the cost and timeliness
of supply of residential property?
or use fiscal policy more actively to stabilise the cycle. In
addition, the implementation of many of the suggestions
would require careful consideration of their impact on
existing institutional frameworks. Given these difficulties,
perhaps the most scope for advancing understanding of
the issues will derive from applying an inter-disciplinary
•
Do taxation structures contribute to the ‘amplifier’
effects of housing demand over the cycle, generating a
approach to the questions, potentially involving several
arms of Government.
propensity for New Zealand households to pay relatively
high real interest rates?
12
Reserve Bank of New Zealand and The Treasury
Macroeconomic policy challenges: monetary policy1
Stephen Grenville, Lowy Institute
1
Introduction: What, if
disaggregated level, there seem to be much more GDP
anything, has gone wrong?
volatility in the tradable sector compared with the nontradable, which might be driven by the wide swings in
New Zealand’s overall economic performance in the upswing
the exchange-rate over the course of the cycle.
from 1999 to 2005 has been exemplary, with GDP growth
averaging 3.5 per cent, taking unemployment down to an
The next section looks at how monetary policy has changed
enviable 3.6 per cent. True, inflation has crept up, but this
over the past two decades. The subsequent section looks
is consistent with the greater flexibility incorporated in the
more specifically at the way monetary policy works now,
inflation target regime. The current account deficit (CAD)
and makes the case that it has become much less potent
has been running at around 9 per cent of GDP, but it can
(at least in its counter-cyclical impact) as the economy
be argued that this is the expected result of international
has become more integrated with international financial
integration – the CAD allows separation of saving and
markets. Section 4 looks at possible problems, while Section
investment decisions.2
5 summarises the issues that policy might address. The
By 2005 capacity constraints were showing, but by the
second half of the year demand growth was slowing, and
the housing boom seemed to be levelling out. There is the
balance of this paper looks at what might be done, using
monetary policy, using prudential policy and lastly through
greater integration with Australia.
prospect of a “dream rebalance.”3
This is a small very open economy operating much as the
text-books suggest it should. What residual concerns might
2
changed over time?
there be?
•
The CAD is an outlier both historically and internationally.
The underlying mean level may indicate that there is a
structural issue here, with the need to examine saving
incentives. Perhaps greater ability to borrow (equity
withdrawal by house-owners) has reduced savings and
this may have further to run.
•
While inflation has been running at an acceptable
rate, inflation expectations are creeping up, and a
depreciating exchange-rate could put further pressure
on the inflation target at a time when it will be
inconvenient to raise rates to support the currency.
•
1
2
3
4
How has monetary policy
Two environmental changes might be noted. The more
obvious one is the integration of New Zealand with world
financial markets, with the exchange-rate float of 1985
marking the watershed. The other factor is a shift in thinking
about what monetary policy might (and should) be trying
to achieve.
New Zealand pioneered inflation targeting, which has
been developed and refined to become the “best practice”
approach of most central banks world-wide. Such was the
entrenched nature of the earlier inflation, and so fundamental
was the break from earlier approaches to macro policy, that
there was a perceived need to put monetary policy in a
Internationally, the amplitude of the business cycle
precise and tightly-defined stand-alone role, anchoring the
seems to have moderated.4 Early indications are that
price level, with no responsibilities (or even regard) for the
this has happened also in New Zealand, but at a
cycle.
I am very grateful for the willing help from RBNZ and New
Zealand Treasury staff, and comments and suggestions from
others.
That is, New Zealand has escaped from the Feldstein/Horioka
constraint.
See “The Dream Rebalance?” Stephen Toplis (Bank of
New Zealand) April 2006.
See Cotis and Coppel (2005).
Testing stabilisation policy limits in a small open economy
13
The second evolutionary element in the monetary policy
Since the float, increasing integration with world capital
environment was the gradual relaxation of this view, as
markets has likely raised the importance of the exchange-
inflation was contained and price expectations anchored.5
rate channel of monetary policy to a point where, in recent
This was not, of course, to expect that monetary policy
times, it may dominate the interest rate channel. The text-
could boost growth beyond the constraints of capacity: it
book model envisages that higher interest rates cause a
was, instead, to recognise more specifically that monetary
floating exchange-rate to rise: this creates an expectation of
policy impinges on the course of the business cycle and
depreciation as the exchange-rate reverts to its longer-run
vice versa. The current debate (to which this paper might
equilibrium, and the expectation of depreciation equilibrates
hope to contribute) is a continuation of the exploration
the higher domestic interest rates with international
of the ways in which monetary policy affects not only
rates.8 In this world, monetary policy still has some of its
price stability, but also inter-acts with the real side of the
impact via interest rates, but to the extent that the action
economy. The more comprehensive view of monetary
is now through the exchange-rate, monetary policy does
policy includes consideration of how it interacts with other
not impinge very directly on an excess demand shock to
“arms” of policy (prudential and fiscal), and the search for
slow it, but instead helps to divert or “spill” it into a larger
“supplementary instruments” which can take some of the
CAD through changes between tradable and non-tradable
pressure off monetary policy.
prices.
Acknowledging that the principal objective of monetary
A floating exchange-rate will buffer a terms-of-trade shock
policy is price stability and that this can be (and has
by smoothing exporter incomes over the course of the
been) successfully achieved, the focus here will be on the
commodity cycle. This in itself is a helpful characteristic, but
interaction of monetary policy and the cycle. New Zealand
the exchange-rate movement does nothing to encourage
is subject to a variety of shocks, but the issues might be
extra savings in “good times” when commodity prices
illustrated by looking at one external supply-side shock
are strong: the flexible exchange-rate spreads the benefit
(terms-of-trade, via commodity prices) and one internal
of the shock widely, in the form of cheaper imports. The
demand shock (the housing cycle). Before the float in 1985,
beneficiaries, the majority of whom will have no direct
policy-determined short interest rates (and, for that matter,
economic relationship to the commodities sector, may
direct controls) impinged directly on excess demand shocks
not understand that “seven fat years” would be followed
such as a housing boom. Terms-of-trade shocks produced
by “seven lean years,” so saving may fall in response to a
booms and busts for the commodity sector, which might
favourable shock.
6
be buffered to some modest extent because the income
variations accrued directly to the commodity-producing
sector, encouraging a cyclical savings variation which may
have helped smooth the CAD cycle (savings rose when times
were good for exporters, although there was the possibility
that this might be offset by investment increases).7
5
6
7
14
This evolution was formalised in the introduction of Clause 4(c)
(later Clause 4(b)) in 1999 to the Policy Targets Agreement as
an explicit recognition that unnecessary volatility in output,
interest rates and the exchange-rate is detrimental to economic
welfare, and may have adverse consequences for economic
growth. See also Bollard and Karagedikli (2005), and RBNZ
“The Evolution of Monetary Policy Implementation” (RBNZ
website under “Monetary Policy”).
Which might be seen as home-grown, but might be affected
by New Zealand’s relative cyclical position influencing
demographics.
This might have been reinforced by the prevalence at the time
of single-seller desks and producer boards which aimed to
smooth farmer incomes.
This configuration of interest rate/exchange-rate is less wellsuited to constraining a housing boom, as the interest rate
channel (which might have been effective in an interestsensitive mortgage market) is muted, and the exchangerate channel squeezes the tradable sector, when the driving
cause of the boom is the rapid increase in house prices and
the associated wealth-effects.9
8
9
This effect on the exchange-rate will be stronger, the longer
the rise in the short-term policy rate is expected to last.
One commentator described the process as: “Sacking the All
Blacks’ coach when the New Zealand cricket team loses a
match.”
Reserve Bank of New Zealand and The Treasury
What we might expect to see in this integrated world
Figure 2
(compared with the old fixed rate/ less integrated world) is
Demand “spill-over” into net exports
less impact of monetary policy on the amplitude of the cycle
in demand (because interest rates are not directly affecting
the cycle so strongly, and counter-cyclical movement
%
%
10
10
GDP
Domestic demand
8
8
6
6
in saving may not be present), and larger fluctuations in
4
4
the current account, with a larger current account deficit
2
2
associated with a stronger exchange-rate. When we look
0
0
for these characteristics in the New Zealand data, they are
-2
-2
partly obscured by other factors. The outcomes may have
-4
-4
been muted in the 1990s, first by the influence on interest
-6
rate settings of the “exchange-rate comfort zone” in the
first half of the decade and then by the MCI-based policy
until 1999. But the exchange-rate did largely offset stronger
-6
90 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05
Figure 3
Tradable and non-tradable CPI inflation
%
%
6
commodity prices in 1995-6 and again in 2003-2006 (see
6
Figure 1), although in 2001 this effect is not present (more
5
5
on this later).10
4
4
3
3
Figure 1
2
2
Commodity prices, USD and NZD
1
1
0
0
Index
170
Index
170
-1
160
-2
150
-3
160
150
140
NZD price
World price
-1
Tradable
Non-tradable
CPI
-2
-3
92
93
94
95
96
97
98
99
00
01
02
03
04
05
06
140
130
130
120
120
In terms of price stability, the post-1985 system works well
110
110
(and in particular the inflation targeting system has been very
100
100
effective). The cyclical variation in the CAD acts to counter
90
90
80
80
90 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06
the inflationary impact of the cycle directly through cheaper
imports (see Figure 3, where tradable prices reflect both
world prices and the exchange-rate) and indirectly through
The “spill” of strong demand growth into net imports
extra supply from imports (Figure 2). But, with echoes of
can be seen in 1995-6 and most clearly in 2003-6
the old “financing versus adjustment” debate, both types
(see Figure 2).11
of shock are being “funded” rather than “adjusted.” If we
consider the impact on the cycle, we might be more inclined
to look for alternative or supplementary instruments,
particularly in the case of domestic shocks such as a housing
boom.
10
11
See Chen and Rogoff (2003).
2001 had the unusual combination of a weaker domestic
economy, strong export demand and a weaker exchange-rate
which combined to give a powerful boost to net exports.
Testing stabilisation policy limits in a small open economy
15
3
Where does this leave
Figure 5
monetary policy now?
Fixed and floating loans
International financial integration seems to have made
%
monetary policy much less potent in its impact on the cycle.
90
Borrowers (particularly housing borrowers, see Figure 5)
80
have moved further out on the yield curve, borrowing at
%
100
100
90
80
Floating
70
70
Fixed
60
60
interest rates only modestly influenced by the movement in
50
50
the OCR (see Figures 4 and 6). The most recent tightening
40
40
30
30
20
20
10
10
phase illustrates the issue. In the first seven months of 2004,
the OCR was raised by 100 b.p. (5 per cent to 6 per cent, see
Figure 4) and the RBNZ warned the market in its September
2004 MPS that interest rates would have to rise further. This
movement and clear statement that high rates would be
maintained should have been fully incorporated into the
term structure of interest rates (including the anticipation
0
1998
0
1999
2000
2001
2002
2003
2004
2005
2006
Figure 6
Weighted average rate new borrowers
%
%
18
18
of future higher rates), but in practice had little effect on
16
16
tightenings, the mortgage rate had risen significantly and
14
14
had not fallen appreciably below the floating rate during
12
12
this phase of the cycle, in 2004 and 2005 the prevailing 2-
10
10
8
8
the yield curve (see Figures 7 and 8). Whereas in earlier
year rate was 100b.p. or more below the floating rate (see
Figure 8). At the same time, because existing borrowers are
borrowing for a fixed term rather than at floating rates, the
impact of an OCR rise is lagged, and those borrowers who
got themselves “set” with a fixed rate in the second half
of 2004 (when RBNZ was moving strongly to tighten) have
6
6
91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06
Figure 7
The OCR and the slope of the yield curve
Basis points
300
largely avoided the tightening until now.
%
6.5
150
6.0
9.0
100
5.5
8.5
50
5.0
0
4.5
%
9.0
OCR
Effective mortgage rate
8.5
8.0
7.0
200
OCR and effective mortgage rate
8.0
7.5
7.5
7.0
7.0
-50
6.5
6.5
-100
6.0
6.0
5.5
5.5
5.0
5.0
4.5
4.5
16
OCR (RHS)
250
Figure 4
4.0
1999
%
7.5
1999
4.0
Spread between 5 yr rate
and 90 day rate
2000
2001
2002
2003
2004
3.5
2005
2006
4.0
2000
2001
2002
2003
2004
2005
2006
Reserve Bank of New Zealand and The Treasury
Figure 8
and the NZD bond offshore issuance, which fund the strong
Mortgage rates
growth in banks’ mortgage lending (Figure 10).
%
%
10.0
Floating
2 years
5 years
9.5
9.0
10.0
There is still some arbitrage along the yield curve (consistent
9.5
with the expectations hypothesis), but there is also powerful
9.0
international arbitrage with overseas longer rates (Figure 9).
8.5
8.5
8.0
8.0
As the RBNZ was pushing up the OCR, the longer end was
7.5
7.5
being affected through the expectations channel, but also
7.0
7.0
6.5
6.5
6.0
6.0
5.5
5.5
5.0
1999
5.0
2000
2001
2002
2003
2004
2005
2006
by the low world interest rates.
What about the impact of the monetary tightening on the
exchange-rate? The strong capital inflow, greatly influenced
by the demand for domestic credit and the banks’ need for
foreign funding, determines the current account (an identity
Figure 9
if there is no official intervention). The exchange-rate has
Co-movement of interest rates
Correlation
to move by enough to “spill” demand into net imports,
Correlation
1.0
opening up the current account deficit so that it equals the
0.8
0.8
0.6
capital inflow. This movement in the exchange-rate often
0.6
1.0
0.4
0.4
0.2
0.2
0.0
0.0
-0.2
NZ-US (10-year)
NZ-Australia (10-year)
NZ-US (2-year)
NZ-Australia (2-year)
-0.4
-0.6
-0.2
-0.4
-0.6
-0.8
-0.8
-1.0
1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006
NZD millions
3000
12
10
8
Net offshore NZD bond issuance (RHS)
Credit growth (LHS)
“heavy lifting.” The main driver is the exchange-rate, driven
by capital flows at the longer end (the increase in net foreign
NZD bond issuance in 2005 was big enough to fund the
whole CAD), rather than short-term interest differentials.
(a) Does the exchange-rate move too much?
bond issuance
14
The central point is that interest rates are not doing the
outcome. It does, however, raise some issues:
Credit growth and offshore NZD
%
differential.12
It could be argued that this is an acceptable, even desirable,
Figure 10
16
seems to be much more than is justified by the interest rate
(b) As a consequence of this, are some sectors unfairly or
unnecessarily subject to excessive price shifts?
2500
2000
(c) If monetary policy is having its main impact via
1500
the exchange-rate rather than interest rates, can it
1000
effectively address a domestic demand shock such as a
500
housing boom?
0
6
-500
-1000
4
1997 1998 1999 2000 2001 2002 2003 2004 2005 2006
Borrowers are in effect reaching into the lower interest
rates available in world capital markets (see co-movements
of New Zealand and USD interest rates, Figure 9). This is
reflected in the close movement of domestic credit growth
Testing stabilisation policy limits in a small open economy
(d) Is the CAD, in some sense, too big?
(e) Does the external funding make the banks vulnerable?
12
The contribution which the exchange-rate makes to the “spill”
depends on import elasticity and the price pass-through (this
is probably getting less over time (see Hampton (2001), so
the exchange-rate might have to move further), and casual
observation suggests that price relativities produced by the
exchange-rate change are not doing much of the work.
17
4
The possible problems
Figure 12
(a) Does the exchange-rate move too much?
The nominal USD/NZD exchange-rate moves by around 20
per cent on either side of the average over the course of the
cycle i.e. about 40 per cent between peak and trough.13 The
Australian dollar – also a “commodity currency” – moves
AUD and USD exchange-rates
Index
Index
140
130
140
Real US$/NZ$
Real AU$/NZ$
130
120
120
110
110
by around 30 per cent on the same basis. The periodicity of
100
100
these exchange-rate cycles matches the general economic
90
90
cycle, which also generally corresponds with the world
80
80
commodity cycle.
70
70
Three other characteristics might be worth noting here. First
that the real exchange-rate, while fluctuating over a wide
range, has been remarkably stable over the past 35 years
(see Figure 11). Second, that there is much less fluctuation
against the Australian dollar (see Figure 3 of Munro (2004))
– perhaps because they are both responding to the world
commodity cycle.14 Thirdly, that since the reforms of the
1980s, the earlier persistent slide of the nominal exchangerate has been halted.
60
60
70 72 74 76 78 80 82 84 86 88 90 92 94 96 98 00 02 04 06
Interest rate differentials don’t seem to be enough to
explain the movements. An interest differential moving
typically by around 300 b.p. over the course of the cycle
might, if the cycle lasted say four years between peak and
trough, explain (at most) an exchange-rate swing of 12
per cent. To put the same point slightly more formally, the
relationship found by Huang (2002) was that 100 b.p. of
differential is associated with an exchange-rate change of
Figure 11
6 per cent, which would imply that the cycle has a much
Real exchange-rate
longer periodicity than observed.15
Index
Inde
x 180
180
160
140
Nominal TWI
Foreign/domestic price level
Real exchange rate
160
140
Nor does the commodity price cycle seem to be an adequate
explanation. Certainly, the exchange-rate should respond to
a change in commodity prices (see Blundell-Wignall et al
(1993), and Chen and Rogoff (2005)), but if these are regular
120
120
100
100
80
80
60
60
This remains the puzzle: there seems to be a fairly regular
40
40
opportunity to buy cheap and sell dear, making up to 40
70 72 74 76 78 80 82 84 86 88 90 92 94 96 98 00 02 04 06
(albeit persistent) cycles, rather than permanent shifts, then
movement over the course of the cycle should be limited
by the arbitrage opportunities based on past fluctuations.
per cent over the course of one downswing or upswing
if the timing is right. Of course there is uncertainty as to
the timing of the cycle but perhaps the main constraint on
exploiting this is the periodicity of the cycle – the position
would have to be held for half a cycle – probably four or five
years. Few arbitrageurs can hold a position for that long.16
13
14
18
The IMF measures exchange-rate movements in terms of
persistence and standard deviations (see IMF Special Issues
2005), but the issue here is best captured in a more intuitive
way – the height of the peaks and troughs.
Trade between the two nations does not appear to be enough to
make this a strong anchor of the cross rate – Australia’s share
of New Zealand imports is only around 20 per cent.
15
16
This relationship applies to the period since the OCR was
introduced in 1999. Before that, the relationship seems even
smaller – 100bp being associated with a 2-3 per cent change
in the exchange-rate (see Zettelmeyer 2000).
For a discussion of the same phenomenon in Australia, see
Gruen and Kortian (1996).
Reserve Bank of New Zealand and The Treasury
Overshooting of this kind might seem to strengthen the case
the exchange-rate did not strengthen in the face of stronger
for foreign exchange intervention. Some movement in the
commodity prices.
exchange-rate over the course of the cycle is desirable (to
help spill excess demand into net imports, and to smooth
inflationary pressures over the cycle). For both these reasons,
when we come to look at possible actions below, tightly
constraining the movement in the exchange-rate would be
inconsistent with the current approach to monetary policy,
but actions to “lop the peaks and fill the troughs” would
seem to do little to inhibit the effectiveness of policy, might
reduce the danger that a downward overshoot might unanchor inflation and inflationary expectation, and might
These funding vulnerabilities have usually been discussed
in terms of roll-over and maturities, and given the retail
distribution methods for the financial instruments, this
may be the most relevant issue. But there is at least the
possibility that these investors could change their FX
exposure at any time, by selling NZD (even if they go on
holding the NZD denominated instrument). It has been
argued that this sort of funding, in which the foreigners
take the FX risk, is inherently safer than other capital flows
where the FX risk is carried by New Zealanders (which
reduce the damage to non-commodity exporters (see next
are characterised as being subject to “original sin” (see
section).
Eichengreen and Hausmann (1999)). While this means that
This case for intervention is strengthened if we explore a
foreign investors (rather than New Zealanders) are holding
low-probability high- impact scenario. Nearly 40 per cent
the balance sheet risk if the NZD falls, their actions could
of the funding for banks’ credit expansion comes ultimately
potentially be a powerful influence on the exchange-rate
from NZD-denominated foreign raisings. These instruments
in the event of a reassessment of New Zealand’s rating. The
are held by unsophisticated investors (mainly Japanese
amounts are now much larger than in 2000 (more than $50
retail investors) with little knowledge of the New Zealand
billion outstanding – or 30 per cent of annual nominal GDP
economy, attracted by the high running yields (a situation
or 115 per cent of annual export earnings) and if these retail
enhanced by the temporary factor of low domestic interest
investors shift out of NZD, it may require a big shift in the
rates in Japan). If, during the downward phase of the NZD,
exchange-rate before other non-residents take over the FX
they decide to cover their FX exposure, the adjustment could
exposure from them.
be substantial. Most of the adjustment is likely to take place
It could be argued that these transactions are being entered
in the exchange-rate rather than the interest rate, with the
into by “consenting adults” who understand the risks and
new equilibrium reached when there are alternative non-
have factored them into their calculations. But in this case
resident investors willing to hold the FX exposure, because
the risk is of a “macro-financial” nature, where the individual
they are fairly confident that there has been enough
decisions may be rational and largely risk-protected, but
exchange-rate overshooting to make this worth the risk.
the impact of an adjustment on the economy will be felt
We know that there are not many currency arbitrageurs,
more widely than on those involved in the transaction
because the exchange-rate moves so widely over the course
itself. The intermediaries – the New Zealand banks – feel
of the terms-of-trade cycle, so a large overshooting seems
possible. It may be that we have already seen a small episode
of this. In the 2000-02 period there was a run-down in
outstanding Uridashi-type funding from around $20 billion
to around $12 billion. This coincided with a downward
overshooting of the exchange-rate which was not only
much larger than can be explained by equations involving
interest differentials and commodity prices (see Munro
well protected because their FX position is balanced;
the borrowers face only the roll-over risk if they want to
continue their mortgages beyond the borrowing period;
the ultimate lenders are relaxed either because of their
ignorance, or because this is only a small part of a diversified
portfolio.17 The risk lies to macro stability. This might come
in the form of drastic cutting back of credit because new
funding is not available. Or it might come in the form of
(2005)), but it is the one example in recent experience when
17
Testing stabilisation policy limits in a small open economy
Neither of these factors would prevent them from cutting their
exposure if the exchange-rate moves unexpectedly.
19
substantial pressure on inflation during an exchange-
that the cyclical adjustment in the tradable sector is on the
rate overshoot: the greatest threat to price stability in the
import side, the export sector is unaffected: in the most
targeting era seems to have been in the period of Uridashi
recent upswing (between 2002 and 2005), imports rose
maturities in the early 2000s, when the exchange-rate
from around 30 per cent of GDP to around 40 per cent.
weakness pushed up tradable prices in 2000-01 (see Figure
3). It would be inconvenient, to say the least, if exchangerate weakness forced a tightening of interest rates at a time
when “headwinds” in credit-provision, a cyclically-weaker
housing sector, declining consumer confidence and weaker
commodity prices were working together to weaken
growth.18,19
Figure 13
Six-year rolling standard deviation of GDP
Standard deviations
0.045
0.04
0.035
0.035
0.030
0.03
0.025
0.025
0.020
0.02
12
0.015
Tradables GDP
Aggregate GDP
Non-tradables GDP
0.010
%
14
Standard deviations
0.045
0.040
0.015
Interest rates defending the exchange-rate
NZD/USD
0.8
Figure 14
0.01
0.005
0.005
0.000
0
90
0.7
91
92
93
94
95
96
97
98
99
00
01
02
10
0.6
8
6
0.5
4
NZD/USD
90-day rate (RHS)
0.4
0.3
1991
1995
by the variability of output in the two sectors. Figure 14
shows that tradable-sector variability is greater than either
non-traded or total (as measured here, by a six-year rolling
2
0
1993
One measure of the degree of disruption might be given
1997
1999
2001
2003
2005
(b) Are the tradable and non-tradable sectors
annual average growth standard deviation). Leaving aside
the issue of whether this maps the cycle closely, even if
we accept that the tradable sector is more variable, this
could be accounted for by the normal nature of the sector
excessively disrupted?
– subject to shocks other than the shock administered by
All that can be offered here is a rough idea of the magnitudes
exchange-rate variability. Greater stability of the non-traded
involved. We know that over the course of the cycle the
sector may be an intrinsic, universal characteristic, reflecting
CAD widens by around 3 percentage point of GDP. But
the stability of the services and government components of
the impact will be felt differently in different parts of the
this sector.
tradable sector. For commodity exporters (45 per cent of
total exports), the movement of the exchange generally
coincides with the commodity cycle, leaving the NZD price
of commodity exports stable (see Figure 1). To the extent
Buckle et al (2001) address some of these issues, and find
that the greater stability in aggregate GDP observed over
the past twenty years results from the absence of the
policy-making shocks of the 1980s (the “Think Big” era).
There are no indications that the sector which might have
benefited from more stable NZD prices (the primary sector)
is less cyclical than before.20
18
19
20
Nor could there be strong reliance that the weaker NZD would
quickly strengthen net exports. This is reflected in changes
made to the RBNZ model FPS: “the impact on the economy
is now smaller and slower. This change was made in response
to the surprisingly slow impact the weak New Zealand dollar
had on New Zealand activity between 1998 and 2000.” RBNZ
Forecast and Policy System August 2004.
The RBNZ might “look through” the pressure on inflation and
resist raising rates, but it is worth noting that rates were raised
during the last two period of exchange-rate weakness.
20
Hours worked in the traded sector (and in the manufacturing
component of this sector) don’t seem all that sensitive to the
cycle.
Reserve Bank of New Zealand and The Treasury
Figure 15
Central bankers are ambivalent about the role of monetary
Goods exports as a per cent of GDP
policy in asset booms/bubbles, but the damaging effects
% of GDP
% of GDP
25
of the housing cycle seem at least as clear in New Zealand
as in any of the countries where this has occurred (see
24
24
23
23
22
22
21
21
20
20
19
19
the building sector) and sets the stage for an uncomfortable
18
18
adjustment.
17
17
90 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06
Hebling 2005). The dominance of housing in household
balance sheets, the rise in household debt and the absence
of a broad capital-gains tax all suggest that this has been
unhelpful (with much wider impact on demand than simply
As against this, house-building (as distinct from house
price inflation) is needed to cope with the demographics of
Based on this rather flimsy evidence, we might conclude
that generally (i.e. when the exchange-rate moves with
commodity prices) the primary sector as a whole is not
a large net migration. On balance, there seems a case for
exploring measures that would impinge more directly on
sector-specific booms.
seriously disadvantaged by the large swings in the exchangerate (although this would not be true for every commodity
sector).
(d) Is the CAD “too big”?
Even if this is true, it is still possible that these exchange-
The cyclical component the CAD is the “spill” of excess
rate swings have a structural impact – discouraging
demand, so is a part of the stabilisation mechanism. Less
investment (and production) in the tradable sector, through
spill means more pressure on domestic resources (including
an uncertainty premium. It is even harder to get a handle
inflationary pressure), so unless measures can be found
on this. One characteristic of the New Zealand tradable
to smooth domestic demand more effectively, the cyclical
sector is that it has grown much less, as a share of GDP,
variation in the CAD probably has to be accepted.
than the OECD average. But this may simply be a different
But the mean level of the CAD is another matter. Sustainability
intrinsic characteristic of the New Zealand economy (the
calculations for CADs are notoriously subjective: common
contrast between the close integration that has occurred in
“rules of thumb” such as sustaining current levels of net
Europe, compared with the commodity-base comparative
foreign debt as a per cent of GDP can be interpreted
advantage of New Zealand). What could be said, however,
differently (see Edwards, this volume).21 But if the underlying
is that goods (and total exports/GDP) has risen only slowly.
mean level has shifted from the 3-4 per cent of GDP to 5-6
The secular slower rise in tradable prices (which is a universal
per cent, the likelihood that it is unsustainable (in the Herb
phenomenon) may suggest that there is less producer
Stein sense of requiring adjustment) seems high, and the
price-power and more competition in this sector, perhaps
policy question is: “how painful will the adjustment be?”
discouraging investment.
Any correction needs to come largely on the savings side,
as there is a generally-accepted need for greater capital
(c) Should policy aim to operate more directly
deepening to raise productivity.22
on the source of the shock?
In two of the last three cycles the housing sector has been
21
an important element of the shock, and to respond to
this by squeezing the tradable sector seems to apply the
instrument of policy to an area well-removed from the
source of the problem.
Testing stabilisation policy limits in a small open economy
22
To give an example of this – with nominal GDP growth of 5
per cent, the CAD would have to be held to 3.8 % of GDP
to keep the net foreign debt at its current level, which (with
investment income around minus 6 per cent of GDP) implies a
positive balance of goods and services of more than 2 per cent
of GDP, compared with minus 2 in 2005.
A fall in housing investment has some potential to contribute
to the re-balancing.
21
This isn’t the place to debate whether New Zealand saving
banks cannot take this exposure for prudential reasons and
is inadequate, but three points might be made. First, that
the existence of foreigners who are ready to take the NZD
the provision of a universally-available retirement pension
exposure is fundamental to this type of capital flow.
seems very likely to distort individual choice in the direction
of lower-than-optimal household saving. Second, the
housing boom seems to have influenced household saving
markedly, and it is not clear that households are correctly
assessing the effect of this on their wealth. Third, if the
Figure 16
New Zealand bond issuance offshore
Maturing/issued
(NZ$billion)
5
Issues
4
underlying saving were raised (say, through mandatory
3
retirement contributions),23 then the exchange-rate would
2
be lower, on average, than currently.24 While the peaks of
the CAD are likely to get the publicity, shifting the average
level lower would give an on-going and sustained different
price incentive (and profitability) to the tradable sector.
Outstanding (NZ$billion)
55
Maturities
Outstanding (RHS)
50
45
40
35
1
30
0
25
-1
20
-2
15
-3
10
-4
5
-5
1994
0
1998
2002
2006
2010
2014
(e) Vulnerability of financial sector
One unusual characteristic (although shared by Australia)
is that the banks draw so heavily on overseas sources of
funding. We noted above that the ultimate suppliers of
the funding (typified by Japanese retail investors) may not
be well-informed investors, creating at least the potential
for them to change their position: their lack of detailed
knowledge means that the arrival of a small amount of new
information may represent a large increase in their overall
understanding, and trigger a change of investment holding.
In terms of conventional prudential issues, the banks seem
safe enough: they have covered both their FX and interest
exposure, and mortgage lending has a low-risk history. The
issue is a macro-financial one:25 what is the wider effect of
any adjustment that the banks might need to make? The
main adjustment might be the loss of such an important
funding source, and the “headwinds” this might create, as
banks reduce the growth in their lending. It is worth noting
that the banks cannot compensate for the loss of the FX
cover by simply continuing their current USD-denominated
The issue could be seen this way: the banks themselves
have protected their prudential positions by shifting risk
elsewhere. In a financially integrated world with large
international flows, some party to the transaction bears
an FX exposure: this risk cannot be removed by hedging,
only shifted to another party. Will the risk-takers (the
Japanese investors) go on taking this exposure? The New
Zealand household borrowers bear other risks, mainly rollover risk as they have borrowed shorter than their planned
asset-holding period. A macro-financial view would take
into account not only the impact of events on the financial
intermediaries which are the direct responsibility of the
prudential authorities, but the macro implications of the
unfolding events if some of these risks came to pass. The
case for policy action, bearing on the banks, is motivated
not by concern for the short-term prudential health of the
banks’ balance sheets, but by concerns to ensure that the
banks can continue to play their role as the predominant
intermediary in the New Zealand economy.
foreign borrowing, without the NZD cover: the New Zealand
23
24
25
22
A mandatory retirement contribution seems to have given
added impetus to the development of a deeper funds
management sector in Australia, and a deeper body of
managed investment funds might help financial stability (c.f.
Greenspan’s “spare tire”).
This is true in the short and medium term, although in the
long run, cumulated changes in national wealth may alter
this.
See White (2005b).
Reserve Bank of New Zealand and The Treasury
5
Summary of the issues
time period may change over time. At present, however,
We have identified a number of issues. Before seeing
the “wiggle room” seems to have been pretty fully used,
whether policy is in a position to address these (and which
as price expectations (at least as measured by financial
“arm” of policy might be most suited), we might summarise
markets commentators) have moved up in response to
them:
the cyclical pressures on prices.
•
•
When monetary policy is tightened, effective borrowing
•
The role of asset prices is under debate among central
rates don’t move as much as might be expected, but
bankers.26 New Zealand practice seems to be centrally
the exchange-rate moves much more.
positioned in the spectrum of opinion (see Bollard
(2004a)), with recognition that wealth effects will
Monetary policy does not impinge very strongly or
influence demand and hence pressure on prices, but
directly on demand shocks (such as a housing boom).
•
without giving monetary policy the task of pricking
The banks’ heavy reliance on foreign funding leaves
them vulnerable, not to a prudential risk of failure,
but to a risk that they would have to sharply alter the
asset bubbles.
•
The single decision-maker role of the RBNZ Governor
was identified by Svensson (2001) as the major departure
growth of their lending.
•
of the New Zealand system from usual international
The large average CAD suggests the need for adjustment
best practice. This issue was comprehensively addressed
that will involve higher national savings.
in the RBNZ submission at the time, and there seems
little more to add. For a counter-view, Blinder and
6
Morgan (2000) argue for better decision-making
What might be done with
by committees, a finding backed up by the Bank of
monetary policy?
England’s experimental work reported by Mervin King
The inflation targeting framework
As noted, New Zealand has a variant on current “world best
(2002).
•
Measurement of CPI. Asset prices play some role in the
practice” for monetary policy and it would require more
New Zealand CPI, because of the way housing costs
serious problems than are currently being experienced to
enter the index. This might not be as “pure” as some
justify a radical departure. Marginal modifications or re-
would like, but if it helps the RBNZ take more account of
balancing could be explored, but there are no obvious
housing price movements, it seems on balance no bad
candidates for radical change:
•
A central issue in any inflation targeting regime is
the degree of flexibility around the inflation target.
Sometimes this is specified in terms of a range, but
more usefully it is specified in terms of the time period
in which the target must be maintained – how quickly
thing.
Figure 17
Official interest rates
%
8
Australia
UK
Canada
USA
7
%
8
NZ
Japan
Sweden
Euro
7
6
6
should the recorded rate return to the specified band.
5
5
Given the re-specification of the target in September
4
4
2002, the RBNZ has the freedom to use not only the
3
3
range, but the specification “over the medium term.”
2
2
The central issue here is price expectations. The relevant
1
1
0
2000
0
2001
2002
2003
2004
2005
2006
Source: Central banks
26
Testing stabilisation policy limits in a small open economy
See White (2006) and references in this paper.
23
•
Is the RBNZ “too active”? In the current cycle,
stance of policy. For example, even if it had been felt that the
New Zealand moved the OCR up by 225 b.p., which
appreciation of the NZD was excessive in the second half of
doesn’t seem out of line with international experience
2005, the OCR setting was restrictive, and foreign exchange
(see Figure 17). In any case, if the shock that New Zealand
intervention might have been seen as inconsistent.30
was encountering was larger than those experienced
elsewhere, large movements might not be described as
“more active”.27 Huang (2002) argues that if the implicit
Taylor rule used by the Australian and US authorities
had been applied to New Zealand, interest rates would
have moved much as they did in practice.
•
This possibility of intervention seems a valuable addition to
the policy arsenal. The RBNZ has described it as “another
monetary policy tool,” in addition to the setting of the OCR.
Like all tools, it has to be used with discretion, but the hurdles
that have been put in front of its use, and the historical
record, may inhibit its use. This would be a pity, as the
Transparency. The RBNZ is more transparent than most
Australian experience suggests that it is an instrument with
central banks, particularly on forecasts. On one aspect
more opportunity for beneficial use than is implied by RBNZ
there might be room for more public dialogue. The RBNZ
commentary.31 If markets are working well and delivering
(and the Governor) have been active in talking about
the correct price signals, there is no case for attempting
the housing boom,28 but might have been much more
to change this outcome. But the evidence of excessive
boldly explicit in pointing out the dangers to individuals
movements and unexploited arbitrage opportunities looks
and the economy. A rise in the price of an asset that
to be quite strong. The “alternative tool” argument is
an owner-occupier will go on using has only a limited
particularly strong if the currency is falling sharply during
wealth effect, properly assessed. Of course, calling the
the downward phase of the cycle. In the exchange-rate falls
housing cycle is a tricky business and there are issues of
of 1997/8 and 1990/91, interest rates were raised around
reputation at stake. At least, however, there might be
200bp (see Figure 13). If the exchange-rate falls sharply
more extensive official factual commentary next time
during the impending weaker phase of the cycle, there
around, and for this the statistical base needs to be
may be opportunity to use the intervention instrument
developed further.
rather than the OCR.32 The power of commentary on the
exchange-rate, used sparingly, again should not be underrated.
Foreign exchange intervention
We have discussed the low-probability high-impact
The RBNZ has, since March 2004, had a specific policy of
scenario that New Zealand’s ability to attract large NZD-
being ready to intervene in the foreign exchange market to
denominated foreign inflows might change sharply. The
“trim the peaks and troughs of the exchange-rate cycle.”
issue is not whether New Zealand would be able to borrow
The circumstances in which this may occur have been set
internationally: the issue is whether there might be moments
out.29 It is envisaged that these actions will be “probably
rare” and no intervention of this type has yet occurred. In
routine practice the main constraint may be that, even if
•
•
the exchange-rate movement is judged to be excessive,
intervention may be inconsistent with the then-current
27
28
29
24
Perhaps the only episode that might be described as overlyactive is the reduction in rates in 2003, reversed shortly
afterwards.
Bollard (2004b) addressed the issue, but in a relatively lowkey way.
“Specifically, before intervening the Bank will need to be
satisfied that all of the following criteria are met:
•
the exchange-rate must be exceptionally high or low;
•
the exchange-rate must be unjustified by economic
fundamentals;
30
31
intervention must be consistent with the PTA; and
conditions in markets must be opportune and allow intervention
a reasonable chance of success” (Eckhold and Hunt (2005) ).
One important issue not dealt with in this article is whether
the profits on intervention at one phase of the exchange-rate
cycle are distributed to the government. As these “profits”
may be needed to offset against the balance sheet revaluation
losses that occur when a long FX position is held during an
appreciation phase of the NZD, an accounting treatment that
allows some revaluation reserves would be desirable.
Not because of any concerns that this would have been
unsterilized intervention – intervention is always sterilized,
in the sense that liquidity operation will always be carried out
to leave system liquidity at the appropriate level.
See Becker and Sinclair (2004).
Reserve Bank of New Zealand and The Treasury
when foreigners who have taken a currency exposure
attraction (as noted in the SSI Paper) is that it is a price-
change their minds. We could explore the quantities of
based mechanism. This point might be strengthened by
intervention that might offset this (see Gordon (2005)), but
observing that it would seem to be doing what the market
this suggests that reserve holding is the key, whereas the
is failing to do – incorporating the implication of the OCR
main issue may be preparing the decision-making mind-set.
into the whole interest rate structure. For those who see
Substantial intervention might run reserves down, but the
this as interference in the workings of the market, it might
FX borrowing capacity of the New Zealand Government
be noted that such a levy would just do what monetary
(and the RBNZ) is greater than any likely requirement.
policy routinely does – shift the rate of interest away from
If foreigners are reluctant to hold NZD exposure, one
its “natural” (in a Wicksellian sense) rate.
response would be to encourage more New Zealanders to
The SSI Paper is on balance against this idea, largely because
hold FX exposure.33 Arguments supporting this are:
of administrative problems, the need for discretionary
(a). Borrowing in NZD is more expensive than borrowing in
USD. New Zealand interest rates, along the yield curve,
are routinely significantly higher than in Australia or the
USA.
decisions, and because it is seen as a “significant departure
from the current approach to monetary policy.” The
argument against discretionary action, and the difficulty of
recognising booms in advance, could be addressed if the
levy were to be in place continuously, with the rate set by
(b). New Zealanders know more about NZD than others,
a mechanical rule. It could be set at a level to ensure that
and are therefore more likely to be stable position-
the actual borrowing rates further out on the yield curve
holders or (better still) counter-cyclical arbitrageurs.
fully reflect the current intention of policy. When the RBNZ
This might include a readiness on the part of the Government
foresees that it will maintain the OCR at current rates for
to switch currency exposure of its own debt.
longer than the market has built into the yield curve (as is
the case at present), the levy could be set to correct what
could legitimately be seen as a market imperfection. If
Instruments that might restore the potency of
interest rates in impinging more directly on
interest-sensitive sectors
The weakness of monetary policy stems from the ability
of mortgage borrowers to shield themselves from OCR
this is too complex, it could be set at a rate based on the
slope of the yield curve, with the tax imposed whenever
the yield curve is negative, at a rate which would ensure
that borrowing costs further out on the yield curve rise at
around the same rate as the OCR.
increases by going further out on the yield curve. So there
A levy seems the simplest approach, but if there are
are attractions to the mortgage interest levy explored in
legislative problems with this, it might be more feasible to
the Supplementary Stabilisation Instruments (SSI) Paper
use a mandatory deposit requirement, as used by a number
(section 3.6, RBNZ and New Zealand Treasury (2006). It is
of countries (Australia in the 1970s, Chile in the 1980-90s),
described as “a wedge...established between the interest
which has the same effect of putting a wedge between
rate paid by domestic borrowers and those available to
the foreign and domestic interest rates. The difficulties of
foreign savers”. In the SSI Paper, this is presented as a way
this (e.g. disintermediation) are explored in the SSI Paper.
of achieving the same degree of policy restraint with less
The core of the suggestion made here, compared with the
34
effect on the exchange-rate, but it might also be presented
as a method of restoring some of the power of monetary
34
policy to affect the housing cycle. Perhaps the main
32
33
West (2003) suggests that large movements of the interest rate
are required to shift the exchange-rate appreciably.
There seems to be a mind-set that prefers NZD-denominated
foreign borrowing – a pernicious influence of the “original
sin” argument.
Testing stabilisation policy limits in a small open economy
It is presented in the SSI Paper as impinging on both
borrowers and savers, but it seems more likely that foreign
borrowing rates are set internationally, and the effect would be
predominantly on domestic borrowers. The SSI Paper holds
out the possibility that it could be effective without pushing
up borrowing rates much, but unless rates go up enough to
reduce borrowing, the capital inflow will be much the same,
and the pressure on the exchange-rate much to same.
25
SSI proposal, is to replace the discretionary element with
The low-probability scenario envisaged here is different
a rule.
– where non-residents are not prepared to supply the
NZD-denominated leg of the funding in sufficient quantity
7
What might be done with
to maintain the outstanding loans and at the same time
prudential policy?
to provide a minimal expansion of intermediation. To the
There are two constraints if prudential policy is to be used
to help macro-issues. First, the Act requires that measures
are primarily for prudential objectives35 – but this is broad
enough to include “promoting the maintenance of a sound
and efficient financial system.” Secondly, administrative
difficulties arise if prudential rules change over the
course of the cycle in a discretionary way. The additional
practical current issue is that banks are grappling with the
complexities of Basle II and will strongly resist any additional
measures.
That said, two points might be made:
•
do things they would not otherwise do, this needs to be
justified in terms of some distortion or some externality,
and the externality in this case is that the banks do not
suffer the full effects of a sharp cyclical reduction in their
intermediation role – extra costs are borne by the actual and
potential borrowers. Without specific contingency plans in
place, the RBNZ prudential authorities would be justified
in limiting the extent of the foreign funding and taking
measures to “lop the top” of the housing boom driving the
foreign borrowing.
This could be done in a variety of ways. Policy might focus
There is a growing awareness in the international debate
on the funding side by imposing direct limits on each bank’s
of the importance of “macro-financial” issues (see Borio
access to foreign funding, or might make this funding
and White (2004), and Borio and Lowe (2002), and
more expensive through taxes, or by additional capital (or
the papers by White cited in the bibliography). There
liquidity) requirements against this more volatile form of
is little doubt that excessive growth in housing credit
funding. Or policy might focus on the mortgage lending
has significantly exacerbated the boom in a number of
volume, imposing additional capital requirements or a
countries (see BIS Papers No 21 April 2005), and this
Loan/Valuation Ratio (LVR) against this type of lending.
problem falls uncomfortably between the objectives
of monetary policy (CPI price stability) and prudential
policy (ensuring systemic financial stability).
•
extent that prudential requirements cause the banks to
Two housing-lending-specific measures are explored in the
SSI Paper. The first suggestion is to link bank capital more
closely with risk factors, specifically acknowledging that
There is an inherent pro-cyclicality in the financial
these risks rise with an asset bubble. The second possibility
sector: the greatest prudential risks are incurred in the
explored in the SSI Paper is to make discretionary changes
mature phase of the upswing.
in a comprehensive LVR. The SSI Paper acknowledges that
The issue is to find a way of addressing these concerns with
discretionary measures are likely to be tardy and inadequate.
a prudential instrument. The key here might be to achieve a
Even a fixed (non-discretionary) LVR would, however, help
general acceptance that the heavy dependence on foreign
as it is only in the last phase of the cyclical expansion that
funding presents a clear problem for the banks and their
LVRs tend to push into new territory36. In any case, such
customers. The scenarios run in the course of the FSAP (IMF
measures can be made both variable and non-discretionary,
2004) took a different tack. They explored the possibility
based on a mechanical calculation (similarly to the approach
that maintaining this funding might require a substantial
to the mortgage levy, above). What might be the basis of
rise in borrowing costs, which might put pressure on
such a mechanical rule? If the measures are directed at the
borrowers. The main concern in the FSAP was the health
mortgage lending side of banks’ balance sheets, the rule
of the banks, rather than any macro-financial concerns.
should reflect the abnormal rise in the asset price which is
35
26
This may be reflected also in the philosophical approach
of many policy makers, who see advantage in keeping the
elements of policy making separated and “pure.”
securing the loan (as this is the element which is most at risk
if asset prices show mean reversion over the course of the
Reserve Bank of New Zealand and The Treasury
cycle). Goodhart (2005) suggests that capital requirements
debt experience would indicate.38 Whatever the appropriate
should be based on the rise in housing and property prices,
risk discounts on average over the cycle, they should be
compared with the underlying inflation. Analogous logic
significantly higher in the mature phase of the boom. A
would suggest that LVR requirements should be based on
modest further step would be to ask each of the banks to
an average price level of the asset over, say, the previous five
set out its contingency plan in the event that the FX cover
years, rather than the current valuation.
they need for their foreign funding is not available.
The case against housing-specific measures is that, so far at
Consideration might be given to two further measures on
least, the bad-debt experience is not adverse. One counter-
the funding side. First, to impose a limit on the share of
argument is that investment-housing has become more
foreign funding in any financial institution’s balance sheet
important, and that the repayment experience on this is
(on the grounds that this is potentially more volatile),
yet to be tested. But the more fundamental point is that
together with a liquidity ratio that would require institutions
this is too narrow a view of prudential risk. The measures
to keep a modest percentage of the borrowed funds
discussed here are macro-prudential, designed to ensure
in a liquid form. Second, to impose a more substantial
that the banks remain efficient financial intermediaries
withholding tax on these funds. At present such funds pay
throughout the cycle, and in the face of low-probability
only 2 per cent (around 10 b.p.). Current arrangements
events such as a “sudden stop” of foreign funding sources.
seem to give foreign providers of funds a substantial tax
Constraining their activities at the height of the asset cycle
advantage compared with New Zealand sources of funds
will improve their ability to perform their role over the full
(e.g. depositors), who presumably bear the normal tax
course of the cycle.
rates. It goes without saying that the timing of introduction
37
A modest first step here might be to collect much more
of measures like this needs careful consideration.
comprehensive data (there are, for example, no data which
separately identify investment housing loans), including
details of LVRs and debt-service/income ratios, and give
8
Other possibilities
these data widespread and critical public coverage and
Greater integration (with Australia?)
commentary. The process of collection, in itself, is a signal
Note the greater convergence of AUD and NZD (Figure
to the banks of official concern. A second modest step
12). In some ways it is surprising that interest rates and
would be to require insurance for loans with LVR above
the exchange-rate show the variation that they do, in two
80% (which is encouraged in Australia by the application
economies so well integrated internationally (including,
of a higher capital requirement if this insurance is not in
unusually by international standards, a high degree of
place). Given that bank lending margins have fallen during
labour mobility). New Zealand pays a significant premium
the upswing of the housing cycle, it may be helpful to
for borrowing overseas in its own currency. There is
undertake a vigorous dialogue with the banks on whether
an interesting contrast with, say, Switzerland which
their dynamic provisioning over the course of the cycle is, in
similarly retains full sovereignty and its own currency, but
fact, reflecting the indisputable fact that loans given in the
whose interest rates are closely linked to the European
mature stage of the cycle are more risky than average bad-
Community, as is the exchange-rate. While only 20 per cent
of New Zealand imports come from Australia, the degree
36
37
For a description of the Australian experience, see “Recent
Developments in Low Deposit Loans” RBA Bulletin October
2003, which indicates that 100% LVRs are three times as
likely to default as 80% LVRs, and that 2% of high LVR loans
are “past due” compared with 0.7 % of all securities loans.
High LVR loans are, however, a small proportion of total
housing loans – only 2% have LVRs of 95-100% and 1⁄2%
are over 100%.
International experience suggests that some borrowers do
walk away from mortgages, although this proportion is small.
Testing stabilisation policy limits in a small open economy
38
BIS research suggests that the loan-provisioning experience is
concentrated in the downswing of the cycle, whereas it might
be more logically concentrated in the upswing when the loans
are taken on in overly-optimistic circumstances. See Table 1
of “The importance of property markets for monetary policy
and financial stability” Haibin Zhu BIS Papers 21 April 2005.
See also T. Helbling “Housing Price Bubbles” BIS Papers 21
April 2005 .
27
of integration in trade and finance is very high (85% of
how dependent it is on reputation and psychology. Price
New Zealand banking is done in Australian subsidiaries/
expectations have to be kept stable without leaning too
branches). Greater integration would provide more
hard on activity. The aim is to avoid having to do what
protection against the sort of “sudden-stop” capital-flow
Volcker did in the US in 1979-82, necessary though that was
changes discussed above, as Australian investors (with their
in the circumstances. How much more difficult for a small
greater knowledge of New Zealand) might be readier to fill
internationally-integrated economy, where the ability to
the gap without the exchange-rate moving so far.
influence the central bank’s usual instrument – the interest
That said, full integration seems hard to achieve in practice.
Even in the banking area, with substantial commonality of
institutions and common laws, the regulatory approach
(particularly in relation to crisis resolution) seems
fundamentally different. Similarly, differences in taxation
treatment seem, unfortunately, to stand firmly in the way of
close integration. Governments on both sides of the Tasman
put high priority on a continuation of the integration process,
but seamless integration still seems a long way off.
rate - is greatly circumscribed by foreign capital flows. It goes
without saying that central banks need to go on maintaining
the primacy of the price stability objective. But they need to
go beyond this, with the aim of pre-emptively minimising
imbalances which sooner or later need a period of painful
slow growth to resolve. Some measures are within RBNZ’s
current mandate (more active FX intervention). Some are in
the RBNZ remit, but require some widening of the thinking
about prudential policy, so that it covers macro-financial
issues. Other ideas explored here would require a high level
of inter-institutional co-operation, particularly with the
9
Conclusion
Treasury. This forum provides an opportunity to take this
forward.
New Zealand’s admirably high level of economic debate,
its vigorous and open institutions, and the quality of the
bureaucracy ensure that foreigners will be hard-pressed to
make much of a contribution to the policy debate. Policy
has, of course, evolved and will continue to do so. The
tightly specified approach to policy, with each element
given a narrowly-defined objective, served well for a time,
and since then the sharp edges have been softened. What
an outsider might be able to offer is some encouragement
to complete this unfinished process, to achieve a pragmatic
mix in which the arms of policy reinforce each other to tackle
problems which fall outside a narrowly-defined task or fall
uncomfortably between the remits of the different arms
of policy. So monetary policy looks beyond price stability,
prudential policy looks beyond the narrow task of keeping
the banks solvent, fiscal policy looks beyond the immediate
needs of the government finances, and each “arm” of policy
asks what it might do to address issues such as the swings
of the housing cycle, the fluctuating price signal from the
exchange-rate, and the adjustment issues presented by the
current account deficit.
In defining their role, central banks need to assert how
subtle (some might say feeble) is their instrument and
28
Reserve Bank of New Zealand and The Treasury
Bibliography
Chen, Yu-chin and Kenneth Rogoff (2003), “Commodity
Allsopp, C and D Vines (2005), “The Macroeconomic Role
Currencies,” Journal of International Economics 60, 133-
of Fiscal Policy,” Oxford Review of Economic Policy Vol. 21
160.
No 4.
Cotis, J-P and J Coppel (2005), “Business cycle dynamics in
Becker, C and M Sinclair (2004), “Profitability of Reserve
OECD countries: evidence, causes and policy implications,”
Bank Foreign Exchange Operations: Twenty Years after the
RBA Conference Proceedings.
Float,” RBA RDP 2004/6.
Drage, D, A Munro and C Sleeman (2005), “An Update on
Blinder, A and J Morgan (2000), “Are Two Heads Better
Eurokiwi and Uridashi Bonds,” RBNZ Bulletin September.
than One?” NBER Working Paper 7909.
Eichengreen, B and R Hausmann (1999), “Exchange-rates
Blundell-Wignall, A, J Fahrer and A Heath (1993), “Major
and Financial Fragility,” Federal Reserve of Kansas City,
Influences on the Australian Dollar Exchange-rate” in The
Proceedings of a Conference.
Exchange-rate, International Trade and the Balance of
Eckhold, K and C Hunt (2005), “The Reserve Bank’s New
Payments RBA Conference Proceedings.
Foreign Exchange Intervention Policy,” RBNZ Bulletin
Borio, C and P Lowe, (2002), “Asset prices, Financial and
March.
Monetary Stability: Exploring the Nexus,” BIS Working
Goodhart, C (2005), “Financial regulation, Credit Risk and
Paper 114.
Financial Stability,” National Institute Economic Review
Borio, C E V, C Furfine and P Lowe (2001), “Pro-cyclicality of
192:118-127.
the financial system and financial stability: issues and policy
Goh, K (2005), ”Savings and the Household Balance Sheet,”
options,” BIS Papers No 1, March, pp 1-57.
RBNZ Bulletin June.
Borio, C E V, W White (2004), “Whither monetary and
Gordon,
financial stability? the implications of evolving policy
Management,” RBNZ Bulletin March.
regimes,” BIS Working Papers No 147, February.
M
(2005),
“Foreign
Reserves
for
Crisis
Gruen, D and T Kortian (1996), “Why does the Australian
Bollard, A (2005), “Imbalances in the New Zealand
Dollar move so closely with the terms of trade?” RBA
Economy,” speech on 14 October RBNZ Bulletin December.
Research Discussion Paper 96/1.
Bollard, A (2004a), “Asset Prices and Monetary Policy,”
Hampton, I (2001), “How much do import price shocks
speech on 30 January RBNZ Bulletin March.
matter for consumer prices?” RBNZ Discussion Paper
Bollard, A (2004b), “What’s Happening to the Property
2001/06 http://www.rbnz.govt.nz/research/discusspapers/
Sector?” speech on 2 September RBNZ Bulletin September.
dp01_06.pdf.
Bollard, A and O Karagedikli (2006), “Inflation Targeting:
Hebling, T (2005), “Housing price bubbles – a tale based on
The New Zealand Experience and Some Lessons,” paper
housing booms and busts,” BIS Papers 21.
presented at a conference at the Central Bank of Turkey
Hodgetts, B (2006), “Changes in the Inflation Process in
19-20 January.
New Zealand,” RBNZ Bulletin, March.
Briggs, P (2004), “Currency Hedging by Exporters and
Huang, A (2002), “Is Monetary Policy in New Zealand
Importers,” RBNZ Bulletin December.
Similar to that in Australia and the United States?” RBNZ
Buckle, R, D Haugh and P Thomson (2001), Calm after the
Briefing Note September.
Storm?: Supply-side contributions to New Zealand’s GDP
International Monetary Fund (2004), Financial System
volatility decline, Treasury Working Paper 01/33.
Stability Assessment New Zealand. May 2004.
Testing stabilisation policy limits in a small open economy
29
King, M, (2002), Speech to the Business Economists 22
White, W R (2002), “Changing views on how best to
May.
conduct monetary policy,” BIS Speeches, 18 October.
Munro, A (2005), “UIP, Expectations and the Kiwi,” RBNZ
— (2004a), “Are changes in financial structure extending
Discussion Paper 2005/5, October.
safety nets?,” BIS Working Papers No 145, Basel, January.
Munro, A (2004), “What Drives the New Zealand Dollar?”
— (2004b), “Making macroprudential
RBNZ Bulletin June.
erational,”
Svensson, L (2001), “Independent Review of the Operation
of Monetary Policy,” see RBNZ website.
in
Proceedings
of
the
concerns opsymposium
on
financial stability policy – challenges in the Asian era, De
Nederlandsche Bank, Amsterdam, 25-26 October, pp 4762.
Toplis, S (2006), “The Dream Rebalance?” Bank of
New Zealand report April.
— (2005a), “Procyclicality in the financial system: do we
need a new macrofinancial stabilisation framework?,” Kiel
Reserve Bank of Australia (2005), “The Changing Nature of
Economic Policy Papers, 2, September.
the Business Cycle,” proceedings of a conference July.
— (2006), “Is Price Stability Enough?” BIS Working Papers
RBNZ “Monetary Policy Statement” (various issues).
RBNZ “Financial Stability Report” (various issues).
RBNZ “The Evolution of Monetary Policy Implementation.”
205, Basel, April.
Zettlemeyer (2000), “The Impact of Monetary Policy on the
Exchange-rate,” IMF Working Paper WP/00/141.
RBNZ and New Zealand Treasury (2006), “Supplementary
Stabilisation Instruments,” Initial Report to Governor,
Reserve Bank of New Zealand and Secretary to the Treasury
10 February.
West, K (2003), “Monetary Policy and the Volatility of Real
Exchange-rates in New Zealand,” RBNZ Discussion Paper
2003/9 November.
30
Reserve Bank of New Zealand and The Treasury
Macroeconomic policy challenges: monetary policy
by Stephen Grenville
Discussion by Christopher Allsopp, University of Oxford
1
Introduction
the trilemma (or “impossible trinity”) between monetary
The paper by Stephen Grenville admirably sets the scene for
policy, a free exchange-rate and free capital movements,
this policy forum.
that a choice has been made to emphasise the first and third
New Zealand’s inflation targeting approach is widely
regarded as an example of international “best practice”.
– and that the exchange-rate consequences should just be
accepted.
Fiscal policy is under control with budget surpluses and
public debt approaching zero. New Zealand’s economic
performance has been impressive, in many ways, exemplary.
2
Growth has been relatively rapid and inflation, though rising
The issues, many of which are discussed by Stephen
a bit at present, has been well controlled.
Grenville, seem to me to be of two types.
But problems have emerged in a number of areas, raising the
First, there is the question as to whether the inflation
concerns which underlie this policy forum. In particular:
targeting system could usefully be improved or modified
•
The exchange-rate has fluctuated over the cycle to an
extent that many regard as excessive.
•
The external current account deficit is large at the
moment, and its average level over the last decade is,
as Stephen Grenville notes, an “outlier, both historically
and internationally”.
•
New Zealand, like many other countries, faces the
problem of what, within an inflation targeting system,
to do about asset prices, particularly, housing and real
estate prices.
•
The developing “imbalances” raise issues not just for
monetary policy, but also for longer term structural
policy and for macro prudential policy.
One response to New Zealand’s experience is simply to
The issues
to take into account the developing characteristics of the
New Zealand economy.
Second there is the question of whether overall performance
could be improved by other policies (e.g. fiscal policy or
specific measures to affect the housing sector). Within this
second group would be measures to control or constrain
some of the macro financial risks discussed by Stephen
Grenville.
Broadly speaking, a consensus view would be that the (best
practice) inflation targeting system should be maintained (it
could be modified perhaps in certain respects) and that the
focus of discussion should be on other supporting policies.
If, however, other policies are ruled out, on grounds of
feasibility, cost, or ideology, then really difficult questions
about the monetary policy framework could well arise.
accept it. There are limits to how much stabilisation can be
achieved. Some shocks are “real”, and cannot or should not
be cancelled by monetary policy. Arguments here include
the view that the commodity sector fluctuates anyway, as
3
Monetary and other policies
well as arguments based on characteristics of New Zealand,
in the “New Consensus
such as a high degree of openness (particularly with regards
Assignment”
capital flows if not trade) or the impact of immigration
What is best practice inflation targeting? A common view
and capital market integration on the domestic housing
is that it involves delegation of medium term control over
sector. There may be implications for policy – but perhaps
inflation to an independent central bank using interest
not for monetary policy. Or it might be argued that given
rates as its instrument to influence inflationary pressure.
Testing stabilisation policy limits in a small open economy
31
In practice, the interest rate policy reaction function is
both inflation control and (subject to that) stabilisation
embedded in an institution, operating with ‘constrained
to a single institution. Usually the institution has a single
discretion’. Legal mandates, such as those of the Bank of
instrument of policy (normally the short term policy rate,
England and the ECB are usually hierarchical or lexicographic,
such as the OCR in New Zealand). This is a complex task,
with the primary objective being price stability and subject
involving commitment and expertise, and the institution
to that a responsibility for stabilisation.
needs to be credible and trusted. This assignment implies
In practice, the system is often described as flexible inflation
targeting, or as inflation forecast targeting. At its simplest,
that other arms of policy do not have that responsibility for
inflation control and for stabilisation.
the system can be represented with just three equations;
Likewise, the inflation forecast targeting authority does
an IS relationship, a Phillips type relationship and the policy
not have responsibility for other important economic
reaction function. The reaction function needs to have certain
variables, such as the current account deficit, household
properties - such as the Taylor principle. Variants abound,
indebtedness, or, for a purist, financial stability. Of course,
ranging through simple rules, such as the Taylor Rule, via set-
central banks may have multiple responsibilities – but the
ups involving assumed loss functions to fully micro-founded
inflation targeting arm needs to be effectively separate,
forward-looking approaches.1 The connection between
even though it may/should take into account what the other
theory and practice is close enough to be highly productive
arm is doing.
– but is not exact. Given the importance of expectations it is
widely argued that central bank behaviour should be “rule
like”, in the sense of Taylor, (1993) and “predictable” in the
Role of the monetary authority
sense of Woodford, (2003).
The role of the monetary authority is to react to shocks
I take all this as well known. I want to make a
few points.
(such as exchange-rate changes, commodity trade impacts,
house price rises or falls, etc) in such a way as to make
sure that inflation homes in on the target in the medium
First, there is nothing in the theoretical structure that
term, whilst minimising, as far as possible, fluctuations in
demands that the instrument of policy should be the
output and prices. Policy shocks, such as fiscal impacts,
policy interest rate or that the agent should be the central
or the effects of intervention on (say) the housing market,
bank. Both of these are pragmatic choices (sensible ones,
also have effects on output and on inflation prospects, and
of course). In principle, other control instruments could be
are taken into account, along with other influences, by the
used within a similar framework, e.g. a fiscal instrument,
monetary authority.
or multiple instruments. In principle, other institutions
could be assigned responsibility for the complex tradeoffs in achieving inflation control and as much stabilisation
as is compatible with it. To illustrate, Singapore uses the
exchange-rate as instrument within an inflation forecast
targeting framework; fiscal policy committees have been
proposed for countries in EMU (Wyplosz 2002).
So long as policy shocks and their effects are anticipated,
they are “internalised” in the monetary policy reaction
function. Thus, in principle, if the system were functioning
perfectly, fiscal policy should have no effect on inflation
or on output gaps. The same is true of other policy
interventions or shocks (e.g. specific measures to affect the
real estate market).
The most important innovation, and, it may be argued,
the one that accounts for the success and popularity of
inflation targeting, is the assignment of responsibility for
1
32
The Reserve Bank of New Zealand’s main macro model,
used to prepare the published projections, embodies such a
reaction function (see Black et al. 1997).
Reserve Bank of New Zealand and The Treasury
Changes in the transmission mechanisms, for example,
inflation and output gaps, which are the responsibility of
towards greater effects of interest rates on the exchange-rate
the monetary authorities.
and hence onto the economy, should also be ‘internalised’,
within the institutionally embedded reaction function,
leaving inflation and output gaps effectively unchanged.2
In fact, however, under this assignment fiscal policy does
matter (short of Ricardian equivalence). In particular, under
this set up, it is the fiscal authorities, by their actions,
that have a major influence over the longer-term natural
The role of other policies (especially fiscal
policy): coordination
In turn, under this assignment, the monetary authority
constrains or influences the actions of the fiscal authorities.
Fiscal policy action is taken in the knowledge of the likely
reaction of the monetary authority. For example, in principle
(or neutral) rate of interest and the exchange-rate. For
large shocks, a fiscal instrument may be preferred as less
costly than an interest rate response. Moreover, specific
instruments, close to the problem or a diagnosed market
failure, may be preferred to more general policies, whether
fiscal or monetary.
this would remove any temptation for a fiscal boost before
The consensus assignment leaves a lot of room for other
an election, since it would be cancelled and just lead to
policies – for good or ill. It is true that fiscal policy “does
higher interest rates.
not matter very much” for inflation and output gaps. But
The UK Treasury describes this interaction as analogous to
a “Stackleberg” game, with the fiscal authority as leader.
(See also Bean, 1998). It is claimed that coordination under
it does matter for the exchange-rate, for imbalances (such
as the current account), and for the course of interest rates
over the cycle.
this assignment works very well H M Treasury (2002). (In
The view that the active use of instruments other than
the UK a non-voting Treasury representative attends MPC
monetary policy to help to manage the economy would
meetings to ensure that information about fiscal actions is
damage the credibility of the system is generally false. For
known).3
example, if a housing boom were controlled by specific
The assignment of the main traditional functions of
macroeconomic policy to central banks (and the interest
rate policy reaction function) has led to the view, often
characterised as the ‘new consensus macroeconomics’
(NCM) that the fiscal authorities are (a) free to concentrate
measures, monetary policy would have less to do, and
probably would be more rather than less credible. What
is true is that the other arms of policy should not try to
usurp the (limited) role assigned to the monetary (or, more
generally, the inflation targeting) authority.
on sustainability and ‘good housekeeping’ (as well as the
microeconomic and distributional aspects of public finance)
and (b) that they should do so. The perceived role of fiscal
4
Should monetary policy
policy in New Zealand appears to fit well with this paradigm.
attempt to deal with
Under this assignment, fiscal policy appears to have little
imbalances, excessive
macroeconomic role. This view of the unimportance of
fiscal policy is true if attention is focused only on aggregate
fluctuations, housing booms
etc?
2
3
This demonstrates that the central bank should not be
committed to a particular rule, but to a set of procedures
connecting the instruments of policy with the complex
objectives. It is a moot point as to whether a change in the
transmission mechanism should be described as leaving the
reaction function unchanged or as changing the reaction
function.
It is of course necessary that the monetary authorities and the
fiscal authorities should not try do the same job (with different
preferences). In practice, it is helpful that monetary policy is
relatively high frequency.
Testing stabilisation policy limits in a small open economy
The conventional answer is no. Monetary policy should not
be diverted from its inflation targeting role. If other policies
can moderate the shock structure, that is normally first
best. If well designed, these other policies should enhance
the credibility of monetary policy, rather than threaten it.
Longer term imbalances should be dealt with elsewhere,
33
e.g. by fiscal policy or by more structural changes designed
exchange-rate are regarded as costly, then a policy (by
to change private sector behaviour.
other agencies) of offsetting shocks by other policy action,
The difficulties arise if there are no other feasible policies,
perhaps because they are regarded as too costly, or because
of ideological concern with “good housekeeping”. Then
there are real difficulties about whether the monetary policy
such as fiscal policy, could be beneficial. If the central bank
were assigned a fiscal instrument as well, as has been
suggested by Wren-Lewis for the UK, then preferences
would determine its use.4
system should be changed to deal with other problems as
well. There is a presumption that it should not be.
(b) Are the tradable and non-tradable sectors
In the absence of alternatives, there may, however, be a
excessively disrupted ?
cost-benefit case for diverting monetary policy away from
This would be a reason for seeking to alter the mix away
inflation targeting in particular circumstances, such as
from interest rates and exchange-rates.
potential situations of boom and bust in asset prices, the
build up of financial risk, etc.
(c) Should policy aim to operate more directly
If the asset price problem is thought of in terms of negatively
on the source of the shock?
serially correlated shocks (a boom followed by bust) then
In principle yes – so long as the instrument used is itself not
it is tempting to try to moderate the upturn in order to
too distortionary and costly. The principle behind this kind
moderate the downturn. The purpose of the diversion is
of recommendation is, of course, to spot the market failure
to change the expected shock structure (e.g. by pricking
and correct it. Identifying the market failure is often not easy.
the bubble). The cost is the potential loss to the credibility,
It is not the responsibility of the inflation targeting central
transparency and predictability of monetary policy. This is
bank to do this. Note that it may be the responsibility of
a potentially serious cost, since credibility and reputation
some other arm of the central bank, such as the department
are often hard won. The consensus is right – most of the
concerned with financial stability. If this is the case, then it
time. The central bank should concentrate on meeting
needs to be made publicly very clear that monetary policy is
its mandate to control inflation in the medium term, and
not being diverted from its primary responsibilities.
not involve itself in pursuing other objectives. (In extreme
circumstances, with very large movements in say the
exchange-rate or in real estate prices, no central bank could
stand idly by if diversion of monetary policy from its primary
aims were the only option).
(d) Is the CAD too big?
If it is, this is not a job for the monetary authorities. What is
the “failure” that justifies policy concern? The “consenting
adults” point made by Stephen Grenville is rather powerful.
The obvious candidate policy is action by the fiscal authorities.
5
The possible problems
Stephen Grenville lists five potential problems or issues.
It is perhaps useful to run through them in terms of the
assignment issue.
An enlarged public sector surplus (or reduced deficit) should
result in lower interest rates, a lower exchange-rate and a
lower current account deficit, at least over a medium-term
horizon. However, in the case of New Zealand, the public
sector already runs significant surpluses, and there are
limits on how much further this could be pushed. Instead,
(a) Does the exchange-rate move too much?
policies to change the private sector’s savings-investment
From the point of view of the monetary policy system, the
balance (for example, by increasing the incentive to save,
answer is surely “no”. The exchange-rate is part of the
or by changes in pension arrangements) may be the most
transmission mechanism to be taken into account by the
4
central bank. However, if the implied movements in the
34
The political likelihood of this suggestion being adopted does
not, at present, appear great.
Reserve Bank of New Zealand and The Treasury
appropriate response. These policies may well be needed,
the system. Not least, since it is impossible to meet a
but under the inflation targeting framework are not the
point target over time, the idea of the authorities using
responsibility of the central bank.
their constrained discretion to home in on a particular
inflation target is readily communicated.
(e) Vulnerability of the financial sector?
(2) There appears to be a danger in New Zealand that
monetary policy will try to do too many things.
If so, the same principles apply. Other policies, including
This is key. The monetary authority should not take
regulation, may be beneficial. The difficult cases arise when
responsibility for economic outcomes which are beyond
the other policies are not available. Diversion of interest
its remit – and beyond its competence. This could only
rate policy to regulatory concerns would normally be highly
lead to a loss of credibility and trust in the system.
damaging to the credibility and transparency of the inflation
target system, although it has been practiced in the US. The
(3) In principle, the monetary authorities might want to use
fact that such diversion would be damaging strengthens
another instrument, such as exchange-rate intervention,
rather than weakens the case for better regulation.
within their limited mandate. (As in New Zealand,
the MPC in the UK has some limited power to use
intervention, although it has never been used). Some
6
scepticism is needed as to whether it would work.
