Transcript THE EFFECT OF CAPITAL MARKET LIBERALIZATION IN EASTERN

Academy For Economic Studies, Bucharest - Doctoral School of Finance and Banking (DOFIN)
THE EFFECT OF CAPITAL
MARKET LIBERALIZATION
IN EASTERN EUROPE:
ECONOMIC GROWTH OR
FINANCIAL CRISIS
- Dissertation Paper -
MSc Student: LAVINIA CRISTESCU
Coordinator: PhD. Professor MOISĂ ALTĂR
Bucharest, July 2008
CONTENTS
I.
II.
III.
IV.
Introduction
Literature Review
Model Specifications
Empirical Analysis
a.
b.
The Data
Testing The Financial Liberalization Effect
V. Conclusions
VI. References
I. INTRODUCTION
Openes international financing path
Decreases the cost of capital
Increases investment
FINANCIAL
LIBERALIZATION
OF EQUITY
MARKETS
Leads to a more rapid economic growth
Also, may lead to:
A decline in credit’s portfolio quality
An increase in financial fragility
Macroeconomic volatility to external
shocks
FINANCIAL CRISES AND LOSSES
II. LITERATURE REVIEW

Bekaert, Harvey and Lundblad (2005) – Capital market liberalization leads to 1%
increase in the economic growth rate.

Kaminski and Reinhart (1998), Glick and Hutchinson (2001) – Banking and
currency crisis propensity increases in the aftermath of financial liberalization.

Dell’ Aricia and Marquez (2004) – Financial liberalization helps developing the credit
sector by reducing the bank’s incentive to monitor potential debtors.

Martin and Rey (2005) – In normal circumstances, liberalization has the positive role
to generate capital inflows, to create diversification opportunities and to stimulate
economic growth; in certain circumstances, liberalization can lead to financial crashes
and a decrease in economic growth.

Ranciere, Tornell and Westermann (2006) – Financial liberalization has an positive
influence on economic growth, although it increases the probability of financial crises.

Henry (2000) – Liberalization leads to an investment boom associated with a decrease
in the cost of capital.
III. MODEL SPECIFICATIONS

GROWTH MODEL (Panel, linear):
yi,t = αXi,t + βFLi,t + γIi,t + εi,t
Where:

yi,t – is the real GDP per capita growth (in logarithm)

Xi,t – is a set of standard control variables

FLi,t – is a dummy for financial liberalization, taking the value 1 if the country i
is liberalized in year t and zero otherwise

Ii,t – is a dummy for crisis, taking the value 1 if there is a banking or currency
crisis in the year t and zero otherwise

εi,t
– is a random, gaussian component.
III. MODEL SPECIFICATIONS

CRISIS MODEL (Panel, probit)
1 if W*i,t > 0
Ii,t =
W*i,t = aZi,t + bFLi,t + ηi,t
0 otherwise.
W*i,t is a latent, unobserved variable (the crisis probability) who depends on:
- Zi,t – a set of control variables
- FLi,t – dummy financial liberalization
- ηi,t – random, gaussian variable
1 with probability P(W*i,t > 0) = Φ(aZi,t + bFLi,t)
Ii,t =
0 with probability P(W*i,t ≤ 0) = 1 - Φ(aZi,t + bFLi,t)
Φ = cumulative distribution function of a standard normal
III. MODEL SPECIFICATIONS
GROWTH MODEL
CRISIS MODEL
Two step estimation
procedure
TREATMENT EFFECT MODEL
(Maddala (1983))
(Heckman (1978))
It measures the average causal effect of a binary variable (the treatment) on an output variable.
CRISIS DUMMY
=
The treatment
GROWTH REGRESSION =
Output Equation
CRISIS REGRESSION
Treatment Equation (represents the
probabiliy of receiving the treatment)
=
III. MODEL SPECIFICATIONS

