Emerging Market Liberalization and Monetary
Control
Bill B. FrancisRensselaer Polytechnic Institute
Delroy M. HunterUniversity of South Florida
Patrick J. KellyNew 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)]
Patrick J. Kelly Emerging Market Liberalization and Monetary Control
2/27
trilema
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 3/27Emerging Market Liberalization and Monetary Control
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 Emerging Market Liberalization and Monetary Control
5/27
Calvo et al. (2002, 2003)
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– Cross country results consistent: more developed and more
internationally integrated markets are more sensitive to local policy.
Patrick J. Kelly Emerging Market Liberalization and Monetary Control
10/27
Data and Methodology
Patrick J. Kelly Emerging Market Liberalization and Monetary
Control
11/27
The Data
Patrick J. Kelly Emerging Market Liberalization and Monetary Control
12/27
• 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)
• Liberalization Dates– Bekaert, Harvey, and Lumsdaine (2002)
11
11
ItGt
ItItGtGtNt MVMV
RMVRMVr
Summary Stats
Patrick J. Kelly Emerging Market Liberalization and Monetary Control
13/27
Start
Date Obs Mean Std Obs Mean Std Obs Mean Std
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
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
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 0.2 8.4 203 0.3 8.6 203 0.4 8.1
Pakistan 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
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 1.7 10.5 71 1.3 11.1 71 2.6 11.3
Slovakia Jan-96 106 0 7.7 56 -1.6 7.8 56 -1 7.7
Turkey Oct-89 195 0.3 15.8 195 0.3 15.8 125 -8.5 64.3
Europe
Aggregate Local Market Investable Non-investable
Central and South America
Middle East and Africa
Asia
Longest
Shortest
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– Even if not the only tool, market-based interest rates will
reflect these changes [Obstfeld et al. (2005)]1. the interbank interest rates, 2. discount rate
3. Treasury bill rate4. money market rate5. 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)):
Patrick J. Kelly Emerging Market Liberalization and Monetary Control
16/27
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)
Many emerging markets actively manage their exchange rates
LMP
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 Response of Returns to
Local Monetary policy U.S. Monetary policy
Patrick J. Kelly Emerging Market Liberalization and Monetary Control
18/27
a one standard deviation shock to local monetary policy results in a 3% decline in stock prices Tech detail
s
Local authorities meaningfully influence local markets
Patrick J. Kelly Emerging Market Liberalization and Monetary Control
19/27
CountryLocal Monetary Policy Proxy
Local Monetary Policy
U.S. Monetary Policy Country
Local Monetary Policy Proxy
Local Monetary Policy
U.S. Monetary Policy
Argentina IB -3.079♠ 0.189 Israel TB -2.296♠ 0.793
Brazil IB -2.000♠ -1.107♠ Jordan DR -0.068 -0.351
Chile IB -1.628♠ -1.076♠ South Africa GB -2.324♠ -1.262♠
Colombia DR 0.19 -1.859♠
Mexico IB -0.713♠ -0.945 Czech IB 0.712 -0.685
Peru DR 0.118 -0.285 Greece TB -2.641♠ -0.246
Venezuela MM -2.308♠ -1.949♠ Hungary TB -1.165♠ -1.674♠
Poland MM -2.708♠ -1.342♠
India DR -0.458 -1.011♠ Portugal DR -0.505 2.442♠
Korea MM -1.142♠ 0.493 Russia IB -2.999♠ 0.549
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 -0.676 0.324
Thailand IB -1.290♠ -1.908♠
Middle East and Africa
Response of Returns toResponse of Returns to
Central and South America
Asia
Europe
Whole market sensitive to local policy: 18/25 markets (average -2.07%)U.S. Policy: 11/25 markets (average -1.32%)
robustness
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