Ways forward
An impression formed from the background documentation
(4) Housing market intervention is problematic but it is easy
as well as Stephen’s paper is that whilst many aspects of
to conceive of situations where it would be needed. It
the policy assignment are clear in the case of New Zealand,
would seem better if this were not the responsibility of
there is considerable doubt and disagreement over which
the monetary authority.
institution should be responsible for dealing with the
(5) Prudential policy is probably needed anyway.
problems that have emerged. With the monetary authority
concentrating on inflation and (subject to that) aggregate
stabilisation and the fiscal authority concentrating on fiscal
7
Concluding remarks
rectitude and control over the public finances, there is not
This is a very useful paper. It is my job to disagree. The
much room for manoeuvre. This is especially true if the
argument ends up with a plea for a pragmatic mix of
problems (if they are problems) emanate from the private
policies, reinforcing each other. My view is different.
sector.
Monetary policy should be left alone. If anything, the
credibility of the inflation targeting system needs to be
strengthened, and made more, rather than less, separate
Monetary policy
from other policies. Prudential policy is needed anyway.
A few points only.
Fiscal policy and policies to affect private behaviour (e.g.
(1) The paper discusses the inflation-target band. This
is superficially similar to the UK system. But the
interpretation is different in an important way. In the
UK the there are perhaps more formal mechanisms for
allowing discretion, for example, in face of commodity
price shocks.
savings behaviour) have important roles to play. A focus
exclusively on public sector control and responsibility,
laudable as it is, is not enough. An overall macroeconomic
view is needed, taking into account the role assigned to the
monetary authority and characteristics of the private sector.
Where does the responsibility for overall policy lie? Surely
with The Treasury.
There are major advantages to the UK’s point target
system. It is clear what the target is. It is symmetrical,
which is important in gaining public acceptance of
Testing stabilisation policy limits in a small open economy
35
References
Taylor (1993), “Discretion Versus Policy Rules in Practice,”
Bean, C (1998), Macroeconomic Policy under EMU Oxford
Carnegie-Rochester Conference series on Public Policy,
Review of Economic Policy Vol 14 No 3 Autumn. pp41-53.
39(1), 195-214.
Black, Richard, Vincenzo Cassino, Aaron Drew, Eric Hansen,
Woodford, M (2003), “Interest and Prices: Foundations of a
Ben Hunt, David Rose and Alisdair Scott, ‘The Forecasting
theory of Monetary Policy”, Princeton University Press.
and Policy System: the core model’, Reserve Bank of New
Wren-Lewis, S (2000), “The Limits to Discretionary Fiscal
Zealand Research Paper No. 43, August 1997.
Policy.” Oxford Review of Economic Policy Vol. 16 No.4
H M Treasury (2002). Reforming Britain’s Economic and
Winter. pp 92-105.
Financial Policy. (Eds. Balls, E and O’Donnell, G.) Palgrave
Wypolosz C (2002), “Fiscal Policy: Institutions versus Rules.”
2002.
CEPR Discussion Paper No. 3238.
36
Reserve Bank of New Zealand and The Treasury
Stabilisation policy in New Zealand:
Counting your blessings, one by one†
Willem H. Buiter, Professor of European Political Economy, European Institute, London School of Economics and
Political Science††
1
Introduction
Responsibility Act 1994 (FRA). Together they provide a
This paper is written in response to an invitation from the
framework which, if adhered to consistently, now and in
Reserve Bank of New Zealand (RBNZ) and the New Zealand
the future, will ensure fiscal sustainability.
Treasury, to take part in a “project to investigate policy
When in 1982, during a spell as a visitor in the Fiscal Affairs
options for a smoother evolution of the New Zealand
Department of the IMF, I wrote a paper (Buiter (1983), see
business cycle, particularly in respect to swings in the
also Buiter (1985) and Gray, Merton and Bodie (2003))
exchange-rate and shifting pressures on the externally
advocating the collection of data that would permit the
exposed sectors of the economy.” My contribution consists
construction of a comprehensive, ‘marked to market’
of a paper that focuses on the role of fiscal policy, but is also
balance sheet for the consolidated public sector (general
expected to consider the overall macro-mix of monetary,
government, central bank and all other agencies for which
fiscal, external and structural policies.
the general government is ultimately financially responsible),
New Zealand has pioneered more important changes in the
including a full accounting for deferred contingent liabilities
design of monetary and fiscal policy and in the institutional
and for the uncertain future cash flows associated with
arrangements through which monetary and fiscal policy
state assets, I did not anticipate that any country would
are designed and implemented, than any country since
ever contemplate actually implementing anything remotely
(at least) the second half of the 20th century. The modern
like it. New Zealand has progressed further, indeed as far
independent central bank was born in New Zealand in
as data availability permits, on the road to the construction
19891,2. Inflation targeting was invented in New Zealand
of a comprehensive balance sheet for the public sector,
with the Reserve Bank of New Zealand Act of March 1989
including the measurement of the comprehensive net worth
and the first Policy Targets Agreement (PTA) in March
of the state than any other country.
1990.
At least as important as these two widely-acclaimed
institutional innovations has been the medium and long-
1
term fiscal-financial sustainability framework developed
and adopted by New Zealand through the Public Finance
Act of 1989, as amended by the Public Finance (State
Sector Management) Bill of 2004, and especially the Fiscal
†
††
Paper written for and presented at the ‘Macroeconomic Policy
Forum’ on June 12, 2006 in Wellington, New Zealand. I would
like to thank my discussant, Pierre Siklos, the participants
at the Forum and the many knowledgeable observers of the
New Zealand economy who freely provided me with facts and
insights. These include, but are not limited to, Chris Allsopp,
Mark Blackmore, William Cline, Steve Dunaway, Brian
Easton, John Edwards, Sebastian Edwards, Peter Harris, John
Janssen, John McDermott, Stephen Grenville, Val Koromzay,
Roger Procter, Klaus Schmidt-Hebbel, Ulf Schoefisch, Graham
Scott, Grant Spencer and Bryce Wilkinson. Bob Buckle and
Aaron Drew provided comprehensive and detailed comments
on an earlier draft of the paper. None of the aforementioned
bear any responsibility for the contents of this paper.
Houghton Street, London WC2A 2AE, UK.
Tel. + 44 (0)20 7955 6959; Fax + 44 (0)20 7955 7546;
Email: [email protected] ; web: http://www.nber.org/~wbuiter
Testing stabilisation policy limits in a small open economy
2
A case can be made for the Deutsche Bundesbank, established
in 1957 as the sole successor to the two-tier central bank
system which comprised the Bank deutscher Länder and the
Land Central Banks of the Federal Republic of Germany, as
the first modern independent central bank.
The Reserve Bank of New Zealand Act 1989 specifies that
the primary function of the Reserve Bank shall be to deliver
“stability in the general level of prices.” The Act says that the
Minister of Finance and the Governor of the Reserve Bank
shall together have a separate agreement setting out specific
targets for achieving and maintaining price stability. This is
known as the Policy Targets Agreement (PTA). A new PTA
must be negotiated every time a Governor is appointed or
re-appointed, but it does not have to be renegotiated when a
new Minister of Finance is appointed. The current PTA, is the
seventh since the passage of the 1989 Act. Finance Minister
Michael Cullen and Dr Alan Bollard signed the current
PTA on 17 September 2002, a week before Dr Bollard took
up his role as Governor. The Act requires that the PTA sets
out specific price stability targets and that the agreement, or
any changes to it, must be made public. The PTA can only
be changed by agreement between the Governor and the
Minister of Finance. Thus, neither side can impose unilateral
changes. Note, however, that under the Reserve Bank Act the
Government has the power to override the PTA. This override
power, akin to the UK Treasury’s Treasury Reserve Powers,
has not been invoked thus far.
37
The outline of the paper is as follows: Section 2 reviews
mid-point of the applicable target range, although it was
New Zealand’s macroeconomic record over the inflation
volatile from year to year, and strayed outside the target
targeting period. Section 3 discusses fiscal policy. Section
range on three occasions, including the first half of 20063.
4 reviews monetary policy. Section 5 considers some recent
proposals for using foreign exchange market intervention
as a stabilisation instrument. Section 6 considers some
alternative stabilisation instruments. Section 7 concludes.
Figure 2
Inflation and inflation target range
6
%
%
6
CPI
Lower bound of inflation target range
Upper bound of inflation target range
5
2
5
4
4
Is there a problem with
3
3
New Zealand’s economic
2
2
performance?
1
1
0
1988
New Zealand’s remarkable economic
0
1990
1992
1994
1996
1998
2000
2002
2004
achievements
Source: Statistics New Zealand
Reserve Bank of New Zealand
A glance at some of the key macroeconomic indicators
The March 2006 figure is a year-on-year inflation rate of
shows the turn-around in New Zealand’s economic
3.3%. The second quarter 2006 year-on-year inflation rate
performance since 1985.
was 4.0% – a figure that should ring alarm bells. It is notable
that all three times inflation fell outside the target range, it
Inflation, inflation targeting and central bank
independence
exceeded the upper bound of the range. Inflation never fell
The post-World War II inflation record is a story of virtue lost
below the lower bound of the target range.
and painfully regained, as is clear from Figure 1.
Real GDP growth and unemployment
Figure 1
The conquest of inflation does not appear to have had a
CPI inflation since WWII
20
%
%
20
15
15
10
10
5
5
0
0
negative effect on the trend growth rate of GDP. There is
even some evidence (see Figure 3) that the volatility of GDP
growth has been reduced since the 1980s, although the
level of volatility remains quite high.
-5
-5
1945 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005
Source: Statistics New Zealand
During the inflation targeting period (since March 1990),
shown in Figure 2, inflation was, on average close to the
3
38
With inflation initially well above the target range of 0 – 2 %,
the original March 1990 PTA specified the target as something
to be achieved by end-December 1992. In December 1990,
the period for achieving the 0 to 2 per cent inflation target
was extended by 12 months to the year ended 31 December
1993.
Reserve Bank of New Zealand and The Treasury
Figure 3
There have been few reversals in the liberalisation processes
Real GDP growth
in the labour markets and in the product markets. Only in
%
%
the last few years have there been some signs of slippage,
20
20
15
15
10
10
5
5
Figure 4
0
0
Unemployment rate
-5
-5
12
-10
-10
1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005
for instance the amendments to the Employment Relations
Act in 2004.
%
%
12
10
10
Source: Statistics New Zealand
8
8
The “eye-ball” identification of reduced volatility in
6
6
4
4
2
2
New Zealand’s real GDP growth since the mid-1980s is
confirmed by detailed statistical studies of the variance of
GDP growth over time and of its decomposition in terms
0
1985
of the sectoral variances and covariances (see e.g. Buckle,
Haugh and Thomson (2003)). Another detailed study of
the thirty-year period prior to 2000 was conduction by the
RBNZ (Reserve Bank of New Zealand (2000b)). It concludes
that output volatility in New Zealand has declined over the
past three decades due, in part, to a more benign world
environment. The economic reforms and restructuring
during the 1980s and early 1990s undid much of the damage
done by over-regulation and macro-mismanagement during
0
1990
1995
2000
2005
Source: Statistics New Zealand
As regards dynamic efficiency, measured by trend
productivity growth, the true record is in a number of
respects rather stronger than is suggested by the growth of
per capita GDP, shown in Figure 5. Average per capita GDP
growth since 1988 has been 1.5% per annum, a number
somewhat below the OECD average (OECD (2005)).4
the 1970s and the first half of the 1980s, permitting a less
Figure 5
vehement business cycle pattern in output to become
Growth rate of real GDP per capita
%
15
apparent. In line with international experience, contractions
in New Zealand appear to be becoming less severe. Many
%
15
10
10
5
5
0
0
-5
-5
-10
-10
of these developments are shared by the other advanced
industrial nations, and although output volatility has been
declining, New Zealand’s relative performance does not
appear to have changed much.
There has been a marked improvement in the functioning of
the labour market, with unemployment coming down from
a high of about 11% in 1991, to a level below 4% in 2005
(Figure 4). There remain cyclical swings in the unemployment
-15
-15
1955 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005
Source: Statistics New Zealand
rate, which rose from 6.0% in 1996Q4 to 7.6% in 1998Q4.
These swings are smaller than they were before inflation
targeting was introduced. The level of static efficiency of
4
the economy has clearly been improved very significantly
since 1990. That is, the economy makes better use of its
existing human resources and other physical resources.
Testing stabilisation policy limits in a small open economy
New official measures of labour, capital and multifactor
productivity were released in the first quarter of 2006, and
cover the March years 1988 to 2005 (see e.g. http://www.
stats.govt.nz/developments/productivity.htm and Statistics
New Zealand (2006a,b) and Treasury of New Zealand
(2005b).
39
The growth rate of GDP per person employed was even
equal to that of labour input growth and capital income
lower, at just under 1.3 % per annum (see Figure 6).
growth combined (see Figure 7), and since 1993, the main
However, Figure 6 also shows the behaviour of productivity
contributor to measured sector labour productivity growth
growth in the “measured sector of the economy”, the
has been multifactor productivity growth, with only a
roughly two thirds of total GDP for which there are
modest contribution of capital deepening (Figure 8).
5
independent estimates of output and inputs. It leaves
out most of the general government (the non-market part
of the public sector) and some private market services for
which output is also measured mainly by its inputs.
Figure 7
Contribution to measured sector real
GDP growth
6
Figure 6
5
4
4
3
3
2
2
7
1
1
6
0
0
and the measured sector
%
7
6
5
5
-1
-1
Multifactor productivity
Capital input
Labour input
-2
4
4
3
3
2
2
1
1
Source: Statistics New Zealand
0
0
Figure 8
-1
Growth rate of GDP per person employed
Growth rate of measured sector labour productivity
-2
1988
1992
1996
2000
6
5
Productivity growth for the whole economy
%
%
%
-1
-2
2004
Source: Statistics New Zealand
For the period 1989-2005 for which we have data for both
series, the growth rate of whole-economy real GDP per
-3
multifactor productivity growth is the main contributor
-3
-4
-4
1988-93
1993-05
1988-05
Contribution to measured sector labour
productivity growth
Annual average percentage change
%
%
4
4
person employed is 1.3% per annum and measured sector
labour productivity growth is 2.7% per annum. Interestingly,
-2
Multifactor productivity
Capital deepening
3
3
2
2
1
1
to measured sector GDP growth, with a contribution
5
40
The ‘measured sector’, consisting of industries for which
estimates of inputs and outputs are independently derived in
constant prices, excludes those industries (mainly government
non-market industries whose services, such as administration,
health and education, are provided free or at nominal charges)
whose real value-added is measured in the national accounts
largely using input methods, such as numbers of employees.
Also excluded are a number of private sector market industries
that similarly use some form of input measure to estimate real
output, for example the residential and commercial property
industries whose output is measured by the growth in property
assets. In accordance with this definition, the measured sector
excludes the following Australia New Zealand Standard
Industrial Classification (ANZSIC) divisions: L: Property
and Business Services, M: Government Administration and
Defence, N: Education, O: Health and Community Services
and Q: Personal and Other Services. The measured sector
accounts for about 65% of total GDP and 69% of total paid
hours over the period 1988 - 2005. Its real value-added has
closely tracked total GDP and has grown faster than the nonmeasured sector. Measured sector employment has grown
more slowly than non-measured sector employment.
0
0
1988-93
1993-05
Source: Statistics New Zealand
1988-05
Unfortunately, comparable data for measured sector
productivity growth are not available for most OECD
countries, so we cannot get a sense of how unusual this is.
Reserve Bank of New Zealand and The Treasury
A comparable exercise for Australia over the same period
Figure 10
shows that its “market sector” (the Australian name for the
Real exchange-rate and current account
same concept as New Zealand’s “measured sector”) has a
balance
slightly lower growth rate of multifactor productivity than
Index
160
New Zealand’s.
%
0
-1
140
-2
120
-3
100
Unbalanced growth?
-4
80
-5
New Zealand’s on balance remarkable record of economic
60
-6
performance since the stabilisation of the late 1980s is
40
characterised by continued wide swings in key endogenous
20
-8
Real TWI exchange rate (LHS)
0
1980
economic variables, and especially in variables that represent
the interface of New Zealand and the global economy.
-7
-9
Current Account Balance (%GDP) (RHS)
1985
1990
1995
-10
2000
2005
Source: Statistics New Zealand
Reserve Bank of New Zealand
Nominal and real exchange-rates
Second, over the period since 1988, the current account has
Figures 9 and 10 show the behaviour of the trade-weighted
(or effective) nominal and real exchange-rates (or TWI) since
1980 (an increase in the index represents a strengthening of
the New Zealand currency).
been in deficit every year. In 1997 and 2000, the current
account deficit exceeded six per cent of GDP. Towards the
end of 2005, it was approaching nine per cent of GDP.
Over the period 1988-2005, there is a positive correlation
There is a mild upward trend in the level of the real TWI
between the level of the real exchange-rate and the size of
exchange-rate, both over the full 25-year period and since
the current account deficit.7 The current account balance by
1990. The downward trend of the nominal TWI exchange-
definition equals the excess of national saving over domestic
rate over the period 1980 to 1985 is followed by a gradual
capital formation. We see in Figure 11 that the increase in
6
appreciation over the subsequent decade.
the current account deficit over the past three years reflects
a combination of stronger investment and weaker saving.
Figure 9
Figure 11
NZD nominal and real exchange-rate
Index
160
Index
160
140
140
120
120
100
100
Saving, investment and the current account
balance
30
%
%
30
25
25
20
20
80
80
60
60
40
40
5
20
0
0
0
-5
-5
-10
-10
Nominal TWI exchange rate
20
Real TWI exchange rate
0
1980
1985
1990
1995
2000
2005
Source: Statistics New Zealand
Two facts stand out about the current account of the
15
15
Gross National Saving
Gross Investment
Current Account Balance
10
-15
1988
10
5
-15
1990
1992
1994
1996
1998
2000
2002
2004
Source: Statistics New Zealand
balance of payments. First, it has been subject to massive
fluctuations, as is evident from Figure 10.
7
6
The two time trends in Figure 9 are second order
polynomials.
Testing stabilisation policy limits in a small open economy
The contemporaneous correlation between the level of the
trade-weighted real exchange-rate and the current account
balance as a percentage of GDP for the period 1988Q1 2005Q4 is -0.39.
41
The low (net) national saving rates, rarely above four per
investment position of New Zealand, as shown in Figure 14,
cent of GDP, aggregates very different sectoral saving rates
was minus 91.6 per cent of GDP at the end of 2005. With
(see Figure 12).
continuing large scale current account deficits, net external
indebtedness will continue to grow unless favourable
Figure 12
valuation effects come to the rescue.
Sectoral saving rates
%
%
8
Figure 14
6
6
Net international investment position
4
4
2
2
8
-78
%
%
-78
-80
-80
-82
-82
-84
-84
-86
-86
-88
-88
-90
-90
Source: Statistics New Zealand
-92
-92
Net household saving rates have been on a downward
-94
0
0
-2
-2
-4
-4
Net Household Saving
Net National Saving
Net Business Saving
Net Public Saving
-6
-6
-8
-8
1987
1989
1991
1993
1995
1997
1999
2001
2003
2005
trend since 1988 and reached minus 7% of GDP in 2005.
Part of the explanation is undoubtedly the very significant
capital gains on housing experienced over the period, which
contributed to a steady increase in the ratio of household
financial wealth to disposable income despite negative
household saving rates (see Figure 13).
2001
2002
2003
2004
2005
Source: Statistics New Zealand
The reflection in the net foreign investment income
account of the large negative external investment position
is a sizeable and growing negative net stream of foreign
investment income, reaching close to 7% of GDP by the end
Figure 15
Household assets as a percentage of
Balance on foreign investment income
disposable income
500
2000
of 2005, as shown in Figure 15.
Figure 13
600
-94
1999
%
%
600
Direct domestic & overseas equities
Other financial assets
Housing stock valuation
500
400
400
300
300
200
200
100
100
0
%
%
0
-1
-1
-2
-2
-3
-3
-4
-4
-5
-5
-6
-6
-7
-7
-8
-8
-9
-9
0
0
2001
2002
2003
2004
2005
1996 1997 1998 1999 2000 2001 2002 2003 2004 2005
Source: Statistics New Zealand
Source: Statistics New Zealand
Although the growing net external indebtedness of
Both business saving and government saving have been
New Zealand has no obvious implications for the conduct
volatile, but positive. Net government saving went from
of stabilisation policy, it does mean the country is highly
zero in 2000 to five per cent of GDP in 2005.
exposed to global financial market developments that are
The persistent current account deficit has made New Zealand
a net external debtor to a degree that is without parallel
among advanced industrial countries. The net external
42
quite beyond its control. There also are potential political
problems associated with the transfer of a large and growing
share of GDP to foreign owners of domestic assets.
Reserve Bank of New Zealand and The Treasury
In other countries with large current account deficits and
Large swings in the current account balance and in the real
a large net external debt position, domestic agents, from
and nominal exchange-rates are often attributed to shocks
banks and other financial intermediaries to non-financial
transmitted from the world economy to New Zealand. It is
corporates and households, have built up significant balance
therefore interesting to note that, as shown in Figure 17,
sheet exposures to exchange-rate risk, by taking on foreign
the amplitude of swings in the (proximately exogenous)
exchange-denominated liabilities for which they have no
terms of trade has been significantly smaller than that of the
natural or artificial hedges. Some of the new EU members
(endogenous) real exchange-rate. Over the period 1980Q1-
from central and eastern Europe are particularly vulnerable
20056Q4, when normalised to have the same average value
to the balance sheet effects of exchange-rate shocks.
of 100, the nominal TWI exchange-rate had a variance of
New Zealand appears to have avoided this problem thus
289.9, the real TWI exchange-rate had a variance of 95.4
far, although a rigorous monitoring of the foreign exchange
and the terms of trade a variance of 57.2.
exposures not just of banks, but also of their customers,
remains a key part of safeguarding financial stability.
Figure 17
Terms of trade and nominal and real
Another bit of evidence that supports the view of a twospeed economy, with a sheltered non-traded sector and a
traded sector that is subject to the full force of international
competition is the difference between the inflation rates of
exchange-rates
Period average = 100 for all three indices
Index
130
120
tradable and non-tradable goods, shown in Figure 16.
Index
150
Real TWI exchange rate
Terms of Trade
Nominal TWI exchange rate
140
130
110
Figure 16
CPI tradable and non-tradable inflation
%
%
120
100
110
90
100
90
12
80
10
10
70
8
8
6
6
60
1980
4
4
Source: Statistics New Zealand
Reserve Bank of New Zealand
2
2
This does not, of course, mean that the fluctuations in
0
0
the terms of trade have not been a significant (or possibly
12
-2
-4
1988
-2
CPI tradable inflation
CPI non-tradable inflation
1990
1992
1994
1996
1998
2000
2002
2004
-4
2006
80
70
60
1985
1990
1995
2000
2005
even the most important) cause of fluctuations in the real
exchange-rate. It does, however, mean that there appears
Source: Statistics New Zealand
Reserve Bank of New Zealand
to be, in the transmission mechanism from exogenous
Over the period 1988Q1 to 2006Q1, the average annual
prices (and in economic activity), a shock-enhancing and
rate of inflation of traded goods prices in the CPI was 2.0%,
amplitude-magnifying domestic “filter”. It is even possible
while for non-traded goods prices it was 5.7%. It is likely
that part or even most of the volatility of New Zealand’s
that the Balassa-Samuelson effect (the productivity growth
real exchange-rate and GDP growth cannot be attributed
differentials between New Zealand and the average of its
to exogenous terms of trade shocks at all, but instead
trading partners in the traded goods sector exceed that in the
reflects domestic supply shocks and demand shocks that
non-traded goods sectors) accounts for part of this inflation
are amplified through the asset markets, especially the
differential. It is unlikely, however, to account for all of it.
housing market.
shocks and impulses to domestic fluctuations in key relative
Cyclical disinflation may have impacted disproportionately
on the traded goods sectors.
Testing stabilisation policy limits in a small open economy
43
The available empirical evidence (see e.g. Wells and Evans
International labour mobility is higher for New Zealand than
(1985) Buckle et. al. (2002) and Grimes (2006), is consistent
for other developed nations of its size. There is, effectively,
with the view that terms of trade shocks (or import and
an integrated regional labour market with Australia. Inward
export price shocks separately) are important determinants
and outward FDI are both high. One manifestation of
of New Zealand’s GDP growth and of its stability. None of
this is that most of New Zealand’s banking sector now is
these models directly addresses the proposition that the
Australian-owned.
high volatility (by international standards, although not by
New Zealand historical standards) of the (endogenous) real
exchange-rate and of real GDP growth, cannot be attributed
mainly to the volatility of New Zealand’s (exogenous) terms
of trade, but requires a shock-enhancing and amplitudemagnifying domestic filter. Only the Buckle et. al. (2002,
2006) SVAR models include any domestic financial variables
– the prime suspects as shock enhancers and amplitude
Financial openness, whether measured by gross and net
financial stocks and flows or by co-movements between
domestic and external yields, is much higher than openness
to trade in real goods and services. New Zealand has been
on the radar of every macro hedge fund and international
investment bank since at least the beginning of this decade.
With the disappearance of many high quality currencies
through the creation of the Eurozone in 1999, the remaining
magnifiers. However, the two domestic financial variables
small developed countries with national currencies,
(a 90-day nominal interest rate and the real rate of return
Switzerland, Norway, Sweden, Denmark, New Zealand,
on New Zealand equities) do not include the prime suspect:
Iceland, Australia, (and to some extent also Canada and the
house prices.
UK), have become disproportionately important sources
Further evidence to support the view that the external
and destinations of actively managed funds; New Zealand’s
environment may not be the main cause of New Zealand’s
recent role in the Japanese yen carry trade is a dramatic
wide economic fluctuations is the fact that New Zealand is
example of this.
not, for an economy of its size and per capita income level,
Finally, New Zealand’s structure of production, that is,
highly open to international trade.
the sectoral composition of its GDP is often mentioned
Exports plus imports as a share of GDP, a common index
as a reason for the volatility of its GDP. Primary industries,
of trade intensity, are just below 60% in 2005, and have
ranged over the past 20 years between just over 50% to
just under 70%. This is much lower than countries of similar
size and level of development, like Norway (73.1% in 2005),
Finland (73% in 2005), Sweden (89.5% in 2005) or Ireland
(145.5% in 2005). It is about the level of trade intensity of
the UK economy (56.1% in 2005), a much larger economy.8
Clearly, transportation costs and other international
transaction costs make New Zealand, for a country of its
especially agriculture, fishing, forestry and mining are viewed
as significant flexible-price, competitive sectors where
global demand shocks, manifesting themselves as termsof-trade shocks, and domestic supply shocks play a unique
role that is different from the global and domestic shocks
affecting the more “New-Keynesian” goods producing and
service industries. Shocks to these primary sectors are then
transmitted, through income and expenditure effects, to
the rest of the economy.9
size and wealth, a rather closed economy as regards trade
in real goods and services.
The geographical distribution of New Zealand’s international
8
trade is not highly concentrated. Australia is the most
important destination for New Zealand exports and the
most important source of imports, with a share of just over
20% for both exports and imports.
44
9
Trade for these other countries, especially for Ireland, has a
much higher ‘entrepot’ content than for New Zealand. That
is, much of Irish imports are raw materials and intermediate
inputs that are processed and re-exported. The value of exports
is therefore significantly higher than the value of value added
in exports.
The ‘size of the primary sector’ argument is distinct from the
terms of trade argument. Terms of trade changes affect not just
the primary sector, but also the export prices faced by nonprimary exporters and the import prices faced by non-primary
producers and by consumers.
Reserve Bank of New Zealand and The Treasury
Table 1
view go so far as to assert that the true importance of the
GDP by production group
natural resource-based sectors is measured by the sum
2005: % of Total
Finance Insurance & Business Services, etc.
Manufacturing
Personal & Community Services
Transport & Communication
Wholesale Trade
Retail, Accommodation, Restaurants
Construction
Agriculture
Government Administration & Defence
Fishing, Forestry, Mining
Electricity, Gas & Water
24.0%
15.2%
12.0%
10.5%
8.7%
7.6%
4.9%
4.5%
4.1%
2.3%
2.0%
GDP
100.0%
of the value added in the natural resource-based sectors
and the value added in the sectors that use the output of
the natural resource-based sectors as their principal input.
That argument only makes sense if without the natural
resource-based sectors, the productive resources (capital,
land and labour) now utilized in the natural resource
processing industries would not produce any value added
at all. Not only would it not be possible to obtain natural
resource inputs for the natural resource processing sectors
from outside New Zealand on terms that would make the
natural resource processing sectors economically viable,
Primary Industries
Goods producing Industries
Service Industries
6.9%
22.1%
67.0%
the resources currently employed in the natural resource
processing sectors would have no economically viable
alternative uses whatsoever in manufacturing or services.
Source: Statistics New Zealand
Given a choice between this extreme hypothesis and the
As is apparent from Table 1, the validity of the argument
alternative extreme hypothesis that without New Zealand’s
that the primary sectors in New Zealand play a unique and
natural resource-based sectors the resources currently
significant role as a source of domestic supply shocks and as a
employed processing the outputs of New Zealand’s natural
transmission/propagation mechanism for global commodity
resource-based sectors would be producing an equal
shocks, is not immediately evident from the data on the
amount of value added, the second seems likely to be closer
sectoral composition of output. New Zealand’s structure of
to the truth.11
production looks like that of any advanced post-industrial
economy. The service sector is overwhelmingly important,
with 67% of GDP.
The key parameters determining the appropriate ‘weight’
of the New Zealand primary sector over and above its
GDP share in the transmission of terms of trade shocks or
Just over 22 per cent of GDP is accounted for by goods
domestic primary sector supply shocks are (a) the terms
producing industries. All primary industries combined
on which substitutes for the outputs of the New Zealand
account for 6.9% of GDP. By contrast, Ireland’s primary
primary sector can be sourced from abroad and (b) the
sectors accounted for 2.5% of GDP in 2005. The primary
sector in New Zealand is perhaps twice as large as a share
of GDP as it is in most advanced industrial countries (other
than major oil or gas producers like Norway).10 It is still a
rather small tail to be a plausible candidate for wagging the
entire dog.
It is frequently argued that the natural resource-based
sectors are more important to the New Zealand economy
than is indicated by their GDP share, because a sizeable
primary processing industry uses the outputs of the natural
resource-based sectors as inputs. Some proponents of this
10
The oil and gas sectors accounted for just under 24% of
Norway’s GDP in 2005.
Testing stabilisation policy limits in a small open economy
11
The thought experiment I consider is that of the complete
disappearance of New Zealand’s primary sector, not a
marginal reduction in its size. The cost to the New Zealand
economy includes the cost of reallocating quasi-sunk factors
of production. Such transitional costs will be higher if the
disappearance of the primary sector is unexpected and sudden
rather than anticipated and gradual.
45
terms on which the inputs of capital, land and labour
3
currently employed in the New Zealand primary processing
Throughout the world, including the USA and much of
12
Fiscal policy
industries can be reallocated to alternative uses.
Western Europe, the authorities are keen to use fiscal policy
The fact that most New Zealand economists are quite happy
actively for stabilisation purposes - or at any rate keen to use
to model the country as a price taker in its international
tax cuts or public spending increases to boost aggregate
import and export transactions of primary commodities
demand during periods of cyclical weakness. The long-
suggests that a competitive foreign supply of primary
term sustainability of the fiscal-financial programme of the
materials would indeed be available to the domestic food
state is often questionable, unless future spending and/or
processing industry, even if there were no primary domestic
taxation programmes turn out to be significantly different
food growing industry. Note that this argument in no way
from what is in the pipeline today. Fiscal incontinence has
affects the importance of terms of trade shocks (which affect
been a worldwide problem for a significant period. General
all traded goods and services and not just primary sector
government deficits are often excessive and general
products) for the New Zealand economy as a whole. It does,
government debt-to-GDP ratios are rising.
however, suggest that the primary sector in New Zealand is
In New Zealand, this pattern is reversed. The long-term
probably no more significant in the transmission of these
sustainability and solvency of the state is beyond doubt if
shocks to the domestic economy, than is indicated by the
the rules adopted since in the past two decades continue to
size of that sector, as measured by its value added. Whether
be implemented in spirit as well as according to the letter of
these primary products are sold on the world market or
the law. The Public Finance Act of 1989 (recently amended
to the domestic food processing industry, or whether the
to include the key parts of the Fiscal Responsibility Act 1994)
domestic food processing industries procure their inputs
provides a framework which, if adhered to consistently, now
from New Zealand primary producers or on the world
and in the future, will ensure fiscal sustainability in the face
markets, is a matter of little significance.13
of the demographic challenges (population aging through a
combination of lower birth rates and higher life expectancy),
which New Zealand faces together with all other developed
nations. The PFA requires government to set both short-term
12
13
46
If a country in its not too distant past had a primary commoditybased economy, its inhabitants tend to greatly overestimate
the continued size and significance of the natural resourcebased sectors today, no matter how much the structure of
the economy may have changed. Icelandic economists were
as unwilling to contemplate the lack of macroeconomic
significance of their natural resource sector (fisheries) in
1999, when I wrote a paper for the Icelandic Central Bank
as New Zealand economists are today (see Buiter (2000)).
The Icelandic fish processing industry could survive very
well without an Icelandic fishing industry, as long as nonIcelandic fishermen and women could and would supply it
with fish on terms comparable to that of the domestic fishing
industry. Since Iceland on a number of occasions almost went
to war with the UK to ban British fishing vessels from what
Iceland considered to be its fishing grounds, it is apparent
that competitively priced fish would have continued to be
available to the Icelandic fish processing industry, even if
there had been no Icelandic fishing industry at all.
The exact determination of the impact on aggregate value
added in New Zealand of the disappearance of the primary
sectors would have to take into account a number of features
that would have the effect of making the impact somewhat
larger than the value of the primary sectors’ value added. For
instance, the world markets could not supply commercially
viable substitutes for the fresh produce supplied by
New Zealand farms. Resources currently employed in
processing such produce would have to be re-allocated to
alternative uses which might create less value added.
fiscal intentions and long-term fiscal objectives consistent
with the “principles of responsible fiscal management” set
out in the Act. It requires a comprehensive accounting for
all government assets and liabilities, including deferred and/
or contingent claims on the state budget. Its five principles
of responsible fiscal management compel governments to
take the long view and to explicitly consider future fiscal
risks. The five principles can be summarised as follows:
(1) Reducing total Crown debt to prudent levels so as
to provide a buffer against factors that may impact
adversely on the level of total debt in the future, by
ensuring that, until those levels have been achieved,
total operating expenses in each financial year are less
than its total operating revenues in the same financial
year.
Reserve Bank of New Zealand and The Treasury
(2) Once (1) has been achieved, maintaining these levels
The comprehensive reporting requirements (the annual
by ensuring that, on average, over a reasonable period
pre-Budget Budget Policy Statement of the government’s
of time, total operating expenses of the crown do not
short-term fiscal intentions and the annual Fiscal Strategy
exceed its total operating revenues.
Report at the time of the Budget which details the long-
(3) Achieve and maintain levels of Crown net worth that
provide a buffer against factors that may impact
term (minimum 10 years) fiscal objectives) impose on the
government the highly desirable discipline of having to
demonstrate that both the short-term fiscal intentions and
adversely on total net worth in the future.
the long-term fiscal strategy are consistent with the five
(4) Manage fiscal risks prudently.
Principles outlined above.
(5) Ensure predictable and stable tax rates in the future.
While the golden rule (cyclically adjusted or not) by itself
Principle (1) has surely been met by now. Principle (2) can
does not rule out explosive behaviour of the public debt (the
be interpreted as a version of the golden rule, according
Crown could engage in massive investment programmes
to which the Crown should only borrow to finance net
that do not yield a financial return at least equal to the
investment. The phrase “over a reasonable period of time”
Crown’s cost of borrowing) the combination of the golden
is vague, but would cover the kind of “cyclically corrected”
rule and the requirement that Crown debt be reduced to
golden rule that the UK is supposed to pursue. I would
and kept at prudent levels together do ensure that Crown
favour a slightly more general interpretation that permits
borrowing to finance public investment cannot get out of
transitory/reversible deficits and surpluses driven by any
control.15
temporary factors, not only conventional business cycles, in
order to smooth public consumption, distortionary tax rates
and private consumption.
New Zealand’s fiscal-financial framework “takes care of
the long run”, including intergenerational distributional
objectives
and
population
aging:
fiscal-financial
Principle (3) means that the asset side of the Crown’s
sustainability is built into the framework, without recourse
balance sheet must be taken into account, as well as the
to arbitrary numerical debt and/or deficit ceilings like the
liability side covered by (1). It is key that assets be “marked-
three per cent of GDP deficit ceiling and the 60 per cent
to-market”, that is valued as the present discounted value
of annual GDP gross debt ceiling of the EU Stability and
of their future contributions to the Crown’s budget. Neither
Growth Pact. Governments are still required, however, to
historic, nor replacement costs are relevant if the assets
specify long-term objectives.
cannot be realised and/or do not yield a future cash flow
to the government. Principle (4) has no obvious operational
content, although the PFA guide cites two examples that,
if they indeed represent the spirit in which Principle (4) will
14
be applied, would be very positive indeed.
The framework also appears to have delivered what it set
out to achieve. Figure 18 shows the almost perfect volcanic
cone described by New Zealand’s net public debt as a
percentage of GDP since 1971.16
Principle (5)
imposes tax (rate) smoothing (something like a constant
average marginal tax rate) as a desideratum. When the
growth rates of tax bases and spending commitments are
not uniform, tax rate smoothing provides a non-Keynesian
rationale for counter cyclical deficits.
14
The first example relates to fiscal risk that can arise in relation
to the Crown’s financial position, for example, changes in the
value of assets and liabilities and the potential for off-balance
sheet items such as guarantees to give rise to liabilities. The
second example concerns risks to operating flows, for example,
changes in the tax base and the risk of certain expenditures
exceeding the amounts budgeted.
Testing stabilisation policy limits in a small open economy
15
16
In the UK the combination of the cyclically adjusted golden
rule and a sustainable investment criterion (net general
government debt cannot exceed 40% of GDP) ensure that
debt-financed public investment cannot become a source of
fiscal unsustainability.
OBERAC = Operating Balance excluding revaluations and
accounting policy changes.
47
Figure 18
“insurance” premia, as there is no true individual insurance
Gross and net government debt
element involved. New Zealand has this absolutely right.17
%
%
70
70
This does not mean that the assets built up in the NZS
60
Fund would be there to be raided. It means that instead of
50
50
earmarking the current NZS Fund and any future planned
40
40
prefunding for a specific expenditure commitment, the
30
30
Crown would have a Fund (and could contribute to that Fund
20
20
in the future) as part of a policy of prefunding future Crown
10
10
Net debt
Gross debt
60
0
0
1972
1976
1980
1984
1988
1992
1996
2000
2004
expenditure in general. Fiscal virtue cannot be achieved
by earmarking, that is, by ring-fencing a particular pot of
Note: March years
money for a particular future expenditure commitment. It
Source: Statistics New Zealand
can only be realised by achieving the appropriate profile
The Crown has also been building up its stock of financial
of current and future Crown revenues of all kinds, given
assets, both in the form of foreign exchange reserves held
the totality of the spending commitments and plans of the
at the RBNZ and though contributions to the NZS Fund
Crown. The New Zealand Treasury has recently produced a
which in 2006 reached NZD10bn or six per cent of GDP and
number of documents (Treasury of New Zealand (2006a))
is expected to more than double by 2010.
that are quite close in spirit to the approach I advocate
It is not immediately obvious, that building up financial
assets through the specific mechanism of pre-funding
future state pension (superannuation) commitments
makes sense. It would be more transparent to “fiscalise”
NZS completely, by abolishing the specific pre-funding of
the state pension and replacing it by a general prefunding
here. Even so, in the introduction to one of the key
documents (Treasury of New Zealand (2006b, p5)), the
Secretary to the Treasury referred to “…the establishment
of the New Zealand Superannuation Fund, which invests
a proportion of current taxes to contribute to the costs of
New Zealand superannuation in the future.”
commitments.
The figures for the stock of debt are mirrored in the
New Zealand is already most of the way there. Unlike all
persistent budgetary surplus of the general government, as
other OECD countries, New Zealand does not levy payroll
shown in Figure 19. The operating balance has been positive
taxes on employers or employees to fund its state pension
since 1993/94, and so have net lending and the primary
programme. As there is no link between individual life-time
(non-interest) budget balance (except for tiny deficits for
contributions during the working life and the state pension
the last two measures in 1999/2000). For the last available
that is paid during retirement, formally turning NZS benefits
year, 2004/5, all three budgetary balances are close to 5
into a state pension based on residence, and perhaps on
per cent of GDP.
of
all
future
government
spending
other characteristics, uncoupled completely from life-time
contributions and financed out of general government
revenues without NZS-specific pre-funding would be
preferable. Likewise, publicly funded health and disability
benefits and programmes should be funded out of general
government revenues, not out of phoney health or disability
17
48
It is a mystery, however, why the sensible public funding of
health care and disability is coupled, in New Zealand, with
virtually exclusive public provision of health care services.
There could be great efficiency gains, to be shared by patients
and tax payers, if much of the public provision of healthcare
were privatised.
Reserve Bank of New Zealand and The Treasury
Figure 19
Not only has the fiscal-financial position of the Crown been
transformed since the late 1980s. In addition to this massive
Government budget balances
%
%
8
8
6
6
Operating balance
Net lending
Primary balance
intertemporal redistribution, there has been a reduction in
the size of the public sector, as measured by the share of
4
total public spending in GDP. From over 43% of GDP at
2
2
the end of the 1980s, (well above the OECD average) this
0
0
ratio has come down to just over 36% in 2004, somewhat
-2
-2
below the OECD average (see Figure 22). This downward
-4
-4
trend appears to have come to a halt or even to have been
-6
-6
reversed in the last couple of years, with obvious implications
-8
for future tax burdens and tax rates.
4
-8
1972
1976
1980
1984
1988
1992
1996
2000
2004
Source: Statistics New Zealand
A comparison of New Zealand with the average of the
OECD countries, gives one a sense of just how remarkable
an achievement this is. Figure 20 shows the comparison for
net lending, Figure 21 shows it for net debt.
Total general government disbursements in
New Zealand and the OECD average
55
%
%
55
NZ
Figure 20
OECD
50
General government net lending in
%
%
6
General government net lending OECD
General government net lending NZ
2
50
45
45
40
40
4
35
35
2
30
New Zealand and the OECD average
4
Figure 22
6
30
1987
1990
1993
1996
1999
2002
2005
0
0
-2
-2
Bathing in the gentle glow of fiscal sustainability, the
-4
-4
New Zealand authorities and economics community wonder
-6
whether fiscal policy could make a larger contribution
Source: OECD
-6
1987
1990
1993
1996
1999
2002
2005
to cyclical stabilisation. This issue can be broken down
Source: OECD
into three sub-questions. The first concerns the role
Figure 21
and effectiveness of the automatic fiscal stabilisers, and
General government net financial liabilities in
possible fiscal or regulatory changes required to enhance
New Zealand and the OECD average
their capacity for damping cyclical fluctuations. The second
%
50
%
NZ
OECD
45
40
50
concerns the more active use of discretionary fiscal policy
45
for macroeconomic stabilisation. The third concerns better
40
coordination between monetary and fiscal policy.
35
35
30
30
25
25
20
20
15
15
10
10
The automatic fiscal stabilisers are one of three automatic
5
5
stabilisers that can dampen the response of income to
0
0
1980
1984
1988
1992
1996
2000
2004
Source: OECD
Testing stabilisation policy limits in a small open economy
The automatic stabilisers
exogenous demand shocks when at least some households
are Keynesian consumers rather than permanent income
49
consumers. Keynesian consumers are those for whom a
Figure 23
change in current disposable income has an effect on current
Primary balance and output gap
consumption that is much larger than the marginal propensity
to consume out of permanent income (approximately one)
times the effect of a change in current disposable income
on permanent income (typically small). Liquidity constraints,
cash-flow
constraints,
wedges
between
household
%
6
%
6
4
4
2
2
0
0
-2
-2
-4
-4
borrowing and lending rates and other capital market
imperfections, like the poor collateralisability of human
wealth, lie behind the Keynesian consumption function and
µ
(holding constant interest rates, the exchange-rate, other
asset prices and money prices and wages) can be written
1
µ=
 1 − (τ d + τ s ) 
1 − (1 − s)(1 − m) 

 1+ τ i
s
where
s
m
s
τ
2002
2005
Source: OECD
The tax on personal income taxes all income other than
capital gains. The structure is progressive; for wage and
salary income a 15% rate is paid on income up to NZD 9,500
and salaries and to interest income and dividends. Fringe
low-income families with dependent children. The highest
marginal rate is among the lowest in the advanced industrial
countries. The progressive income tax structure makes for
is
real fiscal drag (the share of taxes in income increases as
the social contribution
real income grows). There may also be some nominal fiscal
the indirect tax rate (in New Zealand this is
drag, because of imperfect effective indexation of the
the direct or income tax rate,
rate and
1999
benefits are taxed separately. Tax credits are available to
is the marginal propensity to import
(the share of total consumption spent on imports),
i
1996
income above 60,000. Withholding taxes apply to wages
i
is the marginal propensity to save out of current
disposable income,
1993
33% on income between 38,000 and 60,000 and 39% on
 1 − (τ + τ ) 
1 − (1 − s)(1 − m) 
 >0
 1+ τ i

d
1990
per annum, 21% on income between 9,500 and 38,000,
0 ≤ s < 1;0 < m < 1;τ + τ < 1;τ > −1;
d
-8
1987
In an open economy, the simple Keynesian multiplier,
(1)
-6
Primary Balance
-8
the role it creates for the automatic stabilisers.
as follows:
Output Gap
-6
τ
s
τ
d
mainly the Goods and Services Tax or GST, a value-added
different tax rate bands.
tax).
The structure of the income tax suggests the following
New Zealand’s revenue structure is very simple. For the
means of strengthening the automatic fiscal stabilisers
automatic stabilisers, the degree to which taxes and transfer
while reducing distortions. Since there is at this moment no
payments co-vary with the level of economic activity is
apparent need to increase total tax revenues, extending the
key. In what follows, these aspects of the tax and transfer
tax bases, as proposed below should be accompanied by
structure are emphasized. Figure 23 shows that there is a
lowering the average and marginal tax rates, so as to make
reasonably pronounced counter-cyclical behaviour of the
the whole exercise revenue-neutral.
government primary (non-interest) deficit.
Proposal 1. Tax capital gains at the same rate as all other
income. Make the income and capital gains tax index-linked
to the CPI. Index-link the corporate profits tax (only real
interest costs should be deductible etc).
This prevents erosion of the tax bases for the labour income
and capital income, as it is far too easy to turn capital gains
into capital income and, in the unincorporated sector,
capital income into labour income. There is an efficiency
50
Reserve Bank of New Zealand and The Treasury
case for not taxing capital income at all (whether it be
of total social security and welfare spending. There do not
profits, dividends or capital gains), but if any capital income
seem to be any easy angles for augmenting the automatic
is taxed, all capital income should be taxed and at the same
stabilisers from the government transfer payment side of
rate. This will raise
τ
d
and thus reduce the size of the
the budget.
multiplier.
Corporate profits are taxed at a 33 per cent rate, which
More active countercyclical use of the GST
is becoming a rather high number among the advanced
Even at a constant proportional rate of 12.5%,
industrial countries. As noted earlier, unique among OECD
New Zealand’s GST contributes to the conventional
countries, New Zealand has no dedicated employee and
Keynesian automatic fiscal stabilisers. It is, however, also
employer contributions to a pay-as-you go state pension
clear that the magnitude of the multiplier is smaller when
scheme (like the NZS) or other social security funds. This
the indirect tax rate, or indeed any tax rate or any “marginal
means that
τ s = 0 in equation (1).
leakage from the circular flow of income and expenditure”
The main indirect tax is the GST, applied at a uniform rate of
is higher. Raising the GST rate during a boom and lowering
12.5%. Financial services and housing rentals are exempt.
it during a downturn could therefore enhance the working
There also are excise duties on alcohol, tobacco, petroleum
of the automatic stabilisers. It would, however, increase the
and gaming. This immediately suggests another way of
microeconomic dead-weight losses associated with time-
strengthening the automatic fiscal stabilisers:
varying marginal tax rates. If there are more Keynesian
Proposal 2. Broaden the GST base by eliminating the
exemptions for financial services and housing rentals,
including the imputed consumption of housing services by
owner-occupiers.
This increases
τ
i
consumers (who spend their disposable income) in the
downturn than in the upturn of the cycle, a variable GST
could have stabilising effects even if it were revenue-neutral
over the cycle. During the downturn the lower taxes would
relax the current disposable income constraint on more
and reduces the value of the multiplier.
It also eliminates a distortion. The argument that there are
insuperable practical obstacles to the inclusion of financial
services in the tax base (see e.g. Jack (2000) and Auerbach
and Gordon (2002)) because of the inability to measure the
value added of a financial service makes no sense to this
non-specialist. Firms selling financial products create value
added which can be measured as the sum of wages/salaries
plus profits plus rents. These are all reported as part of the
income and profit tax filings.
The exemption of the imputed rental income of owneroccupiers from both the GST and the income tax makes no
consumers than it would incrementally subject to a binding
current disposable income constraint in the upturn.
There is, however, another reason why the New Zealand
authorities might be interested in a more active
countercyclical use of the GST, which does not depend on
the GST contribution to the Keynesian automatic stabilisers,
but on its ability to enhance the neoclassical intertemporal
substitution effect. The Euler equation of a standard
competitive household intertemporal optimisation problem
with an efficient financial market is
(2)


 1 + τ ti 
u '(Ct ;...) = Et  β (1 + rt ,t +1 ) 
 u '(Ct +1 ;...) 
i
 1 + τ t +1 


sense. It should either be taxed as consumption under the
GST or as income under the income tax.
where u (C ;...) is the period utility function, C is
On the transfer payments side of the budget, social security
consumption,
and welfare are just under one third of total public spending.
is the real interest rate between periods t and t+1 and
However, much of this, including superannuation (just
is the expectation operator conditional on information at
under half) is not countercyclical. Unemployment benefits,
time t.
β
is the subjective discount factor,
rt , t +1
Et
an obvious countercyclical item, are only about six per cent
Testing stabilisation policy limits in a small open economy
51
It is clear from that, ceteris paribus a lower GST rate,
in period t or a higher expected GST rate,
τ
i
t +1
τ ti ,
, in period
rise to immediate comparable changes in the household’s
consumption of the service flows of the durables.20
t+1 lower the consumer’s effective real rate of interest at
market prices between periods t and t+1, just as much as a
cut in the real interest rate at factor cost between periods t
Implementing a more countercyclical GST policy:
technical issues
and t+1, rt ,t +1 , would. A-fortiori, the combination of a cut
The key technical issues associated with a more
in the period t indirect tax rate and credible commitment
countercyclical use of the GST rate are the same as for the
to an increase in the period t+1 indirect tax rate that keeps
use of all other discretionary fiscal and monetary policy
total indirect tax revenue constant in present discounted
instruments, especially those that work at least in part
value, would represent a cut in the consumer’s effective
through expectational channels: timing, magnitude and
real interest rate.
credibility. We are considering the adoption of a rule aimed
We can now rephrase this as a combination of a GST rate
at switching consumption expenditure from upturns in
cut during a downturn of the business cycle and a GST rate
the business cycle to downswings. It does not, of course,
increase during the upswing of the cycle, which is revenue-
follow that the GST rate should be varied only in response to
neutral over the cycle in present discounted value. Since the
measurements/estimates of the current output gap or some
proposed changed in the GST tax rule is revenue-neutral
other cyclical variable. It should be targeted at the future
over the cycle, there will be no income effect, and the pure
expected output gap.
intertemporal substitution effect of a lower real interest
Policy is bedevilled by three kinds of lags: two types of
rate will stimulate consumption during the downturn and
inside lag and one type of outside lag. The first inside lag is
dampen consumption during the upturn.18
the recognition lag, the length of time between objective
The thought experiment just carried out kept constant the
occurrence of an event making a rate change appropriate
factor cost (or before-tax) real interest rate, rt ,t +1 . For a small
and the subjective recognition of that occurrence. The
open economy like New Zealand, which is highly integrated
second inside lag is the decision lag, the length of time
into the international financial markets, the assumption that
between recognition of the need for a change in the GST
changes in domestic indirect tax rates have no effect on
rate and its actual implementation. The outside lag is the
the before-tax real interest rate seems quite reasonable. The
length of the time interval(s) between the implementation
ceteris paribus effect of the GST tax changes is therefore
of a GST rate change and its (distributed) effects on
likely to be the same as the total effect.19
consumer demand.
The consumer demand effect of changes in the real
Given these inside and outside lags, the rule should be
interest rate induced by changes in the time pattern of
designed to ensure that the GST rate responds appropriately
indirect tax rates are enhanced for consumer durables,
to anything that helps predict (Granger-causes) movements
compared to consumption of non-durables and services.
in the output gap. The list of feedback variables in the GST
Changes in indirect tax rates can prompt significant shifts
rate decision rule will in general not be restricted to current
in the timing of purchases of new consumer durables
and past observations on the output gap itself. The length
(household investment goods), even when they do not give
of the recognition lag can make both the decision lag and
the outside lag irrelevant: if it takes too long to recognise
that the event making a rate change desirable has occurred,
18
Strictly speaking, the proposed tax changes would have to be
expected utility-neutral rather than revenue neutral in order
for the income effect to be absent.
19
With a constant elasticity of intertemporal substitution γ > 0
and
u (C ) =
1
C1− γ ,
1− γ
uncertainty, that
52
Ct + 1
Ct
equation (2) implies, for the case of no

 1 + τ t 

i
 1 + τ t +1 
=  β (1 + rt , t + 1 ) 

i
γ
−1
.
i.e. the horse has bolted, it does not matter that the decision
20
The policy is therefore similar in its effect on aggregate
demand to a cyclically revenue-neutral sequence of temporary
subsidies and temporary taxes on business investment.
Reserve Bank of New Zealand and The Treasury
to shut the gate is taken promptly and that the gate is shut
There is no time to address the general issue of when
immediately afterwards.
and how, in democratic societies, policy decisions can
Consider the following simple implementation of the rule.
During a boom, the rate is set at 15.0%, during a slump it
is set at 10% and during the neutral phases of the cycle
it is set at its current level of 12.5%. The effect during a
slump on current consumer purchases of a cut in the GST
rate to 10.0% will, under the rule, depend on the length of
the period for which it will be in effect (the rule is known
to be revenue-neutral over the cycle, so rational, forwardlooking consumers know that rates will eventually normalise
to 12.5% and rise to 15.0%). It seems likely that, especially
for consumer durables, the length of the outside lag would
be short: the duration of the downturn is uncertain, and
there is no option value of waiting for a further cut in the
tax rate: with “use it or risk losing it” incentives to spend
now, consumer durables expenditure should strengthen
and so, albeit to a lesser extent, will the actual consumption
of non-durable goods and services.
be delegated by the constitutionally elected government
of the day to operationally independent committees of
appointed technical experts. There are obvious attractions
to depoliticising the operation of a cyclically variable
GST rate. Turning the decision over to a committee
of independent experts would remove the scope for
opportunistic manipulation of the GST rate for electoral or
other party-political advantage. The technical competency
of the committee (the GSTC?) would be a non-trivial issue,
as the size of the population from which plausible domestic
candidates could be drawn (The Treasury, the RBNZ, the
Ministry of Economic Development, the universities and a
few think tanks) is small, so flying in external experts would
soon become necessary, just as it would be if New Zealand
were to create a Monetary Policy Committee to take OCR
decisions. Technical and administrative support for the GSTC
would presumably have to be drawn from the Treasury, the
RBNZ and the Ministry of Economic Development. There are
Thus, even if the duration of the different business cycle
phases is uncertain, the impact of a change in the GST rate
no realistic alternatives, even though it may undermine at
least the appearance of independence.
on demand is likely to be swift (the outside lag is likely to
be short and predictable). This, however, does not mean
the GST rate change will be well-timed and appropriate
in magnitude. The key technical issues are the combined
length of the two inside lags and the trade-off between
the duration of these lags and the quality of the GST
rate decision that is taken. This brings us to institutional
implementation issues.
It would not be desirable to merge the GSTC with the
RBNZ into an independent integrated monetary and
fiscal stabilisation committee. Doing so might undermine
the substance of operational independence for both
institutions, since fundamentally unaccountable economic
decision-making authority can only be legitimate, and
therefore acceptable, if the domain of discretion is strictly
circumscribed. Monetary policy in New Zealand, as in the UK
and the Eurozone, is formally accountable in the sense that
Implementing a more countercyclical GST policy:
institutional issues
The reasonably effective and successful implementation
of monetary policy in many types of market economies
through operationally independent central banks has
created some momentum in favour of delegating other
aspects of stabilisation policy to operationally independent
technical experts. An example is Charles Wyplosz’s
proposal for national Fiscal Policy Committees to play a
key role is determining the appropriate size of the general
government’s financial deficit (Wyplosz (2002)).
there are reporting obligations. Those to whom authority
has been delegated must explain and justify their actions to
the legitimate political authorities. There is no effective or
substantive accountability, however, since typically there are
no adverse consequences for the operationally independent
authorities, other than naming and shaming, if they make
a mess of things, or if their explanations fail to convince.
Typically, dismissal can only occur for incapacity and gross
misconduct, which does not include gross incompetence!
This makes it essential that not too much substantively
unaccountable economic decision making power be
Testing stabilisation policy limits in a small open economy
53
concentrated in one institution. This is certainly the case
to complicate the conduct of monetary policy by conducting
in New Zealand, where the monetary policy decisions are
fiscal policy with transparency, which means that most major
taken not by a committee, as is the case almost everywhere
discretionary changes in fiscal policy are announced well in
else, but by the Governor alone. The GSTC and the monetary
advance. Frequent communication between the Treasury
authority should therefore be independent from each other
and the RBNZ is the rule. Given the New Zealand institutional
as well as from the government, although cooperation and
set-up this is the only arrangement that makes sense. I do
coordination among them would of course be desirable.
not advocate switching to a form of coordination that
The foregoing discussion prompts the following proposal:
would involve the explicit joint determination of monetary
policy by the monetary and fiscal policy agencies (see
Proposal 3: Consider the active use of the GST rate as a
countercyclical stabilisation instrument. Delegate this policy
to an operationally independent GST Committee.
Reserve Bank of New Zealand (2001)). This view matches
that of Lars Svensson (2001) in his Independent Review of
the Operation of Monetary Policy in New Zealand: Report
to the Minister of Finance.
Other discretionary fiscal stabilisation policy
Other than the limited discretionary use of the GST rate for
countercyclical policy, I cannot see a role for discretionary
4
Monetary policy
fiscal policy as a stabilisation instrument. Public investment
As the first country to implement a formal inflation targeting
expenditure cannot be switched on or off, or varied in scale
policy, New Zealand had to invent the necessary institutions,
according to the cycle, without causing serious efficiency
instruments and policy rules from scratch. It is therefore no
losses and cost overruns. The inside and outside lags are
surprise that since 1990 there have been changes both in the
also bound to be so long, variable and uncertain as to
operational inflation target and in the wider characterisation
make this an inappropriate stabilisation instrument. Mutatis
of the RBNZ’s remit, as shown in Table 2.
mutandis, the same holds for other changes in government
An important change of emphasis in the remit occurred
spending and tax programmes. Clearly, if ever New Zealand
in December 1999, when the Bank was instructed to
found itself in a deep, 1930s-style slump, or in a Japanese-
have regard for “unnecessary volatility” in interest rates,
style liquidity trap, expansionary fiscal policy, combined with
output and the exchange-rate, in the course of conducting
expansionary quantitative-easing-style monetary policy,
monetary policy. This concern for volatility in the real
would be the appropriate response. Such exceptional, self-
economy was, however, subject to or without prejudice
evident conditions calling for discretionary fiscal policy are,
to, the primary target of establishing and maintaining price
however, quite unlike the modest cyclical fluctuations that
stability in the medium term.
have characterised New Zealand since the beginning of
inflation targeting.
Monetary and fiscal coordination
New Zealand has an operationally independent central
bank. Monetary policy instrument changes occur at a
higher frequency than fiscal policy changes. The monetary
authorities act effectively as Stackelberg followers vis-à-vis
the fiscal authority: the fiscal actions and fiscal rules are
taken as part of the economic environment within which
the RBNZ pursues its mandate. The fiscal authorities try not
54
Reserve Bank of New Zealand and The Treasury
Table 2
A Summary of evolution of the PTA
March 1990
Initially, the government and Reserve Bank agreed to a phased move towards the initial inflation
target of 0-2 per cent, with the original target date being December 1992.
December 1990
The target date was extended to December 1993.
December 1996
The target band was widened to 0-3 per cent in December 1996 to enable a somewhat greater
degree of inflation variability.
December 1999
A clause 4(c) was included requiring the Reserve Bank to have regard for `unnecessary volatility’
in interest rates, output and the exchange-rate, in the course of conducting monetary policy.
September 2002
The lower bound of the inflation target was raised to 1 per cent, on the grounds that at
extremely low or negative rates of inflation, the volatility trade-off probably worsens. In addition
clause 2(b), specifying the inflation target, was amended from `12-monthly increases in the CPI’
to keeping future CPI inflation outcomes within the target band `on average over the medium
term’. This change made explicit the medium-term focus for price stability, further enhancing
monetary policy flexibility.
Clause 4(c) was retained with modified wording, as clause 4(b).
Source: Bollard and Karagedikli (2005)
There has been a considerable evolution in the RBNZ’s
a little less frequent than in many other countries. I support
operational approach since the exchange-rate was floated
the relatively low frequency of RBNZ meetings. Those who
on March 4, 1985. From June 1997 till March 1999, the
advocate frequent meetings (outside periods of crisis), more
Monetary Conditions Indicator (MCI) was the proximate
often than, say, once a month (with a month off for good
target of monetary policy.
21
The interest component of the
behaviour during the summer) appear to confuse motion
MCI was the 90-day Bank Bill rate, which was not itself an
with action. Most of the time, news relevant to rate setting
instrument of monetary policy. The policy instrument during
tends to accrue slowly and ambiguously. The occurrence of
the MCI period, and for much of the floating rate period,
crises is identified easily and ad-hoc meetings can be called
was the target amount of settlement cash in the banking
immediately.
system. Since March 1999, when the Bank abandoned the
policy of targeting the MCI, the official cash rate or OCR
has been the sole instrument set by the RBNZ to achieve the
inflation target (see Reserve Bank of New Zealand (2000a)
and Woodford (2000)).
The specification of the RBNZ’s inflation target has a nonstandard feature: the target is defined as a range for the
inflation rate, currently between one and three per cent per
annum. There appears to be no presumption, in principle
or in practice, that the centre of that range, two per cent,
The choice of instrument, the overnight rate, is now in line
is the effective point target of monetary policy. I believe
with common practice among central banks of advanced
this feature to be undesirable, because it leaves the market
industrial countries. The OCR is reviewed by the Bank eight
confused. Is policy (the OCR) likely to be different when
times a year, approximately every six weeks, although the
the inflation rate is (or is expected to be over the horizon
RBNZ reserves the right, under exceptional circumstances,
relevant to policy making) close to the upper bound of the
to change the OCR at any time. This is consistent with the
range than close to the lower bound of the range? Is there
practice in the United States and, more recently, Canada, but
a discontinuity in the reaction function when the (expected)
The MCI was a weighted arithmetic average of the 90-day
b
Bank Bill yield i and the proportional appreciation of the
TWI nominal exchange-rate ε . The weight on the interest
rate was twice that of the exchange-rate appreciation:
MCI = i b + 0.5ε .
inflation rate crosses the limits of the range? It would add to
21
Testing stabilisation policy limits in a small open economy
clarity in monetary policy making if the range were replaced
by a point target for inflation, to be achieved in the medium
and long-term (or over whatever horizon the monetary
55
authority is believed to be able to have a systematic effect
Has monetary policy lost its punch and bite?
on the rate of inflation). The target should be symmetric:
A widely shared concern (at all levels) is that monetary
upwards and downwards deviations of inflation from target
policy (changes in the OCR) has lost much of its capacity
of equal magnitude should be viewed as equally costly.
to affect the real economy because changes in short-term
Larger deviations (in absolute value) should be viewed as
nominal interest rates (which are strongly influenced by the
proportionally more costly than smaller deviations. The usual
OCR) are largely neutralized either by a decline on long-
quadratic loss function penalising deviations of inflation
term rates or by borrowers moving along the downward
from the point target provides an adequate representation.
sloping yield curve to cheaper maturities. The lengthening
The range could either be dropped altogether or be given
of household mortgage maturities during the latest series of
some operational meaning along the lines of the “Open
OCR increases was frequently cited as an example.
Letter to the Chancellor” requirement found in the Bank of
England Act 1998.22
It is clear that a change in the OCR by itself has a negligible
effect on economic activity. In normal times, OCR-setting
Another unusual feature of the New Zealand monetary policy
meetings occur eight times a year, that is, once every
making process is that the decision is taken by the Governor
month and a half on average. What is the likely impact of
of the RBNZ alone, rather than by a committee, as is now
an increase of, say, 25 bps? An increase of 25bps, say, for
the practice in most other central banks. On the whole, the
6 weeks, which is all that is “technically” implied by such
quality of the monetary policy decisions made since 1990
an increase, has a negligible effect of the cost of funds for
suggests that thus far Monetary Policy Committees have
corporate investment, the household’s return to saving
not been greatly missed in New Zealand, although I wonder
and the opportunity cost of lending and borrowing. As a
whether a Monetary Policy Committee might have avoided
member of the Monetary Policy Committee of the Bank of
the 1997-1999 MCI interlude. I recognise that there are
England, I once referred to the effect of a 25bps change in
potential drawbacks to decision-making by committee (see
the Repo rate by the Bank of England as “chicken feed”, and
e.g. Sibert (2006)) that could outweigh its more familiar
that characterisation is appropriate also for New Zealand.
alleged advantages (see e.g. Blinder and Morgan (2005)),
prominent among which is the (unproven) presumption
that committees reduce the risk of an outlandish/extremist
decision being taken.
In a small open economy with a floating exchange-rate,
monetary policy (changes in the current OCR) work through
their effect on longer-maturity nominal interest rates, the
exchange-rate and the prices of other financial and real
On balance, I believe the evidence supports the creation
assets. There may also be effects through a variety of credit
of a procedurally transparent, individually accountable
channels, if financial markets are inefficient, for any of a
Monetary Policy Committee, along the lines of the MPC of
number of informational and enforcement reasons, and
the Bank of England to take future OCR decisions. Given
market segmentation prevails. A necessary condition for
the small population size of New Zealand, this would at
a current OCR change to have an appreciable influence
some point require either appointing foreign residents and
on behaviour is that a change in the current value of the
nationals who become residents of New Zealand for the
OCR affects expectations about future values of the
duration of their MPC appointment, or flying in overseas
OCR. If changes in expected future values of the OCR in
experts for the OCR meetings, or using international video
conferencing.23
22
56
Deviations of more than one percentage point above or below
the target require an Open Letter from the Governor of the
Bank of England to the Chancellor of the Exchequer. This will
explain the cause of the deviation, how long inflation will be
away from target, what action the Monetary Policy Committee
(MPC) of the Bank of England is taking, as well as how this
will be consistent with the Government’s wider economic
policy objectives.
23
The UK is now pioneering flying in an MPC member from
across the Atlantic for 10 to 12 days each month. Given
this precedent, and assuming that jet lag affects the quality
of monetary policy making either not at all or favourably,
New Zealand could create a 5 to 7 member MPC, with the
Governor having the casting vote. When the New Zealand
stock of qualified MPC candidates has been exhausted, the
global stock of qualified candidates could be drawn upon. I
don’t think the world is ready yet for the lower-cost alternative
of outsourcing monetary policy to some MPCC (monetary
policy call centre) in Bangalore.
Reserve Bank of New Zealand and The Treasury
turn affect current behaviour of private agents, monetary
Figure 24
policy is effective. This could be either because changes
Short- and long-term nominal interest rates
in current expectations of future OCR levels influence the
and spread
current prices of other financial instruments, or because
they affect non-price determinants of current household
%
30
%
30
25
25
behaviour (e.g. subjective estimates of permanent income),
20
20
of corporate behaviour and of the behaviour of importers
15
15
and exporters. By leveraging expectations, current changes
10
10
5
5
0
0
in an instrument that in and of itself matter very little, can
influence private behaviour.
-5
-5
90 day bank bill rate
10 year bond rate
10year-90day spread
Clearly, if an increase in the OCR only raises market-
-10
determined rates at very short maturities and lowers rates
-15
1985
at longer maturities, and if there are no other channels
Source: Reserve Bank of New Zealand
through which a current OCR change affects private
Changes in short rates are also associated with changes in
behaviour, then conventional monetary policy is weakened
long rates in the same direction, although the relationship is
or even neutralised completely. This concern is not unique
fairly weak, as is clear from Figure 25.24
to New Zealand. When the Federal Reserve Board started its
sequence of 17 consecutive 25bps increases in its target for
the Federal Funds rate from 100bps to 525bps, long nominal
-15
1990
1995
2000
2005
Figure 25
Changes in short and long nominal yields
1.5
interest rates declined and in due course the yield curve
1.0
became inverted. While this raised concerns among the
Change in 10year bond yield
0.5
members of the Federal Reserve Board, including Chairman
Alan Greenspan, no-one concluded that monetary policy
-10
0.0
-6.0
-4.0
-2.0
-0.5
in the US had lost its capacity for influencing economic
-1.0
activity.
-1.5
0.0
2.0
4.0
6.0
8.0
-2.0
For New Zealand, the recent behaviour of the OCR
-2.5
Change in 90-day bill yield
and longer-term interest rates raises three issues. First,
empirically, is the current inverted yield curve a common
phenomenon? Second, do higher short rates (changes in
short rates) tend to be associated with lower long rates
(with changes in long rates in the opposite direction)? Third,
if the answers to these first two questions is affirmative,
does it mean that monetary policy has lost its punch?
The observation that an increase in short rates is, on
occasion, associated with a decline in long rates need not
imply that monetary policy is ineffective. Take, for instance,
the simplest “Dornbusch overshooting model” in which
the authorities raise the short nominal rate of interest in
response to an unexpected temporary increase in the rate of
Figure 24 shows that, while short and long nominal interest
inflation (reflecting some permanent, cost-push price level
rates have moved together, historically, there have been
shock). Assume that the authorities’ reaction function has
several episodes, in addition to the most recent one starting
the “Taylor property” that the short nominal interest rate
in the second half of 2005, when the term structure was
is raised by more than the short-to-medium-term increase
inverted. The most recent prior periods with a negative term
in the inflation rate. Short real interest rates therefore also
premium were from early 1995 till early 1997 and from mid-
rise. There is a sharp appreciation of the nominal and real
1997 till mid-1998.
