Download Emerging Market Liberalization and Monetary Control

Emerging Market Liberalization and
Monetary Control
Bill B. Francis
Rensselaer Polytechnic Institute
Delroy M. Hunter
University of South Florida
Patrick J. Kelly
New Economic School
Benefits of Financial Market Liberalization
By opening to foreign investment emerging markets can
benefit from:
• Greater access to capital [Henry (2000), Mitton (2006)]
• At a lower cost [Chari and Henry (2004), Bekaert and Harvey (2000), de
Jong and de Run (2005)]
• Spurs economic growth [Bekaert, at al (2001, 2009), Quinn and
Toyoda (2008)]
trilema
Patrick J. Kelly
Emerging Market Liberalization and Monetary Control
2/27
Costs of Liberalization
“A narrower question … is whether the increased openness of
the U.S. economy has in some way affected the ability of the
Federal Reserve to…foster price stability and maximum
sustainable employment.
On this issue, some analysts have argued that globalization
hinders monetary policy--for example, by reducing the ability
of the Federal Reserve to affect U.S. interest rates and asset
prices …”
(Bernanke (2007)).
“We need to re-examine the merits of financial liberalization in
the light of [the concern that it] lead[s] to a loss of monetary
control... ”
(Williamson (1998)).
Patrick J. Kelly
Emerging Market Liberalization and Monetary Control
3/27
Costs of Liberalization
But with access to foreign capital, firms may become
• less sensitive to local monetary policy
– Reducing the ability of monetary policy authorities to
influence macroeconomic targets
• and more sensitive to foreign policy
– Foreign policy may not be the best policy for the local
economy
Patrick J. Kelly
Emerging Market Liberalization and Monetary Control
4/27
Central Question
• To what extent are monetary policy authorities in
emerging markets able to influence their economies
following financial market liberalization?
• Reasons to think retaining control might be a challenge
– The “Impossible Trinity” (Obstfeld, Shambaugh, Taylor, 2005)
• Integration
• Exchange rate stability
• Monetary control
Patrick J. Kelly
Calvo et al. (2002, 2003)
Emerging Market Liberalization and Monetary Control
5/27
Received Knowledge: Impact of Monetary Policy
• Monetary policy shocks affect U.S. stock returns
• One of several channels through which monetary policy impacts the
real economy
– the cost of capital
– Default risk (changing collateral value)
– Investor Consumption and risk of default (Mishkin, 1995)
• They reflect anticipated changes in the economy
– Bernanke and Binder (1992), Bernanke and Kuttner (2005), many more
• Developed market stock prices are affected by U.S.
monetary policy (Conover, Jensen and Johnson, 1999, Wongswan, 2005)
• U.S. Monetary shocks impact emerging market stock
prices (Hausman and Wongswan, 2011 and Ehrmann and Fatzscher, 2009)
Patrick J. Kelly
Emerging Market Liberalization and Monetary Control
6/27
Central Question
• To what extent are monetary policy authorities in
emerging markets able to influence their economies
following liberalization?
– We study the stock price response to monetary policy shocks
• Many emerging markets have segments of their market
that have not fully opened to foreign investment
• Does keeping a fraction their firms closed to foreign
investment help preserve the ability of local monetary
policy authorities to affect the local economy?
Patrick J. Kelly
Emerging Market Liberalization and Monetary Control
7/27
Importance
• For policy makers
– Identifies where policy is effective and in what segment of
the market monetary policy is effective.
– Countries (frontier markets) considering opening their
markets to foreign investment
• Investors
– Identifies a factor affecting returns
– Indicator of market integration
• Identifies markets which may provide greater/lessor diversification
benefits for international portfolio managers
Patrick J. Kelly
Emerging Market Liberalization and Monetary Control
8/27
Gap in literature
1. Effect of local monetary policy on emerging stock
markets
2. Effect of local and U.S. monetary policy on stocks
open to foreign investment and closed
• Does keeping a fraction firms closed to foreign investment help preserve
the ability of monetary policy authorities to affect the local economy?
Patrick J. Kelly
Emerging Market Liberalization and Monetary Control
9/27
Findings:
1. Are local markets influenced by local monetary policy
above and beyond the influence of foreign monetary
policy post liberalization?
• In 18 of 25 markets one standard deviation increase in local policy
rates an average 2.07% decline in the local market.