TWO STEP ESTIMATION PROCEDURE:
1. OBTAINING THE PROBIT ESTIMATES (ae, be)
2. COMPUTING AND ADDING TO THE GROWTH REGRESSION OF A
HAZARD (hi,t) VARIABLE:
θ(ae Zi,t + be FLi,t) / Φ(ae Zi,t + be FLi,t),
hi,t =
if Ii,t = 1
- θ(ae Zi,t + be FLi,t) / [1 - Φ(ae Zi,t + be FLi,t)], if Ii,t = 0
Φ = cumulative distribution function of a standard normal
θ = probability density of a standard normal
ASSUMPTION: the errors are bivariate normal, but not independent
III. MODEL SPECIFICATIONS
TOTAL AVERAGE CAUSAL EFFECT OF FINANCIAL
LIBERALIZATION
Due to a change in the financial liberalization dummy from 0 to 1
E(yi,t | FLi,t = 1) - E(yi,t | FLi,t = 0)
Financial Liberalization
Total Effect
=
βe
+
Direct Effect
γe E[Φ(aeZi,t + be) – Φ(aeZi,t)]
Indirect Effect
IV. EMPIRICAL ANALYSIS
a. The Data
THE DATASET:
TIME PERIOD:
DATASOURCE:



NO.
COUNTRY
13 EASTERN EUROPE COUNTRIES
1995 – 2007 (annual series)
AMECO DATABASE, CENTRAL BANKS’ STATISTICAL
SERIES and BEKAERT and HARVEY’S DATABASE FROM
DUKE UNIVERSITY
FINANCIAL
LIBERALIZATION
YEAR
BANKING
CRISIS YEAR
CURRENCY
CRISIS YEAR
FINANCIAL
CRISIS
EU
MEMBER
FROM
1995, 1996
1995
1995, 1997
2007
1
BULGARIA*,c
1998
2
CYPRUS
2004
3
CZECH REPUBLIK,c
1994
4
ESTONIA,c
1996
5
HUNGARY*,c
1996
1995
1995
2004
6
LATVIA*,c
1996
1995
1995
2004
7
LITHUANIA,c
1996
1995, 1996
1995, 1996
2004
8
MALTA*
2004
1997
1997
2004
9
POLAND,c
1995
10
ROMANIA,c
1997
11
SLOVAKIA, c
1996
12
SLOVENIA, c
1994
13
TURKEY
1989
2004
1997
1997
1997
2004
2004
2004
2000
1998
2000
2007
1998
2004
2004
2000
2000
-
IV. EMPIRICAL ANALYSIS
a. The Data
GROWTH DEPENDENT VARIABLE:

REAL GDP PER CAPITA GROWTH – real_gdp_gr – log-difference of real GDP
per capita (stationary, ADF)
GROWTH DETERMINANTS:

CONTROL VARIABLES:


INITIAL REAL GDP PER CAPITA – real_gdp – the ratio between real GDP (2000 current market
prices GDP in national currency / GDP Deflator) and total population (stationary, ADF)
GOVERMENT SIZE – gov_size – ratio of final government consumption to GDP (2000 current market
prices in national currency) (stationary, ADF)

POPULATION GROWTH – pop_gr – log-difference of total population (stationary, ADF)

INFLATION – inflatia – (log 100 + % National CPI all items) (stationary, ADF)

FINANCIAL LIBERALIZATION DUMMY – dummy_fl – measurement: official change
in regulatory that allows foreigners to invest in domestic securities

FINANCIAL CRISIS DUMMY – dummy crisis – takes value 1 in the year where banking
or currency crisis occurs
IV. EMPIRICAL ANALYSIS
a. The Data
PROBIT DEPENDENT VARIABLE:

Ii,t through the unobserved, latent variable W*i,t
PROBIT DETERMINANTS:


CONTROL VARIABLES:

GOVERMENT SIZE

POPULATION GROWTH

INFLATION (1 LAG)

M2 / (INTERNATIONAL RESERVES – GOLD) – m2_res – the ratio between the monetary
aggregate M2 and international liquid reserves (not stationary, ADF => first difference)