24
Testing stabilisation policy limits in a small open economy
The two largest outliers are in the fourth quadrant (a fall in
the bill rate and an increase in the bond rate) and occurred in
December 1994 and July 1997 respectively.
57
exchange-rates. After the initial increase, the short nominal
Figure 26
interest rate declines steadily and ultimately returns to its level
Perceptions of current inflation
prior to the inflationary shock. This future decline in short
Marketscope Survey
nominal rates is reflected immediately in the long nominal
interest rate (and real rate) which rises less on impact than
5
%
%
5
4
4
flat prior to the inflation shock, it would now be inverted.25
3
3
This movement in the yield curve is accompanied by a jump-
2
2
the short nominal interest rate. If the yield-curve had been
appreciation of the nominal and real exchange-rates. In a
small open economy with a floating exchange-rate and
unrestricted financial capital mobility, the exchange-rate
becomes the principal monetary transmission mechanism.
The successful countering of an inflationary shock using the
monetary policy instrument (the OCR) will be associated
Median current inflation
Mean current inflation
Actual current inflation
1
1
0
0
-1
1995
-1
1997
1999
2001
2003
2005
Source: RBNZ J6 Survey of expectations - a survey of businesses,
March 2006 Report, www.rbnz.govt.nz/statistics/econind/
j6/data.html
with a downward shift, and often an inversion of the yield
curve; it is a sign of monetary policy effectiveness, not of its
lack of effectiveness.
Long-term inflation expectations: the crucial anchor
for monetary policy
Figure 27
Predictions of inflation over next year
Marketscope Survey
%
%
Central to the RBNZ’s decision on the appropriate level of
5
the OCR is the state of long-term inflation expectations.
4
4
3
3
2
2
This is the litmus test of credibility for any inflation targeting
5
central bank. Figures 26 to 29 tell a story of remarkable
success in anchoring longer-run inflationary expectations
around the centre of the Bank’s target range, despite non-
1
negligible deviations of actual inflation from the centre of
0
the target range, and indeed despite three episodes during
which actual inflation exceeded the upper bound of the
-1
1995
Median expected inflation
Mean expected inflation
Actual inflation one year later
1
0
-1
1997
1999
2001
2003
2005
range (see Figure 2).
Figure 28
Expectations of inflation actual and expected
%
5
Headline CPI inflation
1 year ahead expectation (advanced 3 qtrs)
2 year ahead expectation (advanced 7 qtrs)
4
58
4
3
3
2
2
1
1
0
0
-1
1994
25
%
5
-1
1996
1998
2000
2002
2004
2006
2008
For other shocks (e.g. the unexpected announcement of a
reduction in the inflation target) it is even possible that the
long nominal interest rate falls on impact.
Reserve Bank of New Zealand and The Treasury
Figure 26 suggests that the public has a tendency to
above the target range than of inflation below the target
overestimate the current rate of inflation (mean current
range, and more tolerant of inflation above the centre of
inflation exceeds actual inflation by about one half of one
the target range than below the centre), has thus far not
per cent, although for median current inflation the gap
caused inflation expectation to drift significantly above the
is much smaller). We cannot be fully confident of this,
centre of the target range.26 There has been some minor
however, as we do not know whether the public’s subjective
drift, however, and it is therefore important to remind the
price index is the same as the CPI used by the national
markets of two key properties a credible inflation targeting
income statisticians. Allowing for the public’s overestimate
policy should have.
of current inflation, their inflation predictions one year and
two years ahead (reported in Figures 27 and 28) show
these medium-term inflation expectations reasonably well
anchored around the upper end of the target range (3 per
cent).
First, the inflation target is symmetric. Deviations above
the centre of the band (2 per cent) are considered of equal
importance as deviations of equal magnitude below the
centre of the band. Also, excursions outside the band are
equally serious when they occur at the upper bound (3%) as
A longer time-series of survey expectations one year ahead is
below the band (1%). Here the RBNZ may have a growing
given in Figure 29. It shows the dramatic decline in inflation
problem on its hands. There has never been a deviation
expectations from 1988 on, as well as the stabilisation of
of the inflation rate below the lower bound of the target
these expectations around 3 per cent in recent years.
range. There have been several excursions north of the
upper bound of the target range. We are currently in such
Figure 29
an episode, with year-on-year CPI inflation at 4.0% in June
Inflation expectations
16
%
2006, and with the near-certain prospect that the inflation
%
16
rate will exceed four per cent towards the end of the year,
before it starts to decline again. I would be surprised if
14
14
12
12
10
10
under these circumstances. Such drift of the nominal
8
8
anchor is dangerous. This drift of actual inflation toward
6
6
the top end of the band, and now well beyond it, was
4
4
2
2
0
0
1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006
Source: The National Bank, Business Outlook, April 2006.
Expected CPI inflation in 12 months time
Source: RBNZ J6 Survey of expectations - a survey of businesses,
March 2006 Report, www.rbnz.govt.nz/statistics/econind/
j6/data.html
Longer term inflation expectations backed out of 10
year index-linked and nominal bonds (so-called ‘break
even rates’) also show longer-term inflation expectations
inflation expectations did not continue to drift upwards
made easier because the inflation target in New Zealand
is a range rather than a point target. Replacing the target
band by a symmetric point target would help focus market
expectations and would make it more difficult for the RBNZ
to become a victim of “upward nominal anchor drift”.
Second, it is important to resist the siren song of “flexible
inflation targeting”, as advocated by Svensson (2001),
Woodford (2003) and many others.27 The objective
26
anchored around 2.5 per cent, one half of one per cent
below the upper bound of the target range.
It is interesting to note that although inflation has never
fallen below the lower bound of the target range, the
apparent asymmetry in the Bank’s pursuit of the inflation
objective (it appears to be more tolerant of inflation
Testing stabilisation policy limits in a small open economy
27
Econometric testing by Karagedikli and Lees (2004) suggests
that the Bank has not acted asymmetrically in its pursuit
of the inflation target. Their data cover the period 1994Q1
to 2002Q4. It therefore does not include the most recent
inflation overshooting episode. Their starting date of 1994Q1
is at least a year after the high inflation inherited from the
pre-inflation targeting regime had been ‘worked off’. Over the
period 1993Q1 till 2006Q2, the average year-on-year rate of
inflation for the target variable is 2.1%. The average centre of
the inflation target band is 1.5%
For a more extensive discussion of these points, see Buiter
(2006).
59
function of the monetary authority in this approach is
(3) Cov t ( y j , y *j ) = 0 (or the conditional covariance is
represented by the discounted sum of expected future
independent of monetary policy).
squared deviations of inflation from its (constant) target
(4) Vart y j = 0 (or the conditional variance of the efficient
level and of expected future squared deviations of output
from its natural or potential level. The typical period loss
function can be written as
Et (π j − π

(3)
where
πj
*
yj
level of output is independent of monetary policy).
Assumption (4) is pretty standard. Assumption (3) is highly
unlikely to be satisfied in most Old- or New-Keynesian
) + λ( y
2
j
− y )  , j ≥ t ; λ > 0

*
j
2
is the inflation rate in period j,
inflation target,
*
π*
is real output in period j, and
Et
potential output in period j,
λ
economy has the long-run natural rate property; it is
is the
not necessarily satisfied in the short and medium term.
y *j
Assumption (1) is a necessary condition for effective inflation
is
represents expectations
formed at the beginning of period t and
models. Assumption (2) is satisfied in the long run if the
is the weight
put on output gap stabilisation; the weight on inflation
targeting, at any rate in the long run. To assume that it is
automatically satisfied is to assume away all the technical
problems, commitment problems and other political
problems associated with inflation targeting. It is true that
targeting is normalised to 1.
for many of the most popular New-Keynesian and Old-
The period loss function given in (3) is a poor choice of
Keynesian models used to address inflation targeting, there
objective function for the policy authority: neither does
are few technical obstacles to meeting the inflation target
it have robust welfare economics foundations nor is it
on average. Indeed, these models all share the property that,
compatible with the price stability mandate given to the
when the inflation rate is, on average, equal to the constant
RBNZ, which is lexicographic in price stability (see Blinder
target rate of inflation, the output gap is, on average, equal
(2006), Buiter (2006) and Buiter and Sibert (2006)).
to zero. So, with the “first moment” problems of inflation
In addition, the flexible inflation targeting approach
targeting and output gap targeting solved, the monetary
compounds the problem by frequently replacing (3) by
policy maker is left with just the problem of choosing the
(4) Vart π j , j −1 + λ Vart y j
optimal combination of the conditional second moments of
where
Vart
inflation and output.
denotes the variance conditional on
information available at time t.
This trivialises the central problem of inflation targeting,
which is meeting the inflation target on average, going
In fact (3) is equivalent not to (4) but
forward, that is, achieving a zero inflation bias. When
Vart π j , j − 1 + λ Vart y j
E t π j = π * , the key problem of the inflation targeting
(5)
(
+λ Vart y *j + E t π j − π *
(E y
t
j
− E t y *j
)
2
)
2
+λ
− 2λ Cov t ( y j , y *j )
where Cov t denotes the covariance conditional on period t
information. It follows that for (4) to represent a period loss
function for the monetary authority that is equivalent to (3)
or (5), the following assumptions have to be made:
(1)
E t π j = π * : there is no inflation target bias (or the
inflation target bias is independent of monetary policy).
monetary authority, that of creating a credible nominal
anchor, is solved. This is difficult to achieve in practice, and
can never be taken for granted: the first moment problem is
also the first-order problem. Monetary authorities in the UK,
in the Eurozone, in the US, in New Zealand and in Turkey are
concerned, as this paper is being written, about the upward
drift of inflation expectations above their inflation targets or
tolerance ranges. The second-moments period loss function
(4), which assumes that there is no first-moments problem,
is an extremely misleading and dangerous construct to
dangle in front of the monetary authorities: the second
(2) E t y j = E t y : there is no output gap bias: the actual
moments are really of second order importance unless the
and optimal levels of output are the same on average (or
first order first moments problem has indeed been solved.
*
j
the output gap bias is independent of monetary policy).
60
Reserve Bank of New Zealand and The Treasury
The apparent similarity of Assumption (2), E t y j = E t y j
*
(no output gap bias) and Assumption (1), E t π j = π
*
Both these changes are unfortunate, because they put on
(no
a Bank a responsibility that it does not have the capacity to
inflation target bias) hides an important difference which
deliver: to stabilise output, interest rates and the exchange-
can come back to haunt policy makers. For models with
rate over and above what can be achieved as the natural by-
the (long-run) natural rate property, the servo-mechanisms
product of targeting price stability in the medium and long
of the market economy will tend to drive actual output
term. The illusion that monetary policy can systematically
towards potential output, at any rate in the long run, even
trade off inflation stability for output stability (let alone
without any policies designed to achieve that. There is no
output stability, interest rate stability and exchange-rate
such built-in mechanism for ensuring that the actual rate of
stability) will be shattered as surely as the older belief that
inflation will be driven towards the target rate of inflation,
monetary policy can trade off expected inflation for the
unless the policy authorities adopt rules (like the Taylor
expected output gap. I would urge the authorities to re-
rule) that ensure that this will be the case: there may be a
emphasize the primacy of the price stability objective and
natural rate of unemployment, a natural level of output and
not to endanger the long-term nominal anchor by the direct
a natural real rate of interest but there is no natural rate of
pursuit of other objectives, or by softening or widening the
inflation; the long-run equilibrium inflation rate is decided
target bounds on inflation.
by the monetary authorities.
The use of alternative nominal anchors to medium and
The original mandate of the RBNZ, like that of the Bank of
long-term inflation expectations, specifically the growth
England and the ECB, cannot be represented by a period
rate of some monetary aggregate at some frequency, does
loss function that trades off inflation volatility for output
not appear to offer much hope for success. The statement
volatility. That mandate is lexicographic in price stability and
attributed to Governor Gerald Bouey of the Bank of Canada
all other desiderata: only without prejudice to, or subject to,
that “In Canada, we did not abandon money supply targets,
the price stability objective being met, can other objectives,
they abandoned us,” applies equally to New Zealand.28
such as output or exchange-rate stabilisation be pursued.
No positive weight on output stabilisation, however low,
represents the lexicographic ordering of price stability.
Nor does a zero weight on output stabilisation represent
5
Foreign exchange market
intervention
a lexicographic ordering with price stability in the first
position. Output stabilisation can be, and is, valued, but
In 2003 the RBNZ conducted a review of its foreign
only without prejudice to the price stability objective.
exchange market intervention capacity and policies. Since
Unfortunately, the mandate of the RBNZ, as laid down in
successive PTAs, has been, if not watered down, at the
very least made fuzzier. As can be seen from Table 2, a
clause 4(c) was included in 1999 requiring the Reserve
Bank to have regard for “unnecessary volatility” in interest
rates, output and the exchange-rate, in the course of
conducting monetary policy. In addition, in 2002, clause
2(b), specifying the inflation target, was amended from
“12-monthly increases in the CPI” to keeping future CPI
inflation outcomes within the target band “on average over
the medium term”. This change made explicit the mediumterm focus for price stability, further enhancing monetary
policy flexibility.
Testing stabilisation policy limits in a small open economy
the NZD was floated in 1985, the Bank’s policy has been
to intervene only during periods of “extreme market
disorder” when operation of the foreign exchange market
itself is under threat. For this purpose, the Bank maintained
foreign exchange reserves in a target range of SDR 1.45bn
to 1.75bn. At the time of the review the exchange-rate was
NZD 1:SDR 0.45. The Bank has never had to intervene for
crisis management reasons during the past 21 years. As part
of the review the Bank recommended an increase in the
target range of reserves held for crisis management to the
range SDR 3.0-3.3bn.
28
An earlier and even longer version of this paper showed the
wild gyrations in the growth of narrow money (Notes and coin
held by the public and M1) and of broad money (M2 and
M3).
61
The Reserve Bank of New Zealand Act also allows the Bank
capacity to use foreign exchange market intervention in
to operate in the foreign exchange market for reasons
order to influence the level of the exchange-rate.
other than preventing disorderly markets and ensuring the
continued operation of the foreign exchange market. First,
the Bank can deal in the market on its own terms, provided
such intervention supports the implementation of monetary
policy. Second, the Minister of Finance can direct the Bank
to intervene, within set guidelines, to influence the level of
the exchange-rate. Note that while this ministerial power is
meant to be more of an emergency power than a power to
be used routinely, it could, if it were to be abused by the
Minister of Finance, completely emasculate the substance
of central bank independence.
The old policy was not so much about maintaining or
achieving any particular level of the exchange-rate as to
preserve the functioning of the foreign exchange market
in a crisis. The new policy goes beyond that, in that it
permits foreign exchange market intervention to be used
to influence the exchange-rate in support of the monetary
policy objectives set in the Policy Targets Agreement (PTA)
of September 2002. That PTA mandates the Bank to keep
future CPI inflation outcomes within the target band “on
average over the medium term” and to have regard for
“unnecessary volatility” in interest rates, output and the
In his review of the adequacy of the Bank’s reserves for crisis
exchange-rate, in the course of conducting monetary policy.
management, Gordon (2005) produces a table, reproduced
In principle, foreign exchange market intervention is now
below as Table 3, which compares New Zealand’s reserves
viewed as a second instrument of monetary policy, next
in 2004 with a peer group of 35 countries (all OECD
to the OCR. Like the setting of the OCR, foreign exchange
countries and some developed emerging market countries
market intervention would be decided and implemented by
with access to international capital markets.)
the Bank, independently of the Government.
Table 3
The new foreign exchange market intervention policy
Comparison of foreign exchange reserves
requires that before the Bank can intervene, all of the four
Reserves to
Value
NZ Ranking
following criteria must be met:
(1). The exchange-rate must be exceptionally high or low.
(2). The exchange-rate must be unjustified by economic
Annual GDP
6%
20/31
Imports
2.7 months
20/36
M2
6.9%
29/33
(3). Intervention must be consistent with the PTA.
27.3%
20/21
(4). Conditions in markets must be opportune and allow
3.1 days
21/32
Foreign short-term debt
(gross)
Daily foreign exchange
turnover
fundamentals.
intervention a reasonable chance of success.
Source: Gordon (2005)
Objective measures of what constitutes an optimal or even
adequate level of reserves don’t exist. Rules of thumb (most
of them originating from IMF programmes in developing
countries with little capital market access) suggest that
reserves should cover at least three months’ worth of
If foreign exchange market intervention is considered, all four
criteria make sense, although only the last one is objectively
verifiable: presumably, if the conditions in markets must be
opportune and allow intervention a reasonable chance of
success, intervention will be profitable to the Bank. To make
intervention possible and credible, the Bank has built up
financial resources of its own.
imports and at least 100 per cent of foreign debt maturing
This new approach towards foreign exchange market
in less than one year.
intervention raises both technical issues and more
On 30 March 2004, the Government of New Zealand
approved a proposal by the RBNZ that gives the Bank the
62
fundamental issues about monetary and exchange-rate
management in a small financially open economy.
Reserve Bank of New Zealand and The Treasury
Sterilised vs. non-sterilised intervention
“Giving the markets something to think
The analyses of Gordon (2005) and of Eckhold and Hunt
about”
(2005) of the mechanics of foreign exchange market
A frequently heard argument for the Bank building up a
intervention assume that pressure of the Bank’s interventions
capacity for foreign exchange market intervention is that it
on the exchange-rate is measured by the change in the
“will give the markets something to think about”. Speculators
Bank’s net long foreign currency position: an increase in the
will be less willing to take open positions in the NZD, when
net long foreign currency position (an increase in the value
there is a risk that the Bank will intervene in the market.
of the Bank’s foreign currency assets relative to that of its
Even if speculative (open) positions that are not justified by
foreign currency liabilities) weakens the domestic currency.
fundamentals are taken by market participants, the welfare
A reduction in the net long foreign currency position or an
economics of countering such “destabilising” speculation
increase in the net short currency position strengthens it.
by deliberately increasing uncertainty are by no means well
It must follow that the intervention referred to in these
papers is sterilised intervention, and this is confirmed in
official statements by the Bank (e.g. “Selling and buying
worked out. Regardless of the welfare economics, the ability
of the authorities to intervene successfully is doubtful, and
subject to a fundamental asymmetry.
New Zealand dollars might initially seem to involve
When the NZD strengthens to a degree deemed undesirable
changing monetary policy by altering the New Zealand
by the authorities, they can, in principle, engage in foreign
money supply. However, it is important to note that the
exchange market interventions (sales of the NZD) of any
intervention would automatically be “sterilised” to undo
magnitude, as long as the fiscal consequences of such
the effect on the money supply. This is standard practice
interventions are tolerable. The Bank can either sell NZD-
internationally”). Foreign exchange market intervention
denominated Government debt it owns or issue its own
that does not involve a change in any country’s monetary
base but still affects the exchange-rate does so because of
NZD-denominated liabilities and use the proceeds to build
up its stock of reserves.
the imperfect substitutability of otherwise identical non-
It is not possible, however, for the Bank to prevent an
monetary financial instruments denominated in different
undesirable weakening of the NZD by selling arbitrarily large
currencies. Convincing direct or indirect evidence on this is
amounts of foreign exchange reserves. Its stock of foreign
not easy to come by.
exchange reserves is limited and must be kept above the
minimum level (currently SDR 2.45 bn) agreed with the
Minister of Finance to manage situations of “extreme
Intervention in spot market, forward market
market disorder”. Its ability to borrow foreign exchange
or option markets
reserves is limited by the willingness of other governments
The relevant Bank documents (Eckhold and Hunt (2005)) refer
and international financial institutions to take on increased
to a number of alternative ways in which sterilised foreign
NZD exposure during periods of NZD weakness, and by the
exchange market intervention could be implemented. It is
market’s perception of liquidity risk and default risk.
clear, however, that foreign exchange market intervention,
The magnitude of the Bank’s resources versus daily turnover
if conducted through the options markets (say by selling call
in the foreign exchange markets (shown in Figure 30)
options on the NZD with a strike price below the current
gives one reasons to wonder about the effectiveness of
market price when there is (unwanted) upward pressure
intervening to prevent an excessive weakening of the NZD.
on the external value of the NZD), will not be associated
with the same easily identified and measured changes in the
Bank’s net long foreign currency position that, in the view
of the Bank, represents the metric for the pressure of the
Bank’s interventions on the exchange-rate.
Testing stabilisation policy limits in a small open economy
63
Figure 30
get a sense of perspective, before the events leading up
Foreign exchange market daily turnover
to “Black Wednesday”, 16 September 1992, when the
US$m
30000
25000
Spot
Total
Outright forwards
Swaps
US$m
30000
UK was forced to leave the Exchange-rate Mechanism of
the European Monetary System, the Bank of England has
25000
been estimated as having spent £10bn (about $18bn at the
20000
20000
exchange-rate prevailing at that time) of the UK’s foreign
15000
15000
exchange reserves in an unsuccessful attempt to keep
10000
10000
5000
5000
0
Jul-04
0
Sep-04 Dec-04 Mar-05 Jun-05 Sep-05 Dec-05 Mar-06
Source: Reserve Bank of New Zealand
Eckhold and Hunt (2005, p. 20 ,fn 13), report the Bank as
having foreign exchange reserves for crisis intervention of
SDR 1.6 billion. These reserves are scheduled to grow to
SDR 2.45 billion over the next few years. During April 2006,
the exchange-rate of the SDR and the US$ was about 1
sterling in the ERM.
There is an obvious asymmetry in a central bank’s ability
to defend an undervalued and an overvalued parity. This
means that if foreign exchange market intervention to
influence the value of the exchange-rate were to be
effective, its effects would most likely be asymmetric: there
would be more intervention and more effective intervention
to weaken the NZD than to strengthen it. This would impart
an upward bias to the inflationary process.
SDR=1.45US$, current reserves are about US$ 2.32bn and
future target reserves about US$ 3.55bn. Figure 30 shows
Effectiveness
daily turnover in the New Zealand foreign exchange market
Even if reserve adequacy is not an issue, the effectiveness
between July 2004 and March 2006. Spot transactions
of foreign exchange market intervention at the frequencies
involving the NZD regularly top US$ 2.0bn; outright
that matter for macroeconomic policy is an open question.
forwards are rarely more than US$ 1bn; swaps dominate
There is a vast and largely inconclusive literature on the
and have topped US$ 12bn in a day. Total turnover has been
effectiveness of foreign exchange market intervention in
as much as US$16bn in a day.
countries that are highly integrated into the international
Clearly, foreign exchange turnover in the New Zealand forex
financial markets (see e.g. Sarno and Taylor (2001)). My
market (the gross flows of foreign exchange transactions
reading of the evidence prompts the following verdict: if
during a period) may well be a poor proxy for the market’s
foreign exchange market intervention does not convey
willingness to take and hold an open position in NZDollars
news about future monetary policy, that is, news about
(the stock of foreign exchange reserves or NZDollars
the future path of the OCR, then there is no lasting effect
willingly held an any point of time): a given stock can
on the exchange-rate. Such effects as do occur are at very
turn over many times during any given period. I do not
high frequencies (hours, days, a couple of weeks at most).
have data on the capacity of individual private investors
No effect is sufficiently persistent to be of macroeconomic
to hold NZDollars (the size of their balance sheets and any
policy relevance. These high frequency changes in asset
external or internal prudential or internal constraints on
prices and yields are of course of great financial relevance
single currency exposure) or, given that capacity, on their
to profit-oriented market participants with significant net
willingness to hold open positions in the NZDollar. Figure
open positions.
30 is, however, consistent with the view that even the new
Can foreign exchange market intervention act as a signal
target level of foreign exchange reserves would be no more
about the future behaviour of the policy instrument that
than a light lunch for the foreign exchange market. Faced
really does matter – the short-term interest rate set by the
with a determined selling attack on the NZD, the stock of
Bank, that is, the OCR? Logically it can, but why should
foreign exchange reserves would be exhausted swiftly. To
the authorities wish to signal using as a signalling device
64
Reserve Bank of New Zealand and The Treasury
something that only matters if and to the extent that it is
No intervention yet, but how close did
perceived as a signal? The only reasonable explanation for
we get?
using a prima facie spurious signal is that the use of the
I have seen no evidence to suggest that there has been
signal is not cheap talk, but represents a commitment by
any foreign exchange market intervention since the NZD
the Bank that could result in a costly loss if the Bank does
was floated in 1985. However, the accepted view of
not in the future act according to the signal that was given.
market commentators is that New Zealand came close to
Foreign exchange market intervention means taking a larger
intervening in December 2005, when the NZD hit 0.72
open foreign exchange position. Assume the authorities
USD on a couple of occasions. According to these sources,
want to see a weaker NZD and signal future OCR cuts
the RBNZ had decided that at least the first three of the
(the fundamental that will ultimately justify a weaker NZD)
four necessary conditions for foreign exchange market
through purchases of additional foreign exchange reserves.
intervention to influence the exchange-rate had been met,
If the authorities do not follow through on the interest rate
and that the fourth condition (a reasonable likelihood of
cut, the NZD would strengthen again, resulting in a marked-
success) was also likely to be met; apparently the Bank did
to-market loss for the Bank on its increased stock of foreign
not intervene because it hoped that the market would do
exchange reserves. So, to the extent that foreign exchange
the job for them, as it did. The NZD now (July 19, 2006)
market intervention is indeed perceived by the market as
trades at 0.62 US$, after hitting a low of 0.61 US$ in March
“putting your money where your mouth is”, it may be an
2006. I would expect the RBNZ to be quite comfortable
effective signal for announcing future OCR decisions.
with this level and to view some further weakening with
equanimity, provided it does not become a rout.
The danger of creating multiple nominal
targets
The most serious danger associated with having foreign
exchange market intervention as an instrument is that
6
Alternative stabilisation
instruments
instead of being seen and used solely as an additional tool
The KiwiSaver as a contributor to the
(of doubtful effectiveness) to support the achievement
automatic stabilisers
of the inflation target, it will tempt the monetary policy
makers into pursuing two nominal targets, inflation and the
exchange-rate. The pursuit of two nominal targets amounts
to the pursuit of one real target – the real exchange-rate
in this case. Such a development would be disastrous. Not
only is the central bank incapable of successfully targeting
the real exchange-rate, the vainglorious attempt to target
the real exchange-rate is likely to lead to a Wicksellian
inflationary explosion or deflationary implosion. It is difficult
enough to target a single nominal objective. Targeting more
than one, or targeting one nominal objective (the rate of
inflation) subject to another nominal constraint (the nominal
exchange-rate) would be an act of hubris and folly.
The KiwiSaver, planned to commence on 1 April 2007, is
a government-sponsored occupational saving plan. From
that date, all new employees are automatically enrolled
into KiwiSaver, but can opt out if they so wish. Existing
employees and individuals not employed can opt into
KiwiSaver. Contributions will be deducted by employers
(along with PAYE) at a rate of 4% (default) or 8% of gross
wages or salary. Members can choose among KiwiSaver
schemes with different investment risk profiles, with a
default scheme for those who do not want to choose.
Savings are locked-in until the age of eligibility for the state
pension, New Zealand Superannuation (NZS), currently 65
years or five years membership (whichever is longer), with
three exceptions: first home purchase, serious financial
hardship and permanent emigration. The government will
also make a NZ$1000 start-up contribution to each new
account.
Testing stabilisation policy limits in a small open economy
65
The purpose of the KiwiSaver is to boost retirement saving
up crown financial assets, or they could be transferred to
by middle income New Zealanders to help them avoid a
the private sector, preferably in a way that will not lead to
significant drop in living standards during retirement. An
an immediate consumption boom, by paying them into the
assessment as to whether such a drop in living standards
KiwiSaver accounts, say.29
during retirement is likely and, if so, a cause for government
intervention, is beyond the scope of this paper. What
matters for stabilisation policy is the possible effect of the
“soft compulsion” element of the KiwiSaver (the default is
“you’re in”, you have to actively opt out) on the savings
Proposal 4: Use (part of) any revenue windfall to fund
the individual KiwiSaver accounts. The stock of foreign
exchange reserves beyond what is deemed prudent to avoid
disorderly markets could also be transferred to households
in this way.
behaviour of those enrolled in it.
Also, since there is no economic rationale linking retirement
Neoclassical economics with efficient financial markets
implies there will be no effect on aggregate saving from
the introduction of the KiwiSaver; there would be none
even if there were hard compulsion (mandating) rather
saving with the first home purchase, it would make sense
to eliminate the option that the KiwiSaver account can be
drawn on for first home purchase before reaching the age
of retirement.
than soft compulsion. Neoclassical economics with cashflow, liquidity, borrowing and dissaving constraints implies
there will be no effect on aggregate saving if it is not truly
mandatory; those for whom the constraints are binding
will all opt out. Behavioural economics implies it will raise
If the government is seriously concerned about undersaving by low and middle-income workers, it should make
contributions to the KiwiSaver accounts mandatory. Political
unpopularity should not stand in the way of a good idea.
aggregate saving.
If the government’s rationale for the scheme works
Migration policy as an automatic supply-side
out, there will be no full offset through a reduction in
stabiliser
discretionary non-KiwiSaver saving, and the introduction
Migration flows affect fluctuations in the level of activity
of the KiwiSaver itself would, in a modest way, increase
both from the supply side, by changing the supply of labour,
the effectiveness of the automatic stabilisers, by raising s in
and from the demand side, by influencing the demand for
equation (1). The government could do more than making
goods and services, and in particular housing demand.
a one-off NZDollar 1000 initial deposit for each individual
KiwiSaver account. Instead of using revenue windfalls for
paying down the remaining public debt, or for increasing
public sector financial assets (though the prefunding of the
NZS or through some other broad ‘Crown National Fund’
not tied to a specific category of future public spending
commitments) it could make further equal per-account
payments into the KiwiSaver accounts.
This could, for instance, be a good use for the stock of
foreign exchange reserves that has been built up for
possible use in foreign exchange-rate interventions aimed
at influencing the value of the exchange-rate. If my view
that foreign exchange-rate intervention is not an effective
tool for macroeconomic stabilisation is correct, these
reserves are redundant and can be allocated to alternative
uses. They could be used to pay down crown debt or build
66
Net permanent and long-term migration, NI, is the difference
between gross permanent and long-term immigration,
GI, and gross permanent and long-term emigration, GE.
Immigration can be decomposed into inflows of persons
who have prior residence rights and work permits in
New Zealand, GIR, and those who do not, GINR. Emigration
29
The fact that, historically, RBNZ foreign exchange reserves
were funded from public debt has no implications for the
optimal use of these resources should it be concluded that
less of them henceforth need be held as foreign exchange
reserves. The realisation that foreign exchange market
intervention aimed at influencing the level of the exchangerate is ineffective at the cyclical frequencies (several years)
that matter for stabilisation policy, is, from an economic point
of view, like technical progress, or like an improvement in
the economic environment. The same policy outcomes can
be achieved with a lower average stock of foreign exchange
reserves. One would expect that the optimal reallocation of
the now-redundant foreign exchange reserves would involve
both some paying down of outstanding public debt and some
capital transfers to the private sector.
Reserve Bank of New Zealand and The Treasury
can be decomposed similarly into outflows of persons
Figure 32
who have prior residence rights and work permits in
Immigration, building activity and house prices
New Zealand, GER, and those who do not, GENR, although
the last category must be very small.
40
%
30
NI = GI − GE
= GIR + GINR − GER − GENR
The New Zealand authorities can, in principle, control GINR,
Number
Building consents issued, real value % growth rate (LHS)
Work Put in Place, Residential Buildings % real growth rate (LHS)
House price to earnings ratio, % change (LHS)
Net permanent and long-term immigration (RHS)
50000
40000
20
30000
10
20000
0
10000
the inflows of those without established residence rights.
-10
0
The other contributors to net immigration, GIR, GER and
-20
-10000
GENR are not a policy instrument.
-30
-20000
1990
1992
1994
1996
1998
2000
2002
2004
From Figure 31 it is clear that, on balance, net immigration
Source: Statistics New Zealand
has been procyclical.
There appears to be no systematic relationship, however,
Figure 31
between such housing price indicators as the change in the
house price to earnings ratio and net permanent and long-
Immigration and GDP growth
Number
50000
40000
%
Net permanent and long-term immigration (LHS)
Real GDP growth rate (March) (RHS)
term migration. Other immigration measures may, however,
10
8
indicators and net migration. Further investigation of the
6
timing and magnitude of the supply-side and demand effects
30000
20000
10000
reveal a more systematic relationship between house price
4
of net migration, and of the capacity of the New Zealand
authorities to manage the net migration rate would seem
0
2
-10000
-20000
-30000
to be required before a confident recommendation can be
0
made as to the use of pro-cyclical GINR as a supply-side
-2
cyclical stabilisation device.
1988 1990 1992 1994 1996 1998 2000 2002 2004
Source: Statistics New Zealand
If it could be established that the supply-side effects
It can be argued that this procyclical behaviour of net
of migration are indeed significant at business cycle
immigration would dampen the business cycle (defined as
frequencies, this pro-cyclical net migration flows could
the gap between actual and potential output), by raising
be an interesting complement to the automatic fiscal
the effective labour supply and thus potential output. There
stabilisers. Given the high degree of integration of the
is also, of course, an effect of migration on the demand
New Zealand labour market with that of Australia, this
for goods and services and through that an effect on the
supply-side stabilisation mechanism could be effective even
derived demand for labour. A frequently heard view in
without discretionary changes in immigration policy, as long
New Zealand, holds that the effect of migration on labour
as the business cycles of New Zealand and Australia are not
supply is subject to significant delays, while there is an
perfectly synchronised.
immediate effect on aggregate demand, especially through
the demand for housing created by new immigrants (and
the increase in the supply of housing created by emigrants).
The evidence on this, is however, rather patchy, as suggested
by Figure 32, which shows that the growth rate of building
consents issued and work put in place appears to lead net
permanent and long-term migration.
Testing stabilisation policy limits in a small open economy
67
Mitigating asset market boom and bust
Figure 33
I will take as given the view that monetary policy (the
Real house price index
OCR rule) should respond to asset market developments
Index
140
if and only if at least one of the two following conditions
120
120
100
100
80
80
60
60
40
40
is satisfied: First, the asset market developments have
implications for current and future developments of the
inflation target that have to be taken into account because
Index
140
of the Bank’s primary mandate: price stability; second, the
asset market developments have welfare implications other
than current or future inflation, that can be addressed by
monetary policy without prejudice to the primary price
20
20
0
2003 = 100
1990
1992
0
1994
1996
1998
2000
2002
2004
stability objective.
Source: Statistics New Zealand,
QVNZ
On the basis of global experience thus far, it seems unlikely
While large, this increase can be accounted for in its entirety
that monetary policy is an effective instruments for
by the fall in the long-term real interest rate over the period.
preventing or mitigating asset bubbles. It can be reasonably
This combined fall in the risk-free real interest rate and
effective to help clean up the mess that results when asset
in virtually every credit risk premium was a world-wide
bust follows asset boom. The reason monetary policy
phenomenon and therefore at most a global bond market
should not target asset prices is simple. First, monetary
bubble without a New-Zealand specific component. The
policy should not try to influence asset prices that reflect
real price of a house can be written as
fundamentals, even if these asset prices move fast and
furiously. Second, monetary policy is not the appropriate
(6) P =
ρ
r−g
tool for influencing asset price movements that are not
where P is the real house price, ρ is the current real rental
driven by fundamentals, that is, monetary policy is not an
rate of housing, r is the permanent or long-run value of
effective tool for bursting or mitigating bubbles. You don’t
the appropriate risk-adjusted real discount rate and g is the
hunt bubbles with fundamentals. At most, if the monetary
permanent or long-run real growth rate of housing rentals.
authorities are sufficiently confident that a given observed
It follows that
pattern of asset price movements does indeed represents a
(7)
bubble, they should use open mouth operations (warnings
d ln P = d ln ρ −
1
d (r − g )
r−g
about irrational exuberance or irrational despondence) to
If the pre-boom value of the real risk-adjusted discount rate
try and prick the bubble.
for housing was 6%, and the growth rate of real housing
Let me start by saying that I find the evidence that there
rentals 3%, any combination of a reduction in the long-run
have been bubbles in key New Zealand asset markets,
risk-free rate, a reduction in the housing risk premium and
specifically the housing market and the stock market, less
an increase in the growth rate of housing rentals adding
than convincing. As shown in Figure 33, between 2001 and
up to 2%, would raise the housing price index by 66.6%.
2005, the real house price index rose by 60 per cent (this is
The combination of a falling discount rate and an upward-
also the increase in the house-price to earnings ratio over
revision in the growth rate of housing rentals of such a
the same period).
magnitude is not at all implausible. The general health and
growth prospects of the New Zealand economy are strong
and, probably for the first time, New Zealand residential
property (especially coastal property) has become a target
for foreign investors, both from Australia and, increasingly,
from emerging Asia. Such a ‘level’ effect on housing prices
68
Reserve Bank of New Zealand and The Treasury
of becoming part of the global housing market is probably
market value. Taxing realisations rather than accruals
better represented by an increase in ρ than by an increase in
may be administratively easiest. The alternative is to tax
g, but this does not matter for its net effect on valuations.
realisations where possible and to impute capital gains on
property that is not traded during the tax year on the basis
of regional indices of comparable house prices. One could
Fiscal and regulatory instruments to address
even integrate a capital gains tax on land with the general
asset market anomalies
capital gains tax I proposed earlier.
The use of fiscal policy instruments or regulatory measures
for controlling or mitigating asset market booms have been
proposed in New Zealand and elsewhere. I shall focus on
the housing market, or property markets generally, and on
the markets for housing (or property) finance.
A necessary condition for any of these measures to make
sense is that the authorities are capable of distinguishing
asset price bubbles from asset price movements driven by
In all this it is key to tax values or capital gains, regardless
of the kind of property they are attached to, and regardless
of the intentions of the buyers and sellers of the properties.
Focusing the land tax on the taxation of profits on properties
purchased for resale is without economic merit.
Proposal 5: Introduce a land tax on all land, regardless of
what it is used for.
fundamentals, and that policy instruments can respond
with sufficient speed and accuracy to mitigate rather than
Measures to avoid: supplementary
exacerbate asset market bubbles. The relevant lags include
stabilisation instruments
both the “inside lags” of the policy making process and the
“outside lags” of the effects of the policy instruments, once
triggered, on the variables of interest.
The recent Initial Report by the Governor, Reserve Bank
of New Zealand and Secretary to the Treasury (2006),
Supplementary Stabilisation Instruments, is “a little shop
A second necessary condition for any policy measures
of horrors” of regulatory and fiscal interventions in asset
targeted at asset market bubbles to make sense is that the
and credit markets, that would fail to stabilise anything of
cure does not do more damage than the disease. This rules
value while creating massive distortions, disintermediation
out the vast majority of regulatory interventions in the asset
and rent-seeking behaviour. There is just one small pearl
markets. It also rules out most fiscal interventions, whether
among the swine – streamlining the planning process for
in the property markets themselves or in the markets for
house building.
financing property.
Most of the proposals amount to the discretionary use
The one exception of a fiscal measure that could effectively
of the regulatory regime and prudential norms of the
dampen property price fluctuations without causing
financial sector for short-run cyclical stabilisation. Clearly,
distortions is, unfortunately, a political tar baby. It is the land
the permanent or long-term features of the regulatory and
tax, either on the value of the unimproved land or on the
supervisory regimes are important parts of the transmission
capital gains on the value of the unimproved land.
mechanism of shocks that drive the business cycle, just as
A land tax
A tax either on the value or on the capital gains on “land”,
the unimproved value of real estate, has many efficiency
and fairness arguments in its favour, quite apart from its
possible usefulness in mitigating asset market booms in
real estate markets. The contribution of “nature’s gift”
to the value of real estate can be obtained quite easily
by subtracting the insured value of the property from its
Testing stabilisation policy limits in a small open economy
the automatic fiscal stabilisers associated with a given tax
and benefit structure. Optimising the permanent, structural
features of the regulatory and supervisory regimes to
mitigate excessive fluctuations is desirable if this can be done
without prejudice to their primary task: enhancing systemic
stability and preventing financial crises. Attempting to do
more, say by tightening prudential constraints on lending
activity in a manner intended to be countercyclical, would
69
overburden both the regulators and the instruments at their
the problem but excessive borrowing secured against
disposal.
residential property? As opposed to unsecured loans or
loans secured by commercial and industrial property?
I shall briefly touch on the main proposals, to indicate
where and how they amount to an abuse of regulatory
•
Discretionary mortgage interest levy. See the previous
and supervisory processes and norms in doomed pursuit of
bullet point. What is the problem that is targeted?
cyclical stabilisation.
Are house prices to high? Are interest rates too low in
•
general? If so, use monetary policy. Are just mortgage
Tax on property purchased for resale. Why is it more
interest rates too low? Lenders will metamorphose
virtuous/efficient to own and hold property than to buy
mortgages into formally non-secured but de-facto
and re-sell it? This is an expression of an atavistic fear of
secured loans.
speculation. It makes no economic sense and smacks of
•
populist pandering. On balance, speculation enhances
If it were possible to design a tax on all interest (and
efficiency and should be encouraged, within a proper
on all financial contracts that involve payment that are
regulatory framework and with proper prudential
economically/functionally equivalent to interest), this
safeguards.
would clearly give the RNBZ a much bigger club with
which to hit not just the short maturity interest rates
Ring-fencing (preventing operating losses on investment
but the whole term structure of interest rates. Granting
properties being offset against other income). Again,
such an obviously fiscal tool to the central bank would
this is economic nonsense. Losses are losses, as long as
be politically problematic (even if the revenues from the
the activity in question is legal.
interest rate tax were transferred to the government).
•
Linking bank capital to cyclical risk. For this to be counter-
I am also not convinced that the OCR is not sufficient
cyclical, capital requirements would have to be lowered
to do the job of achieving price stability in the medium
in a cyclical downswing when solvency risk and other
term.
prudential risk are highest and raised in the upswing.
This would be Basel II on its head and on steroids. It is
indeed true that Basel II capital requirements imposed
for essential regulatory/prudential purposes inevitably
have some “automatic” linkage that renders them
procyclical. If these rules can be revised to make them
less pro-cyclical, that would be a plus not just for
New Zealand, but globally. Using key prudential ‘ratios’
•
Measures to increase the speed at which new land
and houses are able to be brought onto the market
in response to evidence of rising demand. Simplifying
and speeding up planning and zoning applications and
reducing the red tape faced by the building industry is
desirable regardless of its supply-side contribution to
stabilisation policy.
actively in a countercyclical manner is, however, a bridge
too far.
•
70
Discretionary loan to value ratio limit on loans secured
Other policy measures to avoid
by residential property. If enforced on the lender,
Incomes policy
disintermediation among instruments and lending
Incomes policy, or prices and incomes policy, is the direct
institutions will result swiftly. If enforced on the
control of the state over prices and wages. Obviously,
borrower, it will be highly intrusive, but the scope for
the state plays a role in setting wages for public sector
disintermediation is more limited, because individual
employees and in bargaining over prices for goods and
households are less able to bear the fixed cost of
services sold to or purchased from the private sector by the
disintermediation. It is not clear what the problem(s) is
state. In New Zealand, public sector employment (including
(are) that this is meant to address. Is it a price bubble in
health and education) is around 25% of total employment.
the housing market? Or is it not house prices that are
The state as employer therefore is an important player in
Reserve Bank of New Zealand and The Treasury
the national labour market. Through state procurement
avoided, would only have value as a source of revenue. Its
and through the sale of goods and services produced by
stabilising effect on exchange-rates would, for any realistic
the state, the state is a player with some monopsony and
tax rate, be negligible.
monopoly power in a range of markets. One can only hope
that public sector wages and salaries are set with reference
to the opportunity cost of the appropriate labour categories
in the private sector and that similar market-conform
benchmarks are used for pricing other goods and services
sold to or by the public sector.
7
Conclusion
The most important conclusion is that there is little that is
wrong with the current monetary and fiscal framework.
It has produced a reasonable degree of price stability in
Beyond the direct and unavoidable role of the state as
the medium term and healthy trend growth. Economic
employer, procurer of goods and services or seller of same,
fluctuations, in GDP growth and employment are significant,
there is no case for state intervention in wage or price
but monetary policy should not be used more actively to try
setting in the economy, whether through hard ceilings on
and close the output gap. A retreat from the ambitions of
price or wage increases, through the taxation of price or
flexible inflation targeting seems desirable: the authorities
wage increases above the officially sanctioned norm, or
should not try to trade off inflation volatility for output (gap)
through softer “guide posts”. The history of incomes policy
volatility. Instead a more lexicographic approach to inflation
across the world is a sorry tale of short-term “success”, that
targeting, with the output gap targeted for its own sake
is, a reduction in recorded wage or price inflation without
(as opposed to as a predictor of future inflation) only when
any demand-restricting monetary or fiscal measures,
this can be done without prejudice to the inflation target,
accompanied by a steady accumulation of distortions and
may well result in greater ex-post stability of both inflation
build-up of wage and price pressures, culminating in a wage
and output. Not only does flexible inflation targeting
and price explosion which then has to be painfully corrected
violate the price stability mandate of the RBNZ. Through its
with restrictive monetary and fiscal policy.
focus on trading off inflation volatility for output volatility
it de-emphases the primary goal of meeting the inflation
Capital controls, foreign exchange restrictions and the
Tobin Tax
target on average, going forward. In the most extreme,
but unfortunately quite common, manifestation of flexible
I include these measures only for completeness, and in case
inflation targeting, which focuses exclusively on the trade-
some benighted soul would inadvertently advocate them.
off, now and in the future, between the variance of inflation
Whatever may be the merits of phasing out existing capital
and the variance of real output, the nominal anchor has
controls and foreign exchange controls gradually in an
become completely invisible. This is grist to the mill of those
emerging market or developing country that does not yet
who believe that there is an exploitable trade-off not only
have free international mobility of financial capital, the case
between the conditional second moments of inflation and
for re-introducing such controls in an advanced industrial
output, but also between their conditional first moments.
economy where they have been absent for 20-odd years is
non-existent. Putting the capital mobility genie back in the
The inflation target of the RNBZ should be reformulated as
a symmetric point target.
bottle would be administratively costly, lead to wide-spread
evasion, rent-seeking and corruption and, to the extent that
it would be effective, would distort intertemporal trade and
The creation of a Monetary Policy Committee to take
interest rate decisions would be desirable.
international risk-sharing. The Tobin tax (a tax on foreign
Foreign exchange market intervention to influence the level
exchange transactions), to the extent that it cannot be
and volatility of the exchange-rate would be spitting against
the wind.
Testing stabilisation policy limits in a small open economy
71
Giving up foreign exchange market intervention under
As regards enhanced fiscal stabilisation, only modest
any circumstances other than the prevention of disorderly,
offerings are available. The automatic stabilisers could be
illiquid markets would make sense but would probably not
strengthened by broadening the GST base through the
be credible. The best alternative is to retain the option of
elimination of the exemptions for financial services and
using intervention in extremis, but to be determined never
housing rentals. Further base-broadening could be achieved
to exercise the option.
by taxing capital gains at the same rate as all other income.
The only monetary regime change that could make a serious
Active use of the GST rate would help stabilise demand
contribution to macroeconomic stability would be monetary
both through Keynesian automatic fiscal stabiliser effects
union with the only possible candidate: Australia. As noted
and through classical intertemporal substitution effects.
earlier, while Australia only accounts for just over 20% of
exports and imports, it is the single largest national trading
partner. Furthermore, New Zealand and Australia have an
integrated labour market; if there is significant pro-cyclical
net labour mobility between New Zealand and Australia,
this would reduce the cost of the loss of nominal exchangerate flexibility vis-à-vis Australia.
The introduction of the KiwiSaver may have some positive
effect on the automatic stabilisers, apart from its role in
boosting private saving in the long term. Revenue windfalls
could, in part, be used to further capitalise the KiwiSaver
accounts, over and above the NZD 1000 start-up payment
the government are already committed to make. The (in
my view) redundant additional foreign exchange reserves
All the important New Zealand commercial banks are
accumulated by the RBNZ for a more active foreign exchange
Australian-owned. This means that there is, between
market intervention policy, could also be privatised (in part)
New Zealand and Australia, more than the near perfect
by transferring them (in part) into the KiwiSaver accounts.
financial capital mobility that exists between New Zealand
and virtually every developed nation. There is in addition the
almost full integration of the New Zealand financial sector,
as provider of financial products and services, with that of
Australia.
immigration policy could (by relaxing the labour supply
constraint during booms) help stabilise the economy, despite
the stimulus increased immigration provides to aggregate
demand, especially in the housing market.
The question of the synchronisation of business cycles
between New Zealand and Australia (or the frequency
and significance of asymmetric shocks and the asymmetric
transmission
There should be further study of whether a more procyclical
of
common
shocks)
requires
further
investigation. Traditional optimal currency area theory
considers asymmetric shocks a reason for not having a
monetary union. From a “real portfolio diversification”
Of all the fiscal and administrative instruments that have
been proposed for addressing the housing boom, the only
one that makes economic sense is Henry George’s land tax,
which could be levelled either on the value of the land or on
the capital gains. To avoid creating distortions, rent-seeking
and corruption, it should apply to all land, residential,
commercial, industrial and agricultural.
perspective, asymmetric shocks (especially asymmetric
shocks to labour income) imply the possibility of risk
sharing, and call for greater integration, including monetary
integration. Financial market integration is deeper when
countries or regions share a common currency.
In the final analysis, a small open economy like New Zealand
will be characterised by fluctuations in the level of economic
activity, driven by external shocks (terms of trade shocks,
changes in global interest rates, risk premia and valuations in
global financial markets), by internal supply shocks (weather),
Of course, monetary union is a deeply significant symbolic
political act, involving a transfer or sharing of sovereignty, as
well as a technical economic issue. From a purely economic
point of view, however, monetary union with Australia is the
only alternative to the (slightly fine-tuned) current monetary
shocks to internal animal spirits and by policy shocks, which
can never be avoided completely. New Zealand’s overall
monetary and fiscal framework is among the best thoughtout and most far-sighted I have ever come across. My final
advice is therefore not quite “don’t worry, be happy,” but:
regime that would make sense.
72
Reserve Bank of New Zealand and The Treasury
Don’t change a winning game; don’t raise your ambitions
Buiter, Willem H (1985), “A Guide to Public Sector Debt and
for dampening cyclical fluctuations to the point that you
Deficits,” Economic Policy, 1(1), November 1985, 13-79.
endanger the hard-gained fiscal-financial sustainability
achieved by the public sector and the allocative efficiency
that characterises most of the private sector.
Buiter, Willem H (2000), “Is Iceland an Optimal Currency
Area?,” in Már Gudmundsson, Tryggvi Thor Herbertsson
and Gylfi Zoega, eds., Macroeconomic Policy; Iceland in
an Era of Global Integration, University of Iceland Press,
Reykjavik, 2000, pp. 33-55.
References
Buiter, Willem H (2006), “How Robust is the New
Auerbach, Alan J and Roger H Gordon (2002), “Taxation
Conventional Wisdom in Monetary Policy? The surprising
of financial services under a VAT,” The American Economic
fragility of the theoretical foundations of inflation targeting
Review: Papers and Proceedings, Vol. 92, No. 2, pp. 411-
and central bank independence,” mimeo, European
416.
Institute, London School of Economics and Political Science,
Blinder, Alan S (2006), “Monetary Policy Today: Sixteen
June 2006. Paper presented at the 2006 Central Bank
Questions and about Twelve Answers,” paper prepared for
Governors’ Symposium “Challenges to Monetary Theory,”
Bank of Spain conference, Madrid, June 2006, forthcoming
at the Bank of England, on June 23 2006.
in conference volume.
Buiter, Willem H and Anne C Sibert (2006), “The elusive
Blinder, Alan S and John Morgan (2005), “Are Two Heads
welfare economics of price stability as a monetary policy
Better than One? Monetary Policy by Committee,” Journal
objective: why New Keynesian central bankers should
of Money, Credit and Banking, Ohio State University Press,
validate core inflation,” Mimeo, August.
vol. 37(5), pp 789-811, October.
Drew, Aaron (2001), “Lessons From Inflation Targeting in
Bollard, Alan and Özer Karagedikli (2005), “Inflation
New Zealand,” Central Bank of Chile Working Papers No.
Targeting: The New Zealand Experience and Some Lessons,”
113.
18 January; paper presented at the Inflation Targeting
Performance and Challenges Conference organised by
the Central Bank of Republic of Turkey, held in Swisotel,
Eckhold, Kelly and Chris Hunt (2005), “The Reserve Bank’s
new foreign exchange intervention policy,” Reserve Bank of
New Zealand: Bulletin, Vol 68, No.1, March pp.12-22.
Istanbul. 19-20 January 2005.
Evans, Lewis, Arthur Grimes, Bryce Wilkinson with David
Buckle, R A, K Kim, H Kirkham, N McLellan and J Sharma
(2006), “A structural VAR model of the business cycle for a
volatile small economy,” mimeo.
Buckle, R A, K Kim, H Kirkham, N McLellan and J Sharma
(2002), “A structural VAR model of the New Zealand
business cycle,” New Zealand Treasury Working Papers No
02/26. (Wellington: New Zealand Treasury).
Buckle, R A, D Haugh and P Thomson (2003), “Calm after
the storm? Supply-side contributions to New Zealand’s GDP
volatility decline,” New Zealand Economic Papers, pp 217243, 37 (2).
Buiter, Willem H (1983), “Measurement of the Public Sector
Deficit and Its Implications for Policy Evaluation and Design,”
International Monetary Fund Staff Papers, June 1983.
Testing stabilisation policy limits in a small open economy
Teece (1996), “Economic Reform in New Zealand 1984-95:
The Pursuit of Efficiency,” Journal of Economic Literature,
Vol 34, December, pp. 1856-1902.
Gordon, Michael (2005), “Foreign reserves for crisis
management,” Reserve Bank of New Zealand: Bulletin, Vol
68, No.1, March, pp. 4-11.
Governor, Reserve Bank of New Zealand and Secretary to the
Treasury (2006), Supplementary Stabilisation Instruments,
Initial Report, 10 February.
Gray, Dale F, Robert C Merton and Zvi Bodie (2003), “A New
Framework for Analysing and Managing Macrofinancial
Risks of an Economy,” Moody’s MF Risk Working Paper 103, August 1, http://www.moodys-mfrisk.com .
73
Grimes, A (2006), “A smooth ride: Terms of trade,
How Does It Work?,” Journal of Economic Literature, vol.
volatility and GDP growth,” Journal of Asian Economics.
39(3), September, pp. 839-868.
Forthcoming.
Statistics New Zealand (2006a), “Productivity Statistics:
Henderson, David (1996), Economic Reform: New Zealand
Sources and Methods,” http://www.stats.govt.nz/NR/
in an International Perspective, New Zealand Business
rdonlyres/390D179F-11F3-436B-8D49-645160BA0DD5/0/
Roundtable, August.
ProductivityStatisticsSourcesandMethods.pdf.
Jack, William (2000), “The treatment of financial services
Statistics New Zealand (2006b), “Productivity Information
under a broad-based consumption tax,” National Tax
Paper 1988–2005; Productivity Statistics: 1988–2005,”
Journal, Vol. 53, No 4, part 1, pp. 841-852.
http://w w w2.stats.govt.nz /domino /external /pasfull /
Janssen,
John
(2001),
“New Zealand’s
Fiscal
Policy
Framework: Experience and Evolution,” New Zealand
Treasury Working Paper 01/25.
Karagedikli, O and K. Lkees (2004) “Do inflation targeting
central banks behave asymmetrically? Evidence for Australia
and New Zealand”, Reserve Bank of New Zealand Discussion
pasfull.nsf / 7cf46ae26dcb6800cc256a62000a2248 /
4c2567ef00247c6acc25713e000ab753?OpenDocument.
Svensson, Lars E O (2001), “Independent Review of the
Operation of Monetary Policy in New Zealand: Report to
the Minister of Finance,” http://www.treasury.govt.nz/
monpolreview/IndRevOpMonPol.pdf.
Treasury of New Zealand (2006a), Statements on the
Paper 2004/02.
OECD (2005), Economic Surveys: New Zealand, 4th July.
Long-Term Fiscal Position, http://www.treasury.govt.nz/
longtermfiscalposition.
Reserve Bank of New Zealand (2006), Monetary Policy
Treasury of New Zealand (2006b), “New Zealand’s Long-
Statement, March.
term Fiscal Position,” June, http://www.treasury.govt.nz/
Reserve Bank of New Zealand (2004), “Background
longtermfiscalposition/2006/sltfp-06.pdf.
information on the Reserve Bank’s proposal to extend the
purpose for which it holds foreign exchange reserves,”
17
March,
http://www.rbnz.govt.nz/finmarkets/
foreignreserves/0147894v1.pdf.
Reserve Bank of New Zealand (2001), “Fiscal and monetary
coordination,”
http://www.rbnz.govt.nz/monpol/
Treasury of New Zealand (2005a), A Guide to the Public
Finance Act, August.
Treasury of New Zealand (2005b),Briefing to the Incoming
Government 2005 – Sustaining Growth.
Wells, G and L T Evans (1985), “The impact of traded
goods prices on the New Zealand economy,” The Economic
review/0097145.html.
Reserve Bank of New Zealand (2000a), “The evolution of
Record, 61, pp. 421-435.
monetary policy implementation,” http://www.rbnz.govt.
Woodford, Michael (2000), “Monetary Policy in a World
nz/monpol/review/0096178.html .
Without Money,” International Finance 3, pp. 229-260.
Reserve Bank of New Zealand (2000b), “Output volatility
Woodford, Michael (2003), Interest and Prices; Foundations
in
of a Theory of Monetary Policy, Princeton University Press.
New Zealand,”
http://www.rbnz.govt.nz/monpol/
review/0096453.html .
Wyplosz, Charles (2002), “Fiscal Policy: Institutions versus
Sibert, Anne (2006) “Central Banking by Committee,”
Rules,” Centre for Economic Policy Research Discussion
International Finance, forthcoming, online version De
Paper 3238, February.
Nederlandsche Bank Working Paper no. 91, Feb.
Sarno, Lucio and Mark P. Taylor (2001) “Official Intervention
in the Foreign Exchange Market: Is It Effective and, If So,
74
Reserve Bank of New Zealand and The Treasury
Stabilisation policy in New Zealand: counting your
blessings, one by one by Willem H. Buiter
Discussion by Pierre L. Siklos, Wilfrid Laurier University and Viessmann Research Centre1
Willem Buiter was given the seemingly daunting task of
since 1988, and that in 2005 it is approaching 9% of GDP.2
reviewing the conduct of stabilisation policy in New Zealand.
Since the paper’s remit was not to ask whether such a
His remit included the review of the overall macro mix of
deficit is sustainable, Buiter instead considers whether the
fiscal, monetary, external and structural policies, the state
present situation portends a crisis or is sowing the seeds
of coordination between fiscal and monetary policies,
of an eventual serious economic downturn.3 Put in starker
whether a fiscal stabilisation instrument would be desirable,
terms, is New Zealand on the “comfortable path to ruin”?4
and what New Zealand can do to promote macroeconomic
The bottom line is that this is not the case and here Buiter,
stability. It is not surprising, therefore, that the paper begins
and the other participants and discussants in the Forum, are
by providing a general overview of economic performance in
both persuasive and in broad agreement with each other.
New Zealand, as well as examining the state of institutional
reforms up to the present time. The strategy I intend to adopt
below is to underscore areas of agreement with the paper,
bringing in evidence from another small open economy that
Figure 1
Nominal exchange-rate for Australia, Canada,
and New Zealand: 1993-2005
Per US$
shares an affinity with the New Zealand experience, namely
Per US$
2.5
2.5
Australia
New Zealand
Canada
2.0
Canada, while highlighting areas the paper has neglected.
1.5
2.0
1.5
As will be clear from what follows I share many, but not all,
1.0
1.0
of Buiter’s conclusions about New Zealand’s present state
0.5
0.5
0.0
0.0
-0.5
-0.5
-1.0
-1.0
-1.5
-1.5
-2.0
-2.0
and the direction it should follow in future.
The general conclusion is that New Zealand should
indeed “count its blessings” as the sub-title of the paper
suggests. But all is not sweetness and light and the paper
does suggest that there are clouds looming on the horizon.
-2.5
1993
-2.5
1995
1997
1999
2001
2003
2005
future, however, is very much open to debate. For example,
Source: Data from International Financial Statistics CD-ROM
(Washington, D.C.: International Monetary Fund), May
2006 edition
the paper points out that there are worries about the size
Are the foregoing developments peculiar to New Zealand?
and persistence of the current account deficit, the effects
The answer is clearly not. Figure 1 plots the nominal
of volatile commodity prices, and the complications for
exchange-rate for three “dollar bloc” countries, namely
monetary policy when asset prices are rising quickly, as in
Australia, Canada, and New Zealand, for the period 1993-
the case of housing prices, or are seemingly very volatile,
2005. Broadly speaking, movements in the three exchange-
as in the case of the nominal exchange-rate. Indeed, when
rates parallel each other although, by 2006, there is a
it comes to the current account deficit the paper notes the
growing spread between the three currencies with the
Whether any thunder storms are expected in the near
“massive fluctuations” of New Zealand’s current account
balance, the fact that it has been in deficit in every year
2
3
4
1
I am grateful to Bob Buckle and Aaron Drew for extensive
comments on an earlier draft.
Testing stabilisation policy limits in a small open economy
This makes New Zealand’s current account deficit one of the
largest amongst developed economies. Iceland’s deficit is
higher. Several other countries’ deficits are not far behind. See
“Still Waiting for the Big One,” The Economist 6 April 2006.
Other papers presented in this Policy Forum (e.g., Edwards,
and Schmidt-Hebbel) consider the current account
sustainability question more directly.
This quote is from Martin Wolf (2004) who wrote about the
deleterious effect of the US twin deficits in the current account
and in the government budget.
75
Canadian dollar continuing to appreciate against the US
Figure 4
dollar until the Spring of the same year, while, as this is
Government deficits (surpluses) and the
written, the two other currencies have lost some value
current account, Canada 1993-2005
against the US currency. The widening gap is seen from
% GDP
% GDP
6
Figure 2.
Figure 2
Nominal exchange-rates in Australia, Canada,
Per US$
4
2
2
0
0
1.7
Australia
New Zealand
Canada
1.4
-2
Current account/GDP
Government deficit (general)
Government deficit (federal)
-4
Per US$
1.5
4
-2
and New Zealand: Jan 2005-April 2006
1.6
6
-6
1.6
-8
1.5
-10
-4
-6
-8
-10
1993
1995
1997
1999
2001
2003
2005
1.4
1.3
1.3
1.2
1.2
1.1
1.0
Jan-05
Apr-05
Jul-05
Oct-05
Jan-06
Source: Government deficit (surplus) data are from the Bank of
Canada, Banking and Finanicial Statistics; Current account
data are from the same source as listed in Figure 2
1.1
Nor does one learn much from the behaviour of the real
1.0
exchange-rate over the same period. Again, whereas the
Apr-06
Canadian dollar has appreciated strongly in real terms, the
Source: Data from International Financial Statistics CD-ROM
(Washington, D.C.: International Monetary Fund), June
2006 edition
current account has remained in surplus (since 1999-2000) in
contrast to developments in New Zealand (and elsewhere),
as seen in Figure 3. Rising commodity prices seem to have
improved current account prospects in Canada but the same
Figure 3
pattern has not emerged in New Zealand (Munro 2005).5
The current account balance and the real
Instead, Canada seems to be bucking the trend found in
exchange-rate, Canada 1993-2005
Index
% GDP
130
125
4
Current account balance (RHS)
3
Real exchange rate (RULC) (LHS)
120
115
110
several small open economies, including New Zealand, as it
has moved from a position of “twin deficits” to its current
2
position of having a “twin surplus”, that is, a government
1
budget surplus as well as a current account surplus, as seen
0
in Figure 4.
-1
105
100
-2
In other major areas of economic performance New Zealand
-3
and Canada share fairly similar experiences. Both are
95
-4
experiencing rising housing prices, a phenomenon that is
90
1993
-5
1995
1997
1999
2001
2003
2005
Source: Both series are from the same source as in Figure 2
also being seen in many parts of the world, and one that is
also reflected in the growth of non-financial wealth in the
economy. The recent Canadian experience is shown in Figure
5. Meanwhile, long-term interest rates have fallen, again
a global phenomenon. Perhaps all of these developments
are a reflection of the fact that Canada, for example, is
5
76
The results for Canada are not apparently dependent on the
course of oil prices. A possible explanation for the divergent
patterns between the Canadian and New Zealand nominal
exchange-rates in 2006 stems from the perception that the
‘loonie’, as the Canadian dollar is called, is now also perceived
as a petro-currency.
Reserve Bank of New Zealand and The Treasury
considered to be the small open economy par excellence.
Figure 6
Perhaps surprisingly, as Buiter points out, New Zealand is far
Real GDP growth and the nominal exchange-
6
less open than one might think.
rate: Canada, 1993-2005
Figure 5
7
Housing prices and non-financial wealth,
6
Canada 1993-2005
5
CAN$b
Index
130.0
4.8
Housing prices (LHS)
Rate
1.6
Real GDP growth (LHS)
Nominal exchange rate (RHS)
1.5
1.4
4
1.3
3
Non-financial assets (RHS)
4.4
%
125.0
1.2
2
4.0
120.0
3.6
115.0
1.1
1
1.0
0
1993
3.2
110.0
2.8
105.0
2.4
1993
1995
1997
1999
2001
2003
2005
Source: Both series were obtained from CANSIM, see
6. Housing price index (rented and owned)
is V735396; Non-financial assets series is
V20682520. The original series is in millions
shown here in billions of CAD[/100].
100.0
Figure
series
series
and is
1995
1997
1999
2001
2003
2005
Source: Nominal exchange-rate from same source as listed in
Figure 1, real GDP growth is annual log difference in
real GDP. Data are from CANSIM (Canadian SocioEconomic Information Management System),series
V498943, http://dc2.chass.utoronto.ca.remote.libproxy.
wlu.ca/cansim2/English/index.html.
Some of the concern over the effectiveness of the interest
rate instrument also stems from the relative volatility of
nominal and real exchange-rates. Nevertheless, as has long
Turning to the implementation of monetary policy the
been known, if the desideratum is less volatile exchange-
paper asks whether the overnight cash rate (OCR), the
rates then this can only be accomplished at the cost of
primary instrument of monetary policy, has lost its ability to
more volatile interest rates, and this is likely to be far more
determine the degree of tightness or looseness in monetary
damaging to the reputation of the RBNZ than volatile
policy. As the paper correctly points out, the real question
exchange-rates. There is simply no credible evidence that
revolves around the ability of monetary policy to influence
volatile exchange-rates have negatively affected either
expectations. So long as this is the case, and the Reserve
inflationary expectations or economic performance more
Bank of New Zealand (RBNZ) is prepared to act, there is
generally.7 Figure 6 reveals no obvious association between
little danger that the overnight interest rate will prevent the
overall macroeconomic performance in Canada and the
central bank from meeting its inflation target. The worry
behaviour of the nominal exchange-rate. Buiter’s Figures 3
about the ability of households (and businesses) to move
and 9 can be said to make the same case for New Zealand.
up the maturity structure (i.e., toward the longer end) is
misplaced, and is reminiscent of the concerns expressed
some years ago about the potential for hedging and
derivative financial products to weaken the effectiveness of
monetary policy. The imminent ineffectiveness of monetary
policy did not materialize then, and it is unlikely to emerge
A crucial element that seems to be under-emphasized
in Buiter’s paper is the role of the central bank in
communicating more effectively to financial markets and
the public about how monetary policy is being conducted,
not only in New Zealand, but perhaps more importantly
under the present circumstances.
7
6
Financial openness, however, is high in both countries
and this might explain common interest rate and inflation
developments not only in Canada and New Zealand but
elsewhere in the industrial world.
Testing stabilisation policy limits in a small open economy
Collins and Siklos (2004) find that this is indeed the case
for New Zealand, as well as for Canada and Australia. They
consider the implications for optimal monetary policy rules of
targeting the real exchange-rate, in an objective function that
also includes inflation, the output gap, and permits interest
rate smoothing. Other studies rely on different methodologies
but arrive at precisely the same conclusion (e.g., Lubik and
Schorfheide (2005), and Kam, Lees, and Liu (2006)). Also see
Schmidt-Hebbel’s contribution.
77
to an international audience, how effective it has been
overall consistency of monetary policy in New Zealand.10 In
and how policy makers are working hard to ensure
an attempt to be transparent the RBNZ has gone to great
that it continues on this path in future. There is growing
lengths to justify its policy on when it might intervene in
evidence that an effective communications strategy can
foreign exchange markets (see http://www.rbnz.govt.nz/
complement, but never entirely substitute, for the main
finmarkets/foreignreserves/index.html).
instrument of monetary policy (e.g., see Siklos 2002, Amato
and Shin 2003, and references therein). Clearly, the RBNZ
First, economists regrettably have not succeeded in
developing reliable models that link fundamentals to
pays a great deal of attention to its communications policies
the exchange-rate. Hence, the RBNZ’s view that it might
(e.g., see Jackman 2002) but its successes or failures in
intervene in the event that the exchange-rate excessively
this realm will undoubtedly be a function of the monetary
departs from such fundamentals is not only unclear but
policy strategy being followed. Thus, for example, when the
can be downright confusing. Matters are not helped by the
MCI was being targeted, the effectiveness of the RBNZ’s
statement that such intervention would not necessarily be
communications was impaired, and there is the possibility
disclosed.11 Indeed, the RBNZ risks revisiting the difficulties
that tinkering over the years with the PTA, as well as the
it faced some years ago over the Monetary Conditions Index
threat of foreign exchange intervention, may exacerbate
(see Siklos 2000).
the communications problem faced by the Bank (also see
below).
The difficulties are similar to ones raised by Buiter who
warns, as if such warnings are still necessary in this day and
The reader of Buiter’s paper gets the sense that policymakers
age, that central banking remains fundamentally an art.
in New Zealand were prompted to consider the state of
Hence, mechanistic views of what the monetary authority
the current mix of monetary and fiscal policies perhaps
does may be useful descriptions of reality but are downright
because of some disappointment with the experience with
dangerous if taken literally.12
floating exchange-rates. Policymakers may also feel the
need to renew their defence of the current exchange-rate
system.8 Alternatively, policymakers in New Zealand, as is
true elsewhere, are worried about the “big one”, that is,
the event or shock that will lead to the unravelling of the
present state of “global imbalances”.9 While dismissing
these worries as overblown the paper brings up the issue of
foreign exchange intervention as a mechanism to deal with
what are perceived to be excessive movements or levels
inconsistent with fundamentals. In this connection while I
am not prepared to agree with Buiter that foreign exchange
Second, the evidence on the effectiveness of foreign
exchange intervention is, at best, mixed. Indeed, the
evidence is sensitive according to whether one investigates
the impact of intervention on the levels, the volatility, or
uncertainty in foreign exchange-rate movements (e.g.,
see Rogers and Siklos 2003). Ongoing work of mine, with
Diana Weymark (Siklos and Weymark 2006), suggests that
inflation targeting is a monetary policy strategy that has
contributed to reducing uncertainty in foreign exchange
movements in Australia and Canada.
intervention should never be considered, I do wish to
emphasize that recent RBNZ policy has been harmful to the
10
8
9
78
Such fears are not new to New Zealand as the recent history
of the Reserve Bank of New Zealand (Singleton, Grimes,
Holmes 2006) makes clear. Canada also went through such a
phase when the nominal exchange-rate reached its lowest ebb
in 2002.