• Confirming prior literature, U.S. monetary policy influences 11 of 25.
2. Are firms open to foreign investment investment
influenced by local (and foreign) monetary policy?
•
•
Yes, in 16 of 23 markets local policy affects investable stock
compared to 10 of 23 for non-investable.
consistent with an “efficiency” effect
–
Patrick J. Kelly
Cross country results consistent: more developed and more
internationally integrated markets are more sensitive to local policy.
Emerging Market Liberalization and Monetary Control
10/27
Data and Methodology
Patrick J. Kelly
Emerging Market Liberalization and
Monetary Control
11/27
The Data
• S&P’s Emerging Markets Database (EMDB) through 2006
– Global Index – returns to all stocks in a given market
– Investible Index – returns to stocks open to foreign
ownership
– Bae, Chan and Ng (2004) find that 25% to 35% of the smallest size quintile is
in the non-investable category.
– Non-Investible index following Boyer, Kumagai, and Yuan
(2006)
MVGt 1  RGt  MVIt 1  RIt
rNt 
MVGt 1  MVIt 1
• Liberalization Dates
– Bekaert, Harvey, and Lumsdaine (2002)
Patrick J. Kelly
Emerging Market Liberalization and Monetary Control
12/27
Summary Stats
Start
Date
Obs
Investable
Mean
Std
Obs
Non-investable
Mean
Std
Central and South America
Argentina
Oct-93
147
0.8
10.9
147
0.8
10.9
147
1.6
15.6
Brazil
Nov-94
134
1
9.1
134
1
9.3
134
1.4
9
Chile
Sep-94
135
0.4
5.7
135
0.3
5.8
135
0.5
5.1
Colombia
Apr-91
94
0.9
10
94
1.1
10.1
94
-0.4
12.4
Mexico
Jul-89
198
0.9
7.6
198
1
7.9
198
0.9
10.3
Peru
Feb-93
155
1.2
7.6
155
1.1
7.9
155
1.6
8.5
Venezuela
Feb-96
81
-0.4
12.3
69
-0.5
13
69
-0.1
16.5
Middle East and Africa
Israel
Jan-97
108
1.1
6.4
108
1.1
6.4
108
2.1
11.4
Jordan
Feb-96
119
1.6
5.5
69
0.2
3.6
69
0.3
3.6
S. Africa
Feb-97
107
1
6.9
107
1
6.9
55
0.6
14.2
Asia
India
Jan-93
125
-0.3
7.9
125
-0.3
8.1
125
-0.2
7.9
Korea
Mar-92
166
0.4
10.1
166
0.5
10.2
166
0
9.8
Malaysia
Feb-89
203
Longest
0.2
8.4
203
0.3
8.6
203
0.4
8.1
Pakistan
Patrick J. Kelly
Aggregate Local Market
Obs
Mean
Std
Jul-97
102
1.3
11.7
52
-1
14.5
52
0
11.4
Philippines
Aug-91
173
0
8
173
-0.1
8.9
173
0
7.6
Taiwan
Mar-91
178
0.1
9
178
0.2
9
178
0
9
Thailand
Feb-91
179
0
11.2
179
0
11
179
0
11.5
Europe
Czech
Jan-94
144
0.2
8
144
0.3
9.3
119
-0.7
8.8
Greece
Feb-88
157
1
10.5
146
1.3
11.3
129
2.4
12.3
Hungary
Jan-93
156
0.9
9.9
156
1.1
10.7
156
0.4
7.7
Poland
Jan-93
156
1.3
13.2
156
1.3
13.2
119
1.4
10.1
Portugal
Sep-86
146
1
10.5
118
0.5
6.6
118
0.5
6.6
Russia
Feb-00
71
71
1.3
11.1
71
2.6
11.3
Slovakia
Turkey
Jan-96
Oct-89
106
195
56
195
-1.6
0.3
7.8
15.8
56
125
-1
-8.5
7.7
64.3
1.7Shortest
10.5
0
0.3
7.7
15.8
Emerging Market Liberalization and Monetary Control
13/27
Data: Monetary Policy Proxies
• Interest rates chosen based on Calvo and Reinthart
(2002) survey of policy targets
•
Loayza and Shmidt-Hebble (2002) and Kamin, Turner and Van’t
dack (1998) note that with liberalization short term interest rates
become the primary tool of monetary policy
–
1.
2.
