J. of Multi. Fin. Manag.

21 (2011) 40–54

Contents lists available at ScienceDirect

Journal of Multinational Financial

Management
journal homepage: www.elsevier.com/locate/econbase

Financial integration and portfolio investments to

emerging Balkan equity markets
Theodore Syriopoulos ∗
Department of Shipping, Trade &Transport, School of Business Sciences, University of the Aegean, 2A Korai str., 82100 Chios, Greece

a r t i c l e i n f o a b s t r a c t

Article history: The study investigates the risk and return profile of interna-
Received 27 February 2009 tional portfolios allocated by investors to major Balkan equity
Accepted 14 December 2010
markets, namely Romania, Bulgaria, Croatia, Turkey, Cyprus and
Available online 29 December 2010
Greece against developed markets, Germany and the US. An
error-correction vector autoregressive framework models finan-
JEL classification:
cial integration and investigates causality effects and cointegration
C51
vectors, depicting short- and long-run dynamic linkages. The
G11
G15 empirical findings support the presence of two cointegration vec-
F36 tors, indicating a stationary long-run relationship. Both domestic
O16 and external forces affect equity market behavior, leading to a
long-run equilibrium. These findings are important for interna-
Keywords:
tional asset allocation, since long-run comovements imply that risk
Market comovements
diversification and attainment of superior portfolio returns in the
Dynamic cointegration
Causality effects Balkan equity markets may be limited for international investors,
Portfolio diversification although short-run benefits may be potentially feasible in arbitrage
Balkan equity markets mispricings.
© 2010 Elsevier B.V. All rights reserved.

1. Introduction

The fifth European Union (EU) enlargement phase (January 2007) creates a dynamic and unique
economic landscape in Euroland. The efficient economic integration of the newcoming Member States
with the developed mature economies has critical implications for the long-term growth prospects
of the Eurozone and particularly for the Balkan and South-Eastern European economies. The Balkan

∗ Tel.: +30 22710 35861; fax: +30 22710 35299.

E-mail address: tsiriop@aegean.gr

1042-444X/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.mulfin.2010.12.006
 T. Syriopoulos / J. of Multi. Fin. Manag. 21 (2011) 40–54 41

Table 1
Foreign direct investment inflows (USD mln.)a .

2000 2001 2002 2003 2004 2005–2009b

c
Balkans 3617 4461 4215 8531 9934 7070
Central Europed 19,289 16,548 21,994 10,142 14,671 16,870
Baltics 1179 1120 1600 1370 2489 1560

Source: National statistics; IMF.

a
FDI inflows refer to the change in inward direct investment liabilities (new equity investment, reinvested earnings and
intra-company loans minus any disinvestment/decrease in FDI liabilities).
b
Annual average estimation.
c
Bulgaria, Romania, Croatia, Albania, FYROM, Serbia and Montenegro, Bosnia and Herzegovina.
d
Czech Republic, Hungary, Poland, Slovakia, Slovenia.

capital markets can support the convergence process by facilitating investment fund mobility and
portfolio allocation to the corporate sector.
International portfolio diversification can lead to efficient asset allocation and reduce investment
risk. However, in integrated markets, assets associated with similar levels of risk are anticipated to
have similar levels of return. The presence of common trends between Balkan and mature equity mar-
kets may indicate limited portfolio gains from diversification because the presence of common factors
limits the amount of independent variation. Cointegrated markets exhibit comovements and a sta-
ble long-run behavior, although potential short-run profits should not be ruled out. High correlations
alone are not sufficient to ensure long-run performance, as common long-term trends in prices are not
modelled (Alexander, 2001). Depending on their level of cointegration, the Balkan equity markets can
offer profitable opportunities to international investors’ diversified portfolios. Long-run equity market
comovements can be associated to a number of reasons, including economic ties and policy coordina-
tion, financial innovations and market deregulation, interest rate movements or financial crises and
contagion effects (e.g. Chan et al., 1997; Gelos and Sahay, 2000; Chen et al., 2002; Ratanapakorn and
Sharma, 2002; Samitas et al., 2007; Syriopoulos, 2004, 2006, 2007).
The motivation for this study derives from the growing economic performance of the Balkans
economies over the last years, in terms of income per capita, international competitiveness and foreign
direct investment (FDI) allocation (Kekic, 2005). To access EU, the Balkan countries have promoted
major reform policies, including macroeconomic stabilization, market liberalization, restructuring and
privatizing state-owned corporations (IMF, 2005). However, despite these developments, empirical
studies focusing on the assessment of the Balkan equity markets’ risk-return profile remain limited.
Past financial research in dynamic portfolio allocation, market linkages and cointegrated trends has
concentrated predominantly on mature equity markets; a range of empirical findings, however, has
been rather inconsistent.
The purpose of this study is to expand empirical research on the pattern of dynamic comovements
and linkages between Balkan and international mature equity markets, exploring market reaction to
external shocks and assessing potential portfolio diversification benefits. This exercise in considered
timely and interesting, since the participation of the Balkan economies in the EU has been associated
with substantial amounts in FDIs (Table 1). International investors prefer fund allocation to markets
that exhibit a stable economic environment, greater openness to trade and a smooth transition path
(Clausing and Dorobantu, 2005).
A carefully selected sample of new, old and prospective EU members has been selected for the
study; namely, Romania, Bulgaria, Cyprus, Greece, Turkey and Croatia. These emerging equity mar-
kets are examined and compared against representative mature markets, Germany and the US.
Different equity market patterns are analyzed, potential dynamic linkages and interdependencies
are assessed and cointegrating vectors and lead-lag relationships are investigated. The presence of
any cointegrating vector in the emerging Balkan equity markets would justify possible comovements
between the Balkan and developed equity markets. An eight-dimensional vector error correction
model (VECM) is then employed to test for the temporal causal dynamics in the Granger frame-
work and to gain insight into lead-lag relationships of the sample markets. The dynamic structure
of international market integration based on innovation accounting is also studied in order to iden-
42 T. Syriopoulos / J. of Multi. Fin. Manag. 21 (2011) 40–54