OPENESS TO TRADE – openess_trade – the ratio between (total exports and imports) to GDP –
(not stationary, ADF => first difference)

REAL EFFECTIVE EXCHANGE RATE DETRENDED – rero_hptrend01 – real effective
exchange rates (performance relative to 35 industrialized countries, EU) detrended using Hodrick
Prescott filter, λ=100 (stationary, ADF)
DUMMY FINANCIAL LIBERALIZATION
IV. EMPIRICAL ANALYSIS
b. Testing The Effect of Financial Liberalization
TREATMENT EFFECT MODEL TWO STEPS ESTIMATION (STATA 9.1.)
IV. EMPIRICAL ANALYSIS
b. Testing The Effect of Financial Liberalization
THE ESTIMATORS’ CONFIDENCE LEVEL:
All the regressions’
coefficients are
significant for a 95%
level of confidence
IV. EMPIRICAL ANALYSIS
b. Testing The Effect of Financial Liberalization
TESTING THE PROBIT RESIDUALS:
 Distribution:
The probit residual is not normally distributed
IV. EMPIRICAL ANALYSIS
b. Testing The Effect of Financial Liberalization

Correlograms:
There is no evidence of serial residual
correlation
There is no serial correlation of
residuals squared
IV. EMPIRICAL ANALYSIS
b. Testing The Effect of Financial Liberalization

TESTING THE GROWTH REGRESSION RESIDUALS
Histogram
The growth residuals are not normally
distributed.
Correlogram
The growth residuals are not
autocorrelated
IV. EMPIRICAL ANALYSIS
b. Testing The Effect of Financial Liberalization
TESTING THE GROWTH AND PROBIT RESIDUALS’
DEPENDENCE
New linear regression:
εi,t = Ci,t + ηi,t + ei,t
There is a dependence between the
growth and the probit residuals
=> The two residual series are not
normally bivariate and are not
independent
IV. EMPIRICAL ANALYSIS
b. Testing The Effect of Financial Liberalization
TOTAL AVERAGE EFFECT OF FINANCIAL LIBERALIZATION:
DIRECT EFFECT
βe
0.2197727
INDIRECT EFFECT
γe E[Φ(aeZi,t + be) – Φ(aeZi,t)]
2.10817E-19
TOTAL EFFECT
βe+ γe E[Φ(aeZi,t
Φ(aeZi,t)]
+
be)
– 0.2197727
On average, the total effect of capital market liberalization in Eastern Europe
countries was a positive one
IV. EMPIRICAL ANALYSIS
b. Testing The Effect of Financial Liberalization
ESTIMATES DISCUSSION: GROWTH REGRESSION
 REAL INITIAL GDP PER CAPITA (-0.289648, p < 1%) – economic growth rate is smaller for
countries with a higher initial development level, consistent with Kormendi and Meguire (1985), Barro
(1991, 1997), Sachs and Warner (1995).
 GOVERMENT SIZE (3.9021292, p < 0.1%) – has a positive influence on growth, differs from Barro
(1991, 1997), Sachs and Warner (1995) and is consistent with Caesseli (1996).
 POPULATION GROWTH (7.0823379, p < 1%) – has a positive influence on growth, is consistent
with Barro and Lee (1994) and differs from Kormendi and Meguire (1985), Mankiw (1992), Kelley and
Schmidt (1995), Bloom and Sachs (1998).
 INFLATION (-0.17143785, p < 0.1%) – leads to a decrease in economic growth rate, consistent with
Barro (1997), Bruno and Easterly (1998), Motley (1998).
 DUMMY FINANCIAL LIBERALIZATION (0.2197727, p < 0.1%) – leads to an increase in economic
growth rate, consistent with literature.
 DUMMY CRISIS (0.3893808, p < 1%) – consistent with literature (Ranciere, Tornell, Westermann
(2006)), has a negative influence on growth.
IV. EMPIRICAL ANALYSIS
b. Testing The Effect of Financial Liberalization
ESTIMATES DISCUSSION: PROBIT REGRESSION
 GOVERNMENT SIZE - (27.05248, p < 5%) – increases the crisis probability
 POPULATION GROWTH – (127.7304, p < 5%) – increases the crisis probability
 M2 / (INTERNATIONAL RESERVES – GOLD) – (-0.000115, p < 1%) – reduces the crisis
probability and differs from the economic hypothesis.
 INFLATION (1 LAG) – (1.216772, p < 5%) – increases the crisis probability
 REAL EFFECTIVE EXCHANGE RATE HP DETRENDED (-0.140846, p < 1%) – reduces the
probability of crisis.
From economical hypothesis (Kazaks (2000), Shatz and Tarr (2000) and Ranciere, Tornell and
Westermann (2006), I first included in the probit non-linear regression Real Effective Exchange
Rate Overvaluation (also 1 lag), defined as the percentage difference between Real Effective
Exchange Rate and HP Detrended REER (IMF’s definition). However, it showed no statistical
significant influence within the model. Instead, HP detrended REER has a negative statistical
significant effect.
IV. EMPIRICAL ANALYSIS
b. Testing The Effect of Financial Liberalization
ESTIMATES DISCUSSION: PROBIT REGRESSION
 FINANCIAL LIBERALIZATION DUMMY – (-1.60857, p < 5%)
- decreases the probability of occurring a financial crisis!
- the result differs from the ones obtained in the literature and from the
economic hypothesis considerred.
FINANCIAL LIBERALIZATION HAD AN AVERAGE
POSITIVE EFFECT ON GROWTH, COMPOSED BY:
A POSITIVE DIRECT EFFECT
A POSITIVE INDIRECT EFFECT – by decreasing the crisis
probability
V. CONCLUSIONS