Even if not the only tool, market-based interest rates will reflect these
changes [Obstfeld et al. (2005)]
the interbank interest rates,
discount rate
3.
4.
5.
Treasury bill rate
money market rate
10-year government bond rate
• All from Datastream
–
(changes Winsorized at 5th and 95the percentile)
14
Methodology: Measuring Monetary Policy Shocks
Best Practices
• Structural Vector Auto Regression (SVAR) [Christiano
Eichenbaum, and Evans (1999), Kim and Roubini (2000)]
– Model expectation of monetary policy changes as function of:
• Log Oil prices (oil) x 100
• Log first difference Fed Funds Rate (FF)
• Log first difference Industrial Production (IP) x 100
– Where IP unavailable we use manufacturing. In Argentina and
Venezuela crude petroleum production.
•
•
•
•
Inflation (inf) (Log first difference in CPI)
Log first difference of annualize Local Monetary Policy rate (LMP)
Log first difference of Exchange rate (FX) in US/local x 100
Real market return (Ret): Log first difference of index deflated by
local inflation
– Oil, output and CPI are seasonally adjusted
Patrick J. Kelly
Emerging Market Liberalization and Monetary Control
15/27
Structural Vector Auto Regression (SVAR)
• Model the evolution of each of the 7 measures in a 7
equation system (restrictions Kim and Roubini (2000) and Bjornland and Leitemo (2009)):
é
ê
ê
ê
ê
ê
ê
ê
ê
ê
ê
êë
e
e
oil
FF
e
e
e
IP
Inf
LMP
e
e
FX
Re t
ù é
ú ê
ú ê
ú ê
ú ê
ú=ê
ú ê
ú ê
ú ê
ú ê
ú ê
úû ë
Re tt = q 71Oilt
LMP
+P
1
j
j
j
j
j
j
21
31
61
71
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
1
0
0
0
0
1
0
41
51
0
1
j
0
j
j
62
72
j
j
43
0
63
73
j
j
64
74
j
j
65
75
j
j
56
1
76
0
0
1
ùé
úê
úê
úê
úê
úê
úê
úê
úê
úê
úê
ûêë
n
n
FF
n
n
n
oil
IP
Inf
LMP
n
n
FX
Re t
+ q 72 FFt + q 73 IPt + q 74 Inft + q 75 LMPt + q 76 FXt
ù
ú
ú
ú
ú
ú
ú
ú
ú
ú
ú
úû
Many
emerging
markets
actively
manage
their
exchange
rates
number of lags of each of the seven variables + e Ret,t
Oil is allowed to affect all monetary policy expectations, even for the US following
Bernenke and Mihov (1998) Romer2 (2004) and Kim and Roubini (2000)
Patrick J. Kelly
Emerging Market Liberalization and Monetary Control
16/27
Question 1
1. Are markets influenced by local monetary policy
above and beyond the influence of foreign monetary
policy post liberalization?
Here I report only the first period
• Impulse response of returns to local and foreign
monetary policy
– In paper also the standard deviation of the structural shocks
Patrick J. Kelly
Emerging Market Liberalization and Monetary Control
17/27
Example Impluse Reponse
• Russia
Response of Returns to
Local Monetary policy
Response of Returns to
U.S. Monetary policy
a one standard deviation shock to local monetary policy
results in a 3% decline in stock prices
Patrick J. Kelly
Emerging Market Liberalization and Monetary Control
Tech details
18/27
Local authorities meaningfully influence local markets
Whole market sensitive to local policy: 18/25 markets (average -2.07%)
U.S. Policy: 11/25 markets (average -1.32%)
Response of Returns to
Country
Local Monetary Local Monetary U.S. Monetary
Policy Proxy
Policy
Policy
Response of Returns to
Country
Central and South America
IB
Argentina
IB
Brazil
-3.079♠
0.189
-2.000♠
-1.107♠
Jordan
Chile
IB
-1.628♠
-1.076♠
Colombia
DR
0.19
-1.859♠
Mexico
IB
-0.713♠
Peru
DR
Venezuela
MM
Local Monetary Local Monetary U.S. Monetary
Policy Proxy
Policy
Policy
Middle East and Africa
Israel
TB
-2.296♠
0.793
DR
-0.068
-0.351
South Africa
GB
-2.324♠
-1.262♠
-0.945
Europe
Czech
IB
0.712
-0.685
0.118
-0.285
Greece
TB
-2.641♠
-0.246
-2.308♠
-1.949♠
Hungary
TB
-1.165♠
-1.674♠
Poland
MM
-2.708♠
-1.342♠
Portugal
DR
-0.505
2.442♠
Russia
IB
-2.999♠
0.549
Asia
India
DR
-0.458
-1.011♠
Korea
MM
-1.142♠
0.493
Malaysia
TB
-1.944♠
-0.355
Slovakia
IB
-1.193♠
-2.056♠
Pakistan
MM
-1.753♠
1.118
Turkey
MM
-4.803♠
-1.489♠
Philippines
IB
-1.214♠
-0.673
Taiwan
IB
Thailand
IB
-0.676
-1.290♠
0.324
-1.908♠
robustness
Patrick J. Kelly
Emerging Market Liberalization and Monetary Control
19/27
Decomposing Whole Market Returns
• Are firms open to foreign investment investment
influenced by local (and foreign) monetary policy?
