Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
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