Conclusions:

Capital market liberalization had an average positive effect on
economic growth in Eastern Europe

The other estimators’ influence is related to the economies’
specifications (emerging, most of them post-communist)

The conclusions can only be applied to the analyzed sample, a
generalization is not accurate
Utility


The joint analysis of financial liberalization improves economic
decision making
Further research:

Methodology improvement

Analysis of crises appeared in developed economies

Other determinants selection
VI. REFERENCES
1.
Eichengreen, B. and C. Arteta (2000), „Banking Crises in Emerging Markets: Presumptions and
Evidence”, Institute of Business and Economic Research
2.
Davis, E. P. and D. Karim (2007), „ Comparing Early Warning Systems for Banking Crises”,
Economics and Finance Working Paper No. 07 - 11, Brunel University
3.
Bekaert, G. and C.R. Harvey (2003), „Does Financial Liberalization Spur Growth?”, Journal of
Financial Economics
4.
Glick, R., X. Guo and M. Hutchinson (2004), „Currency Crises, Capital Account Liberalization, and
Selection Bias”, UC Santa Cruz International Economics Working Paper No. 04 - 14
5.
Ranciere, R., A. Tornell and F. Westermann (2003a), „Crises and Growth: A Re-Evaluation”, NBER
Working Paper
6.
(2006b), „Decomposing the Effects of Financial Liberalization: Crises vs. Growth”, Journal of Banking
and Finance
7.
(2007c) „Systemic Crises and Growth”, Quarterly Journal of Economics
8.
Chinn, M. D. (2002), „The Measurement of Real Effective Exchange Rates: A Survey and Applications
to East Asia” NBER Working Paper
9.
Kaminsky, G. S. Lizondo and C.M. Reinhart (1998), „Leading Indicators of Currency Crisis”, IMF
Staff Project
10.
Braun, M. and C. Raddatz (2006), „Trade liberalization, Capital Account Liberalization and the Real
Effects of Financial Development”, Journal of International Money and Finance
11.
Li, K. and N.R. Prabhala (2005), „Self-Selection Models in Corporate Finance”, Robert H. Smith
School Research Paper No. RHS 06 - 020
12.
Manning, A. (2004), „Instrumental Variables for Binary Treatments with Heterogeneous Treatment
Effects: A Simple Exposition”, The Berkeley Electronic Press
VI. REFERENCES
13.
Kaminski, G. and C. M Reinhart (1999), „The Twin Crises: The Causes of Banking and Balance-ofPayments Problems”, The American Economic Review
14.
Ergungor, E. and J. B. Thomson (2005), „Systemic Banking Crises”, Federal Reserve Bank of
Cleveland, Policy Discussion Paper
15.
Bekaert, G., C. R. Harvey and C. T. Lundblad (2003), „Equity Market Liberalization in Emerging
Markets”, Federal Reserve Bank of St. Louis
16.
De Souza, L. V. (2004), „Financial Liberalization and Business Cycles: The Experience of Countries in