tify whether the effect of external shocks is permanent or transitory in the equity markets under
study.
The rest of the paper is organized as follows. Section 2 presents a concise literature review, an
analysis of the data employed and a range of descriptive statistics. Section 3 analyses the modeling
methodology and evaluates the empirical findings. Finally, Section 4 concludes.

2. Literature, data and statistics

2.1. Literature review

Past empirical research in financial integration of international equity markets investigates a diver-
sity of issues, such as short- and long-run interdependencies, impact of financial crises and external
shock transmission, macroeconomic factors, regional cooperation and common policies as well as bear
and bull market behavior. In most cases, empirical evidence of long-run linkages, comovements and
interrelationships was concluded (Syriopoulos, 2004, 2006, 2007).
A number of past studies focus on major European stock markets (e.g. Karolyi and Stulz, 1996;
Koutmos, 1996; Alexakis et al., 1997; Dickinson, 2000; Ejara, 2001; Phylaktis and Ravazzolo, 2002;
Bessler and Yang, 2003; Yang et al., 2003a,b). A body of studies investigates the Asian and Pacific stock
markets and provides evidence of long-run interdependencies (e.g. Corhay et al., 1995; Janakiramanan
and Lamba, 1998; Dekker et al., 2001; Sharma and Wongbangpo, 2002; Yang et al., 2003a,b). A large
number of studies examines stock market cointegration across regions (e.g. Masih and Masih, 1997;
Francis and Leachman, 1998; Chen et al., 2002; Chaudhuri and Wu, 2003; Swanson, 2003). Past empir-
ical findings, however, have not been consistent in all studies, as absence or weak cointegration was
concluded in some cases (e.g. for European markets: Booth et al., 1997; Pynnonen and Knif, 1998;
Gerrits and Yuce, 1999, for world markets: Choudhry, 1994; Kwan et al., 1995; Defusco et al., 1996;
Huang et al., 2000). However, the behavior of emerging markets and their linkages with mature mar-
kets, including the Balkan equity markets, has been rather neglected and empirical findings remain
thin and sometimes inconclusive (Syriopoulos, 2004, 2006, 2007).
Of the few studies focusing on emerging markets, Samitas et al. (2006, 2007) examine major Balkan
stock market dynamics (Romania, Bulgaria, Serbia, FYROM, Turkey, Croatia, Albania) in relation to
mature stock markets (US, UK, Germany, Greece). The study employs linear (error correction vector
autoregressive model) and non-linear (switching regime error correction model) methods to test for
potential linkages between Balkans and developed stock markets. The empirical findings indicate
that the Balkan markets display equilibrium relationships with their mature counterparts; reveal the
presence of interdependencies between emerging and developed markets that limit potential portfolio
diversification benefits in the sample Balkan equity markets; and support active rather than passive
portfolio strategies to attain exceptional returns for international investors.