It is inappropriate here to go into the details. Suffice it to say
that the current state of affairs stems from the large current
account and government budget imbalances between the
US and the rest of the world. A burgeoning literature on the
subject is cited by many of the authors of papers delivered at
this policy forum.
11
12
As an indication of the concern Buiter has expressed over
this issue it is worth noting that the conference paper devotes
almost 5 pages .
This element of the foreign exchange intervention rules
contradicts the stated aim of transparency. It reminds me of
the statement once heard: “Transparency is the key – unless
you know you’re not going to get caught.” It may be of interest
to point out that the Bank of Canada has stated that foreign
exchange interventions will be publicly announced on its
website.
See Poole (2006) in relation to the Taylor rule as a device to
discern what central banks do and why they cannot be used
to describe the Fed’s policy making activities in a purely
mechanistic manner.
Reserve Bank of New Zealand and The Treasury
Whether a return to foreign exchange intervention would
Figure 7
increase uncertainty is unclear but Buiter’s portrayal of
Foreign exchange reserves composition,
intervention amounting to increasing the number of
Canada 1993-2005
instruments of policy to two while retaining a single
US$m
24000
monetary policy objective, namely inflation control, suggests
that intervention under inflation targeting would increase
US$m
24000
Euros (mainly)
US$
20000
20000
uncertainty, albeit temporarily. Moreover, relying on evidence
16000
16000
that such intervention can be profitable also muddies the
12000
12000
8000
8000
4000
4000
waters because it suggests that the central bank might use
such practices to “make a profit” for the government rather
than correct some perceived inappropriate exchange-rate
level.13 There is nothing wrong, in principle, with the central
0
0
1993
1995
1997
1999
2001
2003
2005
bank managing its foreign exchange reserves to obtain the
best possible yields. In Canada, for example, a dramatic
Source: Both series are from the Bank of Canada, Banking and
Financial Statistics.
shift away from holding foreign exchange reserves in US
If New Zealand can count its blessing then does this mean
dollars (a depreciating currency) towards the euro primarily
that, as the saying goes, “If it Ain’t Broke, Don’t Fix it?”
(recently an appreciating currency) is evidence that this kind
The answer is a qualified no and I am largely in agreement
of practice does take place (Figure 7). However, to justify
with the position stated by Buiter. However, improvements
foreign exchange intervention on this basis does not inspire
in the delivery of monetary policy might be possible if the
confidence in the floating exchange-rate regime.
following reforms were to be considered. They are:
Third, the publication of foreign exchange intervention rules
(1). Establishing a permanent or long-run point inflation
suggests something is missing in the current institutional
target, say at 2%. As a result, the PTA would
relationship between the RBNZ and the government.
henceforth emphasize how each government wishes
However, nothing is further from the truth. The original
to see exceptions or relaxations of the existing rule.
directive in the RBNZ Act did not prevent the Minister, or
Adoption of such a policy would serve to better anchor
the Governor with ministerial approval, from engaging
inflationary expectations which, in New Zealand and
in foreign exchange intervention. Hence, it seems little is
elsewhere (e.g., Canada and the US) have shown signs
gained from explicitly stating the possibility that foreign
of drifting upwards (also see Buiter on this point, in
exchange intervention is an option as it gives the illusion
particular his Figures 26 through 29). This proposal is
that could be the beginning of the slippery slope toward
similar in spirit to the one outlined by Buiter. Whether
adopting a second instrument of policy to complement the
this can be characterized as a form of “flexible” inflation
current OCR, even though there is no hint at the present
targeting is unclear. His criticism of flexible inflation
time that this is even being contemplated by the RBNZ (or
targeting a la Svensson, and others, is mostly on the
the government).
mark though his equation (3), in particular, makes no
allowance for interest rate smoothing, a practice that
is widely believed to be followed by central banks.14
There is insufficient space to go into the details but
13
One of the better institutional practices implemented in
New Zealand, namely the Funding Agreement between
the RBNZ and the Treasurer (see http://www.rbnz.govt.nz/
about/whatwedo/0090769.html for the latest version) is not
mentioned by Buiter. It represents an important and clear
constraint on the central bank and one that other countries
could learn from (e.g., as in the recent experience in, for
example, Sweden, Germany; also see Ize 2006).
Testing stabilisation policy limits in a small open economy
eschewing a role for gradual interest rate changes
14
And is reflected in Buiter’s remark that typical interest rate
changes engineered by central banks is “chicken feed”.
79
understanding of the forces leading to financial crises.18
represents another way to tackle the first and second
15
moments problem Buiter refers to.
Buiter may be correct that the occurrence of a crisis is
(2). Since the Governor is already being advised by experts
easily identified but the public would not look kindly
why not establish a formal committee structure between
on a central bank that was unable to see such a crisis
the RBNZ and the government? This kind of reform has
coming. Moreover, by the time the Governor is called
previously been recommended (e.g, as in Svensson’s
upon to deal with the crisis the loss of credibility may
report in 2001; see http://www.rbnz.govt.nz/monpol/
be large. Once lost, credibility is difficult to regain. Of
review/index.html) and it seems like a natural evolution
course, there is also the possibility that the crisis serves
in the delivery of monetary policy in New Zealand.
as the defining moment of a Governorship much as the
The creation of a committee would formalize what is
1987 stock market crash forever gave Alan Greenspan
considered to be part and parcel of good governance in
an aura that would stay with him throughout the time
the private sector.16
he served as Chair of the FOMC. However, it is doubtful
that most central bankers will want to take this kind of
(3). Although improvements have been made in recent
chance.
Financial Stability Reports, the RBNZ is also responsible
for financial system oversight. It remains to be seen
While a significant portion of Buiter’s paper is devoted to
whether the RBNZ is getting closer to establishing a
monetary policy, coordination issues between the monetary
proper metric for financial system stability, and whether
and fiscal authorities, and fiscal policy more generally,
it would be willing to warn markets and the public of
is not neglected. Indeed, the focus is almost entirely
an impending problem or crisis in this sector. While
on proposals to generate some kind of automatic fiscal
suggesting that a separate authority be responsible for
stabilizer. In addition, the possibility of more constraints
financial system stability is not likely to be practical at
on overall fiscal policy, via the creation of some oversight
the present time, the RBNZ should be encouraged to
board, is also mentioned. There is no discussion in the
devote more research to the measurement of financial
paper about the fact that the Fiscal Responsibility Act
system stability.17 It is clear that this is also easier said
(FRA), born out of an earlier fiscal crisis, mitigated in large
than done. Central bankers have increasingly pointed to
part the dangers of a return to fiscal irresponsibility once
the need for monetary policy to prepare for potential
the electoral process was changed from the Westminster
financial instability, either of the home grown variety
system presently used in the UK and Canada to a form of
or imported from elsewhere. Yet, while there is broad
proportional representation. As such, there was no loss of
acceptance, though not universal agreement, on how
fiscal responsibility that is often the hallmark of minority
inflation is measured, there is as yet no widespread
or divided governments. Nevertheless, while fiscal probity
agreement on how to evaluate or measure financial
has no doubt been enhanced by the introduction of the
system stability. There is an urgent need to develop an
FRA, other countries have managed their fiscal affairs quite
well without such additional constraints (e.g., Canada). The
suggestion that the fiscal authority be further constrained is
15
16
17
80
It is considered less than elegant to admit interest rate
smoothing but gradualism does seem to take place partly to
deal with the problem of uncertainty and partly to protect
central bankers from having to reverse themselves too often
and risk losing their reputation.
Buiter’s worry about having to outsource membership on a
monetary policy committee to Bangalore (what about Ireland?)
is somewhat overblown. For one thing he skirts the issue of
committee size which, presumably, would be far smaller than
in the UK or the US and, second, that a de facto committee has
already been in place for some time.
This is a point made recently, and forcefully, by Goodhart
(2006).
not only impractical but would severely weaken the notion
of political accountability. We may not always like what fiscal
policy is currently accomplishing but as long as the public
gets to vote there is sufficient accountability at present.
Therefore, I see little merit in creating a GST committee
18
Perhaps, as de Grauwe (2006) points out, it is because the
pure rational expectations model so favoured by economists
and central banks is a fairy tale world view that does not
permit bubbles and crashes to be observed from the data.
Reserve Bank of New Zealand and The Treasury
(more on the GST below). In addition, the problems faced
the late 1990s. What seemed like a sensible thing at the
in the European context that are mentioned in the paper
time, unless there were fraudulent activities involved as was
are clearly not the same as would arise in New Zealand.
true in several cases, turned out to be disastrous once the
Finally, Buiter does not give sufficient credit to the present
bottom fell out of the stock market. Thousands of workers
attempts at the New Zealand Treasury to deal with long-
are still dealing with the consequences.
run aspects of fiscal policy. Policy makers in New Zealand
should not only be congratulated for their far-sightedness
in this respect but for pioneering the measurement of the
impact of the government sector in the economy through
the publication of financial statements of the Government
of New Zealand (e.g., see http://www.treasury.govt.nz/
publicsector/).
The proposal to tax land has merits but the difficulty is that
this proposal is made in isolation of other proposals and
without any clear discussion of the existing system and the
implications for taxation as a whole and the taxation of
property at the local level. Finally, I have similar concerns
over Buiter’s discussion of immigration. Canada, as is wellknown, is on a per capita basis, one of the largest recipients
As for the proposed automatic stabilizers, Buiter’s
of immigrants in the world and there is an enormous
recommendations are all economically sensible but many
literature on immigration and its effects. It would have been
are “politically incorrect”. Were they adopted it is unclear
more credible and preferable to mention the possibility that
whether these would prevent boom and bust cycles in
immigration policies should perhaps be revisited than to
economic activity, large fluctuations in the exchange-rate,
show charts relating housing prices and immigration which
or whether these could be introduced in a cost-effective
add nothing to the debate. Indeed, when the aggregate
manner. Nevertheless, I will briefly comment on the five
demand effects of immigration are considered it is the
proposals made in the paper. His proposal to tax capital
social welfare costs that have concerned many observers.
gains the same as income and to index-link it to the CPI is
New Zealand is a partner in the Metropolis Project (http://
a sensible one but, in practice, the manner in which labour
international.metropolis.net/index_e.html) and I would
income is earned and valued is not the same as for income
urge instead that this aspect of the debate, while relevant,
from investments and it is likely that for every distortion that
is best dealt with separately.
is addressed with such a move it will prompt tax experts
to find new and clever ways to circumvent them. Ideally,
as Buiter points out, capital gains should not be taxed at
all. As for indexing the tax system this is hardly the way
to constrain fiscal policy and could complicate the inflation
control task of the RBNZ.
The sage advice given by Buiter about the current state of
economic policy in New Zealand should not alter the fact
that New Zealand has done exceedingly well in the past
decade especially. While improvements are possible they
are likely to be marginal. To the extent that more radical
changes in economic policy are to be contemplated these
The proposal to broaden the base of the GST is also a sensible
would appear to be politically infeasible and very likely
one but instead of creating an independent committee it
economically unnecessary. New Zealand should indeed
might simply be more practical to reduce or remove GST
count its blessings but keep an eye out for better ways to
from goods and services that are highly sensitive to the
design and implement fiscal and monetary policies.
business cycle. Buiter’s proposal to use any windfall from
the KiwiSaver as well as the stock of foreign exchange
reserves beyond what is deemed prudent seems designed
to contribute to the creation of automatic stabilizers and
is seemingly admirable but I doubt that it will be easy to
agree on a definitions of “windfall” or “prudent”. Just look
at what companies in North American have done with their
supposed ‘windfalls’ from the growth of their pensions in
Testing stabilisation policy limits in a small open economy
81
References
Poole, W (2006), “The Fed’s Monetary Policy Rule,” Review
Amato, Jeffery, and Hyun Song Shin (2003), “Communication
and Monetary Policy,” BIS working paper 123, January.
Collins, Sean and Pierre L Siklos (2004), “Optimal Monetary
Policy and Inflation Targets: Are Australia, Canada and
New Zealand Different from the US?” Open Economies
Review 15 (October), 347-62.
De Grauwe, P (2006), “A Central Banking Model for Neither
Gods Nor Monkeys,” Financial Times of London 26 July.
Goodhart, Charles (2006), Searching for a Metric for
Financial
Stability,”
available
at
http://www.wlu.ca/
viessmann/Capri/Capri.htm.
Ize, A (2006), “Spending Seigniorage: Do Central Banks
Have a Governance Problem?,” IMF working paper 06/58,
March.
Jackman, P (2002), “The Reserve Bank’s External
Communications,” RBNZ Bulletin 65 (March): 28-33.
of the Federal Reserve Bank of St. Louis 88 (January/
February), 1-12.
Rogers, Jeff, and Pierre L Siklos (2003), “Foreign Exchange
Market Intervention in Two Small Open Economies:
The Canadian and Australian Experiences,” Journal of
International Money and Finance 22 (June): 393-416.
Siklos, Pierre L (2002), The Changing Face of Central
Banking: Evolutionary Trends Since World War II (Cambridge:
Cambridge University Press).
Siklos, Pierre L (2000), “Is the MCI a Useful Signal of
Monetary Policy Conditions? An Empirical Investigation,”
International Finance 3 (November): 413-37.
Siklos, Pierre L and Diana Weymark (2006), “Measuring
the Impact of Intervention on Exchange market Pressure,”
unpublished, Vanderbilt University.
Singleton, John, Arthur Grimes, Frank Holmes (2006),
Innovation + Independence: The Reserve Bank of
Kam, T, K Lees, and P Liu (2006), “Uncovering the Hit-List
New Zealand 1973-2002 (Auckland: Auckland University
for Small Inflation Targeters: A Bayesian Structural Analysis,”
Press).
paper presented at the 2006 Australasian Econometric
Society Meetings, Alice Springs, July.
Lubik, T A, and F Schorfheide (2005), “Do Central Banks
Wolf, Martin (2004), “America On the Comfortable Path
to Ruin,” Financial Times 17 August 2004, available from
http://www.rh.edu/~stodder/BE/Wolf_Ruin.htm.
Respond to Exchange-rate Movements?,” working paper,
Johns Hopkins University, November.
Munro, Annella (2005), “New Zealand’s Foreign Liabilities
and Current Account Deficit: Sustainability Assessment,”
RBNZ Memorandum, July.
82
Reserve Bank of New Zealand and The Treasury
New Zealand’s monetary and exchange-rate policy
In international comparison†
Klaus Schmidt-Hebbel ††, Central Bank of Chile
Overview
rate policy, and the need for developing counter-cyclical
New Zealand’s recent business cycle has been driven by
fiscal and financial policy tools to deal with large exchange-
exceptional international conditions, reflected in high world
rate swings and current account imbalances.
growth and favourable terms of trade, and a booming
domestic housing market. While domestic demand is
currently cooling off, the persistence of a large current
1
account deficit and a real exchange-rate that is still on the
New Zealand’s recent business cycle has been driven by
appreciated side, raise questions about the sustainability and
favourable world growth and external terms of trade, as
downside risks of New Zealand’s current and prospective
well as a strong domestic housing boom. However, GDP
conditions. This paper analyses monetary and exchange-rate
growth has been declining from 4.4% in 2004 to 2.4%
policy in New Zealand from the perspective of comparison
in 2005 and a projected rate of 1.7% in 2006, implying
with various control groups comprised of advanced
a narrowing output gap (Figure 1).1 The housing boom
economies. Trends, cyclical behaviour, and correlations of
has been evidenced by strong construction activity, strong
key macroeconomic variables and a comparative assessment
growth in housing permits, and massive capital gains in the
of New Zealand’s overall macroeconomic policy framework
housing market (Figure 2). Construction and the housing
provide a necessary starting point. Evidence on transmission,
market have also started to cool off since 2005, although
efficiency, and independence of monetary policy reveal
house prices are still high.
important similarities and differences between New Zealand
and comparator countries. Drawing from the international
evidence about exchange-rate misalignments and their
implications for growth allows estimating exchange-rate
Introduction
Figure 1
GDP growth and output gap in New Zealand,
1986q1-2005q4
%
misalignment in New Zealand and making inferences about
%
15
15
its potential growth effects. Reviewing the international
10
10
5
5
0
0
-5
-5
evidence on the role of the exchange-rate in the conduct of
monetary policy and the effectiveness of forex interventions
leads to inferences about RBNZ’s monetary and exchange†
††
I thank Paul Baines, Mark Blackmore, Alan Bollard, Philip
Borkin, Robert Buckle, Peter Bushnell, Andrew Coleman,
David Drage, Kelly Eckhold, Hugh Fletcher, Jane Frances,
Andrew Gawith, Khoon Goh, Arthur Grimes, David
Hargreaves, Bernard Hodgetts, John Janssen, Brent Layton,
John McDermott, Anella Munro, Brendan O’Donovan, Michael
Reddell, Rishab Sethi, Murray Sherwin, Grant Spencer, and
Simon Tyler for insightful discussion and/or data provision,
as well as Louise Abolins for kind support, during my April
visit in Wellington. I also thank Alan Bollard, Robert Buckle,
Willem Buiter, Aaron Drew, John Edwards, Stephen Grenville,
Pierre Siklos, Grant Spencer, and other participants for their
valuable comments on the first draft of this paper presented at
the “Macroeconomic Policy Forum” organized by the Reserve
Bank of New Zealand and The Treasury in Wellington on June
12, 2006. I thank Marcelo Ochoa for outstanding research
assistance and discussion. The views presented herein do not
necessarily reflect those of the Reserve Bank of New Zealand,
the Treasury of New Zealand or the Central Bank of Chile. All
remaining errors are mine.
Central Bank of Chile. Email: [email protected]
Testing stabilisation policy limits in a small open economy
-10
Annual GDP growth
Output gap
-15
1985 1987 1989 1991 1993 1995 1997 1999 2001 2003
-10
-15
Source: Statistics New Zealand
Reserve Bank of New Zealand
1
Reserve Bank of New Zealand (2006)
83
Figure 2
Figure 4
House prices, house permits, and GDP growth
Nominal bilateral exchange-rate and real
in residential construction in New Zealand,
effective exchange-rate in New Zealand,
1986q1-2005q4
1986q1-2006q1
Index
200
%
14
US$/NZ$
0.8
180
12
0.7
160
10
0.7
140
8
120
6
100
4
80
2
60
0
40
20
House prices (2000=100)
House permits (2000=100)
GDP growth in residential construction (RHS)
-2
-4
0
-6
1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005
Source: Statistics New Zealand
Bank lending, particularly the boom in mortgage lending, has
been significantly financed by foreign lending in domestic
currency, the counterpart of a large current account deficit.
While the terms of trade attained in 2004-2005 their
highest levels in two decades, the current account deficit
widened continuously to reach 8.8% of GDP in 2005, one
of the highest external imbalances observed among OECD
Nominal exchange rate US$/NZ
Real Effective Exchange Rate (2000=100) (RHS)
Index
150
140
130
0.6
120
0.6
110
0.5
0.5
100
0.4
1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005
90
Source:
Reserve Bank of New Zealand
It is hard to blame macroeconomic policies for the domestic
spending boom. The fiscal stance has continuously improved
since the introduction of the Fiscal Responsibility Act in
1994, as reflected by public sector surpluses, a reduction
in government debt ratios, and very low (even negative)
implicit country risk spreads over US Treasury bonds.
Monetary policy, guided by the world’s first inflationtargeting framework implemented by the Reserve Bank of
countries (Figure 3). The large excess of domestic spending
New Zealand (RBNZ), has been successful in keeping inflation
over income was reflected in a strong appreciation of the
low and stable during the last 15 years (Figure 5). Monetary
New Zealand dollar that peaked in late 2005, with a partial
policy is consistent with international best practice in flexible
reversion taking place since early 2006 (Figure 4).
inflation targeting, providing some weight to its counter-
Figure 3
cyclical stabilisation role. Rising inflationary pressures have
Terms of trade and current account balance in
led the RBNZ to raise its monetary policy rate since 2003,
attaining a level of 7.25% at the time of writing.
New Zealand, 1986q1-2005q4
Index
115
%
0
-1
110
-2
105
-3
-4
100
-5
95
90
85
-6
-7
Terms of trade (2000=100)
Current Account (% of GDP) (RHS)
-8
-9
-10
80
1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005
Source: Statistics New Zealand
84
Reserve Bank of New Zealand and The Treasury
Figure 5
•
What is the monetary policy transmission process in
CPI inflation and monetary policy interest rate
New Zealand and its efficiency compared to other
in New Zealand, 1986q1-2006q1
countries? What is the extent of exchange-rate
%
20
18
monetary policy transmission to financial markets,
18
16
14
devaluation and oil-price pass-through to inflation;
%
20
Short-term interest rate
Annual inflation rate
16
output, and inflation; inflation target accuracy; and
14
monetary policy efficiency reflected in inflation and
12
12
10
10
8
8
6
6
4
4
is the influence of foreign interest rates on domestic
2
2
rates in New Zealand relative to other countries?
0
1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005
0
output volatility?
•
•
What is the world evidence on the costs of real
exchange-rate volatility and misalignments? Which are
Source: Statistics New Zealand
Reserve Bank of New Zealand
the fundamentals that drive the medium-term behaviour
The exchange-rate has been a key absorber of foreign and
of the real exchange-rate in the world? Is there evidence
domestic shocks, in a framework of a flexible, cleanly-
of past and present real exchange-rate misalignment in
floating exchange-rate regime adopted consistently during
New Zealand, and will misalignment have implications
the last 20 years. While the RBNZ announced in 2005 a
for economic growth?
framework for possible exchange-rate interventions, it has
not apparently intervened to date. The banking sector is well
Is monetary policy in New Zealand independent? What
•
What is the best approach to achieve domestic price
regulated and supervised, and commercial bank soundness
stability, while avoiding cyclical extremes in the
indicators remain healthy.
exchange-rate? How should the secondary concern of
Yet the mix of a large current account deficit (projected by
the OECD at 9.1% of GDP for 2006) and a real exchangerate that is still on the appreciated side of its historical
distribution, largely driven by the housing boom, raise
major questions about the sustainability and downside risks
exchange-rate stability or competitiveness be addressed
by a flexible exchange-rate targeting central bank?
Which counter-cyclical fiscal and financial policy tools
could be developed to deal with persistent exchangerate swings and current-account imbalances?
of New Zealand’s current business-cycle conditions. This, in
I address these questions by comparing New Zealand’s
turn, leads to questions about appropriate policy measures
macroeconomic and policy performance to that of relevant
to correct current imbalances and strengthen the economy’s
groups of comparator countries. I do this by providing
structure to avoid future repeats of the recent experience.
descriptive evidence and empirical results for behavioural
Among these questions are:
relations for New Zealand and two alternative comparator
•
How do New Zealand’s macroeconomic performance
and business cycles compare to advanced open
economies during the last two decades? What about the
trend behaviour, cyclical performance, and correlations
of key macroeconomic variables in New Zealand?
•
country groups. The first set (termed CCG1) is comprised of
the world’s 19 inflation-targeting (IT) countries; the second
group (termed CCG2) is a subset of the former, comprised of
5 small advanced ITers that are, at least in part, commodityexporting countries: Australia, Canada, Norway, Sweden,
and Chile.
How does New Zealand’s overall mix of monetary, fiscal,
external, and structural policy framework compare to
those of other advanced countries? Are there areas for
potential policy improvement?
Testing stabilisation policy limits in a small open economy
85
Section 2 documents trends and business-cycle conditions
2
Macroeconomic performance
and business cycles
of key macroeconomic variables in New Zealand and
compares them to those in country group CCG2. Section
How do New Zealand’s macroeconomic performance and
3 summarizes the distinctive features of New Zealand’s
business cycles compare to the five countries grouped in
overall macroeconomic policy mix in comparison to those
CCG2 during the last two decades? I address this question
of the CCG2 country group, pointing out a few areas of
by focusing on the trend behaviour, cyclical performance,
potential improvement. Section 4 provides an international
and correlations of key macroeconomic variables in
comparison of New Zealand’s monetary policy transmission
New Zealand, compared to those in CCG2 countries. I start
and efficiency, focusing on four dimensions: the extent
by reporting the mean, volatility, and cyclical features of
of exchange-rate and oil price pass-through to inflation;
each quarterly series spanning 1986-2005. To characterize
monetary policy transmission to financial markets, output,
the cycle I follow Harding and Pagan’s (2002) method of
and inflation; inflation target accuracy; and monetary policy
describing the cyclical nature of the selected variables. They
efficiency reflected in inflation and output volatility. Section
adopt a two-step procedure. First, they identify the turning
5 assesses if there is scope for an independent monetary
points in a series, which marks the periods of cyclical
policy in New Zealand and comparator countries that are
expansions and contractions.2 Then, given the turning
closely integrated into world financial markets. There I
points, they compute three cyclical features: cycle duration,
report simple correlation coefficients between domestic
cycle amplitude, and cumulative change.
and foreign short-term interest rates and impulse response
dynamics of domestic short-term rates in response to
foreign rate shocks.
Subsequently I report two measures of co-movement
between different series: simple correlation coefficients and
the degree of concordance index proposed by Harding and
Section 6 starts reviewing the world evidence on the costs
Pagan (2006). The latter index captures the co-movement
of real exchange-rate volatility and misalignments. Then I
between two variables during their phases over the cycle.3
review international evidence on the fundamentals that
drive the medium-term behaviour of real exchange-rates and
provide evidence on the behaviour of New Zealand’s real
exchange-rate. This allows identifying evidence of past and
present real exchange-rate misalignment in New Zealand
and, drawing from the world evidence, possible implications
for economic growth.
Figure 6 depicts the raw data. Table 1 summarizes simple
statistics for each series and country, while Table 2 provides
information about the three cyclical features. Correlations
and degree of concordance indexes are reported in Table 3.
Next I discuss selectively New Zealand’s cyclical behaviour
and how it compares to the other five countries.
Section 7 starts by reviewing the literature and international
evidence on the role of the exchange-rate in the conduct
of monetary policy. Then I review the international practice
and evidence about interventions and their effectiveness.
In the light of the latter reviews, and the findings about
New Zealand’s monetary and exchange-rate policy reported
in preceding sections, I draw policy lessons for New Zealand
at the end of this section, identifying possible fiscal and
financial-policy instruments to deal better with exchangerate misalignments and current-account imbalances.
Section 8 concludes with a brief summary of the paper’s
main findings and policy implications.
86
2
3
A contractionary phase is defined as a peak-to-trough period
and an expansionary phase as a trough-to-peak period.
If two variables are perfectly pro-cyclical then the index is
unity, while a value of zero marks reflects exact countercyclicality.
Reserve Bank of New Zealand and The Treasury
Figure 6
Selected indicators in New Zealand and CCG2 countries, 1986-2005
Testing stabilisation policy limits in a small open economy
87
Figure 6
Selected indicators in New Zealand and CCG2 countries, 1986-2005 (cont.)
88
Reserve Bank of New Zealand and The Treasury
Figure 6
Selected indicators in New Zealand and CCG2 countries, 1986-2005 (cont.)
Testing stabilisation policy limits in a small open economy
89
Figure 6
Selected Indicators in New Zealand and CCG2 countries, 1986-2005 (cont.)
90
Reserve Bank of New Zealand and The Treasury
Table 1
Selected indicators in New Zealand and CCG countries, 1986-2005
New Zealand
Australia
Canada
Chile
Norway
Sweden
Mean
-4.61
-4.65
-1.09
-2.14
5.19
2.39
Std. Dev.
1.69
1.16
2.35
2.07
6.50
3.17
101.70
105.39
98.92
125.22
74.78
103.02
5.71
9.50
5.39
27.09
18.14
6.02
Mean
1.34
0.95
0.76
0.89
1.55
-0.14
Std. Dev.
4.58
6.73
4.44
18.14
14.39
3.71
Mean
119.64
111.21
102.93
91.16
96.10
105.40
Std. Dev.
12.02
9.71
7.05
8.86
14.49
10.62
Mean
1.33
0.06
-0.01
-0.72
2.14
-0.86
Std. Dev.
8.37
8.32
5.36
7.14
4.57
6.77
Mean
2.51
3.29
2.78
5.82
2.73
2.12
Std. Dev.
3.36
1.62
1.97
3.46
1.78
1.97
Mean
-0.17
-0.04
0.00
-0.05
0.02
0.01
Std. Dev.
2.25
1.11
1.33
1.99
1.02
1.31
Mean
3.98
3.96
0.47
0.97
3.09
3.28
Std. Dev.
6.24
6.97
4.73
1.80
10.04
7.48
Current Account (% of GDP)
Terms of Trade
Mean
Std. Dev.
Terms of Trade (annual % change)
Real Effective Exchange-rate
Real Effective Exchange-rate (annual % change)
GDP growth (annual, %)
Output gap
Real House Prices (annual % change)
House Permits (annual % change)
Mean
0.42
-0.12
6.38
-
0.82
0.29
Std. Dev.
18.62
20.57
19.68
-
17.74
42.40
Mean
2.45
3.66
2.23
5.22
1.05
3.84
Std. Dev.
10.78
7.93
5.48
8.31
5.30
5.10
Real GDP in residential construction
(annual % change)
Private investment in housing (annual % change)
Mean
4.09
4.10
3.38
-
2.74
7.37
Std. Dev.
13.46
13.07
9.34
-
12.14
9.48
Mean
1.16
1.37
1.15
1.79
0.64
0.44
Std. Dev.
0.45
0.30
0.13
0.15
0.18
0.20
Population Growth (annual % change)
Note:
Definitions and sources of the data are provided in Annex A.
Testing stabilisation policy limits in a small open economy
91
Before focusing on the comparative descriptions for the
Australia’s and Norway’s, similar to Canada’s and Chile’s,
two decades spanning from 1986 through 2005, a word
and much smaller than Sweden’s.
of caution on this sample period is in order. New Zealand
adopted major structural and macroeconomic reforms
from the late 1980s through the early 1990s, which were
associated with transition costs. Hence the full 19862005 period comprises an earlier sub-period (1986-1993)
of higher inflation and lower growth (and larger output
and inflation volatility) and a second sub-period (19942005) of improved macroeconomic stability. The latter
Regarding expansions, New Zealand’s mean duration
is 12.33 quarters, which is similar to Canada’s but much
shorter than recoveries observed in the other CCG2
countries. The mean amplitude of the recovery is 5.7%
of GDP in New Zealand, much below the output increase
observed elsewhere. New Zealand’s cumulative output gain
during expansionary phases is 64.02% on average, well
below the other countries’ output gain, except Canada’s.
fact is highlighted in section 4 below on New Zealand’s
improvement in inflation and output volatility. However, the
following international comparison in this section spans the
full two-decade period. The subsequent figures depicting
New Zealand’s full 1986-2005 performance suggest a
structural break toward lower inflation and output volatility
in the early 1990s. However, no volatility decline is apparent
in housing market-related variables, the current account
balance, and the real exchange-rate.
Figure 7 depicts the three cyclical features during downturns
and recoveries in New Zealand and the average cyclical
features for four of the CCG2 countries: Australia, Canada,
Norway, and Sweden.4 In comparison to the 4-country
average, New Zealand’s output contractions are shorter
but exhibit a larger output decline. While contractions
last about 2.25 quarters and exhibit a mean amplitude of
-1.56% in New Zealand, the average duration in the four
CCG2 countries is 5.75 quarters with a mean amplitude of
-1.32%. New Zealand’s output loss is smaller than the loss
GDP growth and business cycles
of output experienced by the CCG2 country group, which
New Zealand’s annual GDP growth averaged 2.5% between
presents a cumulative output loss of 3.87% on average.
1986 and 2005, a figure that is close to average growth in
Similarly, New Zealand’s expansions are shorter and exhibit
Australia, Canada, Norway, and Sweden (2.7%) but lower
a smaller output gain than the average in the four CCG2
than Chile’s 5.8% (Table 1). However output volatility is
countries, where expansions last about 20 quarters, with
much higher in New Zealand than in CCG2 countries apart
a mean amplitude of 13.1%, and a mean output gain of
from perhaps Chile, as shown by the standard deviation of
189.65%.
annual GDP growth and the output gap. As suggested by
Next I briefly discuss the business-cycle features of other
Figures 1 and 6, New Zealand’s higher 1986-2005 output
macroeconomic variables, focusing only on the comparative
volatility reflects its large GDP swings during the earlier
indicators depicted in Figure 6.
1986-1993 reform period, as discussed above.
New Zealand’s business-cycle features reveal that output
contractions last on average 2.25 quarters while expansions
reach a mean duration of 12.33 quarters. New Zealand’s
mean contraction duration is similar to Norway’s and Chile’s
and is slightly shorter than Australia’s and Canada’s. The
mean amplitude (or depth) of New Zealand’s contractions
is 1.56% of GDP, similar to Canada’s and Chile’s, somewhat
deeper than Australia’s and Norway’s, but much shallower
than Sweden’s. New Zealand’s cumulative output loss in
contractionary phases is 2.04% of GDP, again larger than
92
Housing market
Real house prices have increased in New Zealand at an
average annual growth rate of 3.98%, similar to the figures
observed in Australia (3.96%), Norway (3.09%), and
Sweden (3.28%). However, the trend rise in house prices
in New Zealand has not been associated with a significant
4
Chile is excluded because it is an outlier in the CCG2 group
in many of its cyclical features. Simple averages of individual
country statistics are computed from the statistics of the other
countries.
Reserve Bank of New Zealand and The Treasury
Figure 7
Cyclical features of selected variables in New Zealand and four industrial countries, 1986-2005
New Zealand
Four CCG2 average
New Zealand
Four CCG2 average
New Zealand
Four CCG2 average
New Zealand
Four CCG2 average
New Zealand
Four CCG2 average
New Zealand
Four CCG2 average
New Zealand
Four CCG2 average
New Zealand
Four CCG2 average
Source: Authors calculations
Testing stabilisation policy limits in a small open economy
93
Table 2
Cyclical features of selected indicators in New Zealand and CCG2 countries, 1986-2005
New Zealand
Australia
Canada
Chile
Norway
Sweden
7.25
7.40
6.00
7.20
4.33
5.50
Current Account (% of GDP)
Peak-to-trough
Mean duration (quarters)
Mean amplitude (percentage points)
-2.61
-2.58
-1.27
-3.64
-2.87
-0.99
Cumulation (%)
-9.33
-12.72
-4.31
-14.74
-7.47
-4.18
Mean duration (quarters)
5.50
7.60
6.33
5.60
7.00
6.67
Mean amplitude (percentage points)
2.10
2.50
2.03
4.18
6.01
2.21
Cumulation (%)
6.61
10.40
9.43
12.11
32.81
10.22
5.60
5.80
9.67
4.50
4.86
3.60
Trough-to-peak
Terms of Trade
Peak-to-trough
Mean duration (quarters)
Mean amplitude (%)
-2.93
-4.06
-4.30
-10.69
-6.41
-1.57
Cumulation (%)
-8.47
-12.25
-29.06
-29.10
-16.39
-3.56
Mean duration (quarters)
6.40
5.50
18.00
3.86
4.29
5.80
Mean amplitude (%)
3.24
3.84
7.64
9.20
8.78
1.66
Cumulation (%)
10.54
14.11
57.30
27.32
25.46
5.72
7.50
6.00
6.50
4.17
8.25
5.00
Trough-to-peak
Real Effective Exchange-rate
Peak-to-trough
Mean duration (quarters)
Mean amplitude (%)
-7.47
-5.17
-3.91
-4.57
-2.72
-4.92
Cumulation (%)
-28.9
-18.53
-16.27
-10.83
-12.74
-14.06
Mean duration (quarters)
6.75
6.00
5.50
6.67
9.00
5.86
Mean amplitude (%)
6.31
6.45
4.42
5.27
6.64
3.43
Cumulation (%)
33.06
24.43
20.92
31.84
30.12
10.31
2.25
4.00
4.00
2.50
3.00
12.00
Mean amplitude (%)
-1.56
-0.54
-1.49
-1.50
-0.72
-2.52
Cumulation (%)
-2.04
-1.28
-2.22
-1.95
-1.46
-10.52
12.33
16.00
13.00
31.00
54.00
-
Trough-to-peak
Real GDP
Peak-to-trough
Mean duration (quarters)
Trough-to-peak
Mean duration (quarters)
Mean amplitude (%)
Cumulation (%)
94
5.70
7.07
5.52
26.70
19.50
-
64.02
64.05
46.17
458.73
536.99
-
Reserve Bank of New Zealand and The Treasury
Table 2
Cyclical features of selected indicators in New Zealand and CCG2 countries, 1986-2005 (cont.)
New Zealand
Australia
Canada
Chile
Norway
Sweden
Real house prices
Peak-to-trough
Mean duration (quarters)
4.33
4.50
21.50
7.00
12.00
7.00
Mean amplitude (%)
-1.37
-2.43
-6.37
-0.87
-14.47
-5.66
Cumulation (%)
-2.91
-6.73
-91.40
-3.73
-161.48
-25.84
Mean duration (quarters)
14.0
12.75
2.00
3.50
38.00
15.50
Mean amplitude (%)
10.48
10.70
0.37
0.62
32.86
11.16
Cumulation (%)
108.45
80.06
0.37
1.33
681.33
138.14
Mean duration (quarters)
6.00
6.33
3.83
-
6.80
3.33
Mean amplitude (%)
-15.27
-20.78
-10.57
-
-16.83
-28.40
Cumulation (%)
-52.80
-67.73
-30.11
-
-62.32
-59.21
Mean duration (quarters)
7.80
11.75
8.00
-
6.20
7.17
Mean amplitude (%)
18.07
20.33
16.33
-
15.26
23.14
Cumulation (%)
78.68
134.88
72.00
-
66.29
109.04
4.80
4.33
7.00
2.80
5.83
3.00
Trough-to-peak
House permits
Peak-to-trough
Trough-to-peak
Real GDP construction
Peak-to-trough
Mean duration (quarters)
Mean amplitude (%)
-7.96
-6.69
-4.27
-4.22
-3.52
-2.93
Cumulation (%)
-26.17
-17.86
-27.95
-7.73
-12.78
-5.92
Mean duration (quarters)
9.00
14.00
9.00
8.80
6.17
11.00
Mean amplitude (%)
11.53
11.03
5.63
10.36
3.88
8.64
Cumulation (%)
72.26
89.98
35.54
41.80
19.86
41.68
Mean duration (quarters)
3.40
6.33
4.40
-
9.75
-
Mean amplitude (%)
-10.86
-12.48
-5.74
-
-10.36
-
Cumulation (%)
-18.35
-37.90
-13.49
-
-104.74
-
Mean duration (quarters)
10.80
12.25
5.75
-
8.33
-
Mean amplitude (%)
15.80
18.34
4.85
-
13.33
-
Cumulation (%)
118.78
123.55
17.68
-
44.39
-
Trough-to-peak
Private Investment in housing
Peak-to-trough
Trough-to-peak
Note:
quarterly data for 1986q1 - 2005q4. The table reports the mean duration, amplitude, and cumulative changes for each cyclical
phase, i.e., from peak to trough and from trough to peak. Calculations methods and definitions follow Harding and Pagan (2002).
Testing stabilisation policy limits in a small open economy
95
increase in house permits; house permits have risen by an
Australia’s, but its standard deviation of population growth
average annual rate of 0.42%, lower than the growth in
is 50% larger than Australia’s and more than three times
house permits in the four-country group (1.84%).5 House
Canada’s.
prices and house permits are, on average, less volatile in
New Zealand than in the CCG2 country group.
On the cyclical features of residential construction, the results
show that during the expansionary phase of New Zealand’s
As depicted in Figure 7, house price cycles in New Zealand
construction activity, lasting about 9 quarters, construction
and the four CCG2 comparison countries are characterized
output rises on average by 11.53% per annum. In the
by short-lived phases of contraction and prolonged periods
subsequent contraction phase, which lasts 4.80 quarters,
of expansion. The mean duration of contraction phases in
output in this sector plunges by 7.96%, translating into a
New Zealand is about 4 quarters, while in CGG2 countries
cumulative construction output loss of 26.71%. This loss
it reaches 11.25 quarters, on average. In the contraction
is larger in New Zealand than in the four CCG2 countries,
phase, the fall in real house prices has been 1.37% per
where the mean amplitude reaches -4.35%, and the
annum in New Zealand, much smaller than that observed
cumulative output loss is equal to 16.13%, on average.
in CCG2 countries (7.23%). Real house price booms last
approximately 14 quarters and prices increase on average
by 10% in New Zealand. In CCG2 countries, periods of
expansion exhibit a mean duration of 17.06 quarters and
price gains of 13.77% per annum. In general, this pattern
of larger increases in expansion periods and smaller dips in
contraction phases is consistent with the larger trend rise of
New Zealand’s expansions of real private housing investment
last almost twice as long as those in the four CCG2 countries
and reflect an amplitude that is almost twice as large as
that in the comparator country group. Residential housing
contraction phases last about 3.4 quarters in New Zealand
(shorter than in the four CCG2 countries) and lead to a decline
in investment of 10.86%, similar to the CCG2 average.
real house prices in New Zealand. However, New Zealand’s
house price trend, volatility, and cycles are very similar to
Australia’s, during both booms and busts.
The timing and shape of the most recent expansionary cycle
(the 2001-2005 housing boom) has been very similar in
New Zealand, Australia, and Canada (Figure 8). Starting in
New Zealand’s residential construction activity has increased
in real terms at an average pace of 2.45% during 1986-2005,
while in the four CCG2 countries construction has risen by
2.7% per year. A similar variable is real private investment
2001-2002, the ratio of private housing investment to GDP
rose by approximately one percentage point in the three
countries until 2004, with some cooling off taking place in
2005.
in housing, which in New Zealand has grown by an average
4.1% per year, also slightly below the rate of growth of
Figure 8
private housing investment in four CCG2 countries (4.4%).
Private residential investment to GDP ratio in
In contrast, New Zealand exhibits the highest volatility in
New Zealand, Australia, Canada, and Norway,
real residential construction activity and real private housing
1970-2005 (% of GDP)
investment among the five countries, a likely reflection of
10
9
New Zealand’s larger population and immigration swings.
8
In fact, New Zealand’s average population growth, at
7
1.16% per year, is identical to Canada’s and slightly below
6
5
4
3
2
1
Aus tra lia
C anada
N ew Zealand
2004
2002
2000
1998
1996
1994
1992
1990
1988
1986
1984
1982
1980
1978
1976
1974
1972
0
Cross-country comparisons for housing permits should be
taken with caution due to methodological differences in data
definition and collection.
1970
5
N orway
Source: RBNZ and OECD Economic Outlook database
96
Reserve Bank of New Zealand and The Treasury
Terms of trade
current account deficit (by 1.93 percentage points) and larger
New Zealand’s terms of trade have risen by 1.34% per year
reductions (3.19 percentage points) than New Zealand.
during the last two decades. This is slightly above average
annual growth of terms of trade in the CCG2 group (0.8%).
New Zealand’s terms of trade are among the least volatile
of the six-country group, with the second lowest standard
deviation, after Canada’s.
Real effective exchange-rate
New Zealand has experienced a trend appreciation of its
real effective exchange-rate (REER) during the last two
decades, at an average pace of 1.33% per year, which is
New Zealand’s terms of trade decline 5.6 quarters on
average during a cyclical downturn while the mean duration
of the recovery phase is 6.4 quarters. Both periods are
shorter in New Zealand than in the four CCG2 country
significantly larger than the very slight average appreciation
experienced by the CCG2 country group (0.12% per year).6
New Zealand’s REER has been more volatile than the average
volatility of REERs in the CCG2 group.
group on average.
The mean duration of depreciation phases of the REER (peakThe mean amplitude of New Zealand’s terms of trade
recoveries is larger than that of downturns, which is
consistent with the trend gain in the terms of trade
during 1986-2005. The mean amplitude and cumulative
movements during contractions and expansions are
also smaller in New Zealand than in the CCG2 group on
average, a reflection of New Zealand’s lower terms of trade
volatility.
Current account balance
During the last two decades New Zealand’s current account
balance has remained negative at an average deficit level
of 4.61 per cent of GDP. This Figure is similar to Australia’s
average current account deficit (4.65% of GDP) over the
same period. Although New Zealand’s current account
deficit is large compared to the CCG2 group, its variability
is smaller.
to-trough) is 7.50 quarters in New Zealand and the mean
appreciation phase (trough-to-peak) lasts 6.75 quarters.
The mean amplitude of the REER shows that during periods
of depreciation the REER falls by 7.47%, while the REER
rises by 6.31% during episodes of appreciation. In contrast,
REER appreciation and depreciation periods are slightly
shorter and less intense in the average CCG2 country than
in New Zealand.
Correlations and the degree of concordance
Table 3, overleaf, reports two measures of co-movement
between different pairs of the selected macroeconomic
indicators for New Zealand and the simple averages of the
two co-movement measures for the four-country CCG2
group, for 1986-2005. The lower diagonal reports the
simple correlation coefficient between the corresponding
variables and its p-value. The upper diagonal reports the
Regarding the cyclical features of New Zealand’s current
account deficit, its expansionary phase, from current account
degree of concordance index developed by Harding and
Pagan (2006).
balance peak to trough, lasts 7.3 quarters. It’s narrowing
from trough to peak is shorter (5.5 quarters). In contrast,
CCG2 countries exhibit shorter periods of a widening current
account deficit (5.81 quarters) and longer periods of a
narrowing current account deficit (6.9 quarters).
Not surprisingly, the output gap and real GDP growth are
positively correlated: both in New Zealand and in the CCG2
country group, they show a positive correlation coefficient
of around 0.50. Periods of high growth are identified by
periods of an increasing gap between actual and potential
During periods of current account deficit widening,
New Zealand’s current account deficit to GDP ratio rises by
2.6 percentage points, while during the opposite periods
of narrowing, the deficit declines by 2.1 percentage points.
The CCG2 country group exhibits a smaller widening of the
Testing stabilisation policy limits in a small open economy
6
This trend real exchange-rate appreciation is largely
specific to the 1986-2005 sample period, which starts at an
exceptionally depreciated value and ends at an exceptionally
appreciated level (see Figure 6 and further discussion in
Section 6).
97
Table 3
Correlations and degree of concordance of selected indicators in New Zealand and CCG2
countries, 1986-2005
(Four CCG2 countries)
Current
account
TOT
RER
Output gap
GDP
Real House
House
GDP
Prices
permits
construction
Current account
1.00
0.54
0.52
0.56
0.55
0.47
0.51
0.48
TOT
0.06
1.00
0.61
0.53
0.56
0.55
0.58
0.59
RER
-0.20
0.54
1.00
0.48
0.52
0.61
0.57
0.57
Output gap
-0.06
0.18
0.11
1.00
0.68
0.58
0.57
0.63
GDP
0.20
-0.05
-0.32
0.50
1.00
0.71
0.68
0.75
Real House Prices
0.22
0.05
-0.10
0.41
0.49
1.00
0.71
0.69
House permits
0.29
0.02
-0.03
0.16
0.34
0.53
1.00
0.68
GDP construction
0.20
-0.04
-0.27
0.40
0.43
0.50
0.38
1.00
TOT
RER
Output gap
GDP
Real House
House
GDP
Prices
permits
construction
(New Zealand)
Current
account
Current account
1.00
0.60
0.35
0.46
0.40
0.28
0.41
0.24
TOT
0.20
1.00
0.40
0.36
0.55
0.50
0.41
0.45
RER
-0.19
0.25
1.00
0.64
0.65
0.80
0.66
0.76
0.09
0.02
1.00
0.69
0.64
0.65
0.68
1.00
0.73
0.64
0.67
1.00
0.71
0.80
1.00
0.68
1.00
0.08
Output gap
GDP
Real House Prices
House permits
GDP construction
Note:
98
-0.13
0.34
0.47
0.24
0.00
0.00
0.09
-0.33
0.11
0.65
0.42
0.00
0.32
0.00
-0.17
0.36
0.57
0.52
0.48
0.12
0.00
0.00
0.00
0.00
0.29
0.17
0.04
0.33
0.46
0.43
0.01
0.13
0.75
0.00
0.00
0.00
-0.27
0.06
0.09
0.39
0.74
0.56
0.58
0.03
0.63
0.45
0.00
0.00
0.00
0.00
quarterly data for 1986q1 - 2005q4. The lower diagonal cells represent bivariate correlation coefficients and, in the case of
New Zealand, p-values are reported under each coefficient. The upper diagonal cells represent the degree of concordance index,
which captures the co-movement between variables during their phases over the cycle. When two variables are exactly procyclical the index takes a value of 1, and when they are exactly counter-cyclical the index takes value of 0. Values close to 0.5
represent no relation. The method is based on Harding and Pagan (2006).
Reserve Bank of New Zealand and The Treasury
output. This result is also supported by the index of
countries, these two variables are positively related, with an
concordance between the output gap and GDP. Housing
average correlation coefficient of 0.54.
market variables are typically pro-cyclical. In New Zealand,
the index of concordance between real GDP and real house
prices is 0.73, while the correlation between the output gap
and real house prices is positive and significant (0.65). House
permits and real activity in residential construction exhibit
similar figures. In the four CCG2 countries, on average,
house prices and real activity in residential construction are
pro-cyclical with an average index of concordance around
0.7 and a positive correlation with the output gap of around
0.4.
Real house prices and residential construction activity
There is no evidence of a significant correlation between
real GDP and the current account balance in either
New Zealand or the CCG2 group. In contrast, New Zealand
exhibits a negative correlation between the REER and the
current account balance, i.e., appreciations are related to
current account deficits. The index of concordance between
the latter variables is 0.35 and the correlation coefficient
is negative and significant. The negative co-movement
between the REER and the current account balance is also
observed in the CGG2 countries, exhibiting a negative
correlation of -0.20.
are negatively related to the current account balance
in New Zealand, with an index of concordance of 0.28
and 0.24, respectively. Correlations between the current
account and housing variables are also negative but only
significant for residential construction activity (-0.27).
Negative correlations are not observed in the average CCG2
country, where the index of concordance suggests no
apparent relationship.
3
Overall macroeconomic
policy mix
Table 4, overleaf, summarizes distinctive features of
New Zealand’s overall macroeconomic policy mix in
comparison to those of the CCG2 country group. It
is apparent that there is little country variation in the
Real house prices and real activity in residential construction
fundamental features of the monetary policy regime
are strongly positively related to the REER, with an index of
(they are all inflation targeters), the exchange-rate regime
concordance equal to 0.80 and 0.76, respectively. However,
(floating), fiscal soundness, domestic financial policy, and
only the correlation coefficient between real house prices
external financial policy (absence of any capital controls).
and the REER is significant (0.57). This evidence is not
limited to the case of New Zealand – it is quite similar in the
average CCG2 country.
However, there are several policy differences between
New Zealand and comparator countries that should be
noted. On one hand, New Zealand’s monetary policy
New Zealand’s terms of trade show a positive correlation
horizon is more imprecise than that in many inflation-
with the output gap (0.33) but the index of concordance
targeting comparator countries. New Zealand’s policy
is not conclusive about the pro-cyclical nature of the terms
aims at attaining inflation at target levels “over the
of trade (0.55). In contrast, the CCG2 countries present
medium term”, in contrast to the more specific horizons
a smaller correlation coefficient between terms of trade
defined by most other central banks. On the other hand,
and the output gap (0.18). New Zealand’s current account
deviations from medium-term target ranges are required
balance shows a mild positive correlation with the terms
to be corrected by appropriate policy and communication
of trade, with an index of concordance of 0.60 and a
actions in New Zealand – an exemption clause stipulated
correlation coefficient that is small but significant at 0.18.
in New Zealand’s current Policy Target Agreement (Reserve
The index of concordance between the terms of trade and
Bank of New Zealand 2002) that reflects in this dimension
the REER is 0.40 in New Zealand, while the correlation
a stricter framework than those in most inflation-targeting
between is small but positive and significant (0.25). In CCG2
countries, where exemption clauses tend to be absent. Yet
New Zealand’s monetary policy credibility and effectiveness
Testing stabilisation policy limits in a small open economy
99
Table 4
Monetary, fiscal, external, and financial policies in New Zealand and CCG2, circa mid-2006
New Zealand
Australia
Canada
Inflation targeting
Monetary policy
Monetary regime
Inflation targeting
Inflation target indicator
CPI annual percent change Headline CPI annual
percentage change
CPI annual percent change
Inflation targeting
Inflation target
1-3 %
2-3 %
1-3%
Target horizon
On average over the
medium term
Average over the business
cycle
Through to the end of
2006 6-8 quarter horizon
Operational target:
Inflation forecast
Publication/indication of
future interest rate path:
Yes
No
No
Exemption clause
When certain shocks
cause actual inflation to
be outside the mediumterm target range,
the RBNZ shall take
corrective policy actions
in a transparent and
accountable manner
None
None
Exchange rate policy
Floating ER
Yes
Under normal conditions
Yes
Interventions
None
Infrequent and under
exceptional circumstances
Under specific
circumstances
Intervention framework
Explicitly announced
None
Explicitly announced
Fiscal policy
Policy stance
Sustainable
Sustainable
Sustainable
Explicitly counter-cyclical
No
Yes
No
Intertemporal saving
No
No
No
Prudential regulation
Strong
Strong
Strong
Banking-sector strength
Large
Large
Large
Capital controls
None
None
None
Currency composition of
Public/external debt
Mostly domestic-currency
debt, net foreign-currency
debt is kept close to zero
Foreign
Foreign
Financial policy
External financing policy
100
Reserve Bank of New Zealand and The Treasury
Table 4
Monetary, fiscal, external, and financial policies in New Zealand and CCG2, circa mid-2006
(contd.)
Chile
Norway
Sweden
Monetary regime
Inflation targeting
Inflation targeting
Inflation targeting
Inflation target indicator
CPI annual percent
change
CPI annual percent
change
CPI annual percent
change
Inflation target
3% plus minus 1%
2.50%
2% plus minus 1%
Target horizon
12-24 months (mostly:
24 months)
Over the medium term,
usually I to 3 years
2 years horizon
Operational target:
Inflation forecast
Exemption clause
None
None
Justified in grounds
of consideration to
developments in the real
economy
Role of exchange rate in MP
In as much affects
inflation (and output)
None
Takes into consideration
exchange rate as other
asset prices
Floating ER
Under normal conditions
Yes
Under normal conditions
Interventions
Under exceptional
conditions
Under exceptional
circumstances
Yes, not clear under what
conditions
Intervention framework
Explicitly announced
at start of intervention
period
Monetary policy
Inflation forecast
Exchange rate policy
In the local foreignexchange market
Fiscal policy
Policy stance
Sustainable
Sustainable
Sustainable
Explicitly counter-cyclical
Yes
Yes
Yes
Intertemporal saving
No
Yes
Yes
Prudential regulation
Strong
Strong
Strong
Banking-sector strength
Large
Large
Large
Capital controls
None
None
None
Currency composition of Public/
external debt
Foreign
Foreign
Foreign
Financial policy
External financing policy
Sources: Central bank websites and author’s assessment
Testing stabilisation policy limits in a small open economy
101
could be strengthened by announcing a more specific time
pass-through from exchange-rate shocks in IT countries,
frame for the policy horizon.
like New Zealand, than in non-inflation targeting (NIT)
Second, like the Bank of Canada and the Central Bank of
Chile, the RBNZ has made public an explicit exchange-market
intervention scheme, in contrast to most other (intervening)
central banks in the world. And the RBNZ’s intervention
scheme spells out the conditions that would trigger a forex
intervention in a more detailed and transparent way than
countries. In particular, I will measure response of inflation
to exchange-rate and oil-price shocks in New Zealand and
three relevant country groups, selected by their monetary
regime and/or a specific period:
(i) New Zealand, 1989-1997,
(ii) New Zealand, 1998-2005,
other central banks.
(iii) ITers after the adoption of IT (excluding New Zealand),
Finally, New Zealand does not have in place an explicit
counter-cyclical fiscal policy, in contrast to Chile, Sweden,
(iv) Industrial ITers after the adoption of IT (excluding
New Zealand), and
and Norway. A strong, explicit counter-cyclical policy
framework could have an important stabilizing influence on
the exchange-rate and the current account balance (among
other key macro variables), as argued in section 7 below.
(v) NITers, 1998-2004.
Following Mishkin and Schmidt-Hebbel (2005), I estimate
the response for each group by estimating panel vector
autoregressive (Panel VAR) models, based on quarterly
data.7 The VAR system includes the following six variables
4
Monetary policy transmission
and efficiency
In this section I assess New Zealand’s relative monetary
policy transmission and efficiency across four dimensions.
First, I measure the extent of exchange-rate and oil price
pass-through to inflation. Second, I identify monetary policy
transmission to financial markets, output, and inflation.
Third, I provide an international comparison of inflation
target accuracy as a partial way of assessing monetary policy
efficiency. Finally, this section focuses on overall monetary
policy efficiency, reporting how New Zealand has reduced
inflation and output volatility over time and in comparison
to other inflation-targeting countries.
ranked by exogeneity: international oil price changes,
international interest rates, the output gap, annual inflation,
domestic interest rates, and the nominal exchange-rate.
Since the model yields similar impulse response functions
using two or more lags, for reasons of parsimony a lag
order of two was selected. To control for possible fixed
effects (correlated with the regressors due to lags of the
dependent variable) I use forward mean-differencing
(Helmert procedure) to remove the mean of all the future
observations available for each country. Lagged regressors
are used as instruments and coefficients are estimated by
General Method of Moments (GMM). Finally, the dynamic
responses to innovations in the system are identified using
the Choleski decomposition of the variance-covariance
matrix of residuals and their confidence intervals are
Exchange-rate and oil price pass-through
constructed by bootstrap methods.8 9
,
in New Zealand, CCG1 countries, and non-
Next I compare the dynamic response to shocks for
inflation targeting countries
New Zealand and the corresponding country groups. I
One way to assess monetary policy efficiency is by comparing
the effect of exchange-rate shocks and oil price shocks
7
Except for New Zealand in (i) and (ii), for which I estimate a
simple country VAR model with the same variable ordering,
frequency, time span, and estimation method.
8
See Holtz-Eakin et al., 1988; Love and Zicchino, 2002; and
Miniane and Rogers, 2003, for applied studies using Panel
VAR estimation.
Confidence intervals for responses of groups i) and ii) are also
estimated by bootstrap.
on inflation in New Zealand and relevant country control
groups. If IT improves the credibility of monetary policy
and helps to anchor inflation expectations, inflation would
9
respond less to oil price shocks and there would be a weaker
102
Reserve Bank of New Zealand and The Treasury
Figure 9
Dynamic response of headline inflation to an exchange-rate shock in New Zealand and CCGs
(7 quarters)
New Zealand vs ITERS and NON-ITers
New Zealand before 1997
New Zealand (1998-2005)
Difference
0.30
0.30
0.30
0.20
0.20
0.20
0.10
0.10
0.10
0.00
0.00
0.00
-0.10
-0.10
-0.10
-0.20
-0.20
-0.20
-0.30
-0.30
-0.30
-0.40
-0.40
-0.40
-0.50
-0.50
-0.50
0
1
2
3
4
5
0
6
New Zealand (1998-2005)
1
2
3
4
5
0
6
ITers after start of IT
0.30
0.30
0.30
0.20
0.20
0.20
0.10
0.10
0.10
0.00
0.00
0.00
-0.10
-0.10
-0.10
-0.20
-0.20
-0.20
-0.30
-0.30
-0.30
-0.40
-0.40
-0.40
-0.50
0
1
2
3
4
5
0
New Zealand (1998-2005)
1
2
3
4
5
6
Industrial ITers after start of IT
0.30
0.30
0.20
0.20
0.10
0.10
0.00
0.00
-0.10
-0.10
-0.20
-0.20
-0.30
-0.30
-0.40
-0.40
-0.50
1
2
3
4
5
6
0
1
2
3
0.00
-0.10
-0.20
-0.30
-0.40
1
2
3
4
5
-0.50
6
0.20
0.10
0.10
0.10
0.00
0.00
0.00
-0.10
-0.10
-0.10
-0.20
-0.20
-0.20
-0.30
-0.30
-0.30
-0.40
-0.40
-0.40
-0.50
-0.50
6
4
5
6
Difference
0.30
5
6
0.10
0.20
4
5
3
0.20
3
4
2
0.20
Non-ITers (1998-2005)
2
6
1
0.30
1
5
0
0.30
0
4
0.30
0
New Zealand (1998-2005)
3
Difference
-0.50
0
2
-0.50
-0.50
6
1
Difference
-0.50
0
1
2
3
4
5
6
0
1
2
3
4
5
6
report significant differences (statistically different from
Figures 9 and 11 depict the dynamic response of inflation
zero) between the response in New Zealand before and
to an exchange-rate shock (an exchange-rate appreciation)
after 1998 (group i vs. group ii), in New Zealand after 1998
and an oil price shock, respectively. Each row of three small
and after the adoption of IT in ITers (group ii vs. group iii), in
Figures focuses on a comparison between two particular
New Zealand after 1998 and after adoption of IT in industrial
sample groups. For instance, the second row of Figure 9
ITers (group iii vs. group iv), and in both New Zealand and
depicts the response of New Zealand in 1998-2005 (first
NITers after 1998 (group ii vs. group v).
column) to the response of ITers after adopting IT (second
Note that the assumption of independence between
my samples might be not appropriate. Hence I also use
bootstrap methods to construct confidence intervals for
differences in impulse-response functions instead of simply
taking their differences.10
column). The corresponding difference in response, and its
confidence interval, are depicted in the third column. Figures
10 and 12 focus on New Zealand alone, contrasting the
differences between dynamic pass-through of an exchangerate shock (Figure 10) and an oil-price shock (Figure 12) to
headline inflation and tradable-goods inflation.
The response of New Zealand’s headline inflation to an
10
If we were simply to assume sample independence, the
corresponding confidence intervals for differences would be
narrower.
Testing stabilisation policy limits in a small open economy
exchange-rate shock is not significantly different from
103
Figure 10
Dynamic response of headline inflation and tradables inflation to an exchange-rate shock in
New Zealand, 1989-1997 and 1998-2005 (7 quarters)
Headline inflation (1989-1997)
Tradables inflation (1989-1997)
0.3
0.3
0.2
0.2
0.1
0.1
0.0
0.0
0
1
2
3
4
5
6
-0.1
0
1
2
3
4
5
6
-0.1
-0.2
-0.2
-0.3
-0.3
-0.4
-0.4
-0.5
-0.5
Headline inflation (1998-2005)
Tradables inflation (1998-2005)
0.3
0.3
0.2
0.2
0.1
0.1
0.0
0.0
1
0
1
2
3
4
5
6
2
3
4
5
6
7
-0.1
-0.1
-0.2
-0.2
-0.3
-0.4
-0.3
-0.4
-0.5
-0.5
zero in the early 1989-1997 period (Figure 9). In contrast,
due to lower credibility of monetary policy, in emerging
during the recent 1998-2005 period, headline inflation
economies. Among NITers the pass-through impulse
responds significantly and negatively to an exchange-rate
response is not significantly different from zero. The
appreciation from the second to the fifth quarter after the
differences in inflation responses between all IT countries
shock. The pass-through coefficient attains a maximum
and New Zealand are negative and statistically significant
value close to 0.15 in the third quarter, an estimate that is
from zero; pass-through coefficients are significantly larger
in line with Hampton’s (2002) estimate of a pass-through
in quarters 1 through 5 in all ITers, on average, than in
of 0.15 within 12 months. The differences in pass-through
New Zealand. However, the difference between industrial
estimates between the second and the first period (depicted
ITers’ pass-through and New Zealand’s is not significantly
in the last column of the first row of Figure 9) are small and
different from zero, except in quarter 1, when it is larger
only significant for quarters 2 and 3.
among industrial ITers. Finally, the pass-through coefficient
The inflation response in comparator groups is negative and
significant in the group of all ITers and among industrial-
is significantly larger in quarters 1 and 2 in New Zealand
than among NITers.
country ITers as well. However the pass-through effect is
Transmission of an exchange-rate shock to inflation may
much larger and takes longer among all ITers (comprising
change considerably if we consider inflation of tradable
both emerging-economy and industrial country inflation
goods. Figure 10 depicts the responses of headline inflation
targeters) than among industrial-country ITers alone. This
and tradables inflation to an exchange-rate for New Zealand
reflects the larger role of exchange-rate shocks, possibly
in both sample periods. Both headline and tradables
104
Reserve Bank of New Zealand and The Treasury
Figure 11
Dynamic response of headline inflation to an oil-price shock in New Zealand and CCGs
(7 quarters)
New Zealand vs ITERS and NON-ITers
New Zealand before 1997
New Zealand (1998-2005)
Difference
1.0
1.0
0.03
0.8
0.8
0.02
0.6
0.6
0.4
0.4
0.01
0.00
0.2
0.2
-0.01
0.0
0.0
-0.02
-0.2
-0.2
-0.4
-0.4
-0.03
-0.6
-0.6
0
1
2
3
4
5
0
6
New Zealand (1998-2005)
1
2
3
4
5
6
-0.04
0
ITers after start of IT
1.0
0.6
0.8
0.8
0.4
0.6
0.6
0.2
0.4
0.4
0.2
0.2
0.0
0.0
-0.2
-0.2
-0.4
-0.4
1
2
3
4
5
New Zealand (1998-2005)
1.0
0.8
0.8
0.6
0.6
0.4
0.4
0.2
0.2
0.0
0.0
-0.2
-0.2
-0.4
-0.4
-0.6
-0.6
0
1
2
3
4
5
1
2
3
4
5
0.8
0.8
0.6
0.6
0.4
0.4
0.2
0.2
0.0
0.0
-0.2
-0.2
-0.4
-0.4
1
2
3
4
5
1
2
6
3
5
6
4
5
6
4
5
6
Difference
0.6
0.4
0.2
0.0
-0.2
-0.4
-0.6
1
2
3
4
5
6
0
1
2
0
1
2
3
4
5
3
Difference
-0.6
-0.6
0
0
6
Non-ITers (1998-2005)
1.0
4
-0.8
New Zealand before 1997
1.0
6
-0.6
0
6
5
-0.4
Industrial ITers after start of IT
1.0
4
0.0
0
6
3
-0.2
-0.6
-0.6
2
Difference
1.0
0
1
6
1.0
0.8
0.6
0.4
0.2
0.0
-0.2
-0.4
-0.6
-0.8
0
1
2
3
inflation are not affected significantly by an exchange-rate
The response of New Zealand’s headline inflation to an
shock in quarters 1 through 6 after an exchange-rate shock
international oil-price shock is positive and significant during
in the earlier period. This result changes radically in the
the first quarter in the 1989-1997 sample (Figure 11). During
second period, when both measures of inflation respond
the 1998-2005 period, the inflation effect of an oil-price
significantly to the exchange-rate shock. Not surprisingly,
hike is similar in magnitude (close to 0.25) but slightly more
the pass-through to tradables inflation is much larger and
persistent, extending from quarters 0 through 1. However,
somewhat quicker than pass-through for headline inflation.
the inflation effect is significantly but just slightly larger in
The maximum pass-through coefficient for tradables is
the first period in comparison to the second period – and
0.31, slightly more than twice the maximum pass-through
that occurs only in quarter 1).
estimate for headline inflation, both attained in quarter 3.
Testing stabilisation policy limits in a small open economy
105
Figure 12
Dynamic response of headline inflation and tradables inflation to an oil-price shock in
New Zealand, 1989-1997 and 1998-2005 (7 quarters)
Headline inflation 1989-1997
Tradables Inflation 1989-1997
1.0
1.0
0.8
0.8
0.6
0.6
0.4
0.4
0.2
0.2
0.0
0.0
1
2
3
4
5
6
7
0
-0.2
-0.2
-0.4
-0.4
-0.6
-0.6
Headline inflation 1998-2005
1
2
3
4
5
6
Tradables Inflation 1998-2005
1.0
1.0
0.8
0.8
0.6
0.6
0.4
0.4
0.2
0.2
0.0
0.0
1
2
3
4
5
6
7
1
2
3
4
5
6
7
-0.2
-0.2
-0.4
-0.4
-0.6
-0.6
The oil shock impact on inflation is also positive and
the 1989-1997 period, the effect is on average larger and
significant in the three country groups considered here: all
more persistent during 1998-2005. Now significant effects
ITers, industrial ITers, and NITers. In the group of all ITers,
are measured during quarters 0 through 3 after the oil
the oil price effect on inflation is quick, short-lived (from
shock, with a coefficient that peaks close to 0.60 at quarter
quarters 0 to 1), relatively small (close to 0.20), and is
1 and averages 0.52 during the four quarters.
very similar to New Zealand’s. However, among industrialcountry ITers the oil price to inflation response is more
persistent (extending from quarters 0 through 4) although
of a similar small size. A similar persistence and size of the
oil-price effect on headline inflation is observed among
NITers but here the oil shock to inflation coefficient peaks at
0.35 at quarter 3. Most interesting, no statistical differences
in response coefficients from oil shocks to headline inflation
are detected between New Zealand 1998-2005 and any of
the three country groups.
Like in the case of the exchange-rate shock, the effect of an
I conclude that exchange-rate pass-through to headline
inflation is significant but relatively small in New Zealand.
It is close to a maximum pass-through coefficient of
0.15 during 1998-2005, a likely result of well-anchored
inflation expectations. The dynamic impulse response
pattern in New Zealand is very similar to that observed in a
representative industrial country under inflation targeting.
Pass-through to tradable goods inflation is much larger and
somewhat quicker than pass-through to headline inflation
in New Zealand, which also accords with international
experience.
oil price shock on inflation is much larger and protracted for
tradables than for headline inflation in New Zealand (Figure
12), as elsewhere. While a large and significant effect of an
oil-price hike on tradables inflation is already apparent in
106
Reserve Bank of New Zealand and The Treasury
The dynamic response of headline inflation to an oilprice hike is also relatively small in New Zealand, close to
a maximum oil-price pass-through coefficient of 0.25 in
1998-2005, again a likely reflection of a credible monetary
policy and stable inflation expectations. The size and
dynamics of the oil-price to headline inflation response
in New Zealand is statistically not different from those
observed in all comparator country groups, both with
and without inflation targeting. New Zealand’s tradables
inflation is much more and more persistently affected by an
oil shock than headline inflation, a result also in line with the
international evidence.
Vector
yt includes the nine following variables: US Federal
*
Funds rate ( it ), US output gap
( yt* ) , US headline inflation
( π t ), domestic output gap ( yt ) , deviation of domestic
*
headline inflation from the inflation target
(π t − π tT ) ,
domestic short-term interest rate (it ) , money deviation
from trend ( mt ), (log) exchange-rate
(et ) , and long-term
interest rate ( Rt ). I assume that the US Federal Funds rate
does not respond to any domestic variable but only to
changes in the US output gap or inflation. This specification
is similar to the VAR models by Eichenbaum and Evans
(1995), Grilli and Roubini (1995, 1996), and more recently
by Scholl and Uhlig (2005).11
To identify the structural policy shocks, I follow Sims (1980)
Monetary policy transmission in New Zealand
in assuming recursive identification using the Choleski
and CCG2
decomposition of the reduced-from residuals. Based
Now I turn to assessing the effects of monetary policy
on previous work, I assume that the US output gap and
innovations in New Zealand, in comparison to the evidence
inflation are contemporaneously exogenous to the US
on the monetary transmission mechanism in the CGG2
monetary policy instrument, and this is also the case for the
country group. For this purpose I use a standard country
domestic economy. Therefore, the variables are ordered,
VAR model of the literature developed for identifying
starting with the most exogenous, as follows:
monetary policy shocks (e.g., Christiano et al. 1996, Evans
yt , (π t − π tT ) , it , mt , et , and Rt .
and Marshall 1998, and Kim 2001). I will focus on the
dynamic effects of monetary policy shocks on short-term
interest rates, long-term interest rates, the exchange-rate,
real output, and inflation.