Patrick J. Kelly Emerging Market Liberalization and Monetary Control
21/27
Country Coefficient Coefficient Country Coefficient Coefficient
Argentina -2.229♠ 0.802 Israel -3.635♠ -0.272
Brazil -1.844♠ -1.199♠ Jordan 0.146 -0.147
Chile -0.378 0.374
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♠
India -0.648 -0.135 Portugal -0.76 -0.069
Korea -1.530♠ 0.195 Russia -2.095♠ 0.936
Malaysia -1.817♠ -0.649 Turkey -1.557 -0.933
Philippines -0.945 -0.905♠
Taiwan -0.662 0.258
Thailand -1.370♠ -2.011♠
Response of Returns to
Local Monetary Policy
Response of Returns to
U.S. Monetary Policy
Middle East and Africa
Europe
Central and South America
Asia
Response of Returns to
Local Monetary Policy
Response of Returns to
U.S. Monetary Policy
10/21 sensitive to local policy7/21 sensitive to U.S. policy
Investable Response to Monetary Policy Shocks
Patrick J. Kelly Emerging Market Liberalization and Monetary Control
22/27
Country Coefficient Coefficient Country Coefficient Coefficient
Argentina -3.090♠ 0.163 Israel -2.293♠ 0.766
Brazil -2.326♠ -1.613♠ Jordan 0.118 -0.370
Chile -1.645♠ -1.074♠ South Africa -2.296♠ -1.261♠
Colombia 0.063 -1.822♠
Mexico -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♠
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 0.353
Thailand -1.229♠ -1.772♠
Middle East and Africa
Asia
Central and South America
Europe
Response of Returns to
Local Monetary Policy
Response of Returns to
U.S. Monetary Policy
Response of Returns to
Local Monetary Policy
Response of Returns to
U.S. Monetary Policy
16/23 sensitive to local policy7/23 sensitive to U.S. policy
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
Response of Returns to Local
Monetary Policy
Response of Returns
to U.S. Monetary
Policy
Response of Returns to Local
Monetary Policy
Response of Returns
to U.S. Monetary
Policy
Country Coefficient Coefficient Country Coefficient CoefficientCentral and South America Middle East and AfricaArgentina -0.846 0.134 Israel -0.846 0.134Brazil 0.804 0.030 Jordan 0.804 0.030Chile 1.218♠ 0.401 EuropeColombia -0.674 -0.339 Czech 0.142 1.000Mexico 1.142♠ 0.150 Greece -0.029 -1.940♠Peru 0.269 0.106 Hungary 1.024♠ 0.956♠Asia Poland 0.887 -0.896India -0.129 0.035 Portugal 0.603♠ 0.868♠Korea -0.433♠ -0.552♠ Russia 1.755♠ 0.484Malaysia 0.065 -0.217 Turkey -1.429 0.324Philippines 0.443 -0.311Taiwan 0.032 -0.068Thailand -0.145 -0.211
Non-Investable Minus Investable
Patrick J. Kelly Emerging Market Liberalization and Monetary Control
24/27
In 5 markets Investable stock more sensitive to local monetary policy shocks
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
Patrick J. Kelly Emerging Market Liberalization and Monetary Control
26/27
Global Stock Returns Response to Local monetary shocks U.S. monetary shocks
N 22 22 22 22 22 22Constant -5.83 -5.64 -5.51 1.38 0.54 1.56
(-2.75) (-2.54) (-2.94) (0.65) (0.26) (0.66)Market Capitalization to GDP -0.01 -0.01 0.00 -0.01 0.00 0.00
(-1.87) (-2.30) (-0.55) (-1.35) (-0.45) (-0.48)Capital Control Intensity 0.03 0.03 0.03 -0.01 -0.01 -0.01
(2.69) (2.42) (3.44) (-1.42) (-0.98) (-0.92)Net In-Flows in Billions 0.74 0.81 0.60 -0.49 -0.81 -0.57
(1.85) (2.14) (1.63) (-1.06) (-2.88) (-1.10)Exchange Rate Regime -0.05 -0.04 0.02 -0.22 -0.24 -0.19
(-0.46) (-0.42) (-0.20) (-3.42) (-3.56) (-2.72)Political Risk 0.09 0.09 0.03 0.04 0.03 0.01
(1.82) (2.16) (1.02) (1.10) (0.75) (0.37)Central Bank Political Autonomy 2.58 2.68 1.45 1.00
(-1.56) (-1.72) (-1.31) (-0.89)Central Bank Economic Autonomy 0.38 1.12 -1.78 -1.36
(0.30) (0.87) (-1.23) (-0.91)
Adj R2 0.096 0.154 0.0215 0.017 0.010 0.028
Somewhat greater in more developed markets
Weaker in more segmented markets
Weaker when more foreign capital in-flows
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