the Baltics and Central Eastern Europe”, Deuche Bundesbank Discussion Paper
17.
Goldstein, M., G. L. Kaminski and C. M. Reinhart (2000), „Assessing Financial Vulnerability”, pag.
11 – 32., Institute for International Economics
18.
Bordo, M., B. Eichengreen, D. Klingebiel and M.S. Martinez-Peria (2000), „Is the crisis problem
growing more severe?”, Blackwell Publishing Economic Policy
19.
Hutchinson, M. M. and I. Neuberger (2002), „How Bad are the Twins? Output Costs of Currency and
Banking Crisis”, Federal Reserve Bank of San Francisco Working Paper No. 02 – 02.
20.
Shatz, H. J. and D. G. Tarr (2000) „Exchange Rate Overvaluation and Trade Protection: Lessons from
Experience”, World Bank Policy Research Working Paper No. 2289
21.
Durlauf, S. N. and D. T. Quah (1998), „The New Empirics of Economic Growth”, Handbook of
Macroeconomics
22.
Durlauf S. N., P. A. Johnson and J. R. W. Temple (2004), „Growth Econometrics”, Handbook of
Economic Growth
23.
Barro, R. J. and J. F. Ursua (2008), „Macroeconomic Crises Since 1870”, NBER Working Paper
24.
Kaminski, G. L. and S. L. Schmukler (2003), „Short-Run Pain, Long-Run Gain: The Effects of
Financial Liberalization”, NBER Working Paper No. 9787
VI. REFERENCES
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
Greene, W. (2002), „Convenient Estimators for the Panel Probit Model: Further Results”, Empirical
Economics, Vol. 29, No. 1
Martin, P. and H. Rey, (2005), "Globalization and Emerging Markets: With or without Crash",
American Economic Review
Dell’Ariccia, G., and R. Marquez, (2004), “Information and Bank Credit Allocation.” Journal of
Financial Economics
Demirguc-Kunt A and E. Detragiache, (1998), "Financial Liberalization and Financial Fragility," IMF
Working Paper 98/83.
http://www.duke.edu/~charvey/country_risk/chronology/chronology_index.htm
http://www.centralbankmalta.com/
http://www.cnb.cz/cnb/stat.arady_pkg.parametry_sestavy?p_csest=3000&p_ind=g&p_lang=en
http://www.centralbank.gov.cy/nqcontent.cfm?a_id=1&lang=en
http://www.bankofestonia.info/frontpage/en/%20
http://www.bank.lv/eng/main/all/statistics/bank_mon_stat/bmsq42006/
http://english.mnb.hu/engine.aspx?page=mnben_statisztikai_idosorok&contentid=7822
http://www.lb.lt/stat_pub/statbrowser.aspx?group=7273&lang=en
http://www.nbp.pl/
http://www.nso.gov.mt/
http://www.statistics.sk/webdata/english/index2_a.htm
http://www.nbs.sk/indexa.htm
http://www.bsi.si/iskalniki/letna_porocila_en.asp?mapaid=711
http://www.bnro.ro
https://statistici.insse.ro/shop/
http://ec.europa.eu/economy_finance/indicators/annual_macro_economic_database/ameco_en.htm
http://www.tcmb.gov.tr/yeni/eng/
THANK YOU FOR YOUR
ATTENTION !