Let yt be a vector containing a set of macroeconomic
variables at time t, and let Rt denote a long-term bond
yield. Following Marshall and Evans (1998), I assume that
the economy is described by the following system of
equations:
a b   yt   A( L) B( L)   yt −1   ety 
(1) 
+ R 

  = 

 c 1   Rt   C ( L) D( L)   Rt −1   et 
where a is a square matrix with ones on the diagonal, b is
a scalar, c is a row vector, A(L) is a matrix polynomial in the
lag operator L, C(L) is a row vector polynomial, B(L) and
y
R
D(L) are scalar polynomials, and the process (et et ) is a
vector of structural disturbances, serially uncorrelated with
variance
Λ . I assume that B( L) = 0 and b = 0 , implying
that neither contemporaneous nor lagged values of the
bond yield enter the other system equations.
Testing stabilisation policy limits in a small open economy
yt*, π t* , it* ,
Figure 13 reports quarterly impulse responses to a
contractionary monetary policy shock (a 25 basis point
rise in the short-term interest rate), with 95% confidence
intervals over six years. The first column shows the results
for New Zealand for the 1990-2005 period, while the
second column restricts observations to the 1998-2005
period. There is a trade-off between the larger sample size
of the 1990-2005 sample, on one hand, and the larger
sample homogeneity and higher relevance for today’s policy
of the shorter 1998-2005 sample. Weighting more heavily
the larger sample size, I tend to favour the results from
the full sample and therefore conduct the cross-country
comparison below on the basis of the 1990-2005 sample,
too. Therefore the subsequent columns in Figure 13 depict
the individual impulse responses for the five comparator
countries, based on 1990-2005 data.
11
For robustness I have also estimated alternative VAR country
models, allowing for the lags of the long-term interest rate to
feed back into the system, as in Bernanke, Gertler and Watson
(1997). However, the country impulse responses derived
from the latter alternative specification are not qualitatively
different from those reported below.
107
Figure 13
Dynamic response to a monetary policy shock in New Zealand, 1990-2005 and 1998-2005 and in
CCG2, 1990-2005 (25 quarters)
New Zealand 1990-2005
New Zealand post-1997
0.4
0.35
0.30
0.3
0.25
Short-term
interest rate
0.20
0.2
0.15
0.10
0.1
0.05
0.0
0.00
-0.05
0
1
2
3
4
5
0
6
1
2
3
4
5
6
-0.1
-0.10
-0.15
-0.2
0.15
0.08
0.06
0.10
0.04
Long-term
interest rate
0.05
0.02
0.00
0.00
0
1
2
3
4
5
-0.02
6
-0.05
0
1
2
3
4
5
6
0
1
2
3
4
5
6
0
1
2
3
4
5
6
0
1
2
3
4
5
6
-0.04
-0.06
-0.10
-0.08
-0.15
-0.10
0.02
0.02
0.02
0.02
Nominal
exchange
rate (US$/
Local Curr.)
0.01
0.01
0.01
0.01
0.00
0.00
0
1
2
3
4
5
-0.01
6
-0.01
Inflation
deviation
from target
-0.01
-0.01
-0.02
0.20
0.40
0.15
0.30
0.10
0.20
0.05
0.10
0.00
0.00
0
1
2
3
4
5
6
-0.05
-0.10
-0.10
-0.20
-0.15
-0.30
0.06
0.15
0.04
0.10
0.02
0.00
Output
gap
-0.02
0
1
2
3
4
5
0.05
6
0.00
-0.04
-0.06
-0.05
-0.08
-0.10
-0.10
-0.12
-0.15
-0.14
-0.16
-0.20
Years
108
Reserve Bank of New Zealand and The Treasury
Figure 13
Dynamic response to a monetary policy shock in New Zealand, 1990-2005 and 1998-2005 and in
CCG2, 1990-2005 (25 quarters) (cont.)
Australia
Canada
0.35
0.4
0.30
0.3
0.25
Short-term
interest rate
0.20
0.2
0.15
0.10
0.1
0.05
0.0
0.00
-0.05
0
1
2
3
4
5
6
0
1
2
3
4
5
6
0
1
2
3
4
5
6
-0.1
-0.10
-0.15
-0.2
0.25
0.3
0.20
0.3
0.2
0.15
Long-term
0.10
interest rate
0.05
0.2
0.1
0.1
0.00
0
1
2
3
4
5
6
-0.05
-0.10
-0.1
-0.15
-0.2
0.03
0.010
0.03
0.008
0.006
0.02
Nominal
exchange
rate (US$/
Local Curr.)
0.0
-0.1
0.004
0.02
0.002
0.01
0.000
0.01
0
1
2
3
4
5
6
0
1
2
3
4
5
6
0
1
2
3
4
5
6
-0.002
0.00
0
1
2
3
4
5
6
-0.004
-0.01
-0.006
0.15
0.100
0.10
0.050
0.05
Inflation
deviation
from target
0.000
0.00
0
1
2
3
4
5
6
-0.05
-0.050
-0.10
-0.100
-0.15
-0.150
-0.20
-0.25
-0.200
0.06
0.060
0.040
0.04
0.020
0.02
Output
gap
0.000
0.00
0
1
2
3
4
5
6
-0.020
-0.040
-0.02
-0.060
-0.04
-0.080
-0.06
-0.100
-0.08
-0.120
-0.10
-0.140
Years
Testing stabilisation policy limits in a small open economy
109
Figure 13
Dynamic response to a monetary policy shock in New Zealand, 1990-2005 and 1998-2005 and in
CCG2, 1990-2005 (25 quarters) (cont.)
Norway
Sweden
Short-term interest rate
0.30
0.35
0.30
0.3
0.25
0.25
0.2
0.20
0.20
0.15
0.1
0.15
0.10
0.10
0.0
0
0.05
1
2
3
4
5
6
-0.1
0.00
-0.05
Chile
0.4
0.35
0
1
2
3
4
5
6
0.00
-0.05
-0.2
-0.3
-0.10
-0.4
-0.20
0.12
0.2
0.10
Long-term interest rate
2
3
4
5
6
0
1
2
3
4
5
6
0.08
0.1
0.08
0.06
0.06
0.04
0.0
0
0.04
0.02
1
2
3
4
5
6
-0.1
0.02
0.00
-0.02
0.00
0
1
2
3
4
5
6
-0.04
-0.2
-0.04
-0.06
-0.3
-0.06
-0.08
-0.08
-0.4
0.010
Nominal exchange
rate (US$/Local Curr.)
1
-0.15
-0.15
-0.02
0
-0.10
0.10
Inflation deviation
from target
0.05
0.008
-0.10
0.03
0.010
0.02
0.008
0.006
0.02
0.006
0.004
0.01
0.004
0.01
0.002
0.00
0.002
0.000
-0.01
0.000
0
1
2
3
4
5
6
-0.002
0
1
2
3
4
5
6
-0.002
-0.01
-0.008
-0.004
-0.02
-0.010
0.140
0.20
0.150
0.120
0.15
0.100
0.100
0.10
0.080
0.05
0.060
0.00
-0.05
0.020
-0.10
0
1
2
3
4
5
6
3
4
5
6
0
1
2
3
4
5
6
0
1
2
3
4
5
6
0.000
0
1
2
3
4
5
6
-0.050
-0.100
-0.150
-0.200
-0.20
-0.040
-0.25
-0.250
-0.060
-0.30
-0.300
0.020
2
0.050
-0.15
0.000
-0.020
1
-0.006
-0.02
0.040
0
-0.004
0.15
0.150
0.10
0.100
0.05
0.050
0.000
Output gap
0
1
2
3
4
5
6
0.00
-0.020
-0.05
-0.040
-0.060
-0.080
-0.100
0.000
0
1
2
3
4
5
6
-0.050
-0.10
-0.100
-0.15
-0.150
-0.20
-0.200
-0.25
-0.250
-0.30
-0.300
Years
110
Reserve Bank of New Zealand and The Treasury
Figure 14
Dynamic response to a monetary policy shock in New Zealand, 1990-2005 and 1998-2005,
alternative VAR model (25 quarters)
New Zealand 1990-2005
Short-term
interest rate
New Zealand 1998-2005
0.50
0.4
0.40
0.3
0.30
0.2
0.20
0.1
0.10
0.0
0.00
-0.1
0
0
1
2
3
4
5
1
2
3
4
5
6
6
-0.10
-0.2
-0.20
-0.3
0.15
0.20
0.15
0.10
0.10
Inflation
deviation
from target
0.05
0.05
0.00
0.00
0
1
2
3
4
5
6
0
1
2
3
4
5
6
0
1
2
3
4
5
6
-0.05
-0.05
-0.10
-0.10
-0.15
0.15
0.15
0.10
0.10
Output
gap
0.05
0.05
0.00
0.00
0
1
2
3
4
5
-0.05
6
-0.05
-0.10
-0.10
-0.15
-0.15
-0.20
-0.20
-0.25
-0.25
Years
Let’s start by focusing on the results for New Zealand from
more than offsets a temporary positive response of inflation
both samples. Short-term interest-rate dynamics are similar
(a price puzzle) in the first quarter after the interest rate
in both periods, although they adjust somewhat more
hike. The output gap shrinks in response to the interest
quickly since 1998, possibly suggesting more monetary
hike. While the negative output effect is persistent but not
policy activism. Long-term interest rates respond significantly
significant in the full sample, it is negative and significant
on impact and, in the case of the full sample results, also
during quarters 2 through 4 in the 1998-2005 sample, with
in the first period after the short-term rate increase. While
subsequent oscillatory dynamics.
the exchange-rate appreciates in response to the short-term
rate increase, the appreciation is small and not significantly
different from zero.
In order to check robustness of the latter results on
monetary transmission in New Zealand, I estimate for both
sample periods an alternative, smaller VAR model. Included
A significant negative inflation response in quarters 4
variables comprise the international price of oil, domestic
through 9 to the monetary contraction is found for the full
output gap, deviation of domestic headline inflation from
sample. This significant and persistent reduction in inflation
Testing stabilisation policy limits in a small open economy
111
the inflation target, domestic short-term interest rate, and
from zero in New Zealand. Inflation declines significantly in
the nominal exchange-rate. The impulse responses for both
New Zealand in the second year after the monetary shock,
sample periods are reported in Figure 14. In comparison to
similar to what is observed in two comparator countries and
the results based on the original VAR, reported in Figure
opposed to the non-significant inflation response in the
13, these new results strengthen the evidence on monetary
other three countries. The output gap responds negatively
policy transmission in New Zealand, in particular for the full
and persistently (and significantly under an alternative VAR
1990-2005 sample. The negative and significant response
specification) in New Zealand, like in most other comparator
of inflation to the monetary policy contraction extends now
countries.
from quarters 5 through 9 and the negative output gap
response is now significant during a long period, extending
from quarters 5 through 13.
Now I turn to comparing monetary transmission in
New Zealand to the other five countries, returning to the
larger VAR specification and the corresponding country
Table 5
Mean absolute deviations of annual inflation
rates from inflation targets in 21 inflation
targeting countries, since start of inflation
targeting
Mean absolute deviation
impulse responses depicted in Figure 13. A short-term
Country
interest rate hike has a positive, significant, and persistent
effect on long-term rates in Australia, Canada, and Norway.
from target
percentage
as % of
points
target
However, in Sweden and Chile the response on long-term
rates is not significantly different from zero. New Zealand’s
case is closer to the first group, showing a positive and
Australia
1.20
47.89
Brazil
4.07
94.91
Canada
0.93
38.91
Chile
1.29
20.53
rates. Like in New Zealand, no significant exchange-rate
Colombia
1.25
13 .46
appreciation is identified in Australia, Canada, and Chile.
Czech Republic
2.03
51.56
However, for Norway and Sweden I find some evidence
Hungary
2.25
56.55
of a temporary significant exchange-rate appreciation in
Iceland
1.62
54.66
Israel
2.43
58.01
Korea
1.02
28.34
significant but more short-lived effect on long-term
response to the monetary contraction.
Like in New Zealand but somewhat more persistently,
Mexico
1.26
24.95
an interest rate rise reduces inflation significantly from
New Zealand
0.94
55.98
quarters 4-5 onwards in Australia and Sweden. However,
no significant effects on inflation are found for Canada,
Norway
1.25
50.15
Peru
1.68
29.52
Phillipines
1.68
32.73
Poland
2.27
44.68
Canada, exhibit a negative, significant, and persistent
South Africa
2.82
62.73
dynamic response of the output gap to an interest rate
Sweden
1.12
56.01
hike.
Switzerland
0.40
39.85
Thailand
0.63
35.80
United Kingdom
0.89
36.07
Average
1.57
44.44
exc/. New Zealand
1.60
43.87
Norway, and Chile. Finally, all comparator countries, except
From this evidence I conclude that monetary policy
transmission in New Zealand is broadly comparable
to, and at least as strong as, that observed in the CCG2
country group. Like in three of the five CCG2 countries,
but exhibiting less persistence, long-term rates respond
Source: Mishkin and Schmidt-Hebbel (2005)
Note:
Sample periods depend on the IT regime starting date.
New Zealand data are for 1990q1-2005q4.
to short-term rates in New Zealand. Like in four of the five
countries, the exchange-rate appreciation in response to
a short-term interest rate rise is not significantly different
112
Reserve Bank of New Zealand and The Treasury
Inflation targeting accuracy
Figure 15
Here I compare the accuracy of New Zealand in hitting
Mean absolute deviations of annual inflation
its official inflation target with the IT accuracy observed
rates from inflation targets in 21 inflation
in a sample of 20 ITers. Following Albagli and Schmidt-
targeting countries, since start of inflation
Hebbel (2005), I report the size, frequency, intensity, and
targeting
persistence of actual inflation deviations from inflation
targets, computing (i) the mean absolute deviations of
%
%
2.0
2.0
1.8
New Zealand
CCG2 *
Others **
1.8
inflation rates from target levels, (ii) the frequency of
1.6
deviations, (iii) the mean duration of deviations, and (iv)
1.4
1.4
the mean amplitude of deviations. Tables 5-7 summarize
1.2
1.2
1.0
1.0
0.8
0.8
quarterly data from the quarter at which the IT regime was
0.6
0.6
started in the corresponding country until the last quarter of
0.4
0.4
0.2
0.2
the latter statistics for New Zealand and the 20 ITers, using
2004, except for New Zealand which contains information
from the first quarter of 1990 through the fourth quarter
of 2005.12
0.0
0.0
1990-1997
Note:
New Zealand exhibits a mean absolute deviation of actual
annual inflation rates from its official inflation target of
0.94 percentage points, which is lower than the average
result for other IT countries and higher only to the absolute
1.6
1998-2004
The data for New Zealand considers data from 1990q1 to
2005q4.
* The sample includes Australia, Canada, Chile and
Sweden for the 1990-1997 period and adds Norway for
1998-2004.
** The sample includes Israel, Peru and the United
Kingdom for the 1990-1997 period. The 1998¬2004
period includes 20 inflation targeters.
inflation deviations observed in Canada (0.93), England
(0.89), Switzerland (0.4), and Thailand (0.63). When
Regarding asymmetry in inflation deviations, New Zealand
comparing New Zealand’s IT accuracy across the earlier and
exhibits a large share of episodes when inflation rates
the more recent sample periods, we note that its accuracy
exceed target mid-points (78.6% of time; Table 6, overleaf).
has improved between 1990-1997 and 1998-2005; the
Regarding large deviations, defined as those exceeding 1
average absolute inflation deviation has declined from 0.96
percentage point in absolute value, New Zealand has
percentage points to 0.79 percentage points (see Figure 15).
experienced large deviations 52.4% of the time, most of
New Zealand has also outperformed the average country
which were positive inflation deviations (48.8%). The
in comparator group CCG2, as well as the average country
experience of New Zealand is similar to that of many IT
in the complementary group of other (non-CCG2) ITers in
countries, which on average experience large deviation
both periods.
from targets 52.7% of the time. But, as opposed to
New Zealand, large deviations are symmetric in an average
IT country; both positive and negative large deviations are
observed 26% of the time.
New Zealand has improved its accuracy in hitting its inflation
12
A few inflation-targeting countries – and some countries
during part of their IT experience – have used inflation
measures other than headline CPI inflation, most frequently
core inflation measures. Among the latter is New Zealand,
which used a core inflation measure for its inflation target
until the late 1990s, when it switched to the headline inflation
measure. While I am aware of the latter exceptions, I compute
inflation targeting accuracy measures in this section (and
inflation deviation measures in other sections of this paper)
using headline CPI inflation for all countries, to maintain
comparability across countries and over time.
Testing stabilisation policy limits in a small open economy
target over time.13 The frequency of large deviations has
declined from 43.8% in 1990-1997 to 31.3% in 19982005. This has left New Zealand in a better position than
the average IT country (Figure 16, p 116).
13
As discussed in the preceding footnote, part of this
improvement in New Zealand may reflect a somewhat distorted
measure of the absolute inflation target deviation before 1999,
when the official target was based on a core inflation measure,
not headline CPI inflation.
113
Table 6
Frequency of deviations of annual inflation rates from inflation targets in 21 inflation targeting
countries, since start of inflation targeting
Country
Australia
Frequency of time
Frequency of time
Fraction of time
above the target below the target
1 p.p. above the target p.p. below the target
(% of time)
(% of time)
1p.p above or
below
the target
52.4%
47.6%
21.4%
21.4%
42.9%
Brazil
87.5%
12.5%
Canada
42.9%
57.1%
70.8%
12.5%
83.3%
12.5%
26.8%
39.3%
Chile
62.5%
37.5%
32.1%
14.3%
46.4%
Colombia
58.3%
Czech Republic
39.3%
41.7%
12.5%
20.8%
333%
60.7%
28.6%
35.7%
64.3%
Hungary
93.8%
Iceland
62.5%
6.3%
75.0%
0.0%
75.0%
37.5%
43.8%
0.0%
43.8%
Israel
50.0%
50.0%
36.5%
40.4%
76.9%
Korea
Mexico
46.4%
53.6%
10.7%
32.1 %
42.9%
75.0%
25.0%
37.5%
12.5%
50.0%
New Zealand
78.6%
21.4%
48.8%
3.6%
52.4%
Norway
25.0%
75.0%
18.8%
50.0%
68.8%
Peru
47.7%
52.3%
20.5%
29.5%
50.0%
Phillipines
25.0%
75.0%
12.5%
56.3%
68.8%
Poland
46.4%
53.6%
35.7%
42.9%
78.6%
South Africa
65.0%
35.0%
50.0%
30.0%
80.0%
Sweden
25.0%
75.0%
0.0%
47.5%
47.5%
Switzerland
50.0%
50.0%
0.0%
0.0%
0.0%
Thailand
40.0%
60.0%
10.0%
15.0%
25.0%
United Kingdom
173%
82.7%
3.8%
32.7%
36.5%
Average
51.9%
48.1%
27.7%
25.0%
52.6%
excl. New Zealand
50.6%
49.4%
26.6%
26.0%
52.7%
Source: Mishkin and Schmidt-Hebbel (2005)
Note:
Periods depend on the IT regime starting date. Data on New Zealand calculated for the 1990q 1-2005q4 period.
114
Reserve Bank of New Zealand and The Treasury
Table 7
Features of deviations of annual inflation rates from inflation targets in 21 inflation targeting
countries, since start of inflation targeting
Mean duration of deviations
Country
above the target
below the target
(in quarters)
Amplitude of deviations
Amplitude of deviations
above the target below the target above the target
(percentage points)
below
the target
(as % of target)
Australia
5.50
5.00
1.25
-1.14
50.1.
-45.5
Brazil
21.00
3.00
4.03
-4.31
100.8
-53.9
Canada
3.43
5.33
0.79
-1.03
33.1
-43.2
Chile
5.00
3.00
1.51
.0.91
17.8
-25.1
Colombia
2.80
2.50
0.72
-1.99
12.7
-14.6
Czech Republic
3.67
8.50
2.44
-1.76
52.6
-50.9
Hungary
7.50
1.00
2.40
-0.06
60.2
-1.2
Iceland
5.00
6.00
2.33
-0.43
77.2
-17.0
Israel
5.20
5.20
2.41
-2.45
49.8
-66.3
Korea
6.50
7.50
0.63
-1.35
21.0
-34.7
Mexico
4.50
2.00
1.35
-0.99
29.3
-11.8
New Zealand
13.20
4.50
1.06
-0.62
67.7
-23.3
Norway
2.00
6.00
1.09
-1.31
43.6
-52.3
Peru
3.50
4.60
2.06
-1.34
23.1
-35.3
Phillipines
2.00
12.00
1.39
-1.78
30.9
-33.4
Poland
4.33
7.50
2.17
-2.36
39.6
-49.1
South Africa
6.50
2.33
2.95
-2.59
65.6
-57.5
Sweden
3.33
10.00
0.54
-1.31
27.0
-65.7
Switzerland
3.33
5.00
0.32
-0.47
32.4
-47.3
Thailand
1.60
2.40
0.69
-0.58
39.6
-33.2
United Kingdom
3.00
14.33
0.88
.0.89
35.3
-36.2
Average
5.38
5.60
1.57
-1.41
43.31
-37.97
excl. New Zealand
4.98
5.66
1.60
-1.45
42.09
-38.71
Source: Mishkin and Schmidt-Hebbel (2005)
Note:
Periods depend on the IT regime starting date. Data on New Zealand calculated for the 1990q 1-2005q4 period.
Testing stabilisation policy limits in a small open economy
115
Figure 16
approach adopted here involves estimating an inflation
Frequency of deviations of annual inflation
and output variability efficiency frontier in order to derive
rates from inflation targets in 21 inflation
measures of economic performance and monetary policy
targeting countries, since start of inflation
efficiency. The performance of monetary policy can be
targeting
assessed using the inflation and output variability tradeoff
%
%
60
60
an efficiency frontier that is known as the Taylor Curve
New Zealand
CCG2 *
Others **
50
faced by the policy maker. This tradeoff allows constructing
50
(Taylor 1979). The inflation-output variability frontier is
understood by considering an economy that is hit by two
40
40
30
30
20
20
move output and inflation in opposite directions, forcing
10
10
the monetary authority to face a tradeoff between inflation
0
0
types of disturbances: aggregate demand and aggregate
supply shocks. As is well known, aggregate supply shocks
1990-1997
Note:
1998-2004
The data for New Zealand considers data from 1990q1 to
2005q4.
* The sample includes Australia, Canada, Chile and
Sweden for the 1990-1997 period and adds Norway for
1998-2004.
** The sample includes Israel, Peru and the United
Kingdom for the 1990-1997 period. The 1998¬2004
period includes 20 inflation targeters.
and output variability. Therefore, the position of the
efficiency frontier depends on the intensity of aggregate
supply shocks: the smaller are such shocks, the closer is the
frontier to the origin (Figure 17).
Figure 17
Monetary policy efficiency frontier and
observed performance
Additional features of inflation deviations from targets
are reported in Table 7, including duration of deviations
above or below targets and the amplitude of deviations.
In New Zealand above-target inflation deviations exhibit
an average duration of 13.2 quarters while below-target
deviations last on average 4.5 quarters. Positive deviations
are also larger than negative deviations: above-target
deviations were on average 1.06 percentage points,
while below-target deviations were on average only 0.62 percentage points. The mean duration of abovetarget deviations in New Zealand is almost three times
larger than in the average IT country but the mean size of
deviations is smaller both above and below inflation target
in New Zealand, in comparison to the average ITer.
The efficiency frontier is also an indicator of the degree
of optimality of monetary policy. When monetary policy
is sub-optimal, the economy exhibits large output and
inflation volatility and is positioned at a significant distance
from the frontier. Shifts toward the efficiency frontier
are an indication of improved monetary policy efficiency.
These features of the efficiency frontier allow constructing
Inflation and output volatility and monetary
measures of economic and monetary policy performance
policy efficiency
in order to examine the contribution of policy efficiency
macroeconomic
and variability of shocks to the observed differences in
performance is by focusing on the stability of inflation
macroeconomic performance between different samples
and output. Following Cecchetti and Krause (2001)
of New Zealand over time and across different country
and Cecchetti, Flores-Lagunes, and Krause (2004), the
groups.
An
116
effective
way
of
measuring
Reserve Bank of New Zealand and The Treasury
Table 8 reports three estimated comparative measures of
between 1990-1997 and 1998-2005. In the most recent
economic performance for each pair of comparisons. L
period, New Zealand also exhibits lower levels of volatility
is a measure of an economy’s performance, in terms of
than different groups of IT countries. One possible
output and inflation variability. A high value of L reflects a
explanation is that New Zealand is currently hit by smaller
poorer performance of the country or set of countries; E
shocks than in the past. Alternatively, the Reserve Bank of
measures the monetary policy efficiency, determining how
New Zealand may be more efficient in implementing policies
close actual performance is to the one under optimal policy
to meet its inflation and stabilisation objectives.
(i.e., the distance to the efficiency frontier). Hence the
smaller is the value of E, the closer monetary performance
is to optimal policy; and S gauges the variability of supply
shocks. The smaller is the variability of the disturbances
that hit the economy, the smaller is this measure.14 Table
8 also decomposes the gains in performance into a gain
in efficiency, ∆ E , reflected by getting closer to the
efficiency frontier, and a smaller variability of shocks hitting
the economy, ∆ S , reflected by a shift of the efficiency
frontier. Figures 18-22 depict actual performance points
and efficiency frontiers consistent with E , for each pair of
comparisons performed.
In the spirit of Mishkin and Schmidt-Hebbel (2005), in this
section I compute performance measures in order to identify
the contribution of different monetary policy strategies to
the observed differences in macroeconomic performance
between New Zealand’s experience before and after 1997. I
disentangle the contribution of changes in monetary policy
efficiency and supply shocks to the observed differences
in macroeconomic performance between New Zealand
post-1997 and four different country groups; ITers after IT
adoption, industrial ITers after IT adoption, non inflation
targeters (NIT) post-1997, and the CCG2 sub-sample of
ITers.15
Figure 18 shows that New Zealand has featured a significant
reduction in the volatility of inflation and the output gap
Table 8
Monetary policy performance and policy efficiency changes in New Zealand and CCG1 countries
Group 1
NZ before 1997
L1
E1
S1
Group 2
L2
E2
S2
L2-L 1
E2-E1
S2-S1
1.383
0.985
0.398
NZ after 1997
0.684
0.304
0.380
-0.699
-0.681
-0.018
71.2
28.8
44.4
55.6
97.4
2.6
4.727
2.204
2.523
NZ after 1997
0.684
0.304
0.380
-4.043
-1.900
-2.142
46.6
53.4
44.4
55.6
47.0
53.0
1.697
0.829
0.867
NZ after 1997
0.684
0.304
0.380
-1.013
-0.525
-0.487
48.9
51.1
44.4
55.6
51.9
48.1
0.938
1.261
0.304
0.380
-0.635
-0.881
38.6
61.4
44.4
55.6
41.9
58.1
0.268
0.303
0.304
0.380
0.035
0.077
47.0
53.0
44.4
55.6
31.3
68.7
(as % of L)
ITers after IT
(as % of L)
Industrial ITers
(as % of L)
CCG2 IT countries
2.199
(as % of L)
NITers after 1997
14
0.571
NZ after 1997
NZ after 1997
Mishkin and Schmidt-Hebbel (2005) provide a detailed
description of the methodology used in this section.
Testing stabilisation policy limits in a small open economy
0.684
0.684
15
-1.515
0.113
This group of countries includes Australia, Canada, Chile,
Norway and New Zealand.
117
Figure 18
Figure 19
Estimated efficiency frontiers and observed
Estimated efficiency frontiers and observed
performances: New Zealand before and after
performance points: New Zealand after 1997
1997
and ITers
Output variability
4.0
Output variability
8.0
3.5
NZ pre-1997
observed point
3.0
6.0
2.5
5.0
2.0
1.5
1.0
ITers after IT efficiency frontier
7.0
NZ post-1997
efficiency
frontier
NZ pre-1997
efficiency frontier
NZ post-1997
efficiency frontier
4.0
ITers after IT observed
point
3.0
NZ post-97
observed point
NZ post-1997
observed point
2.0
0.5
1.0
0.0
0.0
0.2
0.4
0.6
Inflation variability
0.8
1.0
0.0
0.0
1.0
2.0
3.0
Inflation variability
4.0
5.0
6.0
Figure 20
Figure 18 depicts New Zealand’s monetary policy efficiency
Estimated efficiency frontiers and observed
frontier before and after 1997 and the first two rows of
performance points: New Zealand after 1997
Table 8 report the estimated measures of performance.
and industrial ITers
Macroeconomic performance between these periods
Output variability
5.0
has improved, as the volatility of inflation and output has
4.5
declined significantly. This is reflected in an improvement in
3.5
the performance measure L and a negative value of L2 - L1.
2.5
Before 1997, the distance to the efficiency frontier explained
1.5
71.2% of actual macroeconomic performance (E2 - E1 =
0.681) while the variability of shocks explained 28.8% of
performance (S2 - S1 = -0.018). In contrast, during the post-
4.0
NZ post-1997 efficiency
frontier
Industrial ITers efficiency
frontier
3.0
2.0
1.0
NZ post-1997
observed point
Industrial ITers observed
point
0.5
0.0
0.0
0.5
1.0
1.5
2.0
Inflation variability
Note: Industrial ITers sample include Australia, Canada, Iceland, Norway,
Switzerland and England
1997 period New Zealand has been much closer to the
efficiency frontier, reflecting an improvement of monetary
The second and third comparisons are between New Zealand
policy efficiency. The gain in efficiency explains 97.4%
post-1997 and two sets of IT countries: first, all countries
of the improved performance while the decline in shock
(less New Zealand) after their implementation of IT and,
volatility explains only 2.6%.
second, only industrial ITers after their implementation of IT.
New Zealand exhibits actual performance levels, efficiency
frontier positions, and policy efficiency levels that are better
than those of all ITers (Figure 19). The superior performance
of New Zealand is equally explained by enhanced policy
efficiency (47%) and smaller shocks (53%). A similar result
is obtained when comparing New Zealand to industrial
ITers (Figure 20). Even though their performance points
and efficiency frontiers are closer to New Zealand’s,
New Zealand also outperforms industrial ITers in efficiency
(by -0.525, equivalent to a 51.9% contribution) as well as
the magnitude of shocks (by -0.487, equivalent to a 48.1%
contribution).
118
Reserve Bank of New Zealand and The Treasury
5
Figure 21
Estimated efficiency frontiers and observed
scope for an independent
performance points: New Zealand after 1997
monetary policy
and NITers
Is there scope for an independent monetary policy in small
Output variability
4.5
open economies that are closely integrated into world
4.0
financial markets? I address this question for New Zealand in
3.5
3.0
international comparison from three different perspectives.
2.5
NITers post-1997
observed point
2.0
1.5
1.0
Role of global markets and
NITers post-1997
efficiency frontier
First, I report simple correlation coefficients between the
NZ post-1997
observed point
domestic short-term interest rate and the US Federal Funds.
NZ pre-1997
efficiency frontier
0.5
0.0
0.0
0.1
0.2
0.3
Note: NIters sample includes OECD NIters.
0.4
0.5
Inflation
variability
Then I compare impulse response dynamics of domestic
0.6
0.7
short-term rates to the foreign rate for New Zealand and
comparator country groups, from the VAR models developed
Figure 22
in Section 4. Finally, I complement the latter results by
Estimated efficiency frontiers and observed
computing the dynamic impulse response of domestic to
performance points: New Zealand after 1997
foreign rates from the country VAR models developed for
New Zealand and the five CCG2 countries in Section 4.
and CCG2 ITers
Output variability
5.0
CCG2 ITers after IT
efficiency frontier
4.5
4.0
3.5
NZ post-1997
efficiency frontier
How correlated are interest rates in New Zealand with
3.0
2.5
2.0
CCG2 ITers after IT
observed point
1.5
1.0
0.5
0.0
Independence of monetary policy
NZ post-1997
observed point
0.0
0.5
1.0
1.5
2.0
2.5
Inflation variability
Note: CCS2 countries include Australia, Canada, Chile, Norway and Sweden
international interest rates? In this section I start by reporting
simple correlation coefficients to assess the extent to which
the short-term interest rate is linked to the international
rate, namely the US Federal Funds rate. I provide evidence
of change in this coefficient over recent years and contrast
the evidence for New Zealand with the experience of
Figures 21 and 22 depict the results of comparing
New Zealand post-1997 to the control group of 13
successful industrial NITers and the restricted set of
CCG2 ITers. The results show that New Zealand exhibits a
performance that is inferior to that of the NITers (L2 - L1
= 0.11). However, most of this difference in performance
is explained by smaller shocks in industrial NITers (S1 - S2
= 0.08, equivalent to a contribution of 68.7%), while the
difference in monetary policy efficiency is smaller (E2 - E1
= 0.03, equivalent to a contribution of 31.3%). In contrast,
New Zealand post-1997 presents a better performance than
the CCG2 country group, but this difference is mainly due to
comparator country group CCG2.
Figure 23
Moving correlation coefficients between the
New Zealand short-term interest rate and the
US Federal Funds rate, 1997-2005
1
0. 8
0. 6
0. 4
0. 2
0
-0. 2
-0. 4
smaller shocks (51.9 %).
F E D rate
Note:
Testing stabilisation policy limits in a small open economy
F E D rate (adj. ex c hange rate ex p.)
Correlation coefficient calculations are based on sevenyear moving windows.
119
Table 9 reports interest rate correlation coefficients for
expectations until 2002, when they start rising quickly to
New Zealand and the five countries in the CCG2 group.
levels around 0.40.
New Zealand presents a high level of connection between
the domestic short-term interest rate and the US Fed Funds
rate, with a large and significant correlation coefficient
close to 0.7 for the full 1990-2005 sample. However, when
splitting the full sample in two, a massive reduction in New
Figure 24
Correlation coefficient between domestic
short-term interest rates and the US Federal
Funds rate in New Zealand and CCG2,
Zealand’s interest rate correlation is observed over time: the
1990-2005
corresponding coefficient drops from 0.93 in 1990-1997 to
1.0
0.30 (only significant at 10% confidence level) in the most
recent 8 years spanned between 1998 and 2005.
Figure 23 depicts the evolution of the correlation coefficient
in New Zealand using a seven-year moving window. Again
there is strong evidence of a massive reduction in domesticforeign interest rate correlation that took place in the late
1990s, to current levels close to 0.40.
Considering exchange-rate depreciation expectations I
examine evidence of uncovered interest-rate arbitrage
by focusing on the correlation between the short-term
domestic interest rate and the foreign rate augmented
by exchange-rate devaluation expectations. During the
first sub-period, the high correlation between domestic
and unadjusted foreign rates vanishes once we consider
devaluation expectations (Table 9). In contrast, for the
more recent 1998-2005 period, I still obtain a positive and
significant correlation coefficient of 0.41. Similar results
are obtained for moving-window correlation coefficients:
they are close to zero when adjusting for devaluation
%
1990-2005
1998-2005
0.9
1990-1997
%
1.0
0.9
0.8
0.8
0.7
0.7
0.6
0.6
0.5
0.5
0.4
0.4
0.3
0.3
0.2
0.2
0.1
0.1
0.0
0.0
New
Zealand
Australia
Canada
Chile
Norway
Sweden
The positive association between domestic and foreign
interest rates is confirmed for the CCG2 country group
(Table 9 and Figure 24). The average correlation coefficient
for the five countries and the full 1990-2005 sample is 0.62
and declines to 0.38 for the uncovered arbitrage relation.
However, in contrast to New Zealand, most countries exhibit
an increasing association between their domestic interest
rate and the external interest rate. However there is large
country heterogeneity in the relation between domestic
and foreign rates (unadjusted and adjusted for devaluation
expectations) and their changes over time.
Transmission of international interest rate
shocks (1)
As a result of enhanced credibility and lower pass-through
from exchange-rate shocks to inflation, we expect more
independence of central banks in their conduct of monetary
policy. In this subsection I assess the response of domestic
short-run interest rates to shocks in international short-run
interest rates, adopting the method applied in Section 4
above.
Figure 25, overleaf, depicts the dynamic response of the
domestic interest rate to an international interest rate shock,
comparing the response for New Zealand over two different
120
Reserve Bank of New Zealand and The Treasury
Table 9
Correlation coefficients between domestic short-term interest rates and the US Federal Funds
rate in New Zealand and CCG2, 1990-2005
New Zealand
FED interest rate
1990-2005
0.70
0.16
(0.00)***
(0.22)
0.93
0.00
(0.00)***
(0.99)
0.30
0.41
(0.099)*
(0.02)**
0.66
0.22
(0.00)***
(0.09)*
0.84
0.03
(0.00)***
(0.89)
1990-1997
1998-2005
Australia
1990-2005
1990-1997
1998-2005
Canada
1990-2005
1990-1997
1998-2005
Chile
1990-2005
1990-1997
1998-2005
Norway
1990-2005
1990-1997
1998-2005
No CHART supplied
Sweden
1990-2005
1990-1997
1998-2005
Note:
FED interest rate (adj. exchange-rate
Time period
depreciation expectations )
0.46
0.58
(0.01)***
(0.00)***
0.76
0.42
(0.00)***
(0.09)*
0.70
-0.27
(0.00)***
(0.14)*
0.93
0.84
(0.00)***
(0.00)***
0.79
0.67
(0.00)***
(0.09)*
0.59
0.33
(0.00)***
(0.07)*
0.83
0.80
(0.00)***
(0.00)***
0.41
0.26
(0.00)***
(0.05)**
0.16
0.19
(0.39)
(0.29)
0.43
0.16
(0.01)**
(0.42)
0.47
0.31
(0.00)***
(0.02)**
0.24
0.30
(0.19)
(0.10)*
0.36
-0.03
(0.04)**
(0.88)
The source of data for exchange-rate depreciation expectations for New Zealand is the RBNZ, while for all other countries we
calculated expectations from estimations based on AR(1) processes.
p-values are reported in parenthesis.
* Significant at 10%, ** Significant at 5%, *** Significant at 1%.
Testing stabilisation policy limits in a small open economy
121
time periods, and the more recent period in New Zealand to
changes of international interest rates. As in the preceding
the group response of all ITers, industrial-country ITers, and
sub-section, the foreign interest rate is the US Federal Funds
NITers – as I did in Section 4 for other impulse responses.
rate.
The response of short-term rates in New Zealand to an
Recall that the VAR model is comprised by nine variables: US
international short-term interest rate shock is positive and
Federal Funds rate ( it ), US output gap ( yt* ) , US inflation
significant in the first quarters after the shock, both in the
*
( π t ), domestic output gap ( yt ) , domestic deviation of
*
earlier 1989-1997 and the more recent 1998-2005 periods.
inflation from the inflation target (π t − π tT ) , domestic
However, the magnitude and persistence of New Zealand’s
short-term interest rate (it ) , money deviation from trend
response is smaller in the more recent period, and the
(mt ) , (log) exchange-rate (et ) , and long-term interest rate
difference is significant from the third quarter onwards.
( Rt ) . I assume that the federal funds rate responds only to
This suggests that monetary policy independence has
changes in the US output gap and inflation, and we also
strengthened in New Zealand during the last decade.
assume that neither contemporaneous nor lagged values of
In the two country groups comprised by all ITers and
the long-term interest rate enter the other equations in the
industrial-country ITers, the response of the short-run
system.16
domestic interest rate to an international interest-rate shock
Figure 26, overleaf, reports the impulse responses to a
is also positive but its magnitude and significance rise over
US monetary policy shock, with 95% confidence intervals
time. This time pattern is strikingly different than the more
over six years. The US monetary shocks are defined by a
front-loaded response of short-term rates in New Zealand.
Federal Funds rate hike of by 25 basis points on impact,
While at short lags the interest rate response in New Zealand
followed by the sample-specific dynamics of US monetary
is larger, at longer lags the response in other IT countries
policy observed in the data. The first two columns contain
increases while it declines toward zero at longer lags in
the results for New Zealand using two different samples:
New Zealand. The last column of Figure 25 confirms the
the 1990-2005 period and the post-1997 experience.17 The
significance of these differences between New Zealand and
impulse responses in the five other countries during 1990-
other ITers.
2005 are presented in the subsequent columns.18
By contrast, the interest-rate response to international rate
In New Zealand, a higher foreign interest rate is followed
shocks is smaller and less significant in NITers than in ITers.
by an increase in both short-term and long-term domestic
Therefore the differences with New Zealand’s time pattern
interest rates. The rise in both rates is significant for about
are even more significant.
a year, and from then on the short-term interest rate rapidly
I conclude that New Zealand’s monetary independence
declines while the long-term interest rate decreases at a
has strengthened since the late 1990s and is similar to that
slower pace. Looking at the post-1997 results, we still find
observed in other inflation-targeting countries.
that an international interest rate shock leads to higher
domestic interest rates. However, the response is short-lived
and significant only during the first two quarters. From an
Transmission of international interest rate
international perspective, the response of the short-term
shocks (2)
and long-term interest rates in New Zealand is similar to
I complement the latter evidence by applying the same
the responses found in Australia and Canada. Norway, on
VAR model developed and used in Section 4 to assess the
evidence on the transmission mechanism of a shock in the
16
external interest rate in New Zealand and the CCG2 group.
17
Here I am particularly interested in the dynamic response
of the domestic interest rate and the exchange-rate to
122
18
Annex B describes he country data definitions and sources of
variables used in VAR estimations.
Due to the small number of observations for this period, we
should be careful in interpreting these results.
For robustness I also considered here alternative VAR
specifications, which, like in section 4, did not yield
significantly different results.
Reserve Bank of New Zealand and The Treasury
Figure 25
Dynamic response of domestic interest rate to an international interest rate shock in
New Zealand, 1990-2005 and 1998-2005, and in CCG2, 1990-2005 (25 quarters)
New Zealand vs ITERS and NON-ITers
New Zealand before 1997
New Zealand (1998-2005)
%
%
1.5
1.5
1.0
1.0
0.5
1.5
Difference
%
%
%
%
1.5
1.5
1.5
1.0
1.0
1.0
1.0
0.5
0.5
0.5
0.5
0.5
0.0
0.0
0.0
0.0
0.0
0.0
-0.5
-0.5
-0.5
-0.5
-0.5
-0.5
-1.0
-1.0
-1.0
-1.0
-1.0
1
2
3
4
5
6
1
7
New Zealand (1998-2005)
2
3
4
5
6
7
-1.0
1
2
3
ITers after start of IT
%
%
4
5
6
7
Difference
%
%
%
%
1.5
1.5
1.5
1.0
1.0
1.0
1.0
0.5
0.5
0.5
0.5
0.5
0.0
0.0
0.0
0.0
0.0
0.0
-0.5
-0.5
-0.5
-0.5
-0.5
-0.5
-1.0
-1.0
-1.0
-1.0
1.5
1.5
1.0
1.0
0.5
1.5
-1.0
1
2
3
4
5
6
1
7
2
3
4
5
6
-1.0
7
1
2
Industrial ITers after start of IT
New Zealand (1998-2005)
%
%
%
3
4
5
6
7
Difference
%
%
%
1.5
1.5
1.5
1.5
1.5
1.0
1.0
1.0
1.0
1.0
1.0
0.5
0.5
0.5
0.5
0.5
0.5
0.0
0.0
0.0
0.0
0.0
0.0
-0.5
-0.5
-0.5
-0.5
-0.5
-0.5
-1.0
-1.0
-1.0
-1.0
1.5
-1.0
1
2
3
4
5
6
1
7
New Zealand before (1998-2005)
%
2
3
4
5
6
-1.0
7
1
2
Non-ITers (1998-2005)
%
3
4
5
6
7
Difference
%
%
%
%
1.5
1.5
1.0
1.0
1.0
1.0
0.5
0.5
0.5
0.5
0.5
0.0
0.0
0.0
0.0
0.0
0.0
-0.5
-0.5
-0.5
-0.5
-0.5
-0.5
-1.0
-1.0
-1.0
-1.0
1.5
1.5
1.0
1.0
0.5
1.5
-1.0
1
2
3
4
5
6
7
1
2
3
4
5
6
7
1.5
-1.0
1
2
3
4
5
6
7
the other hand, presents a delayed response while Chile’s
depreciation. Looking at the post-1997 sample, we find
increase in the short-term interest rate is significant for at
that the currency depreciates only during the first year;
least two years but the response of long-term interest rates
thereafter the exchange-rate starts appreciating. A similar
in not significant.
pattern is also found in Australia, Canada and Norway, but
The fourth row in Figure 26 shows the impulse response
of the nominal exchange-rate to a US monetary policy
contraction. The results show that the bilateral USNew Zealand nominal exchange-rate depreciates until
only Australia presents similar magnitude of changes in the
exchange-rate, while the rest of the countries experience
smaller swings in the exchange-rate. Our results suggest
that there is evidence of delayed overshooting similar to the
approximately the second year and then starts a path of
Testing stabilisation policy limits in a small open economy
123
Figure 26
Dynamic response to an international interest rate shock in New Zealand and CCG2, 6 years
New Zealand 1990-2005
New Zealand 1998-2005
%
Foreign
interest rate
%
0.7
0.7
0.6
0.6
0.5
0.5
0.4
0.4
0.3
0.3
0.2
0.2
0.1
0.1
0.0
0.0
-0.1
-0.1
-0.2
-0.2
-0.3
-0.3
0
1
2
3
4
interest rate
0.4
0.3
0.3
0.2
0.2
0.1
0.1
0.0
0.0
-0.1
-0.1
-0.2
-0.2
interest rate
3
4
5
0.0
-0.1
-0.1
-0.2
-0.2
-0.3
1
2
3
4
5
6
%
%
0.3
0.2
0.2
0.1
0.1
0.0
0.0
-0.1
-0.1
-0.2
-0.2
0
%
1
2
3
4
5
6
%
%
0.20
0.25
0.25
0.15
0.15
0.20
0.20
0.10
0.10
0.05
0.05
0.00
0.00
-0.05
-0.05
0.15
0.15
0.10
0.10
0.05
0.05
0.00
0.00
-0.05
-0.10
1
2
3
4
5
%
0.20
-0.10
-0.10
-0.15
-0.15
-0.05
-0.20
-0.20
-0.10
-0.25
-0.25
6
0
%
1
2
3
4
5
6
%
%
0.02
0.03
0.03
0.01
0.01
0.02
0.02
0.02
0.00
0.00
0.01
0.01
-0.01
-0.01
0.00
0.00
-0.02
-0.02
-0.01
-0.01
-0.03
-0.03
-0.02
-0.02
-0.04
-0.04
-0.03
-0.03
-0.05
-0.04
-0.05
0
1
2
3
4
5
6
%
0.30
0.25
0.25
0.20
0.20
0.15
0.15
0.10
0.10
0.05
0.05
0.00
0.00
-0.05
-0.05
-0.10
-0.10
-0.15
-0.15
1
2
3
4
5
1
2
3
4
5
6
%
0.35
0.30
0
-0.04
0
%
0.35
Inflation
deviation
from target
0.1
0.0
6
%
0
Nominal
exchange
rate (US$/
Local Curr.)
0.2
0.1
0.30
0.30
Long-term
2
0.2
0.3
-0.3
-0.3
1
0.3
0
0.4
0
0.4
0.3
-0.3
%
%
Short-term
0.4
6
5
%
%
%
0.8
0.8
0.6
0.6
0.4
0.4
0.2
0.2
0.0
0.0
-0.2
-0.2
-0.4
-0.4
-0.6
-0.6
0
6
1
2
3
4
5
6
Years
124
Reserve Bank of New Zealand and The Treasury
Figure 26
Dynamic response to an international interest rate shock in New Zealand and CCG2, 6 years
(cont.)
Australia
Canada
%
%
0.35
0.30
0.30
0.30
0.25
0.25
0.25
0.25
0.20
0.20
0.20
0.20
0.15
0.15
0.15
0.15
0.10
0.10
0.05
0.05
0.10
0.10
0.05
0.05
0.00
0.00
0.00
0.00
-0.05
-0.10
-0.10
-0.05
-0.05
-0.15
-0.15
-0.10
-0.10
-0.20
-0.20
-0.15
1
2
3
4
5
%
3
4
5
6
%
%
0.4
0.4
0.15
0.15
0.3
0.3
0.10
0.10
0.2
0.2
0.05
0.05
0.1
0.1
0.00
0.00
0.0
0.0
-0.05
-0.05
-0.1
-0.1
-0.10
-0.10
-0.2
-0.2
-0.15
-0.15
-0.3
1
2
3
4
5
-0.3
6
%
0
%
1
2
3
4
5
6
%
%
0.30
0.25
0.25
0.25
0.20
0.20
0.20
0.20
0.15
0.15
0.15
0.15
0.10
0.10
0.10
0.10
0.05
0.05
0.00
0.00
-0.05
-0.05
-0.10
0.05
0.05
0.00
0.00
-0.05
-0.05
-0.10
-0.10
-0.10
-0.15
-0.15
-0.15
0
1
2
3
4
5
%
0.02
0.02
0.01
0.01
0.00
0.00
-0.01
-0.01
-0.02
-0.02
-0.03
-0.03
-0.04
-0.04
-0.05
-0.05
-0.06
-0.06
0
1
2
3
4
5
0.25
-0.15
0
6
%
2
3
4
5
6
%
%
0.01
0.01
0.00
0.00
-0.01
-0.01
-0.02
-0.02
0
6
%
%
1
1
2
3
4
5
6
%
%
0.3
0.6
0.3
0.5
0.5
0.2
0.2
0.4
0.4
0.1
0.1
0.3
0.3
0.2
0.2
0.6
Inflation
deviation
from target
2
0.20
0.30
Nominal
exchange
rate (US$/
Local Curr.)
1
0.20
0
Long-term
interest
rate
-0.15
0
6
%
interest rate
0.35
-0.05
0
Short-term
%
0.35
0.30
0.35
Foreign
interest rate
%
0.1
0.1
0.0
0.0
-0.1
-0.1
0.0
0.0
-0.2
-0.2
-0.1
-0.1
-0.3
-0.3
-0.2
-0.2
-0.4
0
1
2
3
4
5
-0.4
0
6
1
2
3
4
5
6
Years
Testing stabilisation policy limits in a small open economy
125
Figure 26
Dynamic response to an international interest rate shock in New Zealand and CCG2, 6 years
(cont.)
Norway
Sweden
%
%
Foreign interest rate
0.35
Short-term interest rate
%
0.35
0.35
0.35
0.30
0.30
0.30
0.30
0.25
0.25
0.25
0.25
0.25
0.25
0.20
0.20
0.20
0.20
0.20
0.20
0.15
0.15
0.15
0.15
0.15
0.15
0.10
0.10
0.10
0.10
0.10
0.10
0.05
0.05
0.05
0.05
0.05
0.05
0.00
0.00
0.00
0.00
0.00
0.00
-0.05
-0.05
-0.05
-0.05
-0.05
-0.05
-0.10
-0.10
-0.10
-0.10
-0.10
-0.10
-0.15
-0.15
-0.15
1
2
3
4
5
6
0
1
2
3
4
5
0.5
0
%
1
2
3
4
5
6
%
%
0.4
0.4
0.3
0.3
0.3
0.3
0.2
0.2
0.2
0.2
0.1
0.1
0.1
0.1
0.0
0.0
0.0
0.0
0.4
0.4
0.4
-0.15
6
%
%
%
0.4
0.3
0.3
0.2
0.2
0.1
0.1
-0.1
-0.1
-0.1
-0.1
0.0
0.0
-0.2
-0.2
-0.2
-0.2
-0.1
-0.1
-0.3
-0.3
-0.3
0
1
2
3
4
5
%
Long-term interest rate
%
0.35
0.30
0.5
0
6
%
1
2
3
4
5
-0.3
6
%
0
%
1
2
3
4
5
6
%
%
0.4
0.25
0.3
0.3
0.20
0.20
0.04
0.2
0.2
0.15
0.15
0.02
0.02
0.1
0.1
0.10
0.10
0.08
0.08
0.4
0.06
0.06
0.04
0.25
0.00
0.00
0.0
0.0
0.05
0.05
-0.02
-0.02
-0.1
-0.1
0.00
0.00
-0.04
-0.04
-0.2
-0.2
-0.05
-0.05
-0.06
-0.3
-0.3
-0.10
-0.06
0
1
2
3
4
5
0
6
%
Nominal exchange
rate (US$/Local Curr.)
%
%
0.35
0.30
0
1
2
3
4
5
%
%
-0.10
6
0
%
1
2
3
4
5
6
%
%
0.02
0.02
0.02
0.02
0.02
0.01
0.01
0.01
0.01
0.01
0.01
0.00
0.00
0.00
0.00
0.00
0.00
-0.01
-0.01
-0.01
-0.01
-0.01
-0.01
-0.02
-0.02
-0.02
-0.02
-0.02
0
1
2
3
4
5
6
0
%
%
Inflation deviation
from target
Chile
1
2
3
4
5
%
%
%
%
1.0
0.3
0.2
0.2
0.8
0.8
0.1
0.6
0.6
0.0
0.4
0.4
0.2
0.2
0.0
0.0
-0.4
-0.2
-0.2
-0.5
-0.4
0.0
0.0
0.1
-0.1
-0.1
0.0
-0.2
-0.2
-0.1
-0.1
-0.2
-0.2
-0.3
-0.3
-0.3
-0.3
-0.4
-0.4
-0.4
-0.5
-0.5
-0.5
6
0.5
0
1
2
3
4
5
6
1.2
-0.4
0
1
2
3
4
5
Years
126
6
0.3
0.1
5
5
1.0
0.1
4
4
0.4
0.2
3
3
0.4
0.2
2
2
1.2
0.3
1
1
0.5
0.3
0
-0.02
0
6
0.02
Reserve Bank of New Zealand and The Treasury
6
results found in Scholl and Uhlig (2005) and Eichenbaum
this view. More recently, Demers (1991) and Franke (1991)
and Evans (1995).
show that the uncertainty about the state of the demand
due to price uncertainty caused by exchange-rate risk
depresses output trade volumes, even in the case of risk-
6
Real exchange-rate volatility
and misalignment
What is the world evidence on the costs of real exchange-rate
volatility and misalignment? Which are the fundamentals that
drive the medium-term behaviour of the RER in the world?
Is there evidence of past and present RER misalignment in
New Zealand – and would it have implications for economic
growth?
neutral firms.
The basic models base their findings on the assumption
of absence of hedging instruments that would allow
ameliorating exposure to exchange-rate risk. Viaene and
de Vries (1992) formally include a mature forward market,
concluding that increased levels of exchange-rate volatility
may act to the detriment or benefit of trade flows depending
on the net currency position of that country. Canzoneri et
al. (1984), De Grauwe (1992), and Gros (1987) relax the
assumption that firms cannot alter factor inputs to adjust
World evidence on the costs of real exchange-
optimally in response to exchange-rate shifts, concluding
rate volatility and misalignment
that increased exchange-rate variability could create profit
Exchange-rate volatility, trade, and welfare
opportunities and rise average investment and output,
The post–Bretton Woods floating exchange-rate period has
as firms adjust to take advantage of high prices and to
been characterized by volatile and largely unpredictable
minimize the impact of low prices.
exchange-rate movements. Moreover, the liberalization
Most theoretical modelling of exchange-rate volatility
of capital flows accompanied by the huge increase in
and trade has taken a partial equilibrium perspective by
cross-border financial transactions has contributed to
focusing on firm decisions. Kumar (1992) develops a two-
exacerbated fluctuations in exchange-rates over the last two
country general equilibrium model to explore the effects
decades. Such unpredictability can be costly, both directly
of exchange-rate volatility on trade. The author argues that
and through the potential for associated exchange-rate
higher exchange-rate volatility lowers technological change
misalignments. On the other hand, the growth of financial
and international trade. This comes as the result of resource
hedging instruments and the rising share of international
reallocation to non-export oriented sectors that are not
transactions undertaken by multinational firms may suggest
exposed to exchange-rate risk.
that the impact and costs of exchange-rate volatility are
now lower than, say, some decades ago. The theoretical as
well as the empirical literature has addressed this ambiguity
by exploring if the major changes in the world economy
over the past decades have operated to reduce or increase
the extent to which international trade is adversely affected
by fluctuations in exchange-rates.
Several authors have also argued that the exchange-rate
volatility not only hinders international trade but also takes a
toll in terms of economic welfare. One of the leading articles
attempting to explore the welfare costs of exchange-rate
variability in general equilibrium is Obstfeld and Rogoff
(1998). The latter authors conclude that exchange-rate
volatility could lower welfare through two channels.
The hypothesis that exchange-rate volatility reduces trade
First, exchange-rate volatility increases fluctuations in
flows found support from the very beginning of the
consumption and leisure, thereby lowering welfare. Second,
theoretical literature. The latter argument focuses on the
risk-averse firms hedge against future exchange- rate shifts
notion that unexpected changes in exchange rates affect
by adding a risk premium when setting their prices to cover
the decisions of risk-averse commodity traders, lowering
from the costs of exchange-rate fluctuations. Higher prices
output and trade volumes (Artus 1983, Brodsky 1984).
lower demand, production and consumption to sub-optimal
Either (1973) and Clark (1973) were the first to formalize
less that are less than the optimal for society.
Testing stabilisation policy limits in a small open economy
127
More recent research, however, argues that exchange-rate
Regarding the costs of exchange-rate volatility on welfare,
volatility may be even beneficial for welfare. This is the case
Tchakarov (2003) finds that welfare effects of exchange-
when prices are not fixed in the currency of the exporter
rate volatility are likely to be very small for many countries.
but of their foreign customers (Devereux and Engel, 2003)
For instance, for the US economy the loss in consumer
and when consumption and leisure are complements, not
utility due to exchange-rate volatility is equivalent to 0.1%
substitutes, in utility (Bacchetta and Van Wincoop, 2000).
of annual consumption.
While theoretical economists are extending the analytical
Therefore this brief review of the literature suggests that
frontier on the effects of exchange-rate volatility on trade,
there is no clear-cut relation between exchange-rate
growth, and welfare, empirical research examines the data
volatility, on one hand, and trade flows or welfare, on the
to quantify actual costs. The early empirical work did not
other. The presumption that trade or welfare are adversely
deliver consistent results. Many studies reported little or
affected by exchange-rate volatility depends on particular
no support for a negative effect. Hooper and Kohlhagen
assumptions and hence does not hold in the general case.
(1978) examined the impact of exchange-rate volatility on
The empirical literature is not conclusive either, reporting
aggregate and bilateral trade flow data for all G-7 countries,
mixed results. However, the latter findings should not be
finding little evidence of any negative effect of exchange-
taken to imply that exchange-rate fluctuations should be
rate volatility. Similar results were found by Cushman (1983),
viewed as beneficial or harmless. As noted by Clark et al.
IMF (1984), and Gotur (1985).
(2004), currency crises (a special case of exchange-rate
Several recent studies report effects of exchange-rate
volatility that range from moderate to negligible.19
Dell’Ariccia (1999) examines the effect of exchange-rate
volatility on the bilateral trade of the 15 EU members and
Switzerland over the 20 years from 1975 to 1994, using
four different measures of exchange-rate uncertainty.
volatility) have required painful adjustments in output and
consumption. In this case, however, what is important is
to take appropriate policy regimes to avoid the underlying
causes of large, unpredictable and damaging movements in
exchange-rates. A floating regime goes a long way toward
this goal.
The paper concludes that eliminating volatility to zero
would have raised trade by 10 to 13%, depending on the
Exchange-rate misalignment
particular measure of variability. Similarly, Rose (2000) uses
Larger volatility in developing countries’ real exchange-
a very large data set involving 186 countries, reporting that
rates has been typically associated to larger exchange-
lowering exchange-rate volatility by one standard deviation
rate misalignments. The IMF (2004) reports that if real
(7 per cent) would raise bilateral trade by 13%.
exchange-rate volatility rises by one standard deviation, the
In contrast, Tenreyro (2003) finds that, controlling for possible
average misalignment (defined as the average deviation of
endogeneity in exchange-rate volatility, the negative effect
the exchange-rate from its trend level) increases by about 5
of exchange-rate volatility on trade vanishes; a result that is
percentage points.
robust on the choice of instruments. Clark et. al. (2004) also
Aguirre and Calderón (2006) argue that misalignments
conclude that for the world as a whole, there is no obvious
are used as a way to predict future exchange-rate changes
association between periods of low exchange-rate volatility
among floaters and to evaluate the required exchange-rate
and periods of fast trade growth. The latter findings suggest
adjustment among countries with fixed or intermediate
that, from the perspective of world trade, exchange-rate
exchange-rate regimes. It has been argued that sustained
volatility should probably not be a major policy concern.
real exchange-rate (RER) overvaluations are an early
warning indicator of possible currency crashes (Krugman,
1979; Frankel and Rose, 1996; Kaminsky and Reinhart,
19
128
De Grauwe, (1987), Rose (2000), Dell’Ariccia (1999),
Anderton and Skudelny (2001), Arize (1998), and Fountas
and Aristotelous (1999).
1999). RER overvaluations also have led to drastic relativeprice adjustment and lower growth.
Reserve Bank of New Zealand and The Treasury
The literature concludes that RER misalignments may affect
both internal and external equilibrium. As discussed above,
growth and welfare (Edwards, 1989). A misaligned RER
deviations from the ERER (RER misalignments) have potential
may create distortions in the relative price of traded to non-
serious economic effects that could depend on their size and
traded goods, causing incorrect signals to economic agents
their direction. In order to assess the latter potential effects,
and hence leading to sub-optimal resource allocation and
I start by estimating a model for ERER for New Zealand,
more economic instability.
that allows computation of a time-series measure of RER
Others have argued that the growth effects of misalignments
could differ if the RER appreciates in excess of the equilibrium
appreciation (overvaluation) or if it depreciates in excess of
the equilibrium depreciation (undervaluation). It has been
misalignment. Based on international evidence on the link
between RER misalignment and growth, I infer potential
growth effects of RER misalignment in New Zealand and its
implication for the conduct of policy.
argued that undervaluation, which could be attributed to
I follow the approach of Aguirre and Calderón (2005) who
competitive devaluations, could encourage higher export
identify the role of four fundamentals in traded and non-
and output growth. On the other hand, overvaluations,
traded goods markets in shaping the RER. A permanent
which may reflect macroeconomic policy inconsistency, are
increase in external liabilities requires running a larger
likely to discourage growth (Razin and Collins, 1999).
trade surplus to service them, requiring a RER depreciation.
RER stability and avoidance of misaligned RERs have been
mentioned as determinants of economic performance in
developing countries (Krueger, 1983; Edwards, 1988).
Sachs (1985) claims that the different development
experiences in East Asia, Latin America, and Africa may be
attributed to their different trade regimes and exchangerate management practices. Unstable and overvalued RERs
provided weak incentives to exports and were supported
by protectionist policies, while persistent misaligned RERs
in Africa caused a severe drop in agricultural output (The
World Bank, 1984).
A permanent rise in the ratio of traded to non-traded
goods productivity (strictly speaking, relative to the rest
of the world) causes excess demand in the non-traded
sector and therefore an appreciation of the RER in order
to restore internal equilibrium (Harrod-Balassa-Samuelson
effect). A permanent increase in the terms of trade boosts
the demand for both traded and non-traded goods but
the excess demand in the non-traded sector causes a RER
appreciation. Finally, considering that government spends a
larger share on non-traded services than the private sector,
a permanent rise in government consumption causes an
excess demand in the non-traded sector, contributing to a
Much more recently, Calderón and Aguirre (2006) evaluate
RER appreciation.20
the growth effects of RER misalignments and volatility.
They find that RER misalignments hinder growth but the
effect is non-linear: growth reductions are relatively larger,
the larger are misalignments. A large undervaluation of
the currency hurts growth but small to moderate levels of
undervaluation enhance growth. They also find evidence of
a negative relationship between economic growth and the
volatility of RER misalignments. The authors suggest that
The first step in estimating the ERER is identifying the longrun effects of fundamentals on the RER. I do this by estimating
a cointegration vector for the RER and its fundamentals,
using annual 1965-2005 data for New Zealand, following
the specification in Aguirre and Calderón (2005):
 yT 
 F
+
β
ln
2
 y N  + β3 ln
 Y  t


(2) qt = β0 + β1 ln 
 PX 
 G
 M  + β 4 ln  Y  + ξt
 t
P

t
this negative relationship is also possibly non-linear.
Equilibrium real exchange-rate and
misalignment in New Zealand
According to the literature, the equilibrium real exchange-rate
(ERER) is the unobserved RER level consistent with achieving
Testing stabilisation policy limits in a small open economy
20
The first and fourth are demand effects, the second is a supply
effect, and the third embodies both demand and supply
effects. Under extreme assumptions (uncovered interest
parity, exogenous labour supply), the RER is determined only
by supply factors.
129
Table 10
Real effective exchange-rate estimations for New Zealand and the world
World sample estimations, 1965-2003 (Aguirre and Calderon, 2005)
Time series
estimation for
New Zealand,
1965-2005
Variable
Panel data
Time series
(median estimator
for 60 countries)
All countries
Industrial
countries
Developing
countries
Constant
2.32**
(0.04)
n.a.
n.a.
n.a.
n.a.
Net foreign
Assets
0.11 *
(0.09),
0.15**
(0.01)
0.10**
(0.01)
0.14**
(0.00)
0.09**
(0.00)
Government
Spending
0.41 *
(0.08)
0.27**
(0.00)
0.28**
(0.00)
0.57**
(0.00)
0.22**
(0.00)
Terms of trade
0.49**
(0.01)
0.23**
(0.02)
0.23**
(0.00)
0.36**
(0.00)
0.20**
(0.00)
Productivity
0.21
(0.38)
0.39**
(0.00)
0.15**
(0.02)
0.30**
(0.00)
0.10**
(0.00)
R2
0.30
Note:
p-values are reported in parenthesis. ** (*) denotes coefficient significant at 95% (90%) confidence level.
where q is the effective RER, F/Y is the ratio of net foreign
Table 11
assets to GDP, yT/yN is labour productivity in the traded
Error correction model for the real effective
sector relative to that in the non-traded sector, PX/PM
exchange-rate (WTI) in New Zealand,
is the terms of trade index, and G/Y is the government
expenditure ratio to GDP.21
1965-2005
Coefficient
Std. Error
In testing for cointegration, I found that the trace test
indicates that there is one (and only one) cointegration
vector at a 95% confidence level. This implies that the
Constant
0.01
0.01
Equation [1] residual, lagged -0.74**
0.14
Dif RER, lagged
0.49**
0.14
and it is possible to infer the long-run effects of RER
Dif government spending
0.27
0.24
fundamentals. In the estimation of the RER equation, the
Dif net foreign assets
-0.05
0.12
problem of reverse causality is addressed by implementing
Dif terms of trade
0.35**
0.15
dynamic ordinary least squares (DOLS).22 Results are
Dif productivity
-0.94**
0.35
Dif government spending
-0.26
0.23
Dif net foreign assets
0.13
0.11
Dif terms of trade
0.08
0.15
Dif productivity
0.67**
0.33
coefficients estimated in equation (1) are super consistent
reported in Table 10. For comparison purposes I also report
the cross-country regression results by Aguirre and Calderón
(2005). Although the R2 is relatively low, all coefficients for
New Zealand exhibit the expected signs. The impact of net
foreign assets on the RER is similar in New Zealand to the
21
22
130
Note: ** (*) denotes coefficient significant at 95% (90%)
confidence level.
For the RER I use RBNZ’s real WTI index, starting in 1970.
For 1965-1970, I use the RER measure constructed by Aguirre
and Calderón (2005). The source for all other variables is
Aguirre and Calderón (2005) until 2003. For the 2004-2005
period, I use RBNZ and National Statistics data. In the case
of the productivity variable, I obtained better results using
a variable that reflects labour productivity for the aggregate
economy
See Siakkonen 1991, Phillips and Loretan 1991, and Stock
and Watson 1993 for details.
Reserve Bank of New Zealand and The Treasury
Table 12
Table 13
Partial-adjustment model for the real
Partial-adjustment model for the real
effective exchange-rate (WTI, RBNZ) in
effective exchange-rate (Reer, Aguirre and
New Zealand, 1990q1 – 2005q4
Calderón 2006) in New Zealand, 1965 – 2005
Ø1
Ø2
AR(1)
AR(2)
AR(3)
AR(4)
0.982
1.321
1.251
1.187
(0.036)*
(0.122)*
(0.131)*
(0.124)*
-0.357
-0.099
-0.159
(0.124)*
(0.212)
(0.199)
-0.202
0.279
(0.133)
(0.199)
Ø3
Ø4
-0.382
Ø1
AR(2)
AR(3)
0.578
0.843
0.830
(0.141)*
(0.142)*
(0.165)*
-0.473
-0.380
(0.144)*
(0.191)
Ø2
Ø3
-0.199
(0.159)
Note:
Standard errors are reported in parentheses.
* Significant at 95% confidence level.
** Significant at 90% confidence level.
(0.127)*
Note:
AR(1)
Standard errors are reported in parentheses.
* Significant at 95% confidence level.
Table 14
Estimates of first-year closing and half-life of deviations of real exchange-rate from equilibrium
or sample average levels in New Zealand, various models
Deviation closed after
one year
Half-life of deviation in
years
1. RER error-correction model
74%
0.68
2. AR(1) model of RER misalignment
42%
1.30
3. Partial adjustment of RER deviation from sample mean
(RBNZ quarterly data)
-23%
4.82
4. Partial adjustment of RER deviation from sample mean
(Calderón and Aguirre 2005 annual data)
76%
0.66
Model
Note:
The estimates are calculated from the error correction model reported in Table 11, an AR(1) model of the deviation of the
effective RER from its time-varying equilibrium in equation (3) , an AR(2) model using the deviation of RBNZ’s TWI RER from
its sample mean for 1990q1-2005q4, reported in Table 12, and an AR(2) model using Calderón and Aguirre’s RER deviation
from its sample mean for 1965-2005 period, reported in Table 13
world sample, but the effects of both government spending
(Table 11). The coefficient size implies that 74% of the RER
and the terms of trade are larger in New Zealand than in
deviation from the ERER is closed after one year.
the world sample, but close to those found for industrial
countries.23
Alternatively, I estimate a partial adjustment model for the
RER deviation from its sample mean, using the following
I also estimate an error correction model to investigate
model:
the short-run dynamics of the RER. The results show that
(3)
the lagged residual of the long-run RER equation is highly
p
(
)
qt − q = ∑ φi qt −i − q + ξt
i =1
significant, which represents additional evidence for the
where q is the effective RER and q is the sample mean. In
existence of a cointegration vector among the variables
Table 12 I report the results for equation (3), using quarterly
data for the TWI real exchange-rate for 1990q1-2005q4,
while Table 13 reports comparable results for equation (3),
23
The effects of productivity are not comparable because
Aguirre and Calderón (2005) use the labour productivity
difference between traded and non-traded sectors..
Testing stabilisation policy limits in a small open economy
using annual data constructed by Calderón and Aguirre
(2006). Using annual data, the estimates imply that 76%
131
of the RER deviation from its sample mean is closed after
The estimated ERER series is depicted for New Zealand’s
one year. In contrast, quarterly data show an initial increase
1965-2005 sample period in Figure 27. The ERER has been
in the deviation of around 23%. Therefore, the estimates
stable during the full sample period, exhibiting a slight
using annual data imply a half-life close to 0.7 years, while
equilibrium depreciation phase in the 1960s and 1970s and
the half-life of the quarterly model is around 4.6 years
a slight equilibrium appreciation period starting in the late
(Table 13). These figures can be compared to the estimation
1990s and throughout 2005. Note that the magnitude size
of a half-life close to 1.3 years using a simple autoregressive
of exchange-rate misalignments has increased after 1985.
model for the misalignment using the time-varying ERER.