• Are firms closed to foreign investment influenced by
foreign (and local) monetary policy?
• Is the sensitivity of the market return to local policy a
driven by non-investable stock?
• Same SVARs only with investable and non-investable
indices
– Too few observations to jointly estimate
Patrick J. Kelly
Emerging Market Liberalization and Monetary Control
20/27
Non-Investable Response to Monetary Policy Shocks
Country
Response of Response of
Returns to
Returns to
Local Monetary U.S. Monetary
Policy
Policy
Coefficient
Coefficient
Central and South America
Argentina
-2.229♠
Country
Response of Response of
Returns to
Returns to
Local Monetary U.S. Monetary
Policy
Policy
Coefficient
Coefficient
Middle East and Africa
0.802
Israel
-3.635♠
-0.272
-1.844♠
-1.199♠
Jordan
0.146
-0.147
-0.378
0.374
Europe
Colombia
0.252
-1.757♠
Czech
-0.239
-1.159♠
Mexico
-0.305
-0.315
Greece
-3.238♠
-1.810♠
Peru
0.461
-0.308
Hungary
-1.716♠
0.325
Poland
-1.219♠
-1.609♠
-0.76
-0.069
Brazil
Chile
Asia
India
-0.648
-0.135
Portugal
Korea
-1.530♠
0.195
Russia
-2.095♠
0.936
Malaysia
-1.817♠
-0.649
Turkey
-1.557
-0.933
-0.945
-0.905♠
-0.662
-1.370♠
0.258
-2.011♠
Philippines
Taiwan
Thailand
Patrick J. Kelly
10/21 sensitive to local policy
7/21 sensitive to U.S. policy
Emerging Market Liberalization and Monetary Control
21/27
Investable Response to Monetary Policy Shocks
Country
Response of Response of
Returns to
Returns to
Local Monetary U.S. Monetary
Policy
Policy
Coefficient
Coefficient
Central and South America
Argentina
-3.090♠
Country
Response of Response of
Returns to
Returns to
Local Monetary U.S. Monetary
Policy
Policy
Coefficient
Coefficient
Middle East and Africa
0.163
Israel
-2.293♠
0.766
-2.326♠
-1.613♠
Jordan
0.118
-0.370
-1.645♠
-1.074♠
South Africa
-2.296♠
-1.261♠
0.063
-1.822♠
Europe
-1.730♠
-0.21
Czech
0.738
-0.644
Peru
-0.012
-0.236
Greece
-2.779♠
-0.127
Venezuela
-0.591
-1.571
Hungary
-1.224♠
-1.630♠
Poland
-2.695♠
-1.335♠
Brazil
Chile
Colombia
Mexico
Asia
India
-0.471
-0.136
Portugal
-0.869♠
-0.571
Korea
-1.070♠
0.466
Russia
-3.354♠
0.743
Malaysia
-1.957♠
-0.311
Turkey
-4.729♠
-1.397
Philippines
-1.584♠
-0.566
Taiwan
-0.654
-1.229♠
0.353
-1.772♠
Thailand
Patrick J. Kelly
16/23 sensitive to local policy
7/23 sensitive to U.S. policy
Emerging Market Liberalization and Monetary Control
22/27
Differences between Non- Investables and Investables
• Are investable stock more responsive to local monetary
policy shocks?