After an undervaluation period in the aftermath of the
The evidence on RER adjustment in New Zealand according
to the different models, sample periods, and frequencies
is summarized in Table 14. Using annual data for a long
time period indicate a fast adjustment of the RER toward
its equilibrium level in New Zealand that stands in contrast
Asian Crisis, the RER appreciated steadily between 2002
and 2005, attaining an estimated over-valuation of 14% in
2005. The latter is the largest level of RER over-valuation in
the last four decades, similar to the RER over-appreciations
observed in 1988 and 1996-1997.
to the international evidence, where half of exchange-rate
Figure 28
deviations are closed only after 3 to 5 years (e.g., Rogoff
Real effective exchange-rate misalignment
1996, Calderón and Schmidt-Hebbel 2003, and Cashin
and potential growth effects in New Zealand,
and McDermott 2003). However, the estimates using more
recent quarterly data suggest that the dynamics of RER
misalignment in New Zealand are not so different from the
international evidence.
1965-2005
40%
30%
20%
10%
0%
Now I determine ERER levels based on the coefficient
estimates of equation (2) and long-run levels of the RER
determinants, reflected by the following expression:
-20%
-30%
-40%
0
1965
*
*
-10%
 PX 
 yT 
F
G
q e t = βˆ0 + βˆ1 ln   + βˆ2 ln  N  + βˆ3 ln  M  + βˆ4 ln  
 Y t
 Y t
 y t
 P t
*
(4)
*
1969
1973
1977
1981
1985
1989
1993
1997
2001
2005
Neutral effects on growth
S igni ficant positi ve effects on growth
S igni ficant negati ve effects on growth
where qe is the ERER, the βˆi are the coefficient estimates
of equation (2), and starred variables denote long-run levels
of the corresponding variables. In estimating the latter longrun levels I use the Hodrick-Prescott filter.
Based on Aguirre and Calderón’s (2005) cross-country
analysis, it is possible to make some inference about the
effects of RER misalignment on growth. The latter authors
Figure 27
estimate cross-country growth regressions, including
Real effective exchange-rate (TWI),
RER misalignment as an explanatory variable. Their non-
equilibrium real exchange-rate, and real
linear relation between misalignment and growth allows
exchange-rate misalignment in New Zealand,
1965-2005
to identify varying effects according to the direction and
size of RER misalignment.24 In Figure 28 the estimated
misalignment for New Zealand is contrasted with the
4.8
90%
4.7
intervals of misalignment levels that have significant effects
70%
4.6
50%
4.5
30%
4.4
10%
4.3
-10%
4.2
-30%
1965
132
1969
1973
1977
1981
1985
1989
1993
1997
2001
2005
RER (L HA, L og)
ER ER ( LHA, L og)
RER misalignment ( RHA)
Confidence interv al (RHA, 1 st. dev .)
24
It has to be clear that the effects of misalignments on growth
used in this report are only partial and do not take into
account general equilibrium effects of the variables behind
the misalignment. For example, if an increase in the terms of
trade is the main cause of the misalignment, the effects of the
latter on the GDP growth rate have to be complemented by the
direct effect of the terms of trade on growth.
Reserve Bank of New Zealand and The Treasury
on growth rates. The inference from this world evidence,
The exchange-rate in the conduct of
which may not apply to growth in New Zealand, is that
monetary policy
New Zealand’s estimated overvaluation observed during
Monetary policy rules or reaction functions describe
the most recent years has been within the neutral range,
the response of policy instruments to deviations in key
where growth effects are negative but not statistically
macroeconomic variables, typically deviations of inflation
significant. Moreover, the RER correction that has taken
and output from target and full-employment levels,
place since early 2006 has reduced the likelihood of getting
respectively. The debate about how exchange-rates should
into the range of excessive appreciation that may reduce
be taken into account in simple monetary policy rules is
growth. Finally it is important to note that New Zealand has
relatively new; recent theoretical and empirical research
had many episodes of slight undervaluation in the 1-11%
has started to focus on several important exchange-rate
range (that is, growth-enhancing according to the world
questions. How should monetary policy authority react to
evidence), but these periods were short-lived.
25
the exchange-rate? Should policy makers avoid any direct
reaction and react to the indirect effects of exchange-rate
shocks on inflation and output?
7
How best to achieve domestic
Obstfeld and Rogoff (1995) argue that deviations of the
price stability, while avoiding
real exchange-rate from its long-run equilibrium value calls
cyclical extremes in the
for a monetary policy response. If the real exchange-rate
exchange-rate
With a clear focus on price stability, central banks recurrently
face the issue of how to avoid persistent exchange-rate
misalignments that may be costly, as discussed in the
preceding section. Therefore I start this section by briefly
reviewing the literature and international evidence on the
role of the exchange-rate in the conduct of monetary policy.
If monetary policy is neither adequate nor sufficient to deal
with exchange-rate stability concerns, (sterilized) exchange-
is excessively appreciated, then the central bank should
lower the short-term interest rate, relaxing the monetary
policy stance. Ball (1999) suggests a similar response but
considers more complicated dynamics in the optimal
monetary policy response. Using a model for an openeconomy with sticky prices, the author calls for an initial cut
in interest rates to mitigate the contraction caused by the
appreciation. However, as the appreciation drives inflation
down, monetary policy should not be eased further but the
initial reduction should be partly offset.26
rate interventions may offer an alternative tool to central
banks. Hence I briefly review the international practice
and evidence about interventions and their effectiveness.
In the light of the latter reviews and the findings about
New Zealand’s monetary and exchange-rate policy reported
in preceding sections, I draw policy lessons for New Zealand
at the end of this section, identifying possible fiscal and
financial-policy instrument to deal better with exchangerate misalignments and current-account imbalances.
Taylor (1999) and Svensson (2000) also explore the
performance of these types of monetary policy rules.
Taylor (1999) found that the exchange-rate reaction led
to a better performance for France and Italy but had a
poorer performance in Germany. Svensson (2000) uses a
model with forward-looking agents and more explicit micro
foundations to explore the benefits of having a rule that
reacts to the exchange-rate. His simulations show that
including the exchange-rate as a separate argument in the
26
25
According to the estimations by Aguirre and Calderón (2005),
the misalignment has to be maintained for almost 5 years in
the corresponding intervals, and the volatility of misalignment
has negative effects on growth.
Testing stabilisation policy limits in a small open economy
In Ball’s model, a 10% appreciation of the real exchangerate would call for an initial interest cut of 3.7 percentage
points, followed by a partially offsetting rise of 1.7 percentage
points in the next period. This monetary policy rule implies
a long-run reaction of a 2 percentage point cut in the interest
rate. In this model, such a rule leads to a better performance
than a rule that is insensitive with respect to the exchangerate, reflected in reduced levels of the standard deviation of
inflation.
133
policy rule lowers inflation volatility but increases output
in the exchange-rate have effects on output and inflation.
volatility. Thus a policy rule that reacts to the exchange-rate
This may explain why allowing central banks to react to the
can actually lead to a deterioration of output performance.
exchange-rate may not improve the performance of the
More recently, West (2004) examined the possibility of using
interest rate policy to trade exchange-rate stability against
stability in other variables for the case of New Zealand. In his
model the central bank adjusts interest rates in response to
temporary exchange-rate shocks, by cutting (raising) interest
rates in response to transitory depreciations (appreciations)
of the New Zealand dollar. Using a model consistent with
the recent New-Keynesian literature on monetary policy
in small open economies, he finds that the central bank
could achieve a 25% reduction in the standard deviation
of the real exchange-rate at the price of increasing output
volatility by 10 to 15%, inflation volatility by 0 to 15%, and
economy, to say the least. This argument may also explain
why many authors have found that a closed-economy
policy rule describes closely the actions of the central bank
in small open economies (e.g. Huang, et al. 2000, Lubik and
Schorfheide 2005). Taylor lists two reasons why reacting to
the exchange-rate may not lead to better macroeconomic
performance. First, there may be exchange-rate deviations
from purchasing-power parity that should not be offset
by changes in interest rates. For example, exchange-rate
adjustments may reflect productivity changes that should
not be offset. Second, exchange-rate shocks may have small
costs relative to the costs of smoothing them out.
interest rate volatility by 15 to 40%. However, the author
Taylor’s considerations and the results of the literature
adds two caveats. First, his model assumes that interest-rate
reviewed above strongly suggest that conventional central
adjustments affect exchange rates in a reliable and clearly
bankers’ preference for indirect rather than a direct
understood way. Second, the central bank is assumed to
response of monetary policy to exchange-rate shocks is
know the equilibrium level of the real exchange-rate.
hard to dismiss.
Do central banks actually react to exchange-rate shocks?
Lubik and Schorfheide (2005) estimate a small-scale
structural general equilibrium model for a small open
economy, where the monetary authority reacts to changes
in output, inflation, and the exchange-rate. They focus
on the conduct of monetary policy in Australia, Canada,
New Zealand, and the UK, and test for the hypothesis
whether central banks respond to exchange-rates. Their
estimations suggest that the central banks of Australia and
New Zealand do not respond to the exchange-rate, whereas
the central banks of Canada and the UK do so. Chadha,
Sarno, and Valente (2004) examine empirically whether
exchange-rates are included interest rate rules, using data
for the US, the UK, and Japan. Their findings suggest that
Foreign exchange interventions
Empirical studies and statements by central banks suggest
that central banks intervene in foreign exchange markets
to correct perceived excessive exchange-rate volatility
or misalignments. Their concern is that high short-term
volatility and longer-term swings in exchange-rates that
deviate from equilibrium levels determined by fundamental
conditions may hurt their economies, particularly sectors
heavily involved in international trade. Excessive inflation
pass-through and adverse balance sheet effects provide
additional motivations for central bank interventions,
particularly in developing countries.
the exchange-rate enters only in Japan’s policy rule. Huang,
Non-sterilized interventions may affect the exchange-rate
Margaritis, and Mayes (2000) find that a closed-economy
through changes in liquidity. On the other hand, sterilized
type rule describes quite well the monetary policy of the
interventions may affect the exchange-rate through three
Reserve Bank of New Zealand, finding no role for the
main mechanisms: portfolio, signalling, and information
exchange-rate.
channels.27 The portfolio balance channel assumes that
Taylor (2001) argues that rules with no explicit role of the
exchange-rate in the policy rule are consistent with indirect
27
See Dominguez and Frenkel (1993) or Edison (1993) for a
thorough literature review and Tapia and Tokman (2003,
2004) for an analysis of the information channel.
reaction of interest rates to the exchange-rate, since changes
134
Reserve Bank of New Zealand and The Treasury
investors hold foreign and domestic bonds that are imperfect
A strand of the empirical literature has focused on the
substitutes. Sterilized interventions alter the relative supply
effects of interventions on exchange-rate volatility. Chang
of local bonds, thereby changing the composition of investor
and Taylor (1998), Baillie and Osterberg (1997), Bonse-Neal
portfolios, which alters the exchange-rate accordingly.
and Tanner (1996), Huang (1997), and Dominguez (1993) do
The signalling channel refers to the signals sent by the
not find much or strong support for the notion that central
central bank to the markets regarding the future stance of
bank interventions lead to lower exchange-rate volatility.
monetary policy. For example, a sale of foreign currency
may anticipate a future monetary policy tightening. The
information channel assumes a significant role of forex
market speculators that cause significant exchange-rate
misalignments. In this case interventions aim at stabilizing
the markets, by affecting expectations about exchange-rate
fundamentals.
Since interventions are typically very small relative to the
stock of outstanding assets, many authors, including
Rogoff (1984), have expressed scepticism about the impact
of interventions through the portfolio balance channel.
Dominguez and Frenkel (1992) questioned the conventional
believe that interventions through the portfolio channel are
ineffective. Using data on US dollar-Swiss franc exchangerate expectations, the authors find strong support for the
portfolio balance effect and therefore argue in favour of
internationally coordinated interventions. However, other
studies do not find evidence of this channel and those that
do, including Ghosh (1992) and Evans and Lyons (2001),
suggests it is weak.
The evidence on the effectiveness of interventions through
either the portfolio balance channel or the signalling
channel is still mixed (Sarno and Taylor, 2001). However,
the evidence presented by authors using high-frequency
post-1990 data suggests that official interventions can be
effective, especially if interventions are publicly announced
and concerted. Furthermore, these interventions should be
consistent with the monetary and fiscal policy stance.
The Reserve Bank of Australia (RBA), as many other central
banks, intervenes in the foreign exchange market to influence
the Australian dollar exchange-rate. According to Edison,
Cashin, and Liang (2003), there are four reasons why the
RBA intervenes: to help reverse an apparent overshooting
of the exchange-rate in either direction; to calm markets
threatening to become disorderly; to signal future changes
of monetary policy or calm expectations if monetary policy
is changed unexpectedly; and to maintain an inventory
of net foreign currency assets. The RBA conducts all its
interventions in the spot market vis-a-vis the US dollar, and
sterilizes these operations. The interventions of the RBA are
The impact of intervention through the signalling channel
infrequent, coming mainly at or near the peaks and troughs
has often been found to be substantially stronger than
of the exchange- rate cycle. A recent study by Becker and
through the portfolio balance channel (Dominguez 1987,
Sinclair (2004) evaluates the effectiveness of exchange-rate
1990, and Dominguez and Frankel, 1993). Dominguez
interventions in Australia, concluding that the have had a
(1990) uses daily data for the Japanese yen-US dollar
stabilizing influence on the exchange-rate.
and the German mark-US dollar from January 1985 to
December 1987. He finds different effects for coordinated
and uncoordinated interventions and, in general, the
coefficient on coordinated interventions is statistically
significant and correctly signed. Kaminsky and Lewis (1996)
find that interventions supported by consistent movements
in the interest rate, move the exchange-rate in the expected
direction. However, when the intervention is followed by
inconsistent monetary policy, the exchange-rate tends to
move in the opposite direction.
The Central Bank of Chile (CBCh) has also responded to
exchange-rate deviations from perceived equilibrium levels
through sterilized interventions. After a flexible exchangerate regime was adopted in 1999, the CBCh has intervened
twice in response to massive exchange-rate depreciations.
The peculiarity of this experience is that interventions are
pre-announced at the time they start, when the CBCh
provides information about the length of time it will
intervene (4 months), the maximum amount of overall
interventions, and the financial instruments to be used. De
Gregorio and Tokman (2004) argue that the rationale for the
Testing stabilisation policy limits in a small open economy
135
choice of high transparency is twofold. First, the authorities
By all accounts, the RBNZ’s new framework has not been
commit to intervene in a transparent manner, rather than by
tested yet. (In fact, there is evidence to suggest that the RBNZ
surprising the markets, in order to maximize effectiveness
has not intervened during the last two decades). In light of
through the information channel. Second, interventions are
the empirical evidence on New Zealand dollar exchange-
intended to provide liquidity and stabilize the markets rather
rate volatility, misalignments, and possible implications
than fighting speculators. Tapia and Tokman (2004) provide
for economic growth, reported in the preceding sections,
empirical evidence that the intervention announcements
the RBNZ’s apparent lack of intervention to date seems
by the CBCh had a significant, moderate, and temporary
fully justified. Considering the latter and the international
impact on the exchange-rate, while subsequent actual
evidence
interventions had small, non-significant effects.
effectiveness, I conclude that:
on
policy
best-practice
and
interventions
(1) New Zealand’s RER trend, cycles, misalignments, volatility,
Implications for New Zealand
New Zealand’s monetary and exchange-rate framework
is sound and at the frontier of international best practice.
Monetary policy, consistent with flexible exchange-rate
targeting practice, is effective and efficient. Inflation targets
are met and monetary policy contributes actively to minimize
and correlations with other key variables are similar
to those of other industrial, small, open, commodityexporting, and inflation-targeting economies. From a
cross-country perspective, New Zealand’s RER behaviour
is certainly not anomalous.
(2) New Zealand’s
medium-term
RER
behaviour
is
inflation and output volatility. The RBNZ’s apparent conduct
consistent with the same fundamentals that drive RERs
of monetary policy, consistent with standard central-bank
elsewhere. Misalignments defined as deviations from
practice, tends to react to the exchange-rate indirectly;
the (unobserved but estimated) equilibrium RER driven
only as long it affects inflation and output forecasts. This
tend to occur in New Zealand, as elsewhere too. These
is consistent with the view that responding independently
deviations, from below and from above, recur regularly
to the exchange-rate would be at the cost of causing larger
in New Zealand. The most recent deviation episode
volatility of inflation, output, and monetary policy itself
(a significant RER appreciation in 2004-2005 that has
(West 2004).
partly corrected most recently) does not reach levels at
However, New Zealand’s nominal and real exchangerates exhibit significant and persistent deviations from
average levels, which may be costly for optimal resource
which, inferring from the world evidence (Aguirre and
Calderón 2006), medium-term growth rates would be
affected.
allocation, export success, and economic growth. This
(3) Even when facing a large perceived misalignment that
raises the question about the desirability of exchange-rate
may trigger an intervention (larger than the 2004-2005
interventions.
episode) substantial uncertainty surrounds the measure
The RBNZ put in place in 2005 the world’s presumably
most advanced and transparent framework for possible
of the equilibrium RER and hence the misalignment
estimate.
future interventions (Eckhold and Hunt 2005). Interventions
(4) The reviewed international evidence on intervention
would be triggered if four prerequisites were to be justified:
effectiveness ranges from nil to moderate, at best.
exceptionality (the exchange-rate is outside historical
From the latter points I conclude that interventions are best
norms), disequilibrium (exchange-rate level cannot be
used as an instrument of last resort to correct a situation
justified by market fundamentals), intervention is consistent
of very large and persistent misalignment. The RBNZ’s
with the monetary policy stance, and market opportunity
new intervention policy provides the internationally best
(likelihood of favourable market reaction).
designed framework for such an exceptional circumstance.
136
Reserve Bank of New Zealand and The Treasury
Hence if neither monetary policy (indirectly) nor foreign
exchange intervention policy (directly) are the best ways to
deal with persistent but not extreme RER misalignment, what
should be done to achieve more exchange-rate stability?
While a detailed answer to this question is outside the scope
of this paper, which deals with monetary and exchangerate policies under the control of the monetary authority, I
1. Structural government spending and balance rule
A Chilean-type structural fiscal balance rule would
involve determining a structural fiscal balance according
to structural or permanent long-term estimates of
selected components of spending and revenue,
based on their permanent or trend estimates of their
exogenous determinants, such as trend GDP.
close by listing a few fiscal and financial policy options that
could help in reducing the likelihood of incurring in large
and persistent exchange-rate misalignments and current
account imbalances.
2. Pro-cyclical
tax
rates
or
counter-cyclical
government spending items
As opposed to adopting a comprehensive structural
Counter-cyclical fiscal policy
fiscal balance rule, The Treasury could identify a certain
New Zealand’s fiscal policy, like that in most industrial
levy (for example, the sales tax rate) and vary its tax
countries, is only weakly counter-cyclical. Therefore it plays
rate pro-cyclically, or a particular expenditure item
only a limited role in stabilizing aggregate spending, output,
(for example, public works) and vary its activity level
the current account, and the real exchange-rate over the
counter-cyclically. Another alternative – geared at
business cycle.
avoiding excessive cyclicality in a particular economic
A few countries have put in place strong counter-cyclical fiscal
policies that go beyond automatic stabilizers. One example is
Norway: its Pension Fund acts both as an inter-generational
transfer instrument that smoothens government spending
over the very long term and as a cyclical stabilisation device,
in the limited sense that oil price windfalls are saved and
only their permanent component is spent. Another more
sector or financial activity, such as construction and
mortgage lending, could imply raising pro-cyclical fees
or levies on the corresponding sector transactions, such
as a pro-cyclical mortgage fee. In the case of a procyclical sector tax or transaction fee, it is important to
consider the distortions it imposes on sector resource
allocation.
strongly counter-cyclical fiscal policy has been implemented
by Chile since 2001, where government spending is strictly
Counter-cyclical financial policy
proportional to permanent tax and copper price revenue.
Instead of focusing on fiscal revenue or expenditure flows,
Hence all revenue windfalls caused by positive (negative)
the objective of counter-cyclical financial policy is to
differences between actual and permanent GDP (affecting
implement a government investment management policy
tax revenue) and of actual and permanent copper prices are
geared at reducing the domestic costs of idiosyncratic
saved (dissaved). Casual evidence from both Norway and
shocks (including excessive exchange-rate misalignments
Chile suggests that their counter-cyclical fiscal policies may
and current-account imbalances). This involves considering
help in stabilizing the real exchange-rate and the current
the two following policy alternatives.
account over the business cycle.
New Zealand could evaluate the potential benefits and
costs of the following policy alternatives for strengthening
1. Investment
guidelines
for
international
government funds
the counter-cyclical stance of its fiscal policy.
Investment guidelines for government funds invested
abroad (like the RBNZ’s international reserves and the
New Zealand Superannuation Fund) should weigh
heavily those criteria that minimize the domestic
Testing stabilisation policy limits in a small open economy
137
consequences of idiosyncratic shocks that hit the
correlations, are roughly consistent with those observed in
New Zealand economy, by investing in assets whose
a relevant comparator country group. New Zealand’s overall
returns exhibit low or negative correlation with
macroeconomic policy mix is also broadly consistent with
those domestic and international variables that drive
international best practice. Regarding the inflation-targeting
New Zealand’s cycles, such as its terms of trade.
framework, there is room for some potential improvement
by specifying more precisely the policy horizon for monetary
policy. On fiscal policy, New Zealand lacks a strongly
2. Acquisition of international insurance
Caballero (2002) argues that traditional approaches to
deal with external shocks (accumulating international
reserves or reducing external borrowing) are both
costly and inefficient. In the context of developing
countries, he proposes as an alternative buying assets
(or issuing liabilities) whose returns are correlated
(with the appropriate sign) to the dominant foreign
trade and financial shocks that affect developing
economies. However, this proposal is relevant for any
small open economy that faces external or domestic
idiosyncratic shocks. For insurance against terms-oftrade shocks Caballero (2002) has proposed issuing
public debt indexed to commodity prices for insurance
against “sudden stops”. Caballero and Panageas (2005)
propose including assets indexed to the S&P implied
volatility index (VIX) in emerging-market holdings of
international reserves.
counter-cyclical framework like those successfully adopted
by a few comparator countries.
There
are
many
similarities
between
New Zealand
and comparator countries regarding monetary policy
transmission and efficiency. One is that the inflation response
to an exchange-rate fall, relatively small but significant in
New Zealand, a result quite similar to that observed in other
OECD inflation targets. Pass-through to tradable goods
inflation is much larger and somewhat quicker than passthrough to headline inflation in New Zealand, which also
accords with international experience. The dynamic response
of headline inflation to an oil price hike is also relatively small
in New Zealand. This is likely to be a reflection of credible
monetary policy and stable inflation expectations, as in the
low exchange-rate pass-through. The dynamics of the oil
price impact on inflation in New Zealand is not statistically
different from those observed in other countries. However,
New Zealand’s tradables inflation is much more and more
The implication for New Zealand is that the RBNZ and
persistently affected by an oil shock than headline inflation,
the Treasury could engage in a coordinated financial
a result also in line with the international evidence.
evaluation of the costs and benefits of acquiring or issuing
international insurance instruments that provide explicit
protection against external trade and financial shocks, or
domestic idiosyncratic shocks, that have a large impact on
New Zealand’s main macroeconomic variables, including
output, the RER, and the current account.
Monetary policy transmission in New Zealand is broadly
comparable to, or at least as strong as, that observed in
the comparator country group. Long-term rates respond
to short-term rates in New Zealand, although this response
exhibits less persistence. As in most comparator countries,
the exchange-rate appreciation in response to a shortterm interest rate rise is not significantly different from
zero. Inflation declines significantly in New Zealand in
8
Summary and policy
the second year after a monetary policy shock, similar to
conclusions
what is observed in some comparator countries but not in
In this paper I have presented several findings on
New Zealand’s monetary and exchange-rate policy and
drawn some policy lessons.
others, where the inflation does not respond significantly
to monetary policy innovations. The output gap responds
negatively and persistently in New Zealand to a monetary
policy shock, like in most other comparator countries.
The trend and cyclical behaviour of New Zealand’s key
macroeconomic variables, as well as their volatility and cross138
Reserve Bank of New Zealand and The Treasury
A simple but partial way of gauging monetary policy
ranging from nil to moderate at best. Even when facing a
efficiency under inflation targeting is computing deviations
large perceived exchange-rate misalignment that may trigger
between actual inflation and target levels. New Zealand’s
an intervention, one larger than the 2004-2005 episode, a
accuracy in hitting its inflation target is much higher
high degree of uncertainty surrounds the measure of the
than that of the average IT country but its deviations are
equilibrium RER and hence the misalignment estimate.
skewed: deviations from above are substantially more likely
than those from below. A more comprehensive way of
assessing monetary policy efficiency is in its delivery of low
inflation and output volatility. Here the evidence shows that
New Zealand’s monetary policy efficiency has improved
From the latter points I conclude that interventions are best
used as an instrument of last resort to correct a situation of
very large and persistent misalignment. The RBNZ’s new forex
intervention policy provides the best designed framework
worldwide for such an exceptional circumstance.
massively after 1997, to levels that are better than that of
the representative IT country or even industrial IT country,
but ranks below that of the representative industrial nonIT country in a group comprised of the US, Japan, and
Hence if neither monetary policy (indirectly) nor foreign
exchange intervention policy (directly) are the best ways to
deal with persistent but not extreme RER misalignment, what
should be done to achieve more exchange-rate stability?
European nations.
The answer to the latter question could lie in developing and
The evidence on the scope for an independent monetary
policy (drawn from simple correlation coefficients between
domestic and foreign short-term interest rates and impulse
response dynamics of short-term rates to foreign rate
using counter-cyclical fiscal and financial policy instruments
that could help in reducing the likelihood of exceptionally
large and persistent exchange-rate misalignments and
current account imbalances.
shocks) shows significant degrees of policy independence
in setting short-term rates, both in New Zealand and
comparator countries.
There is a large world literature on the costs of real
exchange-rate volatility and misalignment on trade, growth,
and welfare, including recent work on non-linear effects of
RER misalignment on long-term growth. New Zealand’s
References
Aguirre, A and C Calderón (2005), “Real Exchange-rate
Misalignments and Economic Performance,” Central Bank
of Chile Working Paper 315, April.
RER trend, cycles, misalignments, volatility, and correlations
Artus, J (1983), “Toward a More Orderly Exchange-rate
with other key variables are similar to those of comparable
System,” Finance and Development, 20, pp. 10-13.
inflation-targeting
economies.
From
a
cross-country
perspective, New Zealand’s RER behaviour is certainly not
anomalous. In fact, it is driven by the same fundamentals
Bacchetta, P, and E van Wincoop (2000), “Does Exchangerate Stability Increase Trade and Welfare?,” American
Economic Review 90, pp. 1093-1109.
that drive RERs elsewhere. Misalignments defined as
deviations from the (unobserved but estimated) equilibrium
RER tend to occur in New Zealand, as elsewhere too. These
deviations, from below and from above, recur regularly
Baillie, R and W Osterberg (1997), “Why Do Central Banks
Intervene?,” Journal of International Money and Finance,
vol. 16, pp. 909 – 919.
in New Zealand. The most recent deviation episode, a
Ball, L (1999), “Policy Rules for Open Economies,” in J.
significant RER appreciation in 2004-2005, is not estimated
Taylor (ed.) Monetary Policy Rules, Chicago: University of
to have reached levels at which medium-term growth rates
Chicago Press.
would be negatively affected.
Becker, C and M Sinclair (2004), “Profitability of Reserve
There is also a large international literature on the
Bank Foreign Exchange Operations: Twenty Years After the
effectiveness of forex interventions that suggests effects
Float,” Reserve Bank of Australia, mimeo.
Testing stabilisation policy limits in a small open economy
139
Bergin, P, and I Tchakarov (2003), “Does Exchange-rate Risk
Cushman, D (1983), “The Effects of Real Exchange-rate Risk
Matter for Welfare? A Quantitative Investigation,” NBER
on International Trade,” Journal of International Economics
Working Paper No. 9900.
15, August, pp. 43-63.
Bernanke, B, M Gertler, and M Watson, (1997), “Systematic
De Grauwe, P (1986), “Long Run Exchange-rate Variability
Monetary Policy and the Effects of Oil Price Shocks,”
and International Trade, Chapter 8 in S. Arndt and J D
Brookings Papers on Economic Activity, 1: 91-142.
Richardson, ed. by Real Financial Linkages Among Open
Bonser-Neal, C and G Tanner (1996), “Central Bank
Economies, (London: The MIT Press)
Intervention and the Volatility of Foreign Exchange-rates:
De Gregorio, J and A Tokman (2004), “Flexible Exchange-
Evidence from the Options Market,” Journal of International
rate Regime and Forex Interventions: The Chilean Case,”
Money and Finance 15 (6) 853-878.
Central Bank of Chile Policy Paper no. 11.
Brodsky, D A (1984), “Fixed Versus Flexible Exchange-rates
Dell’Ariccia, G (1999), “Exchange-rate Fluctuations and
and the Measurement of Exchange-rate Instability,” Journal
Trade Flows: Evidence from the European Union,” IMF Staff
of International Economics 16, pp. 295-306.
Papers 46(3), pp. 315-334.
Canzoneri, M B, P Clark, T Glaesner, and M Leahy (1984),
Demers, M (1991), “Investment under Uncertainty,
“The Effects of Exchange-rate Variability on Output and
Irreversibility and the Arrival of Information Over Time,”
Employment,” International Finance Discussion Papers 240.
Review of Economic Studies, vol. 58, pp. 333-350.
Cecchetti, S and S Krause (2001), “Central Bank Structure,
Devereux, M and C Engel (2003), “Monetary Policy in the
Policy Efficiency and Macroeconomic Performance: Exploring
Open Economy Revisited: Price Setting and Exchange-rate
Empirical Relationships,” manuscript, Ohio University.
Flexibility,” Review of Economic Studies 70, pp. 765-783.
Cecchetti, S, A Flores-Lagunes, and S Krause (2004), “Has
Doominguez, K (1987), “Exchange-rate Efficiency and the
Monetary Policy Become More Efficient? A Cross-Country
Behaviour of International Asset Markets,” Yale University,
Analysis,” manuscript, Brandeis University.
unpublished PhD thesis.
Chadha, J, L Sarno, and G Valente (2004), “Monetary Policy
Dominguez, K (1990), “Market Responses to Coordinated
Rules, Asset Prices, and Exchange-rates,” IMF Staff Papers,
Central Bank Intervention,” Carnegie-Rochester Conference
vol. 51(3).
Series on Public Policy, vol. 32, pp. 121-163.
Chang Y and Taylor, S J (1998), “Intraday Effects of Foreign
Dominguez, K (1993), “Does Central Bank Intervention
Exchange Intervention by the Bank of Japan,” Journal of
Increase the Volatility of Foreign Exchange-rates,” NBER
International Money and Finance, vol. 17, pp. 191-210.
Working Paper No. 4532.
Christiano, L M Eichenbaum, and C Evans (1996), “The
Dominguez, K and J Frenkel (1993), “Does Foreign
Effects of Monetary Policy Shocks: Evidence from the Flow
Exchange Intervention Work?,” Washington DC, Institute of
of Funds,” The Review of Economics and Statistics, vol. 78,
International Economics.
pp. 16-34.
Edison, H (1993), “The Effectiveness of Central-Bank
Clark, P (1973), “Uncertainty, Exchange Risk, and the Level
Intervention: A Survey of the Literature after 1982,” Special
of International Trade,” Western Economic Journal 11,
Papers on International Economics, 18, Princeton Univ.
September, pp. 302-13.
Eckhold, K and C Hunt (2005), “The Reserve Bank’s new
Clark, P, N Tamirisa, S Wei, A Sadikov, and L Zeng (2004),
foreign exchange intervention policy,” Reserve Bank of
“Exchange-rate Volatility and Trade Flows - Some New
New Zealand Bulletin, 68(1) : pp. 12-22.
Evidence,” manuscript, IMF.
Edwards, S (1989), Real Exchange-rates, Devaluation, and
Adjustment, Cambridge, Mass.: The MIT Press.
140
Reserve Bank of New Zealand and The Treasury
Eichenbaum, M and C Evans (1995), “ Some Empirical
Harding, D and A Pagan (2006), “Synchronization of
Evidence of the Shocks to Monetary Policy on Exchange-
Cycles,” Journal of Econometrics, 132: 59-81.
rates,” Quarterly Journal of Economics, vol. 100, pp. 9751010.
Ethier, W (1973), “International Trade and the Forward
Exchange Market,” American Economic Review 63, pp.
494-503.
Evans, C and D Marshall (1998), “Monetary Policy and the
Term Structure of Interest Rates: Evidence and Theory,”
Carnegie-Rochester Conference Series on Public Policy no.
49, pp. 53-111.
Evans, M and R Lyons (2001), “Portfolio Balance, Price
Impact, and Secret Intervention,” NBER Working Paper no.
8356.
Franke, G (1991), “Exchange-rate Volatility and International
Trading Strategy,” Journal of International Money and
Finance 10, pp. 292-307.
Gosh, A (1992), “It It Signalling? Exchange Intervention and
the Dollar-Deutschemark Rate,” Journal of International
Economics, vol. 33 pp, 2001-220.
Gotur, O (1985), “Effects of Exchange-rate Volatility on
Trade: Some Further Evidence,” IMF Staff Papers.
Holtz-Eakin, D, W Newey, and H S Rosen (1988), “Estimating
Vector Autoregressions with Panel Data,” Econometrica,
Vol. 56, pp. 1371-1395.
Huang, A, D Margaritas and D Mayes (), Monetary Policy
Rules in New Zealand, 1988-1998,” University of Waikato,
mimeo.
Huang, J H (1997), “Intervention Strategies and Exchangerate Volatility: A Noise Trading Perspective,” Journal of
International Money and Finance 16 (5).
IMF (1984), “The Exchange-rate System: Lessons of the Past
and Options for the Future,” IMF Occasional Paper No. 30
(Washington: IMF).
Kaminsky, G, and K Lewis (1995), “Exchange-rates, Interest
Rates and Current Account News: Some Evidence from
Australia,” Journal of International Money and Finance, vol.
14, pp. 575-595.
Kaminsky, G and C M Reinhart (1999), “The Twin Crises:
The Causes of Banking and Balance of Payments Problems,”
American Economic Review 89(3), 473-500.
Kim, S (2001), “International Transmission of US Monetary
Grilli, V and N Roubini (1995), “Liquidity and Exchange-
Policy Shocks: Evidence from VARs,” Journal of Monetary
rates: Puzzling Evidence from the G-7 countries,” mimeo,
Economics, vol 48, pp. 561-586.
Yale University.
Grilli, V and N Roubini (1996), “Liquidity Models in Open
Economies: Theory and Empirical Evidence,” Journal of
Monetary Economics, vol. 48, pp. 339-372.
Gros, D (1987), “Exchange-rate Variability and Foreign Trade
in the Presence of Adjustment Costs,” Working Paper No.
8704, Département des sciences économiques, Université
Catholique de Louvain.
Frankel, J A, Rose, A K (1996), “Currency crashes in emerging
markets: An empirical treatment,” Journal of International
Economics 41(3-4), 351-366.
Harding, D and A Pagan (2002), “Dissecting the Cycle:
A Methodological Investigation,” Journal of Monetary
Krueger,
A
(1983),
Exchange-rate
Determination.
Cambridge: Cambridge University Press.
Krugman, P R (1979), “A Model of Balance-of-Payments
Crises,” Journal of Money, Credit, and Banking 11(3),
311-325.
Kumar, V and J A Whitt (1992), “Exchange-rate Variability
and International Trade,” Economic Review, Federal Reserve
Bank of Atlanta 37, May/June, pp. 17-32.
Love, I and L Zicchino (2002), “Financial Development and
Dynamic Investment Behaviour,” Policy Research Working
Paper no. WPS 2913, The World Bank.
Lubik, T and F Schorfheide (2005), “Do Central Banks
Respond to Exchange-rate Movements? A Structural
Economics, 49: 365-81.
Testing stabilisation policy limits in a small open economy
141
Investigation,”
Journal
of
Monetary
Economics,
Exchange-rates,” Humboldt University, Discussion Paper
forthcoming.
2005-037.
Mishkin, F and K Schmidt-Hebbel (2005), “Does Inflation
Sims,
Targeting Makes a Difference?,” Document presented at
Econometrica Vol. 48 pp. 1-48.
the 9th Annual Conference of the Central Bank of Chile.
C
(1980),
“Macroeconomics
and
Reality,”
Stock, J H, and M Watson (1993). “A Simple Estimator of
Miniane, J and J H Rogers (2003), “Capital Controls and
Cointegrating Vectors in Higher Order Integrated Systems,”
the International Transmission of US Money Shocks,”
Econometrica 61(4), 783-820.
International Finance Discussion Papers No. 778, Board of
Governors of the Federal Reserve System.
Obstfeld, M and K Rogoff (1995), “The Mirage of FixedExchange-rates,” Journal of Economic Perspectives, vol. 9,
pp. 73-96.
Obstfeld, M, and K Rogoff (1998), “Risk and Exchangerates,” NBER Working Paper #6694.
Phillips, P C B and M Loretan (1991), “Estimating Long-Run
Economic Equilibria,” Review of Economic Studies 58(3),
407-436.
Razin, O, and S Collins (1999), “Real Exchange-rate
Misalignments and Growth,” in Razin, A, Sadka, E, (eds.),
The Economics of Globalization: Policy Perspectives from
Public Economics, Cambridge, New York and Melbourne:
University Press, pp. 59-81.
Reserve Bank of New Zealand (2002), “Policy Targets
Agreement 2002,” Wellington, September 17.
Rogoff, K (1984), “On the Effects of Sterilized Intervention:
An Analysis of Weekly Data,” Journal of Monetary
Economics, vol 14, pp. 133-150.
Rose, A (2000), “One Money, One Market: The Effect of
Common Currencies on Trade,” Economic Policy, April,
9-45.
Sachs, J D (1985), “External Debt and Macroeconomic
Performance in Latin America and East Asia,” Brookings
Papers on Economic Activity 1, 523-64.
Saikkonen, P (1991) “Asymptotically Efficient Estimation of
Cointegration Regressions,” Econometric Theory 7(1), 1-21.
Scholl, A and H Uhlig (2005), “New Evidence on the Puzzles:
Svensson, L (2000), “Open-Economy Inflation Targeting,”
Journal of International Economics, vol. 50, pp. 155-183.
Tapia, M and A Tokman (2003), “Efectos de las Intervenciones
en el Mercado Cambiario: El Caso De Chile,” Central Bank
of Chile Working Paper no. 206.
Tapia, M and A Tokman (2004), “Effects Of Foreign Exchange
Intervention Under Public Information: The Chilean Case,”
Central Bank of Chile Working Paper no. 255.
Taylor, J B (1979), “Estimation and Control of a
Macroeconomic Model with Rational Expectations,”
Econometrica 47, pp. 1267-1286.
Taylor, J (1999), “The Robustness and Efficiency of Monetary
Policy Rules as Guidelines for Interest Rate Setting by the
European Central Bank,” Journal of Monetary Economics,
vol. 43, pp. 655-679.
Taylor, J (2001), “The Role of Exchange-rate in MonetaryPolicy Rules,” The American Economic Review, vol. 91, pp.
263-267.
Tenreyro, S (2003), “On the Trade Impact of Nominal
Exchange-rate Volatility,” Federal Reserve Bank of Boston,
mimeo.
Viaene, J M and C G de Vries (1992), “International Trade
and Exchange-rate Volatility, European Economic Review,”
36, pp. 1311-21.
West, K (2004), “Monetary Policy and the Volatility of Real
Exchange-rate in New Zealand,” NBER Working Paper no.
10280.
World Bank (1984), Toward a Sustained Development in
Sub-Saharan Africa, Washington, DC: The World Bank.
Results from Agnostic Identification on Monetary Policy and
142
Reserve Bank of New Zealand and The Treasury
Annex A
Variable definitions and sources
Variable
Definition
Source
Real effective exchange-rate index
Expressed in terms of US dollars
per unit of the national currencies.
An increase in the index reflects an
appreciation.
IMF International Financial Statistics,
and RBNZ
Current account balance ratio to GDP
Last four quarters
OECD Economic Outlook database,
Central Bank of Chile, and RBNZ
Terms of trade
Price index of exports of goods/ Price
index of imports of goods
IMF International Financial Statistics,
Central Bank of Chile, and RBNZ
Real Gross Domestic Product
OECD Economic Outlook database,
Central Bank of Chile, and RBNZ
Exchange-rate
US$/Local Currency Unit. Therefore,
an increase reflects an appreciation.
OECD Economic Outlook database,
Central Bank of Chile, and RBNZ
International interest rate
FED interest rate
IMF International Financial Statistics
Oil price
IMF International Financial Statistics
Real estate prices
Australia
House price indexes, established
houses (weighted average of 8 capital
cities)
Australia Bureau of Statistics
Norway
House price index of dwellings.
Statistics Norway
New Zealand
Quotable value index for dwellings
RBNZ
Sweden
Real estate price index for one- and
two-dwelling buildings for permanent
living
Statistics Sweden
Canada
Price of land and construction.
Statistics Canada
Chile
Cost of construction
Cámara Chilena de la Construcción
House permits
OECD National Accounts and RBNZ
Real GDP in residential construction
OECD National Accounts, RBNZ, and
Cámara Chilena de la Construcción
Private housing investment
OECD Economic Outlook database
Testing stabilisation policy limits in a small open economy
143
Annex B
Definitions and sources of variables used in VAR estimations
Variable
Definition
US Federal Funds Rate
Source
Federal Reserve Board
US output gap
Difference between actual real GDP OECD economic Outlook database
and potential GDP estimated from a
Hodrick-Prescott filter
Output gap
Difference between actual real GDP OECD economic Outlook database,
and potential GDP estimated from a Central Bank of Chile and RBNZ
Hodrick-Prescott filter
Inflation
Rate of change of headline CPI
Deviation of inflation from target
Difference between actual year-over- OECD economic Outlook database,
year headline inflation and mid-point Central Bank of Chile and RBNZ
inflation target (in IT countries) or
headline CPI inflation trend estimated
from a Hodrick-Prescott filter (in nonIT countries)
Domestic short-term interest rate
Money market rate
OECD economic Outlook database,
Central Bank of Chile and RBNZ
Money deviation from trend
M1 deviation from quadratic trend
OECD economic Outlook database,
Central Bank of Chile and RBNZ
Exchange-rate
Nominal exchange-rate US$/LCU thus, OECD economic Outlook database,
an increase represents an appreciation Central Bank of Chile and RBNZ
of the local currency
Long-term interest rate
10-year government bonds.
144
OECD economic Outlook database,
Central Bank of Chile and RBNZ
OECD economic Outlook database,
Central Bank of Chile and RBNZ
Reserve Bank of New Zealand and The Treasury
New Zealand’s monetary and exchange-rate policy in
international comparison by Klaus Schmidt-Hebbel
Discussion by John Edwards, HSBC
Both the valuable paper by Klaus Schmidt-Hebbel and the
The Treasury pointedly announced that they commissioned
remarkable conference which produced it address what
work on more direct ways of slowing house price inflation
was perhaps a crisis of confidence in the effectiveness
than changes in the cash rate.
of New Zealand’s economic policy instruments, which
deepened over the course of 2005. In the two years from
the beginning of 2004 to the end of 2005 the RBNZ
had increased the cash rate nine times to one of the
highest levels in the OECD. The results were persistently
disheartening. There was some preliminary evidence of a
slowdown in growth, but much of the slowdown was in
exports and there in response to the higher exchange-rate
induced by the higher cash rate. House prices continued
It was perhaps in this crisis of confidence that the RBNZ
conceived the bold idea of inviting a group of foreign
analysts to examine the New Zealand economy from a
completely external perspective. In one way or another,
the papers address the question of whether the monetary
authority in a small developed economy with open and
globally integrated financial markets, a freely floating
currency, and no capital controls could actually run an
effective monetary policy.
to increase, consumer credit growth remained formidable,
household consumption was markedly stronger than GDP
growth as a whole, import growth substantially exceeded
export growth, and the current account deficit ballooned
towards a new record as a share of GDP.
Between the conception of the conference and the actual
event,
however,
New Zealand’s
circumstances
quite
dramatically changed. The June quarter national accounts
published late in September that year showed a sharp
drop in output growth, and it was evident by the end of
While the currency had responded to a higher cash rate,
the long end of the yield curve had not. In the second half
of 2005 five year rates were as low as they had been when
the tightening began. Since New Zealanders borrow more
on fixed rather than variable rates, and are quick to move
where rates are cheaper, the impact of higher cash rates
on the household sector was muted. Oil prices accounted
for most of it, but it was nonetheless disconcerting that
inflation was markedly higher in the third quarter of 2005
than the first quarter of 2004. Even without oil, inflation
the year that the slowdown had continued through the
fourth quarter. After spiking higher earlier in the year,
local government approvals for new home construction
began to slide. In the second half of 2005 the US federal
funds rate first reached 4%, and it was evident the Federal
Reserve would continue to increase it. Once the top of the
New Zealand cash rate tightening episode was signaled at
the beginning of 2006, the currency sharply depreciated.
Both business and consumer confidence dropped, and
house price growth began to slow.
was much the same as it had been.
Through the early months of 2006 it appeared that
In these circumstances the Monetary Policy Statements, cash
rate review announcements and speeches from the Reserve
Bank of New Zealand took on, in my opinion, an edgy,
almost plaintive tone. It was acknowledged that there were
lags in policy, but sooner or later households would have
to roll over their mortgages and at higher rates. The Bank
directly and repeatedly talked down the currency, warning
foreign holders of New Zealand dollar financial instruments
monetary policy was not only effective, but perhaps too
effective. Growth stalled in the fourth quarter, raising the
possibility of a technical recession. Employment fell. The
market began to price in the possibility of an easing of the
cash rate sometime in 2006. It was well into the second
quarter and almost on the eve of the conference at which
this paper was delivered that it became apparent that
growth had probably been a little stronger in the first half of
that currency depreciation was highly likely. The RBNZ and
Testing stabilisation policy limits in a small open economy
145
2006 than the second half of 2005, despite the continuing
The great strength of the paper in my view is that it offers
downturn in residential construction. The downturn in
precisely what was needed to address the controversy
household consumption steadied, business investment
over New Zealand policy effectiveness. It considers it in
firmed, and export growth picked up. There had indeed
the context of other medium sized inflation-targeting
been a slowdown in growth, but New Zealand seemed
commodity-exporting economies, it considers over several
to have dodged recession. Not only had recession been
periods of time, and it considers it over a wide range of
averted, but the gradual strengthening of exports and the
variables. It does it with a daunting display of econometric
flattening of house price inflation suggested the economy
technique. In some respects the conclusions of the paper
was beginning to make the transition to export led growth
suggest New Zealand is much like other economies in
long sought by both the RBNZ and the Government.
roughly similar circumstances, though it has some features
The confirmation that monetary policy did work in
New Zealand (or at all events worked when the rest of the
world decided to go New Zealand’s way) coincided with
another important economic policy discovery. For over
a decade it had been widely believed that New Zealand
labour and multifactor productivity growth were among
which are notably different. The conclusions also suggest
quite strongly that New Zealand’s over all economic
performance has greatly improved over the last decade
compared to the prior decade. To briefly recapitulate what I
take to be the main points, Schmidt-Hebbel shows that:
•
New Zealand has on average performed reasonably well
the lowest in the OECD and well behind Australia. This was
on a test of output growth. At 2.5% over the period
a puzzle, because New Zealand had reformed its economy
1986-2006, annual average GDP growth in New Zealand
in much the same way as Australia and perhaps to a greater
is much lower than Chile’s but close to the other four
extent. It had utilized many of the same technologies, its
inflation-targeting comparable economies – Australia,
workforce was educated to much the same standard in
Sweden, Norway and Canada. He confirms, however,
much the same way, and it was anyway highly integrated
that output volatility is much higher than most of these
with the Australian economy. It may not have invested as
countries. Interestingly, output volatility declined in the
much in capital equipment as Australia but it had certainly
second half of the period, from 1996-2006.
invested a great deal. Why then should its productivity
•
growth be so much less? It was perhaps the lack of scale,
Notwithstanding the recent circumstances, New Zealand
does not have a particular problem with house prices,
some suggested. Or perhaps the elimination of minimum
or at least not in the long run. Though house price
rates awards in New Zealand in 1994 had permitted a
inflation has been quite dramatic in recent years, over
decline in real wages which in turn lowered the capital to
the whole period real house price growth has averaged
output ratio. This considerable controversy was brought to
3.96% annually, similar to Australia’s. Over the whole
a halt at the end of March 2006 when a new methodology
period, house price growth is actually less volatile than
applied by Statistics New Zealand revealed the productivity
comparable countries.
growth gap with Australia had disappeared, qualitatively
confirming some earlier work produced by the New Zealand
Treasury. In fact New Zealand productivity growth was if
•
New Zealand does, however, experience more impact
from the housing construction cycle. Real private
investment in housing (4.1%) is a little below the average
anything somewhat higher than Australia’s.
(4.4%) of comparable countries. But New Zealand has
Within a six month period two big pieces of “conventional”
wisdom
about
New Zealand’s
economy
had
been
undermined by new data. These were the circumstances
immediately preceding the RBNZ conference, at which
Schmidt-Hebbel’s paper delivered another heavy blow to
the highest volatility in real housing investment of the
whole group. Schmidt-Hebbel plausibly suggests this is
caused by immigration swings. The standard deviation
of population growth in New Zealand is three times
Canada’s.
the notion of New Zealand’s economic ungovernability.
146
Reserve Bank of New Zealand and The Treasury
•
With farm products a substantial share of exports, it
add depth to the debate on New Zealand monetary policy
is sometimes supposed New Zealand gets less and less
effectiveness.
for its exports and pays more and more for imports.
I have only a few additional thoughts to offer:
Schmidt-Hebbel shows that far from being persistently
disappointing New Zealand’s terms of trade have risen
•
Schmidt-Hebbel offers but does not I think argue
through a suggestion that New Zealand should
on average 1.34% a year over the whole period –
announce a more precise horizon for the achievement
somewhat above the average of the comparator group.
of the inflation target. It is only in recent years that the
More surprisingly, New Zealand’s terms of trade are
RBNZ has been permitted more flexibility in the time
among the least volatile of the 6 country group.
period over which the inflation target is achieved, and
Turning to the analysis of shocks to inflation Schmidt-
also permitted a somewhat higher target mid-point.
Hebbel finds that:
•
•
Yet this is also the period in which Schmidt-Hebbel
In New Zealand headline inflation did not respond to
finds monetary policy has become more effective,
exchange-rate shocks in the period 1989-1997, but does
target achievement has improved, and the economy
for 1998-2005. Overall Schmidt-Hebbel concludes that
has become less volatile. I would have thought these
exchange-rate pass through to inflation is significant
findings would make New Zealand cautious about
but relatively small in New Zealand because of “well
returning to an earlier and less successful monetary
anchored inflationary expectations.”
targeting regime. That said I think there is great merit in
the suggestion that New Zealand (and for that matter
Headline inflation responds positively and significantly
Australia) look seriously at the Norwegian, Swedish and
to oil price shocks in both periods. The effect is much
Chilean models of explicit counter cyclical fiscal policy.
the same as elsewhere, but stronger in New Zealand.
•
Long term interest rates do respond to changes in
•
It is important to keep a good sense of what the
results do and don’t tell us. The paper relies mostly on
the cash rate and so does the exchange-rate,but the
average annual rates over a long period. For example,
exchange-rate does not respond much.
the results show that on average over the period house
And on the central question of monetary policy effectiveness
price growth in New Zealand is about the same as
he finds that:
•
Australia’s and is less volatile the average of the group.
There is a significant negative response of both inflation
But Schmidt-Hebbel is not I think saying that house
and the output gap to monetary contraction.
price inflation cannot be a big issue in New Zealand
monetary policy. His introduction makes it plain that it
He concludes that “..monetary policy transmission in
is. What we should be concluding is that New Zealand’s
New Zealand is broadly comparable to, and at least as strong
problems in this respect are not unique or especially
as, that observed” in the group of comparable economies.
intractable.
Indeed, he finds that the mean absolute deviation from
the Banks inflation target is lower than average for other
•
So too the exchange-rate may not on average have
IT countries, and accuracy has improved between 1990-97
much impact on inflation but there are certainly periods
and 1998-2005. Comparing the two periods he also finds
in New Zealand when currency appreciation has usefully
a big reduction in inflation and output gap volatility in the
slowed general price inflation, and periods when
latter period.
depreciation has added to it. It is true that in recent
These are valuable findings, the result of bringing an
impressive complexity and variety of techniques to work
on a number of different economies and a wide range of
variables. It is exactly the comparative study needed to
Testing stabilisation policy limits in a small open economy
episodes the pass-through has been less but I doubt this
is because inflation expectations have been changed by
targeting. If this was so then we would expect more
pass-through in the earlier period than the later period
147
instead of the reverse. We would expect that oil prices
•
•
Finally, while both the long term average outcomes and
would also not have much pass-through, but Schmidt-
the recent experience establish that the RBNZ retains
Hebbel finds they do. In the Australian case the import
a reasonable degree of effectiveness, it remains the
price index still responds quite directly to currency
case that the experience from 2003 to 2006 was quite
changes, but retail prices do not. Both economies are
difficult. Long term rates may on average respond to
far more open and competitive than they were. There
short term rates, but on this occasion they did not. The
is as the retailers complain less “pricing power.” These
exchange-rate may not always respond to monetary
structural changes are I think much more important
contraction but on this occasion it seemed to respond
than the inflation targeting regime of the central bank in
quite a lot. Demand and inflation in the long term and
controlling pass through of exchange-rates. One might
on average respond to monetary contraction, but in this
add that having little pass-through from exchange-rates
case it was a long time coming. A particular difficulty
vitiates the function that exchange-rate changes are
in this period was that major central banks in the rest
expected to perform, so it is not necessarily a welcome
of the world imposed unusually low cash rates, and
trend.
global markets were sufficiently confident of continuing
Schmidt-Hebbel calculates that at an annual average
of 1.33% the trend appreciation of the real effective
exchange-rate is much higher than the average, and
low inflation to permit unusually low bond rates. On
average the RBNZ is effective. In any particular policy
episode, however, lots can go wrong.
also more volatile. A glance at a chart confirms that over
the last 20 years the nominal New Zealand exchangerate has cycled around an average of 60 on the trade
weighted index, and is today much where it was in
1986. Real appreciation over the period presumably
reflects higher inflation in New Zealand. Since the
increase in the price level was substantially greater
over the period from 1986 to 1990 than from 1991 to
2006, the real appreciation would have been mostly
in the earlier period. This underlines the very different
episodes caught up in the period 1986-2006. SchmidtHebbel is of course well aware of this distinction and
usefully makes it elsewhere.
148
Reserve Bank of New Zealand and The Treasury
External imbalances in New Zealand
Sebastian Edwards, University of California, Los Angeles and National Bureau of Economic Research†
1
Introduction
Watch, on May 19th 2006, only Brazil, Indonesia, the
During the last three years New Zealand has faced
Philippines and Turkey, among all large countries monitored
increasingly large external imbalances. The current account
by the investment banks, had higher policy interest rates
deficit has increased from 4.3% of GDP in 2003 to almost
than New Zealand.
9.0% of GDP in 2005. During the same period the country’s
Although during the last few months the macroeconomic
net international investment position (NIIP) has gone from a
picture has changed somewhat – the NZD has weakened
negative level equivalent to 78.5% of GDP to negative 89%
and increases in housing prices have moderated – a
of GDP. Also, some of the most important macroeconomic
number of important policy questions remain. Perhaps
variables, including interest rates and the exchange-rate,
the most important one is whether the very large current
have experienced a higher degree of volatility than in other
account deficit of 9% of GDP is sustainable. If it is not, as
commodity countries such as Australia and Canada. Much
many analysts have argued, the next question is what will
of the growth in New Zealand’s external imbalances has
adjustment look like. Will it be smooth and gradual, and thus
been fuelled by a rapid real estate boom that has allowed
with little or no real costs? Or, will it be abrupt and severe?
consumers to withdraw significant amounts of money from
Another way of putting this issue is whether New Zealand
1
2
their homes’ equities, and increase consumption. These
faces a (relatively) high probability of experiencing a
developments have generated concerns among experts
“sudden stop” in capital inflows, and an abrupt reversal in
and observers. According to a recent article in the Financial
the current account deficit.4
Times (March 31st, 2006, emphasis added):
Other important policy issues are related to the relationship
“Countries with large external imbalances such as Iceland
between economic policy and external imbalances. In
and New Zealand, as well as Hungry…Turkey, Australia
particular, has macroeconomic policy contributed to the
and South Africa, are seen as most vulnerable as foreign
creation of these external disequilibria? And, has monetary
investors head for the exits.”3
policy lost some of its power in the last few years? This
In an effort to cool down the economy, and to reign-in
latter question emerges from the fact that, in spite of the
the rapid growth of housing prices, the Reserve Bank of
increase in the OCR policy rate by 225 basis points between
New Zealand has raised its official policy interest rate
January 2004 and December 2006, longer term rates,
(the OCR) several times since January 2003. At 7.25%,
including interest rates on mortgages, have changed with
New Zealand currently has one of the highest policy interest
considerable delay and to a much lesser extent. A central
rates in the world. According to JP Morgan’s Global Data
question, thus, is whether New Zealand should contemplate
some changes in its monetary policy framework, and/or on
†
1
2
3
This a revised version of a paper prepared for the
“Macroeconomic Policy Forum” organized by the New Zealand
Treasury and the Reserve Bank of New Zealand, held on June
12th, 2006 in Wellington. I am grateful to many colleagues
in New Zealand for their help and generosity. In particular,
I want to thank Peter Bushnell, Grant Spencer, Aaron Drew,
Anella Munro, Rishab Sethi, Bob Buckle, Arthur Grimes and
Murray Sherwin. I thank Roberto Alvarez for his excellent
assistance in Los Angeles and Bob Buckle and Aaron Drew
for their comments.
See the IMF’s most recent reports for a broad analysis of
New Zealand’s macroeconomic position and challenges; IMF
(2006a, 2006b). See also IMF (2004a, 2004b).
See, for example, Robinson, Scobie and Hallinan (2006).
Financial Times, “Iceland Acts to Head off Currency Crisis,”
March 31st, 2006. In http://news.ft.com/cms/s/9d6a950ec053-11da-939f-0000779e2340.html. Emphasis added.
Testing stabilisation policy limits in a small open economy
monetary policy implementation. Other specific questions
that have emerged from recent economic developments
and debates include:
4
The most recent IMF reports on New Zealand ask whether the
current account poses macroeconomic risks to New Zealand;
IMF (2006a, 2006b). On “sudden stops” and external
adjustment see, for example, Edwards (2004) and Calvo et al
(2004).
149
•
Is the higher volatility in exchange-rates and interest
deficit is over 9 per cent of GDP. This exercise allows me to
rates observed in New Zealand the result of a lack of
evaluate whether, according to the model, the probability
synchronization between the New Zealand business
of New Zealand experiencing an abrupt and costly reversal
cycle and the business cycle in the major economies (e.g.
has increased significantly in the last few years. The paper
the G-3), or is it a reflection of structural weaknesses in
also deals with monetary policy and its effectiveness in a
New Zealand, including the fact that it is a very small,
context of large external deficits.
very open, commodity-exporting economy?
•
•
The rest of the paper is organized as follows: In Section
Does the close economic relationship between
2 the evolution of New Zealand’s current account balances
New Zealand and Australia play a role in explaining the
during the last two decades is analysed (the starting point
large and persistent imbalances?
of the analysis is 1985, when the NZD was floated). I deal
Should a small country such as New Zealand adopt the
Greenspan view on asset prices, and ignore a property
boom when conducting monetary policy?
with real exchange-rate trends, and with the evolution of
different external accounts. I focus on the recent evolution of
New Zealand’s net international investment position (NIIP),
and discuss some recent computations on the sustainable
The purpose of this paper is to analyse the potential
consequences of New Zealand’s external imbalances.
A particularly important issue addressed in the paper is
the possible nature of future external adjustments. More
specifically, I investigate the probability that New Zealand
will undergo a costly adjustment characterized by an abrupt
and large current account reversal. This is an important
question, since, as I argue in Section 2, there are strong
indications that the current magnitude of the external
level for New Zealand’s current account. In Section 3 an
international comparative analysis of New Zealand’s current
account balance is provided. I show that the persistence
and magnitude of New Zealand’s deficit has virtually no
comparison in the world. I also provide some computations
on the consolidated current account deficit of AustraliaNew Zealand. I show that although this consolidated deficit
is still large from an international perspective, it is smaller
than the current New Zealand deficit.
imbalance in New Zealand is not sustainable through time.
Section 4 asks whether New Zealand’s large external
In order to achieve sustainability, the current account deficit
will have to decline by 3 to 5 percentage points of GDP. It
makes a difference whether this adjustment is gradual or
abrupt; there is ample evidence that suggests that abrupt
current account adjustments (or reversals) are costly, in
terms of lower GDP growth. I deal with the question of
the probability of experiencing an abrupt adjustment in the
following way: I analyse the main characteristics of countries
that in the past have suffered “sudden stops” and abrupt
current account reversals. More specifically, I use randomeffect probit models to estimate the determinants of the
probability of experiencing a major reversal. Following this,
I estimate the probability of reversals using New Zealand
specific data at different points in time. I compute this
probability using New Zealand data for the early 2000s,
when the current account deficit was 2.8 per cent of GDP
(a figure slightly lower than what many analysts consider
imbalances should be a cause for concern. Recent evidence
presented in Calvo et al (2004), Edwards (2004, 2004a,
2005a, 2005b) and Frankel and Cavallo (2004) suggests
that countries that experience sudden declines in capital
inflows and/or abrupt current account reversals have
suffered significant reductions in the rate of economic
growth. In this Section I use a multi-country data set to
evaluate the probability that New Zealand will face an
abrupt reversal in its current account in the near future. A
number of important macroeconomic policy issues related
to the external sector in Section 5. In particular, I analyse
New Zealand’s monetary policy framework and I ask
whether the RBNZ should directly consider exchange-rate
developments when determining the OCR. Finally, I offer
concluding remarks in Section 6, touching briefly on other
policy options, including the merits of New Zealand and
Australia having a common currency.
to be sustainable), and 2006, when the current account
150
Reserve Bank of New Zealand and The Treasury
2
Twenty years of current
Figure 1
account balances and the
Real exchange-rate and current
exchange-rate behaviour in
New Zealand
account balance, 1975-2005
Index
120
In this Section I analyse the evolution of New Zealand’s
110
current account and trade weighted real exchange-rate. The
100
analysis starts with 1985, the year New Zealand adopted
90
-6
-8
-10
80
parts: 5 First, I discuss the evolution of the real exchange-
-12
Current account balance (RHS)
70
rate (RER) and current account during the last two decades.
Real exchange rate
60
I argue that it is possible to divide the last twenty years of
the most recent data on New Zealand’s current account,
-2
-4
a floating exchange-rate. The Section is divided in three
RER behaviour into seven distinct phases. Second, I discuss
% GDP
0
-14
-16
75 77 79 81 83 85 87 89 91 93 95 97 99 01 03 05
Source: Statistics New Zealand
•
including its sources of financing. Here I point out that in
First, it shows that deficits have been a “normal” state
of affairs in New Zealand for the last 20 years. In fact,
New Zealand, as opposed to the US for example, the income
going back for another ten years, one finds that in
account (which measures net interest, dividend, profits
the second half of the 1970s current account deficits
remittances and transfers to the rest of the world) has been
exceeded the 12% of GDP mark!
the main source of disequilibria. More recently, however,
New Zealand has experienced an important deterioration
•
Second, this Figure shows that while recent deficits
have been very large indeed (in the order of 9% of GDP
in its trade account balance. Finally I deal with the recent
in late 2005) they have historical precedents. Current
evolution of New Zealand’s net international investment
account deficits reached that level (briefly) in early
position.
1986.
•
Third, in the last twenty years there have been four
The current account deficit and seven
episodes of retrenchment in the current account
phases of real exchange-rate behaviour in
deficit.
New Zealand
o
In Figure 1 quarterly data for New Zealand’s current
The first of these retrenchment episodes took place
between March 1986 and March 1989, when the
account balance as percentage of GDP and the evolution
deficit shrunk from 8.7% of GDP to a mere 0.7%
of the trade-weighted index of the New Zealand dollar real
of GDP; this has been one of the largest current
exchange-rate are presented for the period 1975-2005. In
account reversals in the modern economic history
this Figure, as in the rest of this paper, an increase in the RER
of advanced countries.
index represents a real exchange-rate appreciation, while a
decline in the index captures a depreciating trend. Several
o
The second external adjustment episode was brief
and modest, and occurred between June 1990 and
interesting features emerge from Figure 1:
December 1991, when the deficit went from 4.2 to
2.8% of GDP.
o
The third retrenchment was in the September 1997June 1999 period; the deficit declined from 6.7 to
5
An interesting exercise, but one that is beyond the scope
of this paper, is to compare exchange-rate volatility (both
unconditional and conditional) in New Zealand to that of
other commodity currencies such as the Australian dollar and
Canadian dollar.
Testing stabilisation policy limits in a small open economy
4.0% of GDP.
151
o
And the final deficit reduction episode took place
of the NZD of 17.3%. During this short phase the
during June 2000 and December 2001, when the
current account deficit was very large.
deficit declined from 6.5% to 2.8% of GDP.
•
•
•
•
Phase 2: December 1985-December 1986. This was
It is interesting to note that two of the current account
also a very short phase. During these 12 months the
retrenchment episodes discussed above were significant,
NZD experienced a 9.4% cumulative depreciation.
exceeding 3.5% of GDP; these adjustment episodes,
During this phase the current account deficit began to
however, were stretched over a period of several years.
decline.
Figure 1 also shows that during the period under study
•
Phase 3: December 1986-June 1988. This is the last of
the RER index experienced significant movements: its
the “short” phases that occurred during the early years
mean was 91.0, its minimum 71.3, and its maximum
of floating. During this period the NZD real exchange-
was 108.0. The standard deviation of the RER index was
rate experienced a rapidly appreciating trend. The
8.9.
trough-to-peak change in the index was 22.3%. Real
exchange-rate volatility, measured as the standard
Figure 1 shows a pattern of mild negative correlation
deviation of the monthly log differences of the RER
between the trade-weighted real value of the New
index, was 0.023. Interestingly, during this phase the
Zealand dollar and the current account balance. Periods
NZD strengthened in real terms at the same time as the
of strong dollar have, overall, tended to coincide
current account deficit was declining in a very significant
with periods of (larger) current account deficits. The
fashion.
contemporaneous coefficient of correlation between the
(log of the) RER index and the current account balance
•
Phase 4: June-1988-March 1993. This is the first of
is –0.22; when lead-lag structures are considered, the
four “long” phases in RER behaviour; it is a depreciating
correlation coefficient declines. This correlation between
phase. As may be seen from Figure 1, between
the trade weighted value of the currency and the
December 1988 and September 1990 the RER was
current account is lower in New Zealand than in the US,
quite stable, having reached a (temporary) plateau of
where the contemporaneous correlation coefficient is -
sorts. At that point, however, the depreciating trend
0.53, and the three quarters lagged correlation is -0.60.
resumed. The peak-to-trough accumulated change in
This may be explained by the fact that in New Zealand
the trade weighted RER index during this period was
the main component of the current account deficit is
-22.4%. During the early part of this Phase the current
the incomes account, while in the US it is the trade
account deficit widened. Starting in late 1990, however,
account. In New Zealand the simple contemporaneous
the deficit stabilized at slightly below the 4% of GDP
correlation between the (log of the) real exchange-rate
mark. During this period the standard deviation of the
and the trade account-to-GDP ratio is -0.41.
monthly log differences of the RER index was 0.022.
An analysis of the data in Figure 1 indicates that it is possible
•
Phase 5: March 1993-March 1997. This phase is
to distinguish seven distinct phases in New Zealand dollar
characterized by a trough-to-peak real exchange-
real exchange-rate behaviour for the twenty-year period
rate appreciation of 28.9%. The strengthening of the
1985-2005. A brief analysis of these seven phases provides
currency was accompanied by a significant widening
a summary of the history of New Zealand’s external sector
of the current account deficit. Interestingly, during this
since the inception of floating in 1985:
phase real exchange-rate volatility declined significantly;
•
Phase 1: March 1985-December 1985. This phase was
very short and includes the early months of floating. It
was characterized by a steep accumulated appreciation
152
the standard deviation of the monthly log differences
of the RER index was 0.011. This is significantly lower
than (real) exchange-rate volatility in other commodity
Reserve Bank of New Zealand and The Treasury
•
countries such as Canada and Australia (Edwards
Figure 2
2006).
Components of the current
Phase 6: March 1997-December 2000. This is phase is
account balance, 1987-2005
Goods and Services
characterized by a trough-to-peak real exchange-rate
depreciation of 32.4%. During the early part of this
% GDP
5
phase the current account deficit retrenched to 3.9% of
4
4
GDP in December 1998. It then widened until it reached
3
3
6.5% in June 2000. During this period unconditional
2
2
1
1
real exchange-rate volatility increased to 0.023.
•
% GDP
5
0
0
Phase 7: December 2000-December 2005: This phase
-1
-1
lasted the longest. During this period the real exchange-
-2
-2
-3
-3
rate appreciated by an impressive 51.5%, and real
exchange-rate volatility increased to 0.029. From the
-4
1987
-4
1989
1991
third quarter of 2001 through December of 2005 the
current account deficit increased steadily from 2.8% of
1993
1995
1997
1999
2001
2003
2005
Investment income
% GDP
% GDP
0
0
-1
-1
-2
-2
-3
-3
-4
-4
-5
-5
-6
-6
-7
-7
Data decomposition
-8
-8
Going beyond the current account, in Figure 2 data from
-9
1987
GDP to almost 9% of GDP. During this phase the real
exchange-rate index experienced its highest degree of
volatility, with a standard deviation of the log difference
of 0.033.
Decomposing the current account balance
-9
1989
1991
1993
1995
1997
1999
2001
2003
2005
1987 through 2004 is presented for: (a) the balance of
trade of goods and services as a percentage of GDP; (b) the
income account, also as a percentage of GDP, and (c) the
Transfers
% GDP
% GDP
1.6
1.6
1.4
1.4
A number of important facts emerge from these Figures.
1.2
1.2
First, until September 2004 the trade account was mostly
1.0
1.0
0.8
0.8
0.6
0.6
0.4
0.4
0.2
0.2
0.0
0.0
transfers account as a percentage of GDP.
in surplus. There were only two brief periods (in 1990 and
1999-2000) when there were small deficits (below 1 per cent
of GDP). However, since December 2004 (and until the time
of this writing) the trade deficit has increased significantly,
reaching its highest level since the adoption of floating
-0.2
-0.2
1987 1989 1991 1993 1995 1997 1999 2001 2003 2005
exchange-rates. This recent emergence and prominence of
Source: Statistics New Zealand
the trade deficit suggests that in the recent years there may
Second, the incomes account has experienced very large
have been a structural change in macroeconomic relations
deficits, and throughout most of the period under study it
in New Zealand. The recent work by Kim, Hall and Buckle
explains, more than fully, the current account deficit (second
(2006) and Munro and Sethi (2006) suggest that a structural
panel, Figure 2). Only in the last year or so the income
change in the economy’s ability to “smooth consumption,”
account deficit has been lower than the overall current
may indeed have occurred. I discuss this issue in greater
account deficit. The historically very large deficit in the
detail in Section 4 of this paper.
Testing stabilisation policy limits in a small open economy
153
Figure 3
Evolution of net savings, 1972-2005
Net National Savings
Households Savings
%GDP
8
NZ$M
2000
%GDP
8000
7000
6
NZ$M
2000
0
0
6000
-2000
-2000
-4000
-4000
-6000
-6000
-8000
-8000
-10000
-10000
5000
4
4000
2
3000
2000
0
1000
0
-2
-1000
-12000
-12000
-2000
-4
71 73 75 77 79 81 83 85 87 89 91 93 95 97 99 01 03 05
71 73 75 77 79 81 83 85 87 89 91 93 95 97 99 01 03 05
Government Savings
NZ$M
12000
Business Savings
NZ$M
12000
NZ$M
7000
NZ$M
7000
10000
10000
6000
6000
8000
8000
5000
5000
6000
6000
4000
4000
4000
4000
3000
3000
2000
2000
2000
2000
0
1000
1000
0
-2000
-4000
-2000
0
-4000
-1000
0
-1000
71 73 75 77 79 81 83 85 87 89 91 93 95 97 99 01 03 05
71 73 75 77 79 81 83 85 87 89 91 93 95 97 99 01 03 05
Source: Claus and Scobie (2002), updated using information from Statistics New Zealand
income account in New Zealand is a reflection of the very
of GDP. More impressive than this, however, is the fact
large negative NIIP, a subject that I discuss in some detail in
that (net) household savings have declined very drastically
the following section. An important question, and one that
since the mid 1990s, and in particular since 2002. This rapid
I explore below, is whether New Zealand’s large negative
collapse in household savings has been partially offset by a
income account balance is related to the close economic ties
rapid increase in government savings (which have recently
between New Zealand and Australia. Finally, the third panel
surpassed 6% of GDP) and by a recovery of corporate
in Figure 2 shows that the transfers account has exhibited a
savings since the mid 1990s.
relatively stable surplus throughout the period under study.
As said above, the drastic decline in household savings
The evolution of savings and the current account
has been related to a rapid increase in housing prices and,
The deteriorating trade balance since around 2002 coincides
thus, in household wealth (See Robinson, Scobie, Hallinan,
with a significant decline in net household savings. In turn,
2006). It is precisely for this reason that a number of analysts
this has been associated with a rapid increase in housing
have argued that a moderation in New Zealand’s current
prices. In Figure 3, data on the evolution of net savings
account deficit will require a decline in housing prices.8
for the period 1972-2005 is presented.7 Several trends
This situation has also prompted the question of whether
are apparent from this Figure. Net national savings have
the Reserve Bank of New Zealand should explicitly take
6
experienced a declining trend. While during the early 1970s
net national savings hovered around the 6% of GDP mark,
during the last few years they have averaged less than 4%
154
6
7
On the recent evolution of housing prices in New Zealand see,
for example, Robinson, Scobie and Hallinan (2006).
The historical series are from Claus and Scobie (2002). I have
updated them using data from Statistics New Zealand.
Reserve Bank of New Zealand and The Treasury
Table 1
New Zealand net international investment position
At 31 March (NZ$ million and Percentages)
2001
2002
2003
2004
2005
Direct Investment Abroad
-35,699
-40,565
-42,676
-54,901
-58,239
40.8
41.0
41.7
49.0
46.2
Portfolio Investment Abroad
-34,400
-33,469
-40,410
-40,086
-43,292
39.3
33.8
39.5
35.8
34.3
Other Investment Abroad
Financial Derivatives
Reserve Assets
-29,916
-32,665
-26,353
-24,686
-31,074
34.2
33.0
25.8
22.0
24.6
3,989
-37
-1,993
-2,510
-2,345
-4.6
0.0
1.9
2.2
1.9
8,566
7,723
9,115
10,093
8,828
-9.8
-7.8
-8.9
-9.0
-7.0
Net International Investment Position
-87,461
-99,013
-102,318
-112,090
-126,121
NIIP as % of GDP
-76.2
-80.1
-79.3
-81.6
-85.4
Source: Statistics New Zealand
into account real estate prices when conducting monetary
9
policy. In the light of low savings, a significant fraction of
expenditure financing has taken place through the offshore
capital market, via the issuance of New Zealand dollar
denominated bonds, sometimes referred as Eurokiwis, NZD
Eurobonds, and NZD Uridashis.10
Figure 4
New Zealand net external position, 1970-2004
% GDP
0
% GDP
0
-20
-20
-40
-40
-60
-60
-80
-80
-100
-100
-120
-120
70 72 74 76 78 80 82 84 86 88 90 92 94 96 98 00 02 04
Source: Lane and Milesi-Ferretti (2006)
The evolution of New Zealand’s net
international investment position and the
financing of recent current account deficits
The counterpart to the large current account deficits of
the last thirty years has been an increasingly negative Net
8
9
10
See, for example, Merrill Lynch, “NZD: The Long Slide,”
Foreign Exchange Strategy, 13 April 2006.
This question is not unique to New Zealand. It has been
addressed several times in recent discussions on US monetary
policy. See, for example, Ben Bernanke’s “The Global
Savings Glut and the US Current Account Deficit,” Speech
delivered on March 10, 2005. It may be found at: http://www.
federalreserve.gov/boarddocs/speeches/2005/200503102/
default.htm.
For details on how the offshore market works, see Drage et. al.
(2005).
Testing stabilisation policy limits in a small open economy
International Investment Position (NIIP). Figure 4 presents
the evolution of New Zealand’s NIIP since 1970. The data
have been taken from Lane and Milesi-Ferretti (2006).
When alternative New Zealand data sources are used the
results are similar: for instance according to New Zealand
official statistics in the period 2001-2005 the NIIP was 76%, -80%, -79%, -82%, and -86%, respectively. These
155
Figures are not very different from those depicted in Figure
11
GDP) in the world. As a point of comparison the NIIP in the
US is currently -30% of GDP, and that of Australia is – 57%
3.
Table 1 provides greater detail on the recent evolution of the
NIIP, as well as of its most important components; naturally,
the year-to-year changes in the different components of the
NIIP provide information on the recent sources of financing
of the current account deficit. Table 2 presents data on the
(see Table 6). The NIIPs of most other advanced countries
are, in fact, positive, denoting that these are net creditor
countries. Figure 4 shows that in spite of some wave-like
movements, New Zealand’s NIIP has exhibited a declining
trend through time, becoming increasingly negative.
recent evolution of this financing. As pointed out above,
In a recent important paper Munro (2005) discusses the
during the last few years an important fraction of foreign
evolution of the NIIP in New Zealand during the last few
financing to cover the current account deficit has been
years. Her most important findings may be summarized as
obtained in the offshore bond market or market for NZD
follows:
denominated Eurobonds (Eurokiwis) or NZD denominated
•
Uradishis, purchased by retail investors in Japan (Drage et.
The increasingly negative NIIP of the last few years has
been the result of private sector investment.
al., 2005; IMF 2006a, 2006b).
•
Table 2
New Zealand’s public sector net international investment
position (including the New Zealand Superannuation) is
Net financial flows, 2003-2005
virtually zero.
(NZ$, million)
•
The importance of bank loans has increased very
Flow
2003
2004
2005
significantly as a source of external liabilities. Indeed,
Direct investment
3,252
4,949
4,123
these higher bank loans have financed the real estate
Equity capital
n.a.
n.a.
n.a.
boom of the last few years.
Reinvested earnings
n.a.
n.a.
n.a.
Other capital
5,306
2,586
1,561
Portfolio investment
1,573
7,332
-150
and liabilities, New Zealand is not subject to significant
-279
-2,518
-1,728
“valuation effects” stemming from exchange-rate
changes.
Equity securities
Debt securities
•
Given the currency composition of international assets
1,851
9,851
1,579
Other investment
630
479
11,708
Trade credits
n.a.
n.a.
n.a.
Loans
-969
-669
11,138
Deposits
1,364
668
1,078
Other instruments
n.a.
n.a.
n.a.
-1,345
-685
-3,475
Special drawing rights
-8
-7
-4
Reserve position in the fund
-304
284
361
level.12 The level at which the NIIP to GDP ratio will stabilize
Foreign exchange
460
-873
-3,627
will depend on the attractiveness of the country’s assets
-205
to international investors. If the international (net) demand
Reserve assets
-1,491
-91
Total
Other reserve asset claims
4,110
12,075 12,206
Current Account Balance
-5,937
-9,385 -13,688
•
In the last few years the maturity structure of
New Zealand’s external liabilities has declined.
Modern analyses of current account sustainability are based
on the notion that in equilibrium the ratio of the NIIP to
GDP (or to some other aggregate) has to stabilize at some
for the country’s securities (including debt and equity)
is high, the NIIP to GDP ratio will stabilize at a high rate.
The opposite will be true if this international demand is
Source: Statistics New Zealand
As discussed in some detail in Section 3 of the paper,
New Zealand’s NIIP is one of the most negative (relative to
11
156
Using the Lane and Milesi-Ferretti data has two advantages.
First, they provide long time series, and second, it is easier to
make comparisons across countries.
low. The sustainable current account to GDP ratio will,
then, depend on this long term stable NIIP to GDP ratio,
and on the country’s long term trend rate of real growth
12
Milesi-Ferretti and Razin (1996), Edwards (2005). For an
illuminating sustainability analysis of New Zealand, see
Munro (2005).
Reserve Bank of New Zealand and The Treasury
and equilibrium rate of inflation. The relationship between
level (relative to GDP), and economic growth is very
the equilibrium and stable ratio of NIIP to GDP, which I will
high, New Zealand will have to go through a substantial
γ,
and the sustainable current account deficit
adjustment process where the current account deficit will
13
( SCAD ) may be written as follows:
have to decline significantly. For instance, if from Table 3
(1) SCAD = γ ( g
one takes the combination of a NIIP of -120% of GDP and
denote as
where ( g
T
T
+ π ),
+ π ) is the nominal rate of growth of trend
T
GDP, g is the long run trend real rate of growth of GDP
π is the long run steady-state inflation rate (which
and
I assume to be equal to the long run international rate
of inflation). According to this simple and yet powerful
equation, the sustainable current account deficit will depend
on both the international demand for the country’s assets
γ
and the country’s nominal rate of growth.
γ , of course,
is not an invariable number; as pointed out above, it is a
variable, whose value changes through time, depending on
the perceived riskiness and/or attractiveness of the country
nominal growth of 5.0% of GDP, the “sustainable” current
account deficit is 5.7% of GDP; this means that adjustment
will have to exceed 3% of GDP. But what is perhaps more
telling is that these figures indicate that under rather small
changes in the key parameters, the magnitude of the
external adjustment required to bring the current account
deficit in line with its long run sustainable level would be
nothing short of brutal. Take, for example, the case where
the steady state NIIP is -80% (still a remarkably high Figure
from international standards) and nominal growth is 5%.
This combination implies a SCAD of 3.8% of GDP, more
than 5 percentage points below its current level!
in question.
A key question that emerges from this analysis, and one that
Table 3
I address in great detail in Section 4 of this paper, is whether
Sustainable current account
this external sector adjustment is likely to be gradual (and
deficit under different scenarios
thus largely harmless from an economic point of view),
Target IIP
Nominal GDP Growth
(% GDP)
4.5%
5.0%
5.5%
5.8%
6.0%
80
3.4
3.8
4.2
4.4
4.5
100
4.3
4.8
5.2
5.5
5.7
120
5.2
5.7
6.3
6.6
6.8
or abrupt and costly. That is, the question is whether
international investors will slowly reduce the rate at which
they add New Zealand securities to their portfolios, or
whether this process will come to an abrupt and sudden end.
Before turning to this important issue, however, I tackle two
Source: Munro (2005)
important questions in the next section of the paper. First, I
Munro (2005) presents calculations for the SCAD under
analyse New Zealand’s external position in an international
alternative values of the long run steady state NIIP ratio
comparative context, and show that New Zealand’s
and nominal rate of growth. Munro’s computations
case is quite unique. Second, I analyse the way in which
are reproduced in Table 3. The results in this Table are
New Zealand’s special economic relationship with Australia
particularly interesting, in that they point out that even if
affects the NIIP and current account statistics.
the NIIP stabilizes at a significantly more negative level than
the current -89%, and if nominal growth is very high by
historical standards (say, 5.5% on average), the sustainable
current account deficit is still significantly smaller than the
current 8.9% of GDP.
The implications of these calculations are simple, and yet
very important: even under an optimistic scenario, where
the (negative) NIIP stabilizes at a significantly more negative
3
The New Zealand current
account in an international
comparative context
International comparisons
How large are New Zealand’s recent current account deficits,
from a comparative point of view? How does the persistence
13
See Edwards (2005) for a detailed analysis along these lines
that incorporates the dynamic effects of changes in γ .
Testing stabilisation policy limits in a small open economy
of deficits compare with that of other countries? And, how
157
Table 4
Distribution of current account deficits
By region, 1970-2004
Region
Mean
Median
1st Perc.
1st Quartile
3rd Quartile
9th Perc.
0.7
-3.8
-1.6
3.0
4.8
A: 1970-2004
Industrialized countries
0.6
Latin Am. and Caribbean
5.4
4.1
-2.5
1.1
8.0
16.9
Asia
3.2
2.7
-7.0
-0.3
6.4
11.4
Africa
6.3
5.3
-3.4
1.2
9.9
16.9
Middle East
0.0
1.4
-18.8
-5.0
6.4
13.6
Eastern Europe
3.9
3.0
-2.4
0.3
6.1
10.7
Total
4.0
3.1
-4.4
-0.1
7.2
13.4
A: 1984-2004
Industrialized countries
0.2
0.3
-4.7
-2.3
2.7
4.8
Latin Am. and Caribbean
5.1
3.7
-2.5
1.1
7.0
17.0
Asia
2.4
2.6
-8.2
-0.8
6.1
10.3
Africa
5.9
4.6
-3.5
0.9
9.1
16.2
Middle East
2.3
1.5
-12.4
-4.0
6.3
14.9
Eastern Europe
4.0
3.1
-2.5
0.3
6.6
10.9
Total
3.9
2.9
-4.5
-0.2
6.7
13.0
Source: Author’s elaboration based on World Development Indicators
large is the (negative) net international liabilities position in
only cases are Ireland in the 1970s and early 1980s; Malta;
New Zealand when compared, from a historical vantage, to
New Zealand; Norway and Portugal.
that of other advanced countries? In Table 4, the distribution
of current account balances in the world economy, as well
as in six groups of nations (Advanced, Latin America, Asia,
Africa, Middle East and Eastern Europe) are seen for the
period 1971-2004. At almost 9% of GDP, New Zealand’s
deficit is very large from a historical and comparative
perspective. It is in the top decile of deficits distribution for
all advanced countries in the first thirty years of floating.
As the data in Table 4 suggest, at this point New Zealand’s
current account balance looks more like a Latin American or
Asian country, than like an advanced nation.
What sets New Zealand truly apart is the historical
persistence of its large current account deficits. I present
a list of countries with “persistently high” current account
deficits for 1970-2004 in Table 5. In constructing this
table, I define a country as having a “High Deficit” if, in
a particular year, its current account deficit is higher than
its region’s ninth decile.14 I then define a persistently high
deficit country, as a country with a “High Deficit” (as
defined above) for at least 5 consecutive years.15 The list
of persistently high deficit countries is extremely short;
only two of them are advanced countries, one of which is
During the last 30 years a number of advanced countries, in
New Zealand during the 1980s. This illustrates the fact that,
addition to New Zealand, have had current account deficits
historically, periods of high current account imbalances have
in excess of 5% of GDP: Australia, Austria, Denmark, Finland,
tended to be short lived, and have been followed by periods
Greece, Iceland, Ireland, Malta, Norway and Portugal. What
is interesting, however, is that very few advanced countries
14
have had current account deficits in excess of 9%: the
15
158
Notice that the thresholds for defining High deficits are year
and region-specific. That is, for every year there is a different
threshold for each region.
For an econometric analysis of current account deficits
persistence see Edwards (2004). See also Taylor (2002).
Reserve Bank of New Zealand and The Treasury
of current account adjustments. At the end of 2006, it
Table 6
is likely that US will be added to this list. This would be
Net stock of liabilities: New Zealand
quite remarkable, since it would be the first large country
and other industrial countries, selected years
– either advanced or developing – to ever make it into this
(Per cent of GDP)
category. It is important to note, however, that even if in
2006 New Zealand still has a very large deficit, it will still
not be classified as a new “persistently high episode.” The
1980 1985 1990 1995
2000 2004
Australia
27.8
37.0
47.1
56.8
52.2
57.8
reason for this is that it requires five years of being in the
Canada
34.2
34.3 34.9
29.9
7.2
12.5
top 10% of deficits.
Denmark
30.9
52.6
41.6
23.8
14.5
12.4
Finland
14.9
19.7
29.1
41.9
151.6 12.1
Iceland
Table 5
List of countries with persistent high
current account deficits
By region, 1970-2004
Region/ Country
Period
25.5
55.0
48.4
51.6
64.3
92.9
New Zealand 30.3
70.9
62.4
103.3 74.8
91.9
Sweden
8.6
19.2
23.7
36.1
0.6
9.5
United States -3.7
-0.3
4.6
5.5
16.8
22.6
Source: Lane and Milesi-Ferretti (2006)
In Table 6, NIIP positions for a group of advanced countries
Industrialized Countries
that have historically had a large negative NIIP position are
Ireland
1978-1984
New Zealand
1984-1988
Latin America and Caribbean
Guyana
1979-1985
Nicaragua
1984-1990 & 1992-2000
1982-1989
Africa
1982-1993
Lesotho
1995-2000
Middle East
2000-2004
represents a unique case in terms of its external position;
NIIP among advanced countries. Moreover, New Zealand’s
nations.18 As pointed out in the preceding Section, the
level at which the NIIP ratio stabilizes determines – jointly
with other variables, such as the potential or trend rate of
growth, and inflation – the sustainable current account
Eastern Europe
Azerbaijan
that emerges from this Table confirms that New Zealand
NIIP is significantly higher than that of other advanced
Guinea-Bissau
Lebanon
compiled by Lane and Milesi-Ferretti (2006). The picture
together with Iceland, it currently has the largest negative
Asia
Bhutan
seen.17 The data are taken from the comparative data set
1995-1999
deficit. According to equation (1) above, if, for example,
New Zealand’s NIIP stabilizes at 100% of GDP, trend growth
Source: Author’s elaboration based on World Development
Indicators. A persistent large deficit is defined as one
that exceeded the ninth decile for the country’s region
for at least five consecutive years.
The importance of the data on persistence in Table 5 is that
they show that countries that run very large deficits don’t
do that for very long periods of time. Countries that move to
is 3.5% and inflation is 1.5%, the sustainable current
account deficit (SCAD) 5% of GDP, four percentage points
below it 2005 level.
16
17
the “High Deficits” category stay there for short periods of
time. Their external accounts adjust, and then move back to
having a more “normal” deficit. A key question is the nature
of this adjustment. As a number of authors have found out,
countries that go through abrupt and sudden adjustments
tend to experience significant declines in growth.16 On the
other hand, countries that experience a smooth adjustment
do not suffer significant costs in their real economies.
Testing stabilisation policy limits in a small open economy
18
Frankel and Cavallo (2004).
For the US the data are from the Bureau of Economic Analysis.
For the other countries the data are, until 1997, from the
Lane and Milessi-Ferretti data set. I have updated them
using current account balance data. Notice that the updated
Figures should be interpreted with a grain of salt, as I have not
corrected them for valuation effects.
During March-May 2006 international investors began to
question the sustainability of Iceland’s external accounts. This
resulted in a decline in the demand for Iceland securities and
in a drastic loss in value of the currency. The central bank was
forced to face this situation by substantially hiking interest
rates. See, for example, Bloomberg, “Iceland’s Central Bank
Raises Key Rate to 12.25%,” May 18, 2006. Story may be
found in: http://www.bloomberg.com/apps/news?pid=100000
85&sid=as0W.Z2_ykUA&refer=europe.
159
New Zealand’s close economic relation with
Table 8 presents the consolidated NIIP for Australia-
Australia and the external accounts
New Zealand. As may be seen, at 61% of GDP the
An important characteristic of the New Zealand economy
is its (increasingly) close relation to Australia. This is
particularly the case with respect to investment in certain
industries and sectors. For instance, Australian investors
are the predominant owners of New Zealand’s banking
sector. An important consequence of this close relationship
is that it has an impact on the external accounts, and may
make the situation appear more difficult than what it really
is. At the heart of this issue is the treatment in Balance of
Payments accounting of reinvested earnings. These are
automatically (and simultaneously) recorded as an outflow in
the investment income account and an inflow in the capital
account. This means that if firms use retained earnings as a
recurrent source for financing their expansion in the normal
course of their business activity, the external accounts will
reflect a large current account deficit.
As a way to gauging the importance of the “Australian
connection” in explaining the magnitude and evolution
of New Zealand’s current account deficit I analysed the
combined NIIP is still negative and large. It is, however,
significantly smaller than New Zealand’s NIIP (89%).21
Figure 5 presents the evolution of the current account
deficit between New Zealand and Australia, and Figure 6
displays the components of the bilateral current account
deficit between New Zealand and Australia. This suggests
that during 2000-2003 the bilateral deficit with Australia
more than explained the aggregate deficit. Also, Figure 6
shows that the bilateral investment income deficit is the
more important component of the bilateral imbalance
between New Zealand and Australia. The main conclusion
of this “consolidated analysis” is that once the trans-Tasman
relationship is taken into account, New Zealand’s external
imbalances don’t look as large; they are still significant, but
not as large as they appear when the aggregate data are
considered.
Figure 5
Current account deficit between New Zealand
and Australia
consolidated Australia-New Zealand NIIP, as well as the
% GDP
2
% GDP
2
behaviour of New Zealand’s current account deficit with
0
0
-2
-2
Australia’s net holdings of New Zealand assets. Three main
-4
-4
points emerge from this table: first, New Zealand’s NIIP vis-
-6
Australia.19
Table 7 presents New Zealand’s NIIP, explicitly detailing
à-vis Australia is negative and equivalent to 24% of GDP;
-6
Current Account Balance
of which Australia
ex Australia
-8
-8
second, the share of the bilateral NIIP relative to Australia
(as a proportion of total NIIP) doubled in merely four
years; and third, the vast majority of Australia’s holdings
-10
-10
1999
2000
2001
2002
2003
2004
2005
Source: Statistics New Zealand
Reserve Bank of New Zealand calculations
of New Zealand assets are FDI (almost 50%). This fact is
particularly important, as it provides support to the notion
discussed above regarding the long-run and ingrained
relationship between the two countries. In particular, the
predominance of FDI suggests that Australian investments
in New Zealand are unlikely to be subject to moody and
knee-jerk reactions, and/or to sudden stops.20
19
20
160
I am grateful to Anella Munro for discussing with me this
issue and, in particular, for providing me with the calculations
on the Australian-New Zealand external accounts.
Whether that is the case of other investments is less clearcut.
21
Naturally, it is larger than Australia’s NIIP of 57% in 2005.
However, since New Zealand economy is smaller than the
Australian economy, the increase in the combined NIIP
relative to Australia’s is not too large.
Reserve Bank of New Zealand and The Treasury
Table 7
New Zealand’s NIIP: total and Australia
2001
2002
2003
2004
2005
New Zealand investment abroad
Direct Investment Abroad
21,198
17,402
17,507
17,413
18,984
of which Australia
9,243
8,396
8,882
9,020
9,847
%
44%
48%
51%
52%
52%
26,191
28,857
24,882
33,254
35,140
of which Australia
3,058
3,612
2,755
5,844
5,826
%
12%
13%
11%
18%
17%
16,322
22,702
23,425
23,289
27,164
Portfolio Investment Abroad
Other Investment Abroad
of which Australia
3,228
1,856
2,792
3,668
5,104
%
20%
8%
12%
16%
19%
Financial Derivatives
12,476
6,074
6,781
6,081
7,841
Reserve Assets
8,566
7,723
9,115
10,093
8,828
Total New Zealand Investment Abroad
84,753
82,757
81,710
90,130
97,957
of which Australia
15,529
13,864
14,429
18,532
20,777
%
18%
17%
18%
21%
21%
Foreign investment in New Zealand
Direct Investment in New Zealand
56,897
57,967
60,183
72,314
77,223
of which Australia
17,779
17,693
21,084
31,017
35,220
%
31%
31%
35%
43%
46%
60,591
62,326
65,292
73,340
78,432
Portfolio Investment in New Zealand
of which Australia
3,129
3,735
6,582
8,655
9,034
%
5%
6%
10%
12%
12%
46,238
55,367
49,778
47,975
58,238
Other Investment in New Zealand
of which Australia
7,642
11,383
11,152
10,021
11,815
%
17%
21%
22%
21%
20%
Financial Derivatives
8,487
6,111
8,774
8,591
10,186
Total Foreign Investment in New Zealand
172,214
181,770
184,028
202,220
224,078
of which Australia
28,550
32,811
38,818
49,693
56,069
%
17%
18%
21%
25%
25%
-87,461
-99,013
-102,318
-112,090
-126,121
Net International Investment Position
of which Australia
-13,021
-18,947
-24,389
-31,161
-35,292
%
15%
19%
24%
28%
28%
Gross Foreign Assets/GDP
74%
67%
63%
66%
66%
Gross Foreign Liabilities/GDP
150%
147%
143%
147%
152%
Net IIP/GDP
-76%
-80%
-79%
-82%
-86%
11%
14%
14%
(of which Australia)
Gross Foreign Assets/GDP
14%
11%
Gross Foreign Liabilities/GDP
25%
27%
30%
36%
38%
Net IIP/GDP
-11%
-15%
-19%
-23%
-24%
Source: Statistics New Zealand
I thank Anella Munro for providing me these data.
Testing stabilisation policy limits in a small open economy
161
Table 8
Consolidated Australia-New Zealand (ANZ) international investment position
2001
2002
2003
2004
2005
Australia-New Zealand investment abroad
Direct Investment Abroad
220,440
270,315
219,087
255,288
294,943
27,022
26,089
29,966
40,037
45,067
203,957
226,923
189,782
244,270
272,830
6,187
7,347
9,337
14,499
14,860
107,492
113,817
101,424
114,507
115,954
of which internal
10,870
13,239
13,944
13,689
16,919
Financial Derivatives
54,896
35,008
47,478
53,753
52,881
Reserve Assets
51,359
47,870
45,190
65,225
60,063
Total ANZ Investment Abroad
638,145
693,934
602,960
733,041
796,671
44,079
46,675
53,247
68,225
76,846
of which internal
Portfolio Investment Abroad
of which internal
Other Investment Abroad
of which internal
Foreign Investment in Australia-New Zealand
Direct Investment in ANZ
of which internal
Portfolio Investment in ANZ
of which internal
Other Investment in ANZ
305,488
325,311
332,744
380,309
448,940
27,022
26,089
29,966
40,037
45,067
615,606
646,163
576,147
721,061
758,120
6,187
7,347
9,337
14,499
14,860
202,505
201,914
198,142
211,426
222,433
of which internal
10,870
13,239
13,944
13,689
16,919
Financial Derivatives
50,557
35,790
52,308
60,533
53,284
Total Foreign Investment in ANZ
1,174,157
1,209,177
1,159,343
1,373,330
1,482,777
44,079
46,675
53,247
68,225
76,846
Net IIP/GDP
-56%
-50%
-56%
-58%
-61%
Gross Foreign Assets/GDP
67%
68%
61%
66%
71%
Gross Foreign Liabilities/GDP
123%
118%
117%
124%
132%
of which internal
(excl internal)
Net IIP/GDP
-56%
-50%
-56%
-58%
-61%
Gross Foreign Assets/GDP
62%
63%
55%
60%
64%
Gross Foreign Liabilities/GDP
118%
114%
111%
117%
125%
Source: Statistics New Zealand, IMF International Financial Statistics, RBNZ estimates
I thank Anella Munro for providing me these data.
4
Figure 6
Components of bilateral current account
Should New Zealand’s large
external imbalance be a cause
deficit with Australia
for concern?
% GDP
1
% GDP
1
In the preceding Sections I have analysed New Zealand’s
0
0
external conditions. Six aspects stand out from this
-1
-1
analysis.
-2
-2
-3
-3
large current account deficits. According to official
-4
-4
New Zealand data the average deficit for the two first
-5
decades of floating was 4.8% of GDP. The smallest
-6
deficit was 0.7% of GDP in March 1989, and the largest
-5
-6
Current Account Balance
Goods Balance
Transfers Balance
1999
162
2000
2001
Services Balance
Investment Income balance
2002
2003
2004
2005
•
First, New Zealand has historically exhibited very
Reserve Bank of New Zealand and The Treasury
•
was 8.9% of GDP, a level achieved in December 2005.
Given the points made above, it is reasonable to ask whether
According to IMF data the average deficit was somewhat
the current very high deficit of the current account is a
larger, at 5.4% of GDP. But deficits have not only been
cause for concern. A number of authors, most notably Max
large, they have also been persistent. As shown in Table
Corden (1994), have argued that very large current account
5, New Zealand has been one of the few countries in
deficits “don’t matter,” as long as they are the result of
the world that has had “persistently high” deficits.
higher (private sector) investment and not the consequence
Second, at this time New Zealand has one of the highest
current account deficits in the world. In 2005, among
the advanced countries, only Iceland and Portugal had
comparable deficits.22
of higher public sector deficits. This is known as the
“Lawson Doctrine,” or as the “consenting adults” view of
the current account. Since for many years New Zealand has
run significant fiscal surpluses, this view implies that the
large current account deficit of the last few years should not
•
Third, the most important component of New Zealand’s
large current account deficit is the investment income
account. In contrast with the US, until recently
New Zealand’s trade balance was in surplus, and only in
2004 did it turn significantly into deficit.23
•
Fourth, New Zealand’s NIIP is one of the most negative
among advanced nations. In part, this negative NIIP
is attributable to the special relationship between
New Zealand and Australia. However, even when data
for these two countries are consolidated the NIIP is very
high from a comparative perspective.
•
•
in sentiments in capital markets is small, as is the probability
of either a “sudden stop” or an abrupt and costly “current
account reversal.”
An elegant way of empirically addressing the question of
whether large external deficits are worrisome is to investigate
if they are consistent with intertemporal optimizing models
that posit that savings and investment decisions (and thus
the current account) are the result of optimal decisions by
with Australia is very high. During 2001-2003 this
of intertemporal models is that, at the margin, changes
bilateral deficit explained more than 100% of the overall
in national savings should be fully reflected in changes in
current account deficit. The most important component
the current account balance (Obstfeld and Rogoff 1996).
of this bilateral deficit is the investment income account.
Empirically, however, this prediction of the theory has been
This reflects the fact that Australian nationals have very
systematically rejected by the data.25 Typical analyses that
large investments in New Zealand, and is (partially) the
have regressed the current account on savings have found
consequence of the accounting treatment given to
a coefficient of approximately 0.25, significantly below the
retained earnings.
hypothesized value of one. Many numerical simulations
Sixth, most analysts believe that New Zealand’s
to know what the precise sustainable level is, most
studies put it at between 4.5% and 5.5% of GDP.24
This number is approximately 4% of GDP lower than
the current account balance in 2005.
24
means that the likelihood that there will be a sudden change
the private sector. An important and powerful implication
than its 2005 level. Although it is almost impossible
23
what they are doing, and thus are unlikely to overreact. This
Fifth, New Zealand’s bilateral current account deficit
sustainable current account deficit is significantly lower
22
be a cause for concern. According to this view adults know
Recent data suggests that in 2006 Spain will be added to this
group.
This assertion refers to the recent time. During 1999-2000 the
trade balance was slightly negative.
See Munro (2005) for a discussion on alternative estimates for
current account sustainability in New Zealand.
Testing stabilisation policy limits in a small open economy
based on the intertemporal approach have also failed to
account for current account behaviour. According to these
models a country’s optimal response to negative exogenous
shocks is to run very high current account deficits, indeed
much higher than what is observed in reality. Obstfeld and
Rogoff (1996), for example, develop a model of a small
open economy where under a set of plausible parameters
25
See, for example, Ogaki, Ostry and Reinhart (1995), Ghosh
and Ostry (1997), and Nason and Rogers (2006).
163
the steady state trade surplus is equal to 45 per cent of GDP,
and the steady state debt to GDP ratio is equal to 15.
26
The common rejection by the data of the intertemporal (or
Present Value) model of the current account has generated
an intense debate among international economists. Some
have argued that there is a group of “usual suspects” that
explain this outcome (Nason and Rogers 2006); others
have argued that the problem resides on the low power of
traditional statistical tests (Mercereau and Miniane 2004).
In a recent paper using New Zealand quarterly data for
1982-1999, Kim, Hall and Buckle (2006) find that the
implications of the intertemporal, present value model, of
the current account cannot be rejected. More specifically,
they find that there is no evidence of consumption-tilting
towards the present in New Zealand. The authors’ main
conclusions from this research are:
New Zealand’s current account using data for 1982-2005.
Their results support those of Kim, Hall and Buckle (2004),
and indicate that the main implications of the present value
model cannot be rejected. However, these new results
by Munro and Sethi (2006) also suggest that the recent
deterioration of the trade account is not consistent with the
long-term solvency condition. An important implication of
this finding is that New Zealand’s external sector will have
to go through a significant correction.
In this Section I take a somewhat different approach to the
question of whether the large current account deficits in
New Zealand should be a cause for concern. I use a broad
multi country data set to investigate the determinants of the
probability that a country experiences a sudden and large
“current account reversal.” I then use New Zealand data to
evaluate how likely it is that the country will face such a
reversal in the near future. I also analyse the evolution of
“(1) Despite substantial deterioration in New Zealand’s
current account deficits during the late 1990s, its
the estimated probability of a current account reversal in
New Zealand during the 1999-2005 period.27
current account movements over our sample period
as a whole have been consistent with its intertemporal
budget constraint and hence its formal external
solvency condition has been satisfied. (2) The data is
not consistent with consumption-tilting towards the
present. (3) The current account paths predicted by our
intertemporal optimisation models have satisfactorily
reflected the actual directions and turning points for
the consumption smoothing component of the current
account.” (p. 25-26).
The importance of analysing the likely nature of
New Zealand’s future adjustment stems from the fact that
abrupt current account reversals have, historically, been
associated with interest rate spikes, higher inflation, rapid
currency depreciation and, more importantly, a significant
decline in the rate of GDP growth.28 According to Edwards
(2005a), reversals have historically been associated with real
depreciation ranging between 15% and 40%, and interest
rates increases in the 240 to 570 basis points range. In
addition, regression analyses in Edwards (2005b) indicate
These empirical findings led the authors to conclude that
the available evidence suggests that the large deficits are
no cause for concern. The large imbalances were the result
that countries that experience large and abrupt current
account reversals have had, on average, a decline in GDP
per capita growth that ranges from 2.5% to 5.5%.
of optimal decisions, and would revert themselves smoothly
in due course.
The Kim, Hall and Buckle (2006) paper, however, did not
include data for the 2000-2005 period, when the current
27
account deficit widened significantly. In a recent paper
Munro and Sethi (2006) revisit this issue, and provide new
results for the estimation of the present value model of
26
164
Obstfeld and Rogoff (1996) do not claim that this model is
particularly realistic. In fact, they present its implications to
highlight some of the shortcomings of simple intertemporal
models of the current account.
28
The latest IMF reports on New Zealand (IMF 2006a, 2006b)
analyse whether the large current account deficit poses risks
for the country. Although there is no empirical investigation,
the authors of the report review work on reversals. On the
bases of that review the IMF (2006b, p. 11) conclude that “the
current account deficit poses no immediate threat to macro
stability.”
Calvo et al (2004), Edwards (2005b), and Frankel and Cavallo
(2004). See the discussion below for a comparison of GDP
growth in New Zealand during reversal and non-reversal
years.
Reserve Bank of New Zealand and The Treasury
Table 9
As may be seen, during the last 35 years New Zealand
Incidence of current account reversals, 1972-2004
experienced abrupt and significant current account
reversals on four occasions. Only Iceland and Portugal have
Region
No Reversal
Reversal
Industrial countries
94.7
5.3
Latin American and Caribbean
80.3
19.7
Asia
82.1
17.9
Africa
77.2
22.8
than the average growth for the “non-reversal” years at
Middle East
83.5
16.5
1.5%.32 Moreover, in New Zealand, average real GDP per
Eastern Europe
83.9
16.1
capita growth was also negative (-0.26%) one year after
Total
82.8
17.2
the reversals.
experienced as many reversals.31 It is interesting to note that
the average rate of growth of per capita GDP in New Zealand
Observations
during the four reversal years (1975, 1976, 1983 and 1988)
was negative at around -1%. This is significantly lower
3.491
In the regression analysis reported in this Section I focus
Pearson
90.58
on countries with a GDP in 1995 of at least USD 52 billion.
Design-based F(5, 14870) 18.11
This allows me to focus on a group of countries that are
P-value
somewhat homogeneous. However, in the discussion
Uncorrected chi2 (5)
0.000
presented below I also discuss results obtained when a
Data and empirical model
large group of countries is included in the analysis. The
In this study I define a “current account reversal” (CAR)
basic empirical model is a variance component probit, and
episode as a reduction in the current account deficit of at
is given by equations (2) and (3):
least 3% of GDP in a one year period.29 Table 9 presents data
on the incidence of current account reversals for six groups
ρ tj
(2)
=
of countries. As may be seen, for the overall sample the
{
1, if
ρ tj* > 0,
0, otherwise.
incidence of reversals is 17.2%. The incidence of reversals
among the advanced countries is smaller, however, at 5.3%.
ρ tj*
(3)
The advanced countries that have experienced current
account reversals during the period under study are:
Variable
ρ tj
=
αω tj + ε tj .
is a dummy variable that takes a value of one if
country j in period t experienced a current account reversal
•
Austria (1978, 1982),
•
Canada (1982, 2000),
•
Finland (1976, 1977, 1993),
•
Greece (1986),
•
Iceland (1978, 1983, 1986, 1993),
•
Ireland (1975, 1982, 1983),
•
Italy (1975, 1993),
•
New Zealand (1975, 1976, 1983, 1988),
•
Norway (1978, 1980, 1989),
•
Portugal (1982, 1983, 1984, 1985),
•
Switzerland (1981).30
31
29
Later I also discuss results obtained when alternative
definitions of reversals are considered in the probit analysis.
32
(as defined above), and zero if the country in question
did not experience a reversal. According to equation (2),
whether the country experiences a current account reversal
is assumed to be the result of an unobserved latent variable
ρ tj* . In turn, ρ tj* is assumed to depend linearly on vector
ω tj . The error term ε tj is given by a variance component
and
mean and variance
30
Testing stabilisation policy limits in a small open economy
ε tj = ν j + µ tj . ν j is iid with zero mean
2
variance σ ν ; µ tj is normally distributed with zero
model:
σ µ2 = 1 . The data set used covers 44
In the analysis the basic cross-country data were obtained
from the IMF’s International Financial Statistics, and from
the World Bank’s World Development Indicators. The Figures
may be slightly different from national sources’ data. See
Edwards (2005b) for alternative definitions of reversals.
In its recent report on New Zealand the IMF (2006b) analyses
whether the reversal in Finland in 1993 (as well as the milder
adjustment in Sweden) offer lessons for New Zealand.
See Edwards (2004) for a treatment of regression analysis of
the effects of reversals on GDP growth.
165
countries, for the 1970-2004 period; not every country has
positive, reflecting the fact that when a similar country
data for every year, however. See Edwards (2005b) for exact
experiences a “sudden stop,” capital flows to the
data definition and data sources.
country in question will tend to decline increasing the
likelihood of a massive current account correction.35
In addition to the random effects model, I also estimated
fixed effects and basic probit versions of the probit model
(d) Changes in the logarithm of the terms of trade (defined
in equations (2) and (3).33 One of the advantages of relying
as the ratio of export prices to import prices), with a
on a probit model, such as the one described above, is that
one year lag.
they are highly non-linear. More specifically, the marginal
(e) The country’s initial GDP per capita (in logs). This
effects of any independent variable on the probability are
measures the degree of development of the country
conditional on the values of all covariates. This means that
in question. If more advanced countries are less likely
if the value of any of the independent variables changes,
to experience a reversal, its coefficient would be
the marginal effect of any of them on the probability of the
negative.
outcome variable will also change.
In addition to the base estimates with the covariates
In determining the specification of this probit model I
followed the literature on external crises, sudden stops and
reversals. In the basic specification I included the following
discussed above, I estimate a number of regressions that
further include (some combination) of the following
covariates:36
covariates, which have data for a large number of countries
and years:34
(f) The one-year lagged rate of growth of domestic credit.
This is a measure of the monetary policy stance.
(a) The ratio of the current account deficit to GDP, lagged
one period.
(g) A dummy variable that takes the value of one if that
particular country had a flexible exchange-rate regime,
(b) The lagged ratio of the country’s fiscal deficit relative to
and zero otherwise.
GDP.
(h) An index that measures the extent to which the country
(c) An index that measures the effect of “contagion.” This
is dollarized. If countries subject to “original sin,” that
index is measured as the relative occurrence of sudden
is, countries that are unable to borrow in their own
stops in the country’s reference group of counties. It is
currency are more prone to experience current account
calculated, for each year and group, as the proportion
reversals, its coefficient should be positive. The data
of countries that experienced a “sudden stop.” In
for this index were taken from Reinhart, Rogoff and
this calculation data for the country in question are
Savastano (2003).
excluded. In that sense, then, this “contagion” index
measures the relative occurrence of sudden stops in the
(i) An index that measures cases of significant real
exchange-rate appreciation. This index takes the value
country’s immediate reference group. For New Zealand
of one if in a three year period the accumulated real
the reference group is the “advanced countries.” In the
exchange-rate appreciation exceeds 30%.
case of New Zealand, for 1970-2004 the contagion
variable has an average value of 0.064, and a standard
(j) And, an index that takes the value of one if the country
deviation of 0.047. The lowest value of the “contagion”
in question is a “commodity country,” and zero
variable for New Zealand is zero (obtained in several
otherwise.
years) and the highest is 0.19 (1973 and 1995). I expect
the coefficient of this “contagion” variable to be
33
34
166
In the ‘basic probit” estimation, the error term is assumed to
have the standard characteristics.
See, for example, Frankel and Rose (1996), Milesi-Ferretti
and Razin (2000) and Edwards (2002).
35
36
There are six groups. Five of them are strictly regional, while
the sixth refers to “advanced” nations and, thus, covers
more than a region. New Zealand belongs to the “advanced”
countries group.
Most of these variables have a lower number of observations
than those in (a)-(e) above.
Reserve Bank of New Zealand and The Treasury
Table 10
Determinants of current account reversals
Random effects probit regressions
(10.1)
Current-Acc. deficit to GDP
(10.2)
0.177
(8.65)***
Fiscal deficit to GDP
Contagion
Terms of trade change
Initial GDP per capita
(10.3)
(10.4)
(10.5)
0.183
0.174
0.171
(8.27)***
(7.82)***
(6.57)***
(10.6)
0.039
0.002
0.012
0.033
(2.56)***
(0.13)
(0.62)
(1.95)*
1.960
2.408
1.731
2.224
1.956
2.360
(2.74)***
(3.60)***
(2.35)**
(2.78)***
(2.20)**
(2.93)***
-0.012
-0.018
-0.012
-0.011
-0.013
-0.020
(2.27)**
(3.59)***
(2.25)**
(1.93)*
(1.77)*
(3.26)***
-0.053
-0.115
-0.062
-0.014
-0.081
-0.115
(1.02)
(2.09)**
(1.17)
(0.23)
(1.06)
(1.94)*
-0.397
-0.398
-0.264
(2.38)**
(2.18)**
(1.62)
Flexible
Commodity
0.089
(0.45)
Domestic credit growth
Dollarization index
0.0002
0.0001
(1.36)
(1.01)
-0.188
(0.82)
Appreciation
-0.280
(1.15)
Pseudo-R2
0.32
0.33
0.3
0.37
0.39
0.33
Observations
881
822
822
741
599
608
Countries
42
40
40
42
35
36
Absolute value of z statistics is reported in parentheses; all regressors are one-period lagged; constant term is included, but not reported.
*** significant at 1%; ** significant at 5%; * significant at 10%.
Unfortunately, it is not possible to analyse formally the way
Basic results
in which the close relationship between two countries, such
The basic results obtained from the estimation of this
as the one between New Zealand and Australia, affects
probit model for a sample of 44 countries are presented
the probability of a current account reversal. There are
in Table 10. In equations (10.1) and (10.2) the coefficients
no readily available data on cross-country assets holdings
of both the current account deficit and the fiscal deficit
such as that discussed in Section 3 of this paper. However,
are significantly positive, indicating that an increase in
it is possible to perform some indirect tests on the way in
these imbalances increases the probability of the country in
which the trans-Tasman relationship between New Zealand
question experiencing an abrupt current account reversal.
and Australia is likely to affect the probability of a hard
All the other regressors in equations (10.1) and (10.2)
landing or abrupt current account reversal. I do this at the
have the expected signs, and are significantly estimated at
end of Section 4, where I discuss the role of FDI on these
conventional levels. The results confirm the presence of a
probabilities.
“contagion” effect, and that a deterioration in the terms of
trade increases the probability of a reversal. These results
also indicate that counties with a higher (log of) GDP per
capita have a lower probability of a reversal. When these
Testing stabilisation policy limits in a small open economy
167
equations were estimated using a fixed effects procedure,
results reported in Table 10 are robust to specification, time
37
period, country coverage, and the exclusion of “extreme
the results were very similar.
In equations (10.1) and (10.2) the fiscal and current
account deficits variables were introduced separately in the
estimation. In equation (10.3) I present estimates when both
values” of the different variables. I also considered
alternative specifications, and included additional variables
that (potentially) capture the extent of external imbalances.
variables are included in the same probit equation. As may
The results presented in Table 10 consider the current
be seen, in this case the coefficient of the (lagged) current
account deficit as the measure of external imbalances, and
account deficit continues to be positive and significant.
don’t control by the country’s initial NIIP. That is, it makes no
However, the coefficient of the fiscal deficit ceases to
distinction between countries with a large deficit and a very
be statistically significant. This result is rather intuitive:
negative initial NIIP, and one with a very large deficit and
higher fiscal imbalances that are not associated with a
a low initial NIIP. When the value of the initial NIIP to GDP
deterioration of the external accounts, do not affect in a
ratio was included as an additional regressor its coefficient
significant way the probability of an abrupt current account
was negative, as expected, indicating that a more positive
38
Equation (10.4) indicates that countries with a
NIIP would tend to reduce the probability of a current
flexible exchange-rate regime have had a lower probability
account reversal. However, the coefficient for this variable
of experiencing an abrupt and significant current account
was statistically insignificant. Moreover, its inclusion did not
reversal.
affect in any way the analysis on marginal effects reported
reversal.
In equations (10.5) and (10.6) I report estimates with
below.
additional covariates. The results are suggestive and confirm
As an additional robustness test I also considered alternative
that countries with flexible exchange-rates have been less
definitions of “current account” reversals. In particular, I re-
likely to experience an abrupt current account reversal;
estimated the probits when a reversal was defined as being
they also indicate that a more expansive monetary policy
a 4% reduction in the current account deficit in one year.
has had a positive, although statistically marginal, effect
The results obtained (available on request) are very similar to
on the probability of a sudden current account reversal.
those reported here. The main difference is that when this
Interestingly the commodity, appreciation and dollarization
stricter definition is used, the estimated coefficient of the
variables are not significant in the estimation of the current
initial (log of) GDP per capita was significantly negative.
account reversal equations. All the estimated models
presented in Table 10 performed quite well; the pseudo-R2
ranged between 0.3 and 0.4.
Evaluating the effect of a larger external
imbalance on the probability of a major
current account adjustment in New Zealand
Robustness analysis
The results reported above show that larger external
Standard robustness tests were performed, including
imbalances – measured by the (lagged) current account to
estimating the equations for alternative time periods and
GDP ratio – have been associated with a higher probability
alternative data sets (larger number of countries). I also
of experiencing an abrupt (and costly) current account
re-estimated the model excluding outlier observations.
reversal. However, the probit estimated coefficients
Generally speaking, the results obtained suggest that the
reported above are difficult to interpret; it is not possible to
know how the recent rapid growth in the current account
37
38
168
In the fixed-effects estimation I used dummies for the
different regions. In this case (the log of) initial GDP became
insignificant. The reason for this is that the regional dummies
capture income per capita differentials.
The significant positive coefficient of the fiscal deficit in
(10.2) is picking up the effect of the omitted current account
variable.
deficit has affected the probability that New Zealand will
face a current account reversal.
Reserve Bank of New Zealand and The Treasury
In order to address the interpretation issue I report the
Table 11
Current account reversals: marginal effects and
estimated marginal effects (and standard error) computed
predicted probability
Marginal effects
from one of the probit regressions reported above (equation
Variable
(11.1)
(11.2)
“Early 2000”
High Imbalance
Current-Account
deficit to GDP
0.012
(2.98)***
0.050
(3.80)***
Contagion
0.148
0.638
(2.59)**
(2.88)**
Changes in terms
of trade
-0.001
(1.51)
-0.003
(1.78)*
GDP per capita
-0.001
-0.004
(0.23)
(0.23)
-0.038
-0.131
(2.27)**
(2.40)**
0.029
0.208
10.4). The marginal effects are estimated as the derivatives
of the cumulative normal distribution with respect to
the corresponding regressor. These derivatives are then
evaluated for given values of the independent variables. An
important property of probit models is that marginal effects
are highly nonlinear and are conditional on the values of
all covariates. If the value of any of the independent
variables changes, the marginal effect of any of them on
the probability of the outcome variable will also change.
In the exercise reported in this Section I attempt to
answer the following specific question: “At the margin,
Flexible
by how much have increases in the current account
imbalances affected the probability of an external crisis in
New Zealand.” In order to address this issue I follow a two
steps strategy. First, I evaluate the marginal effects at the
values of the covariates that prevailed in New Zealand in
Predicted
Probability
Absolute value of z statistics are reported in parentheses.
*** significant at 1%; ** significant at 5%; * significant at 10%.
the early 2000. In particular, I use a value of the current
account deficit of 2.8% of GDP, which corresponds to the
The results obtained from the computation of marginal
year 2001. (For the other covariates I use the following
effects are presented in Table 11. I present two sets of
values: Contagion=0.01; dlogtt=.03; logGDP0=9.43084;
estimates : “Early 2000” and “High Imbalance.” The first
Flex=1). Second, I re-evaluate the marginal effects using a
column contains the marginal effects obtained when
significantly higher value of the external imbalance. More
equation (10.4) is evaluated using the values of the covariate
specifically, I use a value of the current account deficit of
corresponding to New Zealand in the early 2000s.39 Four
9% of GDP, which corresponds to New Zealand’s deficit in
results stand out from Column 1:
2005-06. In order to focus the analysis on the effects of the
•
external disequilibria, in this second evaluation I maintain
All, but one, of the marginal effects are significant at
conventional levels.
the assumed values of the rest of the covariates.
•
The marginal effect of the current account deficit is
significantly positive. Its point estimate, however, is
rather low. A marginal increase in the deficit from its
initial value of 2.8% of GDP increases the probability of
reversal by only 1.2 per cent.
•
For this specific configuration of values of the key
variables, the marginal effect of the contagion is rather
39
Testing stabilisation policy limits in a small open economy
In these estimates the current account deficit – the variable
of greatest interest – is given a value of 2.8% of GDP; this
corresponds to the current account deficit experienced by
New Zealand in 2001. When alternative specifications of the
probit equation are used to evaluate the marginal effects, the
results are very similar to those discussed here.
169
•
large; the point estimate is 0.15, indicating that an
The results discussed above suggest that, although a higher
increase in sudden stops in similar countries increases
current account deficit increases significantly the marginal
the probability of a reversal crisis by 15 per cent.
probability of a reversal crisis, this is not its main effect;
According to the estimate for “flexible exchange-rate”
a country that, with other things given, moves from a
pegged to a flexible exchange-rate regime reduces its
probability of a crisis by 4.4%.
The marginal effects in the second column of Table 11
also correspond to equation (10.4), but they have been
evaluated for a value of the current account deficit of 9% of
GDP. All other covariates continue to have the same values
as in the first column. The differences between the “High
Imbalance” marginal effects in Column 2 and the “Early
2000” marginal effects in Column 1 are very interesting
and may be summarized as follows:
indeed, its marginal effect is only 5%. From New Zealand’s
point of view, the main consequence of the recent increase
in the current account deficit is a very significant increase in
its degree of vulnerability to contagion.
The discussion presented above has focused on the marginal
effects of changes in the current account deficit on the
probability of experiencing a current account reversal. A
related question, and one that is perhaps more relevant
from a policy point of view in New Zealand, is how the
rapid increase in the current account deficit has affected the
overall predicted probability of an abrupt current account
reversal in New Zealand. This question is addressed in the
last row of Table 11, where I report the predicted probability
•
The marginal effect for the current account deficit is four
times higher in the “High Imbalance” case (Column 2)
than in the “Early 2000” case (Column 1). The point
estimate, however, is still on the low side: 0.050.
•
abrupt current account reversal is significant. It goes from
3% in the “Early 2000” case (a scenario associated with
New Zealand in the early 2000s) to 21% under the “High
estimates has to do with the marginal effect of
Imbalance” scenario.
account deficit is significantly more vulnerable to
contagion than a country with only a 2.8% current
account deficit (other things being the same). The
differences in the marginal effect for contagion in these
two estimates are indeed startling: the point estimate
increases from 0.15 to 0.64. Interestingly, these
marginal effects for contagion are not very sensitive
to the assumed value of the contagion variable itself;
when I repeated this exercise using a value of 0.0 for
contagion, its marginal effect was 0.14 for the “Early
2000” case and 0.63 for the “High Imbalance” case.
“Maxi” current account reversals
The results reported in Tables 10 and 11 are for current
account reversals of at least 3% of GDP. Historically,
however, a number of countries have experienced more
severe adjustments of, say, 5% of GDP in one year. This is
usually the case when the international capital market turns
viciously against a country, forcing it to adjust severely. As
Frankel and Cavallo (2004) and Edwards (2004) have shown,
these more severe reversals are more costly in terms of GDP
collapse. In order to address this issue I estimated random
effect probit equations of the type given by equation (3)
for an alternative and stricter definition of current account
The marginal effect of the “flexible exchange-rate”
reversal of 5% of GDP in one year. The regression results are
variable goes from -0.044 to -0.13. That is, the benefits
in Table 12; the estimated marginal effects and predicted
of adopting a flexible exchange-rate regime are three
probabilities computed from equation (12.1) are presented
times higher for countries with (very) large current
in Table 13, overleaf.
account deficits than for countries with moderate
deficits.
170
be seen, the increase in the predicted probability of an
The most important difference between these two
“contagion.” A country with a 9% of GDP current
•
for the “Early 2000,” and “High Imbalance” cases. As may
As may be seen, qualitatively speaking the probit results
are very similar to those in Table 10 for the 3% definition
Reserve Bank of New Zealand and The Treasury
Table 12
Determinants of current account reversals:
reversal 5%, random effects probit regressions
Current-Account deficit to GDP
(12.1)
(12.2)
(12.3)
0.138
0.147
0.144
(5.41)***
(5.25)***
(5.21)***
-0.010
-0.015
Fiscal deficit to GDP
Contagion
Terms of trade change
Initial GDP per capita
Flexible
(0.53)
(0.70)
3.117
2.917
2.896
(3.53)***
(3.14)***
(3.06)***
-0.009
-0.010
-0.009
(1.36)
(1.43)
(1.35)
-0.116
-0.132
-0.195
(1.41)
(1.57)
(2.17)**
-0.455
-0.506
-0.557
(2.10)**
(2.23)**
(2.44)**
Commodity
0.131
(0.57)
Appreciation
-0.215
(0.76)
Dollarization index
-0.406
(1.54)
Pseudo-R2
0.31
0.31
0.34
Observations
741
694
685
Countries
42
40
39
Absolute value of z statistics is reported in parentheses; all regressors are one-period lagged; constant term is included, but not reported.
*** significant at 1%; ** significant at 5%; * significant at 10%.
of reversals. The signs of the estimated coefficients are the
The role of FDI
same, and virtually the same variables are significant.
An interesting question is whether a large FDI component
The marginal effects and predicted probabilities, however,
present some differences. For every covariate the marginal
effect in Table 13 is substantially lower than in the previous
analysis. As an illustration, under the “High Imbalance” case
the marginal effect of the (lagged) current account deficit is
now a mere 1.4%. From a policy perspective, perhaps the
most important result in Table 13 refers to the predicted
probabilities of a “5% current account reversal,” for a
New Zealand-like country. As may be seen, the predicted
probability in the “Early 2000” scenario is less than one
per cent (0.6%); under the “High Imbalance” scenario the
predicted probability of a “5% current account reversal” is
a mere 5%.
Testing stabilisation policy limits in a small open economy
in capital inflows has an effect on the probability of
experiencing a reversal. This is potentially important since
New Zealand has traditionally had a large, positive and
steady flow of FDI, mostly coming from Australia. For the
complete period, for example, the mean FDI to GDP ratio
for New Zealand was 3.0%, and the standard deviation
was 1.72. For all Advanced Countries the mean was 1.80%
with a standard deviation of 3.0%. When the FDI to GDP
ratio is added to the random effects probit equations, its
estimated coefficient is negative and its p-value is 0.08.40
40
This result is obtained when the FDI to GDP ratio is added to
the specification in equation (10.1). When added to the other
specifications in Tables 10 and 12, the results are similar.
Notice that when this variable is added to the regressions the
number of observations falls by approximately 50%.
171
This suggests that, with other things given, countries with
New Zealand discussed in Section 3 of this paper. As may
a higher flow of FDI will tend to face a lower probability of
be seen in Table 7, the stock of Australian FDI represents
experiencing a current account reversal.
almost 50% of all FDI in New Zealand. Moreover, FDI is
Table 13
Current account reversals:
marginal effects
and predicted probability, reversal 5%
more than 60% of all Australian assets in New Zealand.
Variable
Changes in terms of trade
GDP per capita
Flexible
Predicted Probability
probit analysis suggests that the trans-Tasman connection
(13.1)
(13.2)
“Early
2000”
High
Imbalance
Current-Account deficit to GDP 0.002
Contagion
The centrality of Australian FDI in New Zealand, and the
0.014
will, overall, tend to reduce the probability of New Zealand
facing a hard landing.42
5
Monetary policy, external
(1.65)*
(1.83)*
0.052
0.311
(1.77)*
(2.29)**
-0.0002
-0.001
(1.03)
(1.20)
-0.002
-0.011
(1.43)
(1.58)
-0.013
-0.065
(1.80)*
(1.99)**
according to an econometric analysis of the determinants of
0.006
0.047
current account crises, the recent worsening in the current
imbalances and other policy
issues
Monetary policy
In the preceding Sections I analysed the recent evolution
of New Zealand’s external imbalances, and I showed that
this is a unique case, in several respects. I also showed that,
account balance has increased New Zealand’s external
Absolute value of z statistics are reported in parentheses.
*** significant at 1%; ** significant at 5%; * significant at 10%.
In order to investigate further the role of FDI, I computed the
marginal effects and predicted probability of reversal under
two assumptions for FDI behaviour. The first assumption is
that the “high imbalance,” which as before is assumed to
be characterized by a current account deficit of 9% of GDP,
is fully financed by FDI flows. In the second scenario, none
sector vulnerability and, in particular, the probability of being
subject to contagion. Although at this time the predicted
probability of experiencing an abrupt current account
reversal of at least 3% of GDP is not at an overly critical
level, it is estimated at a quite high 21%. On the other hand,
my estimates indicate that even under with a very large 9%
current account deficit, the predicted probability of a much
of the “high imbalance” is financed by FDI flows.41 The
more severe “5% current account reversal” is only 5%.
results obtained highlight the importance of FDI. When the
In this Section I address briefly an important issue related to
deficit is fully financed with FDI the predicted probability of
the relationship between monetary policy and the external
reversal is 12.1%; when FDI declines to zero, the predicted
sector: The question is whether the RBNZ would benefit
probability increases to 27%. There is also an effect on the
from formally considering exchange-rate developments
marginal contribution of the current account deficit: when
FDI fully finances the imbalance, a marginal increase in the
deficit raises the probability of reversal to 4%; when there
are no FDI flows the marginal effect of the deficit increases
to 6%. These results shed some light on the importance
of the trans-Tasman relationship between Australia and
42
41
172
That is, in the first scenario the predicted probabilities and
marginal effects are evaluated at values of the current account
deficit of 9% and of FDI of 9%; in the second scenario, the
deficit is 9% and the FDI ratio is zero.
On the other hand, given the importance of the “contagion”
variable in this analysis, if Australia herself is subject to a
“sudden stop,” New Zealand is highly likely to go through a
hard landing and an abrupt reversal. Assessing the likelihood
that Australia will experience a sudden stop is beyond the
scope of this paper.
Reserve Bank of New Zealand and The Treasury
when deciding on the level of the OCR? 43 The traditional
be framed explicitly in terms of the Taylor rule in a small
literature on inflation targeting, both theoretical and
open economy. In 2001 Taylor himself posed the problem
applied, has ignored this exchange-rate question. Most of
as follows:
this literature has relied on discussions on how the central
“How should the instruments of monetary policy (the interest
bank should adjust the monetary policy interest rates. The
rate or a monetary aggregate) react to the exchange-rate?”
Taylor rule provides a powerful guidance for addressing this
(Taylor, 2001, p. 263. Emphasis added)
issue. The seminal book by Mike Woodford, Interest and
Prices (2003), which provides firm analytical underpinnings
for interest rate-based monetary policy and discusses a
number of Taylor based rules, does not deal explicitly with
In order to address this question more formally, consider the
following equation:44
(4)
i t = fπ t + gy t + h0 et + h1 et −1
exchange-rates; the index has no entries for “exchange-
Where
rate(s),” “devaluation,” or “pass-through.” There is one entry
bank as a policy tool,
for “open economy,” although no open economy model
inflation from its target level (possibly zero),
is presented, and the discussions on optimal policy rule do
deviation of real GDP from potential real GDP (often called
not consider the (potential) role of open economy variables.
the output gap), and e t is the log of the real exchange-
(To be fair, however, one could interpret the discussion in
rate in year t.45
Section 2.1 of Chapter 7, on cost-push shocks, as including
coefficients;
shocks stemming from exchange-rate depreciation.) The
and lagged log of the real exchange-rates in the expanded
chapter on “Design and Implementation” (Chapter 3) of the
Taylor rule, and are the main interest of this discussion. If
influential and pioneering book by Bernanke et al (1999)
h 0 = h1 = 0 exchange-rate developments should not be
does not discuss at the analytical level whether exchange-
incorporated in the policy rule, and the Taylor rule reverts to
rate considerations should be explicitly incorporated into
its traditional form.
the policy rule in an inflation targeting (IT) setting. In the
chapter on Israel, Australia and Spain the authors discuss
how Spain and Israel gradually relaxed exchange-rate bands
when they adopted IT, and they explain that in both of these
countries the authorities decided “not to respond to short
term exchange-rate fluctuations” when making monetary
policy decisions (Bernanke et al, 1999, page 205). There is
no explicit discussion, however, on whether the authorities
should explicitly and directly consider exchange-rate
developments when setting the policy rate. And yet, this is
an important question for central bankers from around the
world, including in New Zealand.
it
is the short term interest rate used by the central
π t is
the deviation of the rate of
yt
is the
f and g are the traditional Taylor rule
h 0 and h1 are the coefficients of the current
It is conceivable, in principle, that in a small open economy
the optimal monetary policy rule (that is the policy that
maximizes the authorities’ objective function) is one where
h 0 and h1 are different from zero. Interestingly,
h0 > 0 and h 0 = − h1 , then the rule implies that
both
if
monetary policy should react to changes in the (real)
exchange-rate. Notice that the formulation in equation
(4) does not imply, even when
h 0 and h1
are different
from zero, that the monetary authorities should defend
a certain level of the exchange-rate. If the optimal policy
calls for intervention, that is for
h0 and h1 different than
zero, and if the monetary authorities do follow this policy, a
From a technical point of view the discussion of the relation
casual observer may conclude that the country in question
between central bank policy and the exchange-rate may
is subject to “fear of floating.” This, however, would be an
It is not my intention, however, to provide a comprehensive
survey on the topic of central bank intervention. The literature
is voluminous country-specific, and continues to grow every
day; interested readers are directed to, among others,
Dominguez and Frankel (1993), Taylor (2004), Kearns and
Rigobon (2005), Neely (2001), Sarno and Taylor (2001).
For an excellent analysis of different central bank policies,
including Chile’s case, see Tapia and Tokman (2004). On
New Zealand’s policy of RBNZ foreign exchange intervention
see Eckhold and Hunt (2005).
incorrect inference, as the country in question would be
43
Testing stabilisation policy limits in a small open economy
practicing “optimal flotation.”
44
45
This is the precise equation presented by Taylor in his
discussion on the subject.
In this formulation an increase in e denotes a real exchangerate appreciation.
173
Recently McCallum (2006, 2005) has argued that very open
some scepticism on the general merits of adding the
small economies may benefit from replacing Taylor’s rule,
exchange-rate into the interest rate equation. This is for,
either with or without an exchange-rate term, with an
at least, two reasons. First, and as pointed out earlier, in
“exchange-rate rule” of the following form:
properly specified models, the exchange-rate already plays
∆et = ∆q − ∆p t + µ 1 (∆p t − π *) + µ 2 ( y t − y ) + η t
(5)
an indirect
through its effect on π t and
∆erole
+ µ 2y( ty;t second,
− y) +η t
t = ∆q − ∆p t + µ 1 ( ∆p t − π *)
e is the nominal exchange-rate, expressed as foreign
adding the exchange-rate (or any other asset price, for that
currency units per unit of home currency; ∆ q is the average
matter) into the Taylor rule is likely to add considerable
long run rate of real exchange-rate appreciation, and ∆ p t
volatility to monetary policy. This conclusion is similar to
where
is actual inflation. Deviations of inflation from target and
the output gap have been written in an explicit way.
a random term, and
µ1
and
µ2
η
is
are greater than zero. In
this model, when inflation exceeds its target the authorities
manipulate the nominal exchange-rate in order to generate
an appreciation, through purchases (or sales) in the money
that of Mishkin and Schmidt-Hebbel (2001) who provide
an extensive discussion on the subject. According to them,
when implementing policy, central banks should consider
the effects of exchange-rate fluctuations on inflation and
the output gap, but should not consider an independent role
for e t . According to them, “targeting on an exchange-rate
market or foreign exchange market.46 It is important to
is likely to worsen the performance of monetary policy.”
notice that in McCallum’s model, the policy rule in equation
Generally speaking, and unless there is strong evidence
(5) is not intended to maintain the exchange-rate at any
to the contrary, I believe that this general principle that
particular level. It is simply an alternative way of achieving
states the exchange-rate should not have a formal and
the monetary authorities’ objectives. Using a simple
independent role in the Taylor rule also applies to the case
simulation model, and parameter values consistent with
of New Zealand.47 This does not mean, however, that the
the case of Singapore (that is for a very open economy)
monetary authorities should be oblivious of real exchange-
McCallum (2006) argues that the exchange-rate based
rate developments. Indeed, if there is evidence that the
policy rule (5) outperforms a more traditional Taylor-type
exchange-rate becomes significantly out of line with
rule. According to these computations, however, countries
respect to the value dictated by fundamentals it is likely to
with a ratio of exports to GDP similar to that in New Zealand
be welfare enhancing to intervene directly in the foreign
(in the vicinity of 0.3) the interest rate based rule will tend to
exchange market, either buying or selling foreign exchange.
yield a greater degree of macroeconomic stability.
In 2005 New Zealand unveiled a specific foreign exchange
Taylor (2002) reviewed 19 recent models developed to analyse
inflation and monetary issues. Of these, only 5 assumed
that the exchange-rate affected aggregate demand, and
only six assumed that exchange-rate change was a factor
in the process of price determination. This illustrates quite
starkly the fact that many influential researchers continue to
think in terms of closed economy monetary models.
intervention policy.48 According to this framework, the
Reserve Bank of New Zealand will intervene if four criteria
are met: (1) the exchange-rate is exceptionally high or
low; (2) the exchange-rate is unjustified by the level of
fundamentals; (3) intervention should be consistent with
the Policy Targets Agreement (PTA), or inflation target;
and (4) conditions in the market should be opportune and
there should be a reasonable probability of success. These
At the end of the road, whether h0 and h1 should indeed
be different from zero is a country-specific empirical
question, that should be dealt with by analysing country
specific evidence – both historical and based on simulation
intervention principles make eminent sense: while they
acknowledge that on occasion exchange-rate misalignment
may be acute and costly, at the same time they recognize
that intervention in the foreign exchange market should be
exercises. After much reflecting on this subject I find it
difficult to disagree with Taylor (2001) when he expresses
46
174
See McCallum (2005b) for a discussion of these issues from
Japan’s perspective.
47
48
Of course, exchange-rate developments do play an important
indirect role through their impact on inflation and the level of
economic activity.
See Eckhold and Hunt (2005).
Reserve Bank of New Zealand and The Treasury
a rare event, and one that should not be taken lightly. I
Australia.49 For all practical purposes this would imply
don’t see reasons at this time for changing this policy.
adopting the Australian dollar, as it is extremely difficult
to think that Australia will give up its own currency. In my
judgement, the answer to both of these questions is “no”;
Further reflections
I believe that the current policy framework should not be
Although this paper has covered a significant amount of
ground, there are many policy issues related to New Zealand’s
external sector and monetary policy that have not been
subject to major changes, and I believe that New Zealand
should continue to have its own currency. In what follows I
elaborate on these two points.
addressed. In the rest of this Section I present some brief
reflections on two of these issues. These remarks are not
exhaustive; my purpose in presenting them is to raise as
many questions as to provide some tentative answers.