– In sufficient power to model both series in an SVAR
• We difference the two series and model:
– Non-Investable minus investable
– Positive means Investable as a stronger effect
– Negative means Non-Investable has a stronger effect
Patrick J. Kelly
Emerging Market Liberalization and Monetary Control
23/27
Non-Investable Minus Investable
Response of Response of
Returns to
Returns to
Local
U.S. Monetary
Monetary
Policy
Policy
Country
Coefficient
Central and South America
Argentina
-0.846
Brazil
0.804
Chile
1.218♠
Colombia
-0.674
Mexico
1.142♠
Peru
0.269
Asia
India
-0.129
Korea
-0.433♠
Malaysia
0.065
Philippines
0.443
Taiwan
0.032
Thailand
-0.145
Patrick J. Kelly
Coefficient
0.134
0.030
0.401
-0.339
0.150
0.106
0.035
-0.552♠
-0.217
-0.311
-0.068
-0.211
Response of Response of
Returns to
Returns to
Local
U.S. Monetary
Monetary
Policy
Policy
Country
Coefficient
Middle East and Africa
Israel
-0.846
Jordan
0.804
Europe
Czech
0.142
Greece
-0.029
Hungary
1.024♠
Poland
0.887
Portugal
0.603♠
Russia
1.755♠
Turkey
-1.429
Coefficient
0.134
0.030
1.000
-1.940♠
0.956♠
-0.896
0.868♠
0.484
0.324
In 5 markets Investable stock
more sensitive to local monetary
policy shocks
Emerging Market Liberalization and Monetary Control
24/27
Summing up Investable and Non-investable results
• Local policy affects both
– Investable
– Non-investable
• Local policy has a more pronounced effect on
investable stock
– Efficiency effect? Consistent with Reese and Weisbach (2002)
– firms enter foreign markets to raise more local capital.
• Shows that the sensitivity to U.S. Monetary policy is not
solely driven by the investable component.
– Contribution over Hausman and Wongswan (2006),
Ehrmann and Fratzscher (2006)
Patrick J. Kelly
Emerging Market Liberalization and Monetary Control
25/27
Cross Country Explanations
Global Stock Returns Response to
Local monetary shocks
U.S. monetary shocks
N
Constant
Market Capitalization to GDP
Somewhat greater in more developed markets
Capital Control Intensity
Weaker in more segmented markets
Net In-Flows in Billions
Weaker when more foreign capital in-flows
Exchange Rate Regime
Political Risk
Central Bank Political Autonomy
Central Bank Economic Autonomy
2
Adj R
Patrick J. Kelly
22
-5.83
(-2.75)
-0.01
(-1.87)
0.03
(2.69)
0.74
(1.85)
-0.05
(-0.46)
0.09
(1.82)
2.58
(-1.56)
0.38
(0.30)
22
-5.64
(-2.54)
-0.01
(-2.30)
0.03
(2.42)
0.81
(2.14)
-0.04
(-0.42)
0.09
(2.16)
2.68
(-1.72)
0.096
0.154
22
-5.51
(-2.94)
0.00
(-0.55)
0.03
(3.44)
0.60
(1.63)
0.02
(-0.20)
0.03
(1.02)
22
0.54
(0.26)
0.00
(-0.45)
-0.01
(-0.98)
-0.81
(-2.88)
-0.24
(-3.56)
0.03
(0.75)
1.00
(-0.89)
1.12
(0.87)
22
1.38
(0.65)
-0.01
(-1.35)
-0.01
(-1.42)
-0.49
(-1.06)
-0.22
(-3.42)
0.04
(1.10)
1.45
(-1.31)
-1.78
(-1.23)
0.0215
0.017
0.010
Emerging Market Liberalization and Monetary Control
22
1.56
(0.66)
0.00
(-0.48)
-0.01
(-0.92)
-0.57
(-1.10)
-0.19
(-2.72)
0.01
(0.37)
-1.36
(-0.91)
0.028
26/27
Conclusion
• Local monetary policy is a factor that affects the entire
market in 18 of 25 markets: 2.07% decline for a 1
standard deviation shock
– Not driven by non-investable stock
• When different – investable are more sensitive
– US policy does not dominate local policy
• It appears there are externalities to liberalization that
– Stocks become more sensitive to local policy, but
– Causes non-investable stock to respond to foreign policy as if
they too were investable in a few (7) markets
Patrick J. Kelly
Emerging Market Liberalization and Monetary Control
27/27