There is significant evidence suggesting that the current
Inflation Targeting framework used by the RBNZ has worked
well. Indeed, New Zealand has been a case that is studied
with care and admiration by central bankers from around the
During the last 3 years there has been a growing perception
among analysts and observers that New Zealand has had
some difficulties with the implementation of monetary
policy. These have been reflected by an increased volatility
of some of the more important macro variables, including
interest rates, exchange-rates and nominal GDP. The rapid
increase in housing prices revealed potentially important
macro imbalances. Heightened volatility during this period
seems to have been higher in New Zealand than in other
commodity-currencies
countries,
including
Australia
and Canada. In addition, there has been a perception
that during the last few years the RBNZ has faced some
difficulties in implementing monetary policy. Increases in
the OCR were not translated into desired changes in longer
term interest rates, and did not seem to have the expected
(and desired) effect on mortgage rates, external imbalances
world. Many of the main characteristics of New Zealand’s
system are considered to be among the most desirable in
countries that adopt an IT system. Having said this, there
are some minor points that deserve attention and further
study. Here by “minor” I don’t mean “unimportant,” I
mean that these issues do not affect the most important
characteristics of the current policy framework. More
specifically, I believe that some idiosyncratic aspects of the
New Zealand economy should be informally considered in
the policy process. By “informally” I mean that these should
be important inputs in the policy making process, perhaps
more important than what they have been until now; at
the same time I mean that they should not be incorporated
formally into a Taylor rule type of equation.50 Some of the
most important aspects that, in my opinion, should be
considered in the undertaking of monetary policy are:
and or aggregate demand. (To be fair, this has not been
unique to New Zealand. In the US, a 400 bps increase in
•
Greater attention should be given to the relationship
between New Zealand’s business cycle and the
the Federal Funds rate has barely been translated into a 20
business cycle in the major economies (e.g. the G-3 and
bps increase in the longer term 10 year rate. Former Fed
Australia). In particular, the increased sensitivity of (short
Chairman Greenspan referred to this as a “conundrum.”)
term) capital flows to interest rate differentials, the soAlthough the macro picture in New Zealand has changed
called “carry trade,” should be taken into account when
somewhat (the NZD has experienced some weakening and
formulating policy.
increases in housing prices have moderated) a number
of important policy questions remain open. The most
important of these questions is whether, in light of the
macro developments of the last few years, there should be
significant changes in the way New Zealand implements
monetary policy. An even more profound question is
whether New Zealand should form a monetary union with
Testing stabilisation policy limits in a small open economy
49
50
The monetary union issue has been analysed by Bjorksten
et al (2004), Drew et al (2003), and Grimes (2005a, 2005b,
2006).
At any rate, it is important to emphasize that the Taylor rule
has always been meant as providing broad guidance to the
policy making process.
175
•
The RBNZ should consider being more “patient” and
6
reducing the number of times when it undertakes
policy action. In that regard, coordinating the number
and dates of policy meetings with those of the most
important foreign central bank makes eminent sense.
“Patience” also means not overreacting to changes in
capital flows in the short run.
•
It is unlikely that a small country such as New Zealand
can successfully adopt the Greenspan view on asset
summary
This paper has dealt with a number of issues related to
New Zealand’s external accounts. I have shown that in a
number of ways New Zealand’s situation is unique in the
world economy. The most important conclusions from the
analysis may be summarized as follows:
•
account deficits.
reason for this is that consumers’ expenditure decisions
•
During the last thirty five years New Zealand has been
one of the few countries with persistently high current
prices, and ignore significant property booms. The main
are significantly more dependent on housing wealth
Concluding remarks and
•
During this period it has also been subject to a number
than in the US. Thus, a rapid increase in housing prices
of adjustments, including some characterized by large
would result in rapid increases in aggregate demand
and rapid current account reversals (1975, 1976, 1983,
and in inflationary pressures.
and 1988).
The increased degree of flexibility in the Inflation
•
The recent levels of the current account deficit are
Targeting framework may have contributed somehow
very large, both from a historical and comparative
to the increased macroeconomic volatility. In particular,
perspective. Indeed, at 9% of GDP, they are larger than
changing the inflation target to 1-3% and adding “over
most estimates of the “sustainable” current account
the medium run” may have given the signal that the
deficit.
RBNZ was becoming weak on inflation-fighting.
•
New Zealand’s
large
negative
Net
International
From a monetary policy point of view, asking whether
Investment Position (NIIP) is currently 90% of GDP.
New Zealand should form a monetary union with Australia
This is a very large Figure, both from a comparative
makes sense. At some level there are good reasons
perspective, as well as when compared with the
for thinking that the answer may be “yes.” After all,
evolution of the NIIP for New Zealand.
New Zealand and Australia seem to satisfy a number of the
•
In contrast with the US the main source of New Zealand’s
so-called Optimal Currency Area (OCA) criteria. However,
current account deficit is not the trade deficit. Indeed,
given the good marks given to New Zealand’s monetary
until recently the trade balance was in surplus. The main
policy, and the fundamentally important role played by
source of New Zealand’s current account deficit is the
flexible exchange-rates in helping accommodate external
investment incomes account.
shocks, the costs of giving up the currency would exceed
the benefits. Many analysts have the view that countries
•
turned into deficit, contributing to the large overall
that belong to a monetary union are not subject to major
current account imbalance.
external crises in the form of “sudden stops” or “current
account reversals.” However, as I show in a recent paper
Having said this, in recent years the trade balance has
•
To an important extent the (very) negative NIIP and
(Edwards, 2006b), historically this has not been the case.
(very) large current account deficit may be explained
Indeed, a number of currency union countries have been
by New Zealand’s very close economic relationship
affected by these types of crises. Moreover, in that paper
with Australia. In particular, the significant presence
I show that many of the costs of these crises, measured as
of Australian FDI in a number of sectors (including
the decline in the rate of growth of GDP, are significantly
banking) explains the large negative investment incomes
larger in currency union countries than in nations that have
account. (Remember that in balance of payments
a currency of their own.
accounting, reinvested earnings of foreign owned
176
Reserve Bank of New Zealand and The Treasury
companies are treated simultaneously as an outflow in
•
•
The evaluation of the “predicted probability” of
the investment incomes account and as an inflow in the
experiencing an abrupt current reversal indicates that
finance account).
the results depend on the magnitude of the reversal
in question. The probability of facing a “3% of GDP”
Once the data are adjusted by the effects of the
reversal has increased to approximately 20%; on the
“Australian (or trans-Tasman) connection,” both the
other hand, the probability of facing a “5% of GDP”
NIIP and the current account look less “threatening.”
reversal as increased to only 5%. In this regard, the
•
However, even after making the “trans-Tasman”
current external imbalances should not be a cause for
adjustment the current account balance appears to be
great concern.
significantly larger than what is sustainable. This implies
that at some point in the future New Zealand will have
to go through an external adjustment process. A key
question is whether this adjustment will be gradual, and
thus costless, or whether it will be abrupt and (very)
costly.
•
An analysis of the framework used by the RBNZ for
conducting monetary policy suggests that this is largely
appropriate. In particular, there appears to be no
compelling reason at this time for including exchangerate developments as an independent factor in the
monetary policy rule.
•
In order to address this issue I estimated a number
of probit models to analyse the determinants of the
probability of facing an abrupt current account reversal.
I evaluated these models using data for New Zealand in
the early 2000s, when the current account deficit was
below 3%, and in 2005-06, when the deficit is 9%.
•
The main result from this analysis is that the rapid
growth in the deficit during the last few years has
(greatly) increased New Zealand’s vulnerability to
“contagion.” It has also increased the advantage of the
country’s current floating exchange-rate regime.
Testing stabilisation policy limits in a small open economy
177
References
Edwards, S (2004a), “Financial Openness, Sudden Stops and
Björksten, N, Karagedikli, Plantier, C, and A Grimes (2004),
Current Account Reversals,” American Economic Review,
“What Does the Taylor Rule Say About a New Zealand–
94(2): 59-64.
Australia Currency Union?,” Economic Record, 80: 34-42.
Edwards, S (2005), “Is the US Current Account Deficit
Bernanke, B S, Laubach, Th Mishkin, F S, and A S Posen
Sustainable? And If Not, How Costly is Adjustment Likely
(1999), Inflation Targeting: Lessons from the International
to Be?,” Brookings Papers on Economic Activity, 0(1): 211-
Experience, Princeton, Princeton University Press.
271.
Calvo, G A, Izquierdo, A and L F Mejia (2004), “On the
Edwards, S (2005a), “The End of Large Current Account
Empirics of Sudden Stops: The Relevance of Balance-Sheet
Deficits, 1970-2002: Are There Lessons for the United
Effects,” NBER Working Paper No. 10520, May.
States?,” NBER Working Paper No. 11669, October.
Claus, I and G Scobie. (2002) “Saving in New Zealand:
Edwards, S (2005b), “Capital Controls, Sudden Stops and
Measurement and Trends,” New Zealand Treasury Working
Current Account Reversals,” NBER Working Paper No.
Paper 02/02.
11170, March.
Corden, W M 1994. Economic Policy, Exchange-rates, and
Edwards, S (2006), “The Relationship between Exchange-
the International System. Oxford: Oxford University Press,
rates and Inflation Targeting Revisited,” NBER Working
and Chicago: The University of Chicago Press.
Paper No. 12163, April.
Domínguez, K and J A Frankel (1993), “Does Foreign-
Frankel, J A, and A K Rose (1996), “Currency Crashes in
Exchange Intervention Matter? The Portfolio Effect.”
Emerging Markets: An Empirical Treatment,” Journal of
American Economic Review, 83(5): 1356-69.
International Economics, 41(3-4): 351-366.
Drage, D Munro, A and C Sleeman (2005), “An Update on
Frankel, J A and E A Cavallo (2004), “Does Openness to
Eurokiwi and Uridashi Bonds,” Reserve Bank of New Zealand
Trade Make Countries More Vulnerable to Sudden Stops, Or
Bulletin, 68(3): 28-38.
Less? Using Gravity to Establish Causality,” NBER Working
Drew, A Hall, V B McDermott, J and R St. Clair (2003),
Paper No. 10957, December.
“Would Adopting the Australian Dollar Provide Superior
Ghosh, A R and J D Ostry (1995), “The Current Account in
Monetary Policy in New Zealand?,” Reserve Bank of
Developing Countries: A Perspective from the Consumption-
New Zealand Discussion Paper, DP2001/03.
Smoothing Approach.” World Bank Economic Review, 9(2):
Eckhold, K and C Hunt (2005), “The Reserve Bank’s New
305-333.
Foreign Exchange Intervention Policy,” Reserve Bank of
Grimes, A (2005a), “Intra & Inter-Regional Industry Shocks:
New Zealand Bulletin, 68(1): 12-22.
A New Metric with Application to Australasian Currency
Edwards, S (2002), “Does the Current Account Matter?” in
Union,” Motu Working Paper 2005-03.
Preventing Currency Crises in Emerging Markets, S Edwards
Grimes, A (2005b), “Regional and industry cycles in
and J A Frankel (editors), The University of Chicago Press.
Australasia: Implications for a Common Currency,” Journal
21-69.
of Asian Economics, forthcoming.
Edwards, S (2004), “Thirty Years of Current Account
Grimes, A (2006), “Are ‘Regional’ Shocks Truly Regional?
Imbalances, Current Account Reversals and Sudden Stops,”
Aggregate, Regional and Industry Shocks in Australasia,”
IMF Staff Papers, Vol. 61, Special Issue: 1-49. International
mimeo, Motu Economic & Public Policy Research.
Monetary Fund.
178
Reserve Bank of New Zealand and The Treasury
International Monetary Fund (2004)a, “New Zealand:
Munro, A (2005), “New Zealand’s Foreign Liabilities and
Selected Issues,” IMF Country Report 04/127.
Current Account Deficit: Sustainability Assessment,”
International Monetary Fund (2004b), “New Zealand Article
mimeo.
IV Consultation and Staff Report,” IMF Country Report
Munro, A and R Sethi (2006), “Modelling New Zealand’s
04/128.
Current Account“, mimeo.
International Monetary Fund (2006a), “New Zealand Article
Nason, J M and J H Rogers (2006), “The Present Value
IV Consultation and Staff Report,” IMF Country Report
Model of the Current Account has been Rejected: Round
06/160.
up the Usual Suspects,” Journal of International Economics,
International Monetary Fund (2006b), “New Zealand
Selected Issues,” IMF Country Report 06/161.
Kearns, J and R Rigobon (2002), “Identifying the Efficacy of
Central Bank Interventions: The Australian Case,” Journal of
International Economics, 66(1): 31-48.
Kim, K, V B Hall, and R A Buckle (2006), “Consumption-
68(1): 159-187.
Neely, Ch J (2001), “The Practice of Central Bank
Intervention: Looking under the Hood.” Federal Reserve
Bank of St. Louis Review, 83(3): 1-10.
Obstfeld M and K Rogoff (1996), Foundations of
International Macroeconomics. MIT Press.
Smoothing in a Small, Cyclically Volatile Open Economy:
Ogaki, M Ostry, J D and C M Reinhart (1995), “Saving
Evidence from New Zealand,” Journal of International
Behaviour in Low and Middle Income Developing Countries:
Money and Finance, forthcoming.
A Comparison” IMF Working Paper: 95/3, January.
Lane, P R and G M Milesi-Ferretti (2006), “The External
Reinhart, C K Rogoff and M Savastano (2003), “Addicted to
Wealth of Nations Mark II: Revised and Extended Estimates
Dollars,” NBER Working Paper No. 10015, October.
of Foreign Assets and Liabilities, 1970–2004,” IIIS Discussion
Paper No. 126.
McCallum, B T (2005), “A Monetary Policy Rule For
Automatic Prevention of a Liquidity Trap? NBER Working
Paper No. 11056, January.
Robinson, M Scobie G M and B Hallinan (2006),
“Affordability of Housing: Concepts, Measurement and
Evidence,” Working Paper 06/03, New Zealand Treasury.
Sarno, L and M P Taylor (2001), “Official Intervention in the
Foreign Exchange Market: Is It Effective and, If So, How
McCallum, B T (2006), “Singapore’s Exchange-rate-
Does It Work?,” Journal of Economic Literature, 39(3): 839-
Centered Monetary Policy Regime and Its Relevance for
68.
China,” MAS Staff Paper No. 43, March.
Tapia, M and A Tokman (2004), “Effects of Foreign Exchange
Mercereau, B and J Miniane (2004), “Challenging the
Intervention under Public Information: The Chilean Case,”
Empirical Evidence from Present Value Models of the
Central Bank of Chile Working Papers N° 255.
Current Account,” IMF Working Papers 04/106.
Taylor, J B (2001), The Role of the Exchange-rate in
Milesi-Ferretti, G M and A Razin (2000), “Current Account
Monetary-Policy Rules,” American Economic Review, 91(2):
Reversals and Currency Crises: Empirical Regularities” in P.
263-267.
Krugman (Ed), Currency Crises, U. of Chicago Press.
Taylor, J B (2002), “The Monetary Transmission Mechanism
Mishkin, F S and K Schmidt-Hebbel (2001), “One Decade of
and the Evaluation of Monetary Policy Rules,” in Loayza
Inflation Targeting in the World: What do we Know? What
N and K Schmidt-Hebbel (eds.): Monetary Policy: Rules
do we Need to Know?,” NBER Working Paper No. 8397,
and Transmission Mechanisms, Series on Central Banking,
July.
Analysis, and Economic Policies, vol. 4. Santiago: Central
Bank of Chile.
Testing stabilisation policy limits in a small open economy
179
Taylor, M P (2004), “Is Official Exchange-rate Intervention
Effective?,” Economica, 71(281): 1-11.
Taylor, A M (2002), “A Century of Current Account
Dynamics,” NBER Working Paper No. 8927, May.
Woodford, M (2003), Interest and Prices, Princeton and
Oxford: Princeton University Press.
180
Reserve Bank of New Zealand and The Treasury
External imbalances in New Zealand
by Sebastian Edwards
Discussion by William R. Cline1 Institute for International Economics
It is a pleasure to be in New Zealand once again and to
stabilize it must equal the marginal ratio. The marginal ratio
comment on the paper by Sebastian Edwards. I thank the
has the change in net external liabilities in the numerator and
Reserve Bank of New Zealand in particular for the invitation
the change in nominal GDP in the denominator. Abstracting
to participate in this important conference. I will first
from valuation changes, the change in net foreign liabilities
address the paper, then consider some additional data and
is the current account deficit. The change in nominal GDP
analytical questions of my own, and then conclude with
in the denominator equals is GDP multiplied by the nominal
policy implications.
growth rate. Dividing both the numerator and denominator
by GDP yields the ratio of the current account deficit as a
The paper
This is another fine Edwards paper. It exemplifies the author’s
usual cogency, empirical painstaking and econometric
dexterity. I agree with the main thrust of the paper, so
my comments will focus on differences of emphasis and
interpretation. In short, the fundamental message of the
per cent of GDP to the nominal growth rate of GDP. With
real GDP growing at about 3.5 per cent and inflation at
1.5 per cent, nominal growth is about 5 per cent. If the
net external liability ratio is to be held to 100 per cent, the
current account deficit (CAD) will need to fall from 9 per
cent of GDP to 5 per cent.
paper is that New Zealand’s large current account deficit
The paper implicitly takes this diagnosis as axiomatic, but
and resulting further increase in already exceptionally high
it is important to recognize that many economists tend
net international liabilities constitute a problem. I agree.
to dismiss the notion of any serious limits on the external
Edwards conducts tests that nonetheless reassure him that
current account deficit or net foreign liabilities. Indeed, the
when the adjustment comes it will be benign. I am not so
Lawson Doctrine holds that so long as the fiscal accounts
sure. As a corollary, he implies that nothing special should
are not in deficit, any external deficit is no cause for concern
be done about the external deficit, which by implication
because it is the result of private market decisions among
will take care of itself; most importantly, nothing should be
“consenting adults.” So it is crucial to recognize up front
done that would endanger the nicely working machinery
that, in sharp contrast to that doctrine, this paper implicitly
of New Zealand’s macro-policy regime centred on inflation
argues that there is some ceiling level on the ratio of net
targeting. I am skeptical of this implied do-nothing
international liabilities to GDP that should not be exceeded.
position, although I recognize that finding feasible tools for
By implication, for economic policy to ignore this ceiling is
constructive action is difficult.
to store up trouble for the future.
One of the most important findings in the paper is Edwards’
Let me make this point more starkly in a fashion that might
forthright recognition of the implications of the simple
be used by a devil’s advocate questioning New Zealand’s
arithmetic of sustainable external debt. The basic premise
remarkable success. Net international liabilities have grown
is that the ratio of net international liabilities to GDP cannot
from 30 per cent of GDP in 1980 to about 60 per cent in 1990,
keep rising without limit. If this ratio is to stabilize at 100 per
75 per cent in 2000, and 92 per cent in 2004. An economist
cent of GDP (up from about 90 per cent now), for example,
visiting from Mars could conclude that New Zealand’s
then the current account deficit as a per cent of GDP cannot
strong record of sustained growth and control of inflation
exceed the nominal growth rate of GDP. This is essentially
has been based on “other people’s money,” and that some
an accounting identity because for the average debt ratio to
other model will need to be found for continued success
in the future because there is a limit to such money. This
1
Center for Global Development and Institute for International
Economics.
Testing stabilisation policy limits in a small open economy
visitor could also make the point that the traditional benign
181
form of a large foreign deficit – use of foreign resources
causal chain from external credit cutoff to a surge in interest
for investment in tradables capable of future debt service
rates, collapse in domestic demand and hence collapse in
– seems to have eroded in recent years, as the rise in the
imports. But it is far from clear that this sequence typically
current account deficit has substantially exceeded the rise
applies to industrial countries even in CAR episodes. Indeed,
in private investment – much of which in turn increasingly
in New Zealand’s most recent CAR, when the current
2
account deficit narrowed by 3.7 per cent of GDP from 2000
More fundamentally, with an extremely large deficit on
to 2001, there was positive growth of 3 per cent rather than
capital services income already, the capital inflow covering
a recession. This favourable external adjustment reflected
the current account deficit is mostly not available to use for
the lagged response to an extremely low real exchange-rate
real investment because it is earmarked to leave the country
in 2000.
seems to have gone into housing rather than tradables.
again immediately upon arrival to pay for income payments.
In other words, large current account deficits comprised
almost wholly of net income payments cannot be benign
manifestations of future growth from present real capital
stock buildup; they are inherently manifestations of the bill
coming due on either such buildups in the past (preferably),
or on past consumption (less favourably).
A parallel question is whether the model really deals with
the differences between industrial and developing countries.
For example, in their cross-country study of debt crises,
Reinhart, Rogoff, and Savastano (2003) found that whereas
the debt to GDP ratio was strongly significant and had the
right sign for developing countries, it did not even have the
right sign for industrial countries. In the tests in the Edwards
The danger of Sebastian’s paper is that some may read his
paper, the 44 countries include many developing countries,
key empirical findings as a diagnosis that there is nothing
and the great majority of the CAR events are in developing
to worry about and that no special changes in policy need
countries. In particular, I wonder whether the high sensitivity
be considered. This is because his focus is on whether the
to contagion in the current variable values for New Zealand is
adjustment when it comes will be a hard landing involving
really appropriate for an industrial as opposed to developing
the type of severe recession in the face of a cutoff of foreign
country. Similarly, it seems quite possible that a larger share
capital witnessed in many international episodes of the
of the CAR events for industrial countries will have been
Sudden Stop. His econometric means for answering this
the consequence of a sharp domestic policy tightening to
question is a cross-country logit model explaining Current
deal with inflation rather than the result of a shut-down in
Account Reversals (CAR), defined as a reduction in a large
availability of foreign capital. Of course, it could be argued
current account deficit by 3 per cent of GDP in one year. The
that if there is insufficient differentiation between industrial
central finding is that even with a current account deficit
and developing countries in the model, there is even less to
of 9 per cent of GDP, the probability of New Zealand’s
worry about for New Zealand than identified in the paper’s
experiencing a CAR is only 21 per cent. So the paper
21 per cent probability, because industrial countries are
concludes the risk of an ugly rather than benign external
more resilient. More fundamentally, however, I will argue
adjustment is relatively low.
below that delaying external adjustment can impose a
Questions can be raised about this framework. Ironically,
a CAR could be compatible with growth stimulus rather
than recession. It is, after all, a surge in exports and decline
welfare loss from undue burden imposed on the populace
in the future relative to the present, even if the eventual
adjustment avoids recession.
in imports, which in the first instance boosts real output.
The paper raises the important question of whether the
The stylized fact of recessionary Sudden Stops hinges on a
vulnerability to the current account deficit is overstated
2
182
Gross fixed capital formation rose by 1.8 per cent of GDP
from the 1991-2001 average (22.0 per cent) to 2005 (23.8
per cent), whereas the current account deficit rose from an
average of 5.0 per cent of GDP to 8.8 per cent. IMF (2006b).
The question of whether a collapse in household saving has
occurred is discussed below.
because of the large role of reinvested earnings of foreign
firms in the capital services deficit. The accounting
convention of treating these earnings as an income outflow
in the current account matched by a capital inflow in the
Reserve Bank of New Zealand and The Treasury
capital account (now “financial” account in IMF parlance)
policy and higher interest rates might curb the housing
may overstate the economic reality of the income payments
boom and thereby at least partially restore private saving
burden because of semiautomatic reinvestment rather than
rates. However, this discussion raises a key point that I will
repatriation. In 2005, New Zealand’s capital services balance
return to below: policymakers do not really know the sign
was -6.8 per cent of GDP. Of this amount, reinvested
of the current account change in response to an interest
earnings by foreign direct investors contributed -2.5 per
rate change. From the standpoint of consumption and
cent of GDP (Statistics New Zealand, 2006). There may thus
investment, a rise in the rate will curb demand and hence
be some merit to the notion that the deficit looks more
reduce imports, raising the current account outcome.
dangerous than it really is. Even so, completely removing
However, from the standpoint of the exchange-rate, a rise
reinvested direct investment earnings would only reduce the
in the interest rate will likely raise the exchange-rate, making
current account deficit from 9 per cent of GDP to 6.5 per
exports less competitive and encouraging imports.
cent, still large by industrial country standards. The paper
addresses this issue indirectly, using the argument as a
motivation for aggregating the net international investment
positions of Australia and New Zealand. When it does so, it
finds the aggregate NIIP is -61 per cent of combined GDP.
This is broadly consistent with the more direct estimate
here, that removing retained earnings from capital income
payments cuts the capital services income deficit by about
one third, or about the same proportion as the shrinkage
from New Zealand’s NIIP relative to GDP to that for the two
countries combined.
A crucial question in diagnosing New Zealand’s rising current
account deficit is whether it has been driven by a collapse
in household saving. Edwards believes that it has. His Figure
3, which requires his own estimates after 2001 for lack of
direct data, shows a plunge in household saving from an
average of 1.2 per cent of GDP in 1984-92 to an average of
-5.5 per cent of GDP in 2003-05, for a downswing of 6.7
per cent of GDP. But the national accounts do not show
a corresponding surge in household consumption, which
actually declined from an average of 59.7 per cent of GDP in
the first period to 58.9 per cent in the second (IMF, 2006b).
In contrast, for the United States, the two series fit each
other like a glove.3 So the household saving data in Figure
3 may be suspect.
In discussing the recent literature, the paper reports the
notion of a shift in the international demand function
for a given country’s assets (the Gamma function in the
paper). This framework makes me uneasy, as it lends itself
to a tautological dismissal of any problem on grounds that
there is a permanent upward shift in foreign demand for
the country’s assets and hence in the sustainable ratio of
net foreign liabilities to GDP. For example, it is difficult to
believe that there was such a shift that was the driving force
in the surge in New Zealand’s current account deficit from
2.8 per cent of GDP in 2001 to 9 per cent in 2005.
Similarly, the paper’s reference to work identifying
New Zealand’s experience as a case of rational consumption
smoothing indicates that whereas earlier work concluded
that large imbalances were optimal consumption smoothing
and would revert themselves, more recent work while
confirming these findings for the past, finds that the recent
deterioration of the trade account violates the long-term
solvency condition and thus that a significant correction
will be needed. This is reassuring about the methodology,
because common sense strongly suggests that today’s
imbalance of 9 per cent of GDP at a time of high rather
than low terms of trade is unlikely to represent optimal
consumption smoothing. If we believed Edwards’ estimate
of household saving at -5.5 per cent of GDP, it would be an
After emphasizing the role of the remarkably large dissaving
open and shut case that today’s consumption is excessive
by households in New Zealand’s external imbalance, the
and is robbing consumption from the future, constituting
paper raises the question of whether tighter monetary
consumption
roughening
rather
than
consumption
smoothing. However, as noted earlier, the national accounts
3
For the same two periods, US personal saving in the national
accounts fell from 5.8 per cent of GDP to 0.8 per cent, while
household consumption rose from 65.5 per cent of GDP to
70.1 per cent (BEA, 2006, and IMF, 2006b).
Testing stabilisation policy limits in a small open economy
series for household consumption shows it no higher as
a share of GDP today than the average for the past two
183
decades, so that data make it more difficult to judge that
Further issues and evidence
consumption roughening is occuring. In any event it may
Let me turn, then, to some informal analysis of my own.
be more appropriate to speak of absorption roughening,
One important fact to recognize is that New Zealand is by
because whether the resources are consumed or placed
no means alone as an industrial country that has slid further
into investment, it seems highly likely that the external
into external deficit and net liabilities over the past dozen
deficit will have to be cut back. Under these circumstances,
years or so. The case of the United States is well known,
blessing the deficit with the benign diagnosis that it is a
but there are four other such economies of note: Australia,
case of consumption smoothing would seem misguided,
Portugal, Spain, and the UK. This year Portugal’s current
especially if the Edwards data on saving are correct.
account deficit as a per cent of GDP will slightly exceed even
Or again, the literature discussion on the Taylor rule and
that of New Zealand, according to IMF projections.
McCallum’s extension incorporating the exchange-rate
Figure 1
seems highly inappropriate for New Zealand today. That
Current account balance as a per cent of GDP
formulation essentially seeks to use the exchange-rate as an
extra tool to curb inflation. But the last thing New Zealand
needs now is an even stronger exchange-rate as a means of
holding prices down. Argentina tried that approach and it
ended badly. This particular Inflation Targeting (IT) treatment
2
%
%
2
0
0
-2
-2
-4
-4
of the exchange-rate considers solely its impact on inflation.
-6
This is a worrisome sign that in IT regimes policymakers
-8
may risk ignoring obvious potential problems – such as an
-10
ever escalating and already high foreign liability position
-12
Australia
Spain
UK
USA
New Zealand
Portugal
-6
-8
-10
-12
1991
1993
1995
1997
1999
2001
2003
2005
– because of the primacy of inflation in the macroeconomic
Source: IMF (2006a)
policy targets.
Although the US deficit has become by far the largest in
In sum, as always Sebastian Edwards has made an important
absolute terms, the aggregate deficit of the other five in
contribution in this paper. It is particularly important,
this group of 6 current account deficit countries has also
however, that his strategic message on the unsustainability
reached a size of global significance, at about $250 billion
of the large current account deficit not be drowned out
annually (Figure 2).
by his tactical message that the eventual adjustment for
Figure 2
New Zealand is unlikely to be malign and of the recessionary
Sudden Stop variety. It would have been nice to see in the
paper some suggestions as to what New Zealand authorities
can do to achieve an earlier rather than later external
adjustment, in part to reduce the probability that the
adjustment turns out malign. But Edwards’ main concern
seems to be that changing the IT regime and the other
features of recent macro policy, including the fiscal surplus,
could jeopardize the overall performance of the economy,
levying too high a price for a move toward earlier external
adjustment.
Current account balance of the United States
and five other industrial countries ($ billions)
US$billion
US$billion
100
100
0
0
-100
-100
-200
-200
-300
-300
-400
-400
-500
USA
-500
CAD5
-600
-600
-700
-700
-800
-800
-900
-900
-1000
-1000
1991
1993
1995
1997
1999
2001
2003
2005
2007
Source: IMF (2006a)
184
Reserve Bank of New Zealand and The Treasury
Unlike the United States, New Zealand’s rising external
Canada; Figure 5). This difference is clearer when the six
deficit has not been accompanied by a rising fiscal deficit.
countries weighted aggregate growth rates are compared
Instead, there has been a rising structural fiscal surplus, a
with those of the CAS5 (Figure 6).
phenomenon also present in Australia and Spain (Figure
3).
Real GDP growth in major industrial
Figure 3
countries (per cent)
Structural fiscal balance as a per cent of GDP
6
Figure 5
7
%
%
6
Spain
USA
Portugal
Australia
UK
New Zealand
4
4
2
2
0
0
-2
-2
-4
-4
-6
-6
-8
-8
1991
1993
1995
1997
1999
2001
2003
2005
%
%
7
6
6
5
5
4
4
3
3
2
2
1
1
0
0
Australia
UK
New Zealand
CAS5
-1
-2
-3
1991
1993
1995
1997
Spain
USA
Portugal
-1
-2
-3
1999
2001
2003
2005
Source: IMF (2006a)
Source: IMF (2006a)
Figure 6
Note, however, that New Zealand’s structural fiscal surplus
Weighted aggregate GDP growth rates of
will have dropped by about 1 per cent of GDP from 2004 to
2006 and is projected to ease further next year.
6 CA deficit and 5 CA surplus industrial
economies (per cent)
In several of the group of 6 current account deficit industrial
countries (CAD6) rising investment has contributed to the
widening external deficit, especially in Spain, Australia, and
New Zealand (Figure 4).
Figure 4
Gross fixed investment as a per cent of GDP
%
%
30
30
25
25
%
20
15
Australia
Spain
UK
USA
New Zealand
Portugal
10
5
4.0
1995
1997
4.0
3.5
3.5
3.0
3.0
2.5
2.5
2.0
2.0
1.5
1.5
1.0
1.0
0.5
0.5
0.0
0.0
-0.5
-0.5
1993
1995
1997
1999
2001
2003
2005
Source: IMF (2006a)
So there is some truth to the argument that the industrial
10
countries that have been running growing current account
deficits are drawing resources from the rest of the world
because they are investing more and growing faster than the
0
1993
4.5
15
5
0
1991
5.0
CAD6
CAS5
4.5
1991
20
%
5.0
1999
2001
2003
2005
2007
Source: IMF (2006a)
Economic growth has also tended to be faster in this set
of industrial countries than in the 5 other largest industrial
countries running current account surpluses or not in
significant deficit (CAS5: Japan, Germany, France, Italy, and
Testing stabilisation policy limits in a small open economy
industrial economies that have instead been net suppliers
of capital, most notably Japan and Germany. Especially in
the case of the United States and arguably for New Zealand
as well, however, it is increasingly questionable to justify
widening external deficits on this developmental argument,
namely that these countries are like developing countries
when it comes to their phase in the international debt
185
cycle. One reason is that the magnitudes are increasingly of
account deficit has typically been nearly fully attributable to
questionable sustainability, especially for New Zealand with
a capital services deficit (Figure 8).
net international liabilities already at -90 per cent of GDP.
Another is that in both the United States and New Zealand it
has been more a collapse of household saving than a surge
in investment that has driven the rising current account
Figure 7
Rates of return on New Zealand’s external
assets and liabilities (per cent)
%
%
deficits.
12
Turning to the case of New Zealand, there is an important
10
disadvantageous feature of the current account and
8
8
6
6
4
4
the opposite pattern from that in the United States, and it
2
2
means that New Zealand’s capital services deficit is larger
0
net foreign liability trajectories that warrants emphasis.
12
Assets
Liabilities
10
New Zealand has shown a higher rate of return on its external
liabilities than on its external assets (Figure 7). This is exactly
0
1991
relative to GDP than would be expected given its net external
1993
1995
1997
1999
2001
2003
liabilities.4 In a sense, the disadvantageous differential rate
Source: IMF (2006b)
of return means that New Zealand’s net external liabilities
Figure 8
are larger, in terms of economic burden, than its net liabilities
Capital services balance (right) and net
in accounting terms. Conversely, for the United States the
international investment position (left) as a
higher return on direct investment abroad than on foreign
per cent of GDP
direct investment in the United States has kept the capital
services balance in surplus until last year despite the plunge
of the net international investment position to a deficit of
22 per cent of GDP by end-2004 (see Cline, 2005). The
adverse rate of return differential for New Zealand is one
reason that, as Edwards emphasizes in his paper, the current
%GDP
%GDP
-60
0
NIIP (LHS)
-65
-1
KSV (RHS)
-2
-70
-3
-75
-4
-80
-5
-85
-6
-90
4
186
Note that, in contrast to the differential return effect, valuation
effects appear to be neutral. At first glance they seem negative
also, because in dollar terms the rise in net international
liabilities in recent years has substantially exceeded the
cumulative current account deficit. It turns out, however,
that this is an optical illusion caused by the strengthening
New Zealand dollar. In terms of the kiwi dollar, the change in
the NIIP from 2000 to 2005 was almost the same as the sum
of the current account deficit over that period. The paradox
can be understood by thinking about the path of the dollar
NIIP if the current account were strictly balanced, there were
no external assets, and all external liabilities were expressed
in New Zealand dollars. Then the large appreciation of the
currency through end-2005 would have caused a large rise in
net international liabilities expressed in foreign dollars but no
change expressed in New Zealand dollars. (I am indebted to
Aaron Drew for pointing out that the valuation erosion in the
NIIP was fully explained by the exchange-rate change.) At
the same time, the close tracking of the NIIP with the current
account in New Zealand dollars does raise the question of
whether the official statistical estimates are capturing price
valuation effects. If New Zealand assets held by foreigners
have been rising in price, then even in New Zealand dollars
the net international liabilities should have risen by more
than the cumulative current account deficit.
-7
-95
-8
1992
1994
1996
1998
2000
2002
2004
Source: IMF (2006b)
These are the trends and the facts. What about the
influence of monetary policy on the external account? I
find it problematical that we really do not know even the
sign of the current account deficit on a change in the key
monetary policy instrument, the interest rate. Consider
the contrast between the traditional elasticities approach
and the absorption approach to the balance of payments.
In the former, the driving influence is the real exchangerate, which in turn is influenced by the interest rate. In
what Paul Krugman has called the “Massachussets Avenue
model,” there are two trade equations and one exchangerate equation in a simple system. Exports are a function of
Reserve Bank of New Zealand and The Treasury
foreign GDP and the real exchange-rate lagged; imports
absorption from the future to the present. If we accept
are a function of domestic GDP and the real exchange-
Edwards’ estimate of large household dissaving, the
rate lagged; and the real exchange-rate is a function of the
external deficits are transferring consumption from the
differential between the domestic and the foreign interest
future to excess consumption at present. They may or may
rate. Thus, in the first set of equations in Figure 9, a rise in
not lead, in addition, to a hard landing for the economy.
the domestic interest rate will translate into a widening of
The fiscal surplus helps reduce the intertemporal distortion
the trade deficit as a consequence of a stronger exchange-
to consumption, but is insufficient to remove it. And the
rate.
structural fiscal surplus is declining. It makes no sense for
Figure 9
The impact of monetary policy on the current
account balance
New Zealand to be running a large current account deficit
when the terms of trade are abnormally high. There is
procyclical borrowing, leading to consumption (or at least
absorption) roughening instead of smoothing over time.
Despite New Zealand’s past success with inflation targeting,
completely ignoring the current account deficit and external
debt as a policy issue implicitly assumes there will be an
early reversal in the CAD back toward lower levels. This can
by no means be taken for granted. The high interest rate
policy currently being pursued under inflation targeting
aggravates the external sector problem by causing an
At the same time, pursuing the absorption approach, a
overly strong New Zealand dollar. The strong Kiwi dollar
rise in the interest rate will contract domestic demand by
encourages imports and discourages exports. The policy
reducing interest-sensitive consumption (such as housing)
remedy will likely involve a weaker exchange-rate and
and investment, and the resulting downward pressure on
maintenance or an increase in the fiscal surplus so long as
demand will reduce imports, resulting in a reduction in the
private dissaving is so large.
trade deficit. So unless we know the relative magnitudes
of the key parameters, it will be uncertain whether
tighter monetary policy will result in a smaller or larger
current account deficit. Most likely the exchange-rate and
elasticities effects will dominate, and a higher interest rate
will aggravate the current account deficit. This in fact is the
classic reason why macroeconomic policy assignment tends
to use fiscal policy rather than monetary policy to curb
demand in dealing with an external deficit. Fiscal restraint
has positive (reinforcing) feedback, because tighter fiscal
policy reduces the interest rate and hence the exchange-rate.
Monetary restraint has negative (undermining) feedback
because although it curbs demand, it also worsens relative
trade prices by bidding up the exchange-rate.
Although the exchange-rate has weakened from its peak
last year, it likely has a considerable ways further to go for
consistency with reducing the current account deficit to
about 5 per cent of GDP. With a base of 2000 = 100, the
IMF’s real effective exchange-rate (deflating by consumer
prices) for New Zealand stood at 139.4 in the fourth quarter
of 2005. In contrast, it was an average of 113 during the
period 1992-2003 when the current account deficit averaged
4.6 per cent of GDP (IMF, 2006b). From the fourth quarter
of 2005 to mid-2006 the currency fell about 10 per cent
against the US dollar. On this basis, it has probably gone
only about half way toward a level consistent with a current
account deficit of 5 per cent of GDP, and the distance to
go may be further after taking account of the higher level
of net external liabilities (and hence capital income deficit)
Policy implications
today than before.
New Zealand’s large current account deficit and external
A weaker exchange-rate could be encouraged by a lower
debt are a problem, in my view. They are transferring
interest rate, but the interest rate appropriately is reserved
Testing stabilisation policy limits in a small open economy
187
for addressing inflation. Jawboning the exchange-rate is
to think about how tax policy can moderate such inflows,
one option the authorities already appear to have used with
as a means of helping moderate continued escalation of net
some success in the recent warnings about risk to holders
international liabilities in a price-based and hence market-
of Uridashi bonds (held by foreigners but denominated in
friendly way. Although foreign investors in principle can
New Zealand dollars). A more active role on exchange-rate
typically take credits against such withholding, it is unlikely
intervention might be considered. The Reserve Bank of
that such offsets would be so easy and universal that the
New Zealand could amend its guidelines for its exchange
effect of a higher withholding rate would disappear.
market intervention, and make it clear that it may engage
in sterilized intervention to help prevent appreciation of
the exchange-rate and/or to facilitate depreciation in
circumstances of persistent large current account deficits.
Sterilized intervention would preserve the principle of
reserving interest rate policy for price stability. It would at
least send a policy signal, and could be effective in affecting
the exchange-rate and/ or exchange-rate expectations
under certain circumstances. New Zealand has a low level
of international reserves, and intervention on the side of
avoiding appreciation or facilitating depreciation would
provide an opportunity to build up reserves.
Finally, New Zealand’s economic policymakers could usefully
seek to arrive at some consensus about the ceiling net
international liabilities relative to GDP they consider safe,
and begin to integrate a serious intention of staying within
this limit into their overall economic policies. It is implausible
that net liabilities should be allowed to rise indefinitely in
the name of sole reliance on inflation targeting as the
macroeconomic framework. For most countries a 100%
of GDP level for net international liabilities would be risky.
It is probably safe for New Zealand, but it would seem
dangerous for policy makers to sit idly by if the ratio begins
to rise much beyond this level.
Private dissaving reflects the surge in property values, as
households see no need to save out of current income when
their assets are rising without doing so. The resulting swing
into large negative saving seems to have been a major factor
driving the external imbalance. It might be thought that an
increase in the interest rate is an appropriate response for
moderating the housing price boom. However, a rise in the
References
BEA (2006), Bureau of Economic Analysis, National Income
and Product Accounts Tables. (Washington: Department of
Commerce). Available at: www.bea.gov.
interest rate likely would aggravate the current account
Cline, William R (2005), The United States as a Debtor
deficit problem by boosting the exchange-rate, as more
Nation. (Washington: Institute for International Economics
capital enters in response to the higher rate.
and Center for Global Development).
Consideration might be given to micro instruments for
IMF (2006a), International Monetary Fund. World Economic
curbing property value inflation. These could include higher
Outlook Database. April.
thresholds required for down payments on mortgages as
well as changes in zoning regulations that currently restrict
IMF (2006b), International Monetary Fund. International
Financial Statistics. CD-Rom.
the supply of suburban land for housing.
Reinhard, Carmen M, Kenneth S.Rogoff, and Miguel A
Another area of possible micro action concerns the carry
trade in Uridashi bonds. Some consideration might be given
Savastano (2003), Debt Intolerance. Brookings Papers on
Economic Activity (Spring) 1: 1-74.
to increasing the withholding tax on capital income on nonresident holdings of these and other New Zealand financial
instruments. The current withholding tax of 10 per cent on
Statistics New Zealand (2006), Balance of Payments and
International Investment Position: December 2005 Quarter.
interest paid to residents of most countries with bilateral tax
treaties is low and reflects a design intended to maximize
inflows of capital. In the future it may be necessary instead
188
Reserve Bank of New Zealand and The Treasury
Comments on the macroeconomic policy forum
Val Koromzay, OECD
I have found this forum to be immensely stimulating: both
with globalization, monetary policy has lost control over the
the presentations and the discussions have been very
yield curve beyond the very short run.
thought-provoking, and I would like to thank the Reserve
Bank and The Treasury for inviting me to participate.
This is clearly an issue to watch, but I would argue that the
apparent failure of hikes in the overnight cash rate to push
This conference has addressed two sets of issues, which I
up rates farther out the yield curve is not a systemic change,
would summarise as follows:
but reflects rather the impact of a coincident global liquidity
(1) Is volatility in the New Zealand economy greater than
it needs to be? If so, what can policy do to reduce it?
Which policies? How?
shock (one, to be sure, that is still not fully understood) that
was simultaneously driving down long rates world-wide,
and compressing risk margins as well. I would expect that,
as this shock unwinds, the impact of the OCR on the yield
(2) How worried should the New Zealand authorities be
curve in New Zealand will also normalize.
about the capital account surplus? What risks does it
pose? What actions could be taken to reduce risks?
Second, I would argue that the main reason that inflation
targeting is now in such high repute internationally is not
A third question which, to my mind, deserves to be included
with the two above, was not a focus in this Conference.
This is to ask what policies might do to reduce the costs
associated with volatility (or indeed the costs of adjusting
to a drying-up of capital inflows.) Essentially I have in mind
policies (largely of a structural nature) that can assure a
maximum of resilience to the real economy. But perhaps
just that it seems to be effective in anchoring inflation
expectations, but that, in the process, it also dampens
volatility in the real economy. This is true, in theory, in the
case of demand shocks. I think it is also true, in practice if
not in theory, in case of a broad range of supply shocks. The
smooth absorption of the oil shock over the past two years
is a case in point.
this is for another conference.
This consideration re-enforces, to my mind, the argument
I won’t seek, in these comments, to give my answers
systematically to these questions, but rather to focus on
selected issues that have come up during the discussions
where there may still be something left to say.
put by Willem Buiter that, among the tasks assigned to
monetary policy, inflation control should have lexicographic
priority over other tasks. If in general a strong anchor
for inflation expectations is stabilizing for output, then
Regarding the overall macro framework, it is important
there is clearly no exploitable trade-off between these
to stress, as others have done, that the New Zealand
two objectives. Of course, one cannot ask everything of
framework stands out in international comparison as a very
monetary policy. If it is to stabilize both prices and (to a
good one indeed. In asking whether it could be adjusted to
considerable extent) output it cannot be held accountable
reduce volatility (in particular volatility of the exchange-rate)
also for the composition of output as between tradables
it is important to keep in mind that such adjustment could
and non-tradables or, for that matter, housing construction
come at the expense of losing what New Zealand presently
and widget production.
has. So caution is appropriate.
Specifically, as regards monetary policy, I would make three
comments:
This leads to my third point, intervention.
I am not sure why this issue has become one that
generates almost ideological battles among proponents
First, we had an interesting discussion, triggered by Stephen
and opponents. Foreign-exchange intervention is hardly
Grenville’s paper, as to whether monetary policy has in fact
a mortal sin, but it has a bad reputation among many
become impotent (or at least much less potent) because,
economists, perhaps because too much is claimed for
it by its proponents. The evidence is that intervention,
Testing stabilisation policy limits in a small open economy
189
particularly uncoordinated intervention (and with whom
financial innovation. Perhaps there are useful changes that
would the RBNZ coordinate?) is a weak and uncertain
could be made in New Zealand to further strengthen the
instrument at best. The risk is, however, that if a central
resilience of the financial sector and reduce system risk. If
bank puts intervention into its central tool kit, markets
it happened that such changes had the effect of reducing
(and the business community) will start to hold the bank
pro-cyclical swings in credit availability, this might be a good
responsible for the exchange-rate – a charge which is
thing. But I would argue that supervision is too important
neither desirable nor achievable. This would, at a minimum,
a matter to subject it to secondary, so called “macro-
greatly complicate the implementation of a clear, effective
financial” considerations. The supervision framework needs
and credible communication strategy. Perhaps the present
to be pretty single-mindedly dedicated to limiting financial
New Zealand approach, which allows intervention in
system risks, while promoting financial development and
principle, but only under a set of restrictions that make it
innovation. From this perspective, adjusting prudential rules
very unlikely in practice, represents a reasonable balance in
or norms to the state of the business cycle would be a bad
that it would not appear likely to allow market sentiment
idea. At the OECD we flirted with this concept in one of our
to start holding the bank accountable for exchange-rate
reviews of an EU-member country: the idea was that maybe
outcomes.
regulation could provide some kind of ersatz monetary
What about Stephen Grenville’s point that there seems
to be a market failure, with insufficient arbitrage across
the commodity cycle leading to excessive amplitude in
the associated exchange-rate cycle? The trouble with
such “regularities” in financial markets is that just when
the evidence for them appears compelling, they tend to
disappear (for obvious reasons). But if one wanted to exploit
this regularity, I would see no great harm in assigning a role
to public debt managers: a rule, for instance, that some
policy in a situation where euro-area wide monetary policy
was out of line with local requirements. The (winning)
counter-argument was that such attempted fine-tuning
would just push financial intermediation across the border.
This argument may have somewhat less apparent force
in New Zealand, but the risk is surely here as well that an
overactive regulatory policy not clearly based on prudential
principles would stifle financial development, and this would
not be good for longer term growth.
fraction of new debt issues would be made in foreign
What about fiscal policy? In principle, (or at least economic
currency if the real effective exchange-rate is, say, 10 per
theory) a fiscal stance that leaned more strongly against
cent or more above its long term trend (with such positions
fluctuations in the terms of trade (or, what is in New Zealand
unwound if it is 10 per cent below) would make money
much the same thing, the business cycle) could limit the
if Stephen is right, and not drag the Reserve Bank into
amplitude of the exchange-rate cycle. If markets recognized
complications it should rather avoid.
that a substantial part of the income gains associated with
If I see little more that monetary policy can or should do to
dampen the exchange-rate, what other policies could be
considered? A few words about regulatory policies, and a
few more about fiscal policy.
rising terms of trade would be captured by the budget (and
conversely for terms-of-trade declines), market pressures
or the exchange-rate should, in theory, be correspondingly
decoupled from terms-of trade changes. This is the whole
point, for example, of the Norwegian fiscal rule (all oil-
As regards financial supervision and regulation, my view
is that the current New Zealand approach, which deemphasizes formal rules in favour of requiring financial
institutions to demonstrate clearly to the supervisors the
adequacy of their own risk-management strategies and
practices, is the right one and should be maintained. Of
revenues go into the pension fund, and only the notional
returns on the fund go to the budget.). Since by now the
fund is quite large relative to the annual flows into it (even
at current oil prices) this rule, if credible, should provide
fairly strong decoupling. The Chilean copper rule has a
similar purpose. But in these two cases, the link between
course financial supervision is an unending game; practices
need to be constantly reviewed and revised to deal with
190
Reserve Bank of New Zealand and The Treasury
the relevant commodity price and government revenues is
ex ante fiscal rule on how the budget will deal with
direct, powerful and relatively easy to calculate.
revenue windfalls/shortfalls. As noted above, the link
For New Zealand, the problem is that while the impact
of commodity prices is relatively important, it relates to a
somewhat diffuse bundle, and the link to the budget is
essentially through profit taxes. Further, these are themselves
subject to lots of other shocks, and so even identifying a
terms-of-trade component in the budget may be difficult.
between terms-of-trade and budget revenues is not all
that tight in New Zealand, but surely the correlation is
positive. Budget projections based on “normal” terms
of trade would thus tend to underestimate (respectively,
overestimate) revenues in the face of terms-of-trade
shocks. Telling the markets exactly how the associated
revenue surprises would feed into spending, tax or debt-
Could fiscal policy nonetheless be made more strongly
countercyclical? I offer three points for consideration:
(1) An activist, discretionary fiscal policy is not to be
recommended. The international evidence suggests to
me that this is little better, given the various lags, than a
crap-shoot. And the political economy of this approach
is awful. New Zealand has a world-class approach to
fiscal responsibility that has yielded admirable results
overall. Don’t undermine it.
(2) Could the “automatic fiscal stabilizers” be strengthened?
Big automatic stabilisers are arguably a good thing (only,
arguably, because of political-economy considerations,
on which more below.) But on this point I disagree
with Willem Buiter. The only real way to buy bigger
automatic stabilizers is to raise tax and spending shares
in GDP; that is, to opt for bigger government. While
the size of government is a basic social choice (and the
empirical evidence linking size of government to overall
economic performance is less clear-cut than popular
debate suggests,) it would be bizarre to make decisions
on something so fundamental and long lasting as size
of government on the basis of something as ephemeral
as short-run stabilisation properties. This point becomes
even more compelling once it is recognized that
political-economy constraints often mean that, beyond
fairly narrow limits, ‘cyclical’ surpluses or deficits will
in fact be neutralized through discretionary choice (by
spending ministries when surpluses get too big; by the
finance minister when deficits look too large.)
adjustment decisions could strengthen the stabilisation
properties of the budget to some extent. (The more
so if the ex-ante rule adopted emphasized debt draw
down, or build-up).
I turn briefly (very briefly because I don’t have any bright
ideas) to the second theme of this conference: external
vulnerability. Clearly, national saving in New Zealand is
very low. National investment is not particularly high in
international comparison (and surely not higher than it
needs to be if New Zealand is to continue catching up to
OECD’s highest-income countries.)
The outcome is a large structural current-account deficit that
is needed to close the gap between low domestic saving
and moderate domestic investment. Is this a problem? I
confess to belonging to the “consenting adults” school of
thought on this issue. The public sector is not borrowing
to consume; if the private sector chooses to finance its
investments by borrowing rather than saving – well, nobody
raises an eyebrow if an individual borrows 80% (or for that
matter 95%) of the purchase price of a house; why should a
country be different? But obviously this is too simple. There
is still a powerful, if diminishing, Feldstein-Horioka effect;
and markets are sensitive to it. Current-accounts remain, in
the medium-term, powerful statistical indicators of country
risk (and hence exchange-rate risk.) And it is probably too
benign to assume, even for an advanced post-industrial
economy like New Zealand, that it is exempt from a
“brutal adjustment” scenario if conventional “sustainability
indicators” are breached.
(3) All that said, it does seem to me that something could
be done within New Zealand’s present fiscal framework
to reduce market uncertainty and perhaps moderate
exchange-rate fluctuations. What I have in mind is an
Testing stabilisation policy limits in a small open economy
191
What to do about it?
(1) I would reject out of hand measures to make foreign
credit more expensive (or more difficult to come
by.) I simply cannot see any real medicinal value in,
for example, putting a distortionary tax on foreign
borrowing. It would simply stimulate unhealthy (because
distorted) financial innovation.
(4) Finally, to end this presentation on a more positive note,
some things could surely be done to shift the national
balance away from housing investment towards
business investment. This would, inter alia, generate
more future income to service foreign debt. As in almost
all OECD countries, tax preferences provided to homepurchase exist in New Zealand but serve no obvious
social purpose insofar as they basically generate rents
(2) Should the fiscal stance be tighter than otherwise
(higher public saving) to compensate for low private
saving? Analytically, the answer is almost surely yes.
(Ricardian equivalence is nowhere 100%, though it is
non-negligible in most countries except, interestingly,
the United States.) But here one runs into political
economy. How large a surplus can a democratic political
system sustain on an argument as abstract as “we need
it to maintain national saving because you (the voters)
aren’t saving enough”? The US experience from the
Clinton/Bush transition is telling: if one administration
tries to save, it provides its successor with the wonderful
opportunity to win friends by cutting taxes!
(3) The question then becomes, what can be done on the
tax side to encourage more private saving? I wish I had
a better answer here. The international evidence is that
tax rules can powerfully influence the composition of
private savings, but not the aggregate (very much.) Of
course, the international evidence doesn’t cover many
(indeed any) radical changes in the tax regime.
to current owners and are rapidly “capitalized out” of
the market through higher land prices, thus leaving
new buyers no better off. Reducing such incentives,
and indeed increasing property taxes, would both be
efficiency-enhancing. The political-economy of such
moves is daunting but not insuperable: some OECD
countries have managed to reduce tax distortions in
favour of housing by phasing-in such changes over a
fairly long horizon and ‘grandfathering’ existing rents
by making new tax laws apply only to new purchases.
But I have no illusions: the required changes are not
electoral winners and would require extra-ordinary
political leadership.
To conclude, just one word on how New Zealand needs to
think about minimizing the potential costs of high volatility
in the exchange-rate. This requires micro-economic policies
that maximize flexibility on product and labour markets. To
the extent that New Zealand is more exposed to macroeconomic volatility than most other OECD countries,
New Zealand cannot afford regulations and institutions that
For instance, moving the tax system entirely from an
generate only “OECD average” results. Substantially more
income base to a consumption base (e.g. by exemption
labour market flexibility than the “OECD average,” and
of all income placed in investment vehicles from income
stronger than average competition on product markets is
taxation while raising the VAT to be revenue-neutral ex
required. This is the situation at present, and this differential
ante) should logically, and even analytically, improve
needs to be maintained
.
both private saving and overall tax efficiency. But the
politics of shifting the burden of taxation from income
to consumption seem to be extremely difficult. Indeed,
it is striking that the new government in Canada opted
to anchor its successful electoral campaign on cutting
the consumption tax – economic efficiency be damned!
As I noted above, half-way measures, such as increasing
the tax value of certain instruments have almost no
impact on aggregate private saving.
192
Reserve Bank of New Zealand and The Treasury
Comments on the macroeconomic policy forum
Steven Dunaway, IMF
With the papers presented, the discussants’ remarks, and
Thus, at the end of the day, there is a lot to lead you to
the participants’ questions and comments, this conference
believe that the economic situation that New Zealand has
has certainly done an excellent job of addressing its key
recently experienced is, to a significant degree, a classic case
issue: “whether external balance and smaller swings in the
of a small economy out of synch with cyclical developments
exchange-rate can be achieved while maintaining and/or
in the rest of the world. Complicating the situation of course
enhancing overall prospects.” It is a rather difficult task to
was the set of unique circumstances in the world economy.
try to find something very insightful that will meaningfully
Nonetheless, it is good to ask the question whether the
add to the discussion. That being the case, I will just focus on
situation represents something else and, if this is indeed the
some of the more practical aspects of the issue, reflecting
case, to explore how similar situations in the future should
on the discussion today.
be dealt with. I have been involved with New Zealand on
I think that Governor Bollard in his opening remarks framed
the key question well when he asked whether New Zealand’s
recent experience was a unique set of circumstances. Recent
experience with very loose monetary policy in the major
developed economies of the world (the United States, Euro
and off over the last fifteen years, and one of things that I
have always admired is the unrelenting push by this country’s
macroeconomic policymakers to ensure that they are on
the cutting edge of policy formulation and implementation.
This conference epitomizes why they truly are.
Area, and Japan) does not seem likely to be repeated in
One development that may have not received enough
the near future (although a needed fiscal correction in the
attention
United States could produce somewhat similar conditions
implementation, particularly the way monetary policy has
with a looser monetary policy offsetting the effects of a
been implemented, appears to have changed the behaviour
fiscal contraction; but the prospects of that happening in
of economic agents, especially households, and affected the
the near term are probably pretty slim). The savings glut
transmission of monetary policy changes. Stephen Grenville
in Asia that has fed world liquidity and held world interest
referred a bit to this in his presentation. Household behaviour
rates down also is unlikely to last. To a significant extent
has been influence by the way monetary policy changes have
it has been fueled by undervalued exchange-rates, and
been implemented in steps. As a result, households have
adjustments in these rates will be forced eventually one way
demonstrated flexibility by moving along the yield curve
or the other. Hopefully this will happen in a constructive
and positioning themselves to mitigate the effects of rising
way, with increased exchange-rate flexibility in the region,
interest rates by shifting to longer-term fixed mortgages
which is an objective we at the IMF are working hard to
when short-term interest rates were rising. Firms too have
achieve.
behaved similarly and used available market instruments
Moreover, the discussion today has focused on developments
in New Zealand as being largely driven by domestic demand,
especially through the housing market. However, there
probably was a significant “push” element behind the large
capital inflows New Zealand has received. With interest rates
higher in New Zealand and growing worldwide recognition
of the basic soundness of New Zealand’s economy and
macroeconomic policy management, there was probably
an element of a shift in portfolio preferences in favour of
New Zealand dollar assets. This was probably a one-off
today
is
how
macroeconomic
policy
to hedge interest rate risk. They also have adapted well to
an environment of increased exchange-rate variability by
hedging their exchange-rate risk for longer periods into the
future. As a result, the influence of monetary policy on the
economy is diluted and delayed. But these methods only
delay, not avoid, the impact of monetary policy changes.
As a result, interest rate movements tend to be larger and
when hedging cover begins to roll off the impact on the
economy is more sudden and stronger. The lesson for
policymakers is that they will have to be nimble; prepared to
factor that drove capital flows into New Zealand.
Testing stabilisation policy limits in a small open economy
193
quickly shift gears from a tightening to a loosening stance
policy can still produce results consistent with aiding
when conditions dictate.
monetary policy in stabilizing economic activity in the short
In these changed circumstances, it is understandable to
look at other policies to help support monetary policy. But
in doing so, you always have to ask what is the cost of doing
this and is it really feasible, particularly in a political economy
term. In the end, I guess, what is always needed is a solid
framework for formulating fiscal policy, but it is essential to
have a prudent fiscal authority implementing that policy,
which fortunately New Zealand has been blessed with.
sense. In the many instances discussed today of fiscal policy
Finally, let me note that financial regulation and supervision
initiatives that could support monetary policy, the key
can also play an important role, especially in helping to
element boils down to how such discretion in fiscal policy
avoid problems or at least limit excesses. I know this type of
might be used to the greatest benefit. That is, how can it
use of regulatory authority tends to be frowned on by some
be made more independent and free of political decisions.
in the profession. They suggest that it is not appropriate to
While such considerations are attractive, at the end of the
use regulatory authority in an asymmetric way—that is, to
day, political reality intrudes. Will political authorities give
tighten rules during economic upturns to avoid excesses,
up control over some aspects of fiscal policy? To a large
but not to loosen them during a downward phase of the
extent, perhaps we should be grateful that the political
business cycle to try to boost the economy, since these rules
authorities in many countries have provided monetary
are intended for prudential reasons. Again, I think practical
policy independence, and we should not push our luck. In
considerations should win out in this debate. Regulatory
the end, maybe the best that can be hoped for is that fiscal
guidance can and has been successfully used to prevent
policy is implemented in a relatively stable and predictable
excesses in the financial sector during economic expansions.
fashion, providing a useful environment in which to operate
The best example of this is in the United States during the
monetary policy.
last business cycle. Throughout the end of the 1990s,
If this is the case, then the best thing to do might be to focus
on broad fiscal policy rules, perhaps some countercyclical
rule of the type that Klaus Schmidt-Hebbel has talked
about. Such rules should basically be designed to ground
fiscal policy appropriately over the medium term and help it
to complement monetary policy actions in the short term.
However, it may be possible to achieve this result without
a formal rule. The approach to budgetary targeting used
by Canada could provide an example. It was described by
the opposition party in Canada as the Minister of Finance
“hiding resources from his spendthrift friends in his own
party.” What it boiled down to in practical terms was a
the Federal Reserve and the Office of the Comptroller
of the Currency frequently issued guidance letters to the
commercial banks warning them about properly factoring
in prospective economic conditions into their lending
standards and practices. The result was that during the
recession of 2000-01, no major financial institution in the
United States encountered substantial financial difficulties,
despite significant losses by some institutions owing to the
major corporate frauds that occurred. This was a sharp
contrast with the situation during the recession in the early
1990s when a few major institutions encountered difficulties
which potentially threatened the US banking system.
prudent, conservative approach to budgeting that tended
to underpredict revenue and run surpluses in good times.
It involved relatively conservative economic projections
being matched to a careful assessment of the “longterm” relationship between revenue and income. And a
somewhat similar prudent, conservative approach in budget
formulation has been followed in New Zealand, with the
same basic results. So, perhaps a formal rule is not needed.
Of its own accord, a rational, medium-term oriented fiscal
194
Reserve Bank of New Zealand and The Treasury
Comments on the macroeconomic policy forum
John McDermott, Victoria University of Wellington
Thank you for the opportunity to speak today. My
effect of commodity price and terms of trade changes are
understanding is that the panel’s function is to offer its
larger than would otherwise be the case.
reflections on today presentations. There was a vast amount
of material presented today; enough for policymakers to
reflect on well after our overseas guests have departed. It
would be impossible to tackle it all, so instead I will focus
on four issues: the importance of commodity prices, foreign
exchange intervention, prudential instruments, and the
current account.
Steven Grenville also discounted the importance of
commodity prices by noting that when commodity
prices move, they move a long way but the movements
are temporary. However, the empirical evidence on the
persistence of commodity price movement does not support
this conclusion. Formal econometric evidence suggests there
is unit root in commodity prices implying that the impacts of
We tell everybody when they arrive that New Zealand is
commodity prices are permanent. At the very least the half-
a big farm and it is a nice place to visit. However, in his
life of any shock is very long, in the order of years rather
paper Willem Buiter suggests that we should be suspicious
than months.1
about the overall importance of the primary sector for
economic performance since agriculture makes up only a
small proportion of production. Despite his suspicion about
the importance of the farming sector I hope he agrees that
it is a nice place to visit.
In addressing the issues of whether intervening in the
foreign exchange market is a sensible and useful policy it
is important to consider the structure of the New Zealand
economy. New Zealand is a small open economy whose
commodity exports make up around 50 per cent of its
Looking at the size of the primary sector is an inadequate
exports of goods and services, so there seems good reason
means to determine the sectors relative importance. First
to suppose that commodity prices would be an important
of all a large portion of New Zealand manufacturing is
determinant of the exchange rate. Again this is an empirical
based on food processing. Second and more importantly,
issue.
what matters is how shocks to the primary sector, typically
changes to commodity prices, are transmitted to the real
economy. The issue of whether commodity prices are
important is an empirical one.
Chen and Rogoff (2003) have reported strong evidence
that the New Zealand dollar, along with the Australian
dollar and Canada dollar, is a commodity currency. That is,
commodity prices have a strong and stable influence over
Borkin (2006) has found that the terms of trade impacts
the real exchange rate. The long swings observable in the
significantly on economic growth in New Zealand.
real exchange rate are fundamentally driven by commodity
Specifically, he finds that economic growth is positively
prices and there seems no room for profitable foreign
related to the growth rate of export prices (but not import
exchange intervention.
price growth) and is negatively related to the volatility of
import prices (but not the volatility of export prices). The
transmission mechanism from terms of trade shocks to the
macroeconomy appears strong in New Zealand because of
its economic structure. Unlike most advanced countries,
However, it is possible that the exchange rate overshoots its
fundamental value from time to time and that the Reserve
Bank could intervene to mitigate this overshooting. But is
this a sensible policy objective? What market failure would
the Reserve Bank resolve?
Cashin and McDermott (2003) found that New Zealand
importables and nontradables are complements implying
that, in response to adverse movements to the terms of
Even if we consider the exchange rate excessively volatile,
exporters and importers can buy financial instruments to
trade, the household sector cannot substitute away from
relatively expensive importables. Hence, the real income
Testing stabilisation policy limits in a small open economy
1
For example see Cashin and McDermott (2002), Cashin,
Liang, and McDermott (2000), and Cashin, McDermott, and
Scott (2002).
195
hedge this currency risk. Unfortunately, the currency cycle is
the owner will live in, a significant proportion are used to
too long relative to the duration of options that are typically
buy investment property or as start-up capital for small
traded. Financial markets will provide long duration options,
businesses. Requiring banks to collect such information
but because they are traded rarely the pricing often deviates
and report it publicly could prove to be hugely beneficial in
from the theoretical pricing from a standard Black-Scholes
developing our understanding of the credit movements in
formula. The deviations favour the sellers of the options
this economy.
who believe they need to be compensated for the illiquidity
of the market. This liquidity effect reduces the demand
for long-term options even more than would otherwise
be the case. If this is the identifiable market failure then
the Reserve Bank could enter the market be selling longduration options at the theoretical fair price, thus making a
market and making an accounting profit at the same time.
Another risk with foreign exchange intervention is what I
refer to as the “Dominion Post” risk. Suppose the Reserve
Bank intervened in the foreign exchange market a little bit
too early and initially started to lose money. There would be
a tremendous amount of interest from the media regarding
the financial losses, creating a huge distraction from the
objectives of monetary and prudential policy. The fact that
it may make a profit over the cycle would be lost in the
The current account balance is the final issue I want to
mention. Stephen Grenville addressed the issue by noting
that it is a safety valve, as was also suggested from the floor.
When there is pressure on domestic resources you can ease
this pressure by importing additional resources. Examining
the cyclical nature of New Zealand’s current account I think
it has been used in that fashion.
Access to global capital markets also allows New Zealanders
to shift consumption through time in more optimal ways.
There is some evidence that New Zealand does actually
consumption smooth through many of its cycles.2 However,
during the current expansion phase of the business cycle
we have not been consumption smoothing. If we had been
the current account deficit would not have grown the way
it did.
furore.
Nevertheless, I do not think that was necessarily a bad
On balance, the exchange rate, to a large degree, moves
with the economic fundamentals suggesting there are little
or no market failures and thus nothing to be gained from
foreign exchange intervention. Moreover, the associated
credibility risk with intervention would more than outweigh
any potential benefits.
outcome. A rational response from the private sector
seeing a rapidly appreciating exchange rate that by most
measures was overvalued would be to bring forward their
purchases of imported durables, and in particular Japanese
second hand cars. In effect Japan was selling cars to New
Zealand cheaply and offering very attractive finance at the
Grenville suggests that it may be useful to introduce some
same time. Now that the exchange rate cycle has turned we
prudential instruments as a complement to monetary policy.
should expect to see the imports of durable goods being
One problem with such an approach is that trying to control
reduced and the early signs are that this is exactly what is
the macroeconomy via the credit channel may prove more
happening. Economic forces are working as they should and
difficult than anticipated since we have no information on
this suggests that direct intervention would not improve the
the likely responses of the private sector to direct controls.
situation.
Another problem is that the distortions introduced by using
direct controls may be worse than the problem they are
trying to cure.
Another issue with regard to the current account is the
ability and willingness to repay. At some point it will be
necessary to repay the loans used to finance the current
One modest but very good idea suggested by Grenville
account deficit. If that deficit financing was used to fund
was that of improving data collection. We do not know
investment and if that investment was sensible then you
the purpose of most loans secured using a house. While
would have the ability to repay the loans. However, if the
the majority would be for the purchase of a house which
196
2
For example see Kim, Hall and Buckle, (2006).
Reserve Bank of New Zealand and The Treasury
deficit financing was used for consumption then that is a
are more incremental in their nature, but improving the
much more troubling development.
implementation and technical aspects of monetary policy
Examining the structure of the current account we see that
some of it went to consumption and some of it went to
investment. The investment part I have no problem with
since it will add to the New Zealand capital stock, thus
improving our potential growth rate and increasing our
ability to repay the loans. I am even comfortable with the
increase in consumption because I think much of it went
into the purchase of durables which can be viewed as an
is as important as any grand scheme and probably a more
fruitful avenue for research and policy making.
References
Borkin, P (2006), “Past, Present and Future Developments
in New Zealand’s Terms of Trade,” New Zealand Treasury
Working Paper 06/09, June. http://www.treasury.govt.nz/
workingpapers/2006/wp06-09.asp.
investment in a future stream of consumption services.
Cashin, P and C J McDermott (2002), “The Long-Run
Moreover, for the reasons stated above this type of spending
Behaviour of Commodity Prices: Small Trends and Big
will naturally slowdown and the exchange rate has started
Variability,” IMF Staff Papers, 49, 175-199.
to depreciate.
Cashin and McDermott (2003), “Intertemporal Substitution
Reflecting on all issues discussed at the forum, there was no
and Terms-of-Trade Shocks,” Review of International
obvious missing instrument that would deliver a home run
Economics, 11, 604-618.
for economic policy. However, a consensus did emerge that
there are marginal improvements to be had and that we can
improve the overall system and I think that that is a sensible
Cashin, P H Liang, and C. J. McDermott (2000), “How
Persistent are Shocks to World Commodity prices?” IMF
Staff Papers, 47, 177-217.
way to look at the issue. Examining our own history and the
evolution of policy tools in New Zealand shows that policy
can always be improved. For example, the introduction of
the Official Cash Rate (OCR) system was an improvement on
Cashin, P, C J McDermott, and A Scott (2002), “Booms
and Slumps in Commodity Prices,” Journal of Development
Economics, 69, 277-296.
the previous regime which used the Monetary Conditions
Chen, Y and K Rogoff (2003), “Commodity Currencies,”
Index (MCI) to implement monetary policy. It is these types
Journal of International Economics, 60, 113-160.
of improvements that we can make and that we should be
searching for. Possibly they are not as exciting because they
Kim, K-H, V B Hall and R A Buckle (2006), “Consumptionsmoothing in a small, cyclically volatile open economy:
Evidence from New Zealand,” Journal of International
Money and Finance, forthcoming.
Testing stabilisation policy limits in a small open economy
197
198
Reserve Bank of New Zealand and The Treasury
Testing stabilisation policy limits in a small open economy
199
200
Reserve Bank of New Zealand and The Treasury