2.2. Data analysis

The Balkan equity markets have a short active presence compared to major mature peer markets,
such as Germany and the US and a brief historical overview now follows. The Bucharest Stock Exchange
(BVB), Romania, was re-opened on April 1995; the BET Index (BET-C) was launched in 1997. The
Bulgarian Stock Exchange (BSE) was established, as in its present form, in 1995, following the merger
of nearly twenty regional exchanges; the BSE-Sofia Index (SOFIX) was launched on October 2000.
The Istanbul Stock Exchange (ISE), Turkey, was established in 1986; the ISE-National-100 Index (ISE-
100) is the representative stock market index. The Zagreb Stock Exchange, Croatia, was established in
1991; the CROBEX share index (CROBEX) was set in 1997. The Cyprus Stock Exchange (CSE) started its
operation on March 1996; the Cyprus-General Index (CYP-GI) is the main stock market index. Finally,
the Athens Stock Exchange (ASE), Greece has exhibited robust growth rates over the past decade; the
ASE-General Index (GR-GI) is the representative stock market index (Syriopoulos et al., 2006).
Table 2 summarizes key stock market figures including capitalization and stock index returns of
major Balkans and established mature equity markets; namely, Romania, Bulgaria, Croatia, Turkey,
Cyprus and Greece against Germany (Frankfurt Stock Exchange – DAX index) and the US (New York
 T. Syriopoulos / J. of Multi. Fin. Manag. 21 (2011) 40–54 43

Table 2
Key equity market figures.

2001 2002 2003 2004 2005 2006 2007

Number of listed companies

US 2798 2783 2591 2642 2707 2764 2805
Germany 984 934 866 818 764 760 754
Greece 349 347 355 360 356 358 294
Cyprus 144 154 152 149 144 141 141
Turkey 278 262 265 275 282 291 292
Romania 65 65 62 60 64 58 59
Croatia 66 73 175 183 194 202 383
Bulgaria 30 31 31 31 31 27 20
Market capitalization (mln. D )
US 17,863,173 12,684,781 13,654,048 9,942,496 12,589,925 12,836,232 11,989,248
Germany 1,021,320 350,321 500,112 550,458 706,660 858,250 933,016
Greece 96,949 65,749 84,547 92,137 123,033 157,928 195,502
Cyprus 9647 7633 7572 7876 10,881 17,047 20,160
Turkey 53,836 32,822 54,974 71,897 137,480 124,141 197,431
Romania 1427 2686 3035 8859 15,178 21,682 24,600
Croatia 3630 3926 5031 8281 10,978 22,024 47,945
Bulgaria 0565 0700 1391 2046 4301 7760 12,720
Turnover (mln. D )
US 11,841,370 9,837,915 7,721,855 8,546,259 12,264,994 12,992,500 14,975,200
Germany 1,014,089 607,616 429,886 449,298 517,196 673,054 822,620
Greece 41,447 24,771 34,887 35,639 52,487 85,339 121,279
Cyprus 3,839 646 263 226 397 2934 3900
Turkey 90,763 67,508 79,801 108,319 170,368 227,128 199,567
Romania 140 208 250 626 2,114 2925 4152
Croatia 136 162 202 355 643 1424 3062
Bulgaria 1.4 7.3 31 243 189 433 959
Market return (%)
US −11.86 −25.50 19.88 8.92 2.95 13.61 6.45
Germany −19.03 −54.21 26.17 5.74 23.95 21.97 22.2
Greece −25.58 −40.72 23.51 22.35 27.95 19.93 17.85
Cyprus −67.35 −32.18 −16.13 −9.96 51.63 128.79 23.59
Turkey 39.99 −31.36 55.79 31.34 46.15 −1.65 41.97
Romania −4.84 126.95 26.04 103.50 38.22 28.49 32.64
Croatia 17.93 11.57 0.72 24.79 27.29 60.70 63.23
Bulgaria 10.62 42.47 91.74 32.50 25.77 48.20 44.42

Source: National stock exchanges. Figures as of respective year-end.

Stock Exchange, NYSE – S&P500 index). The inclusion of Germany and the US in the dynamic mod-
eling of the Balkan equity market behavior is important for two reasons. First, these markets serve
as reasonable proxies for the mature European and North American equity markets respectively in
depicting possible linkages with the emerging Balkan markets. Second, due to significant interna-
tional investment flows and transactions turnover dominated by these developed markets, the US
and Germany are expected to play an influential role in international stock market comovements and
volatility transmission effects including the Balkan equity markets.
Compared to world developed markets, the Balkans equity markets remain thin in terms of listed
companies, market capitalization and turnover values. According to all measurements and as antici-
pated, the NYSE and the Frankfurt Stock Exchange are the largest established equity markets, whereas
Bulgaria is seen to be the smallest Balkan stock market in the sample under study. Following the
implementation of extensive economic reforms, the Balkan economies have embarked upon a robust
economic growth path over the last five years, leading to their accession into the EU. This improved
economic environment has attracted international investor funds at increasing numbers and has pro-
duced exceptionally high portfolio returns in the Balkan equity markets during 2001–2007, albeit
with apparent intense volatility. Over that period, cumulative returns in the Romanian stock market
were higher than 350% (ranging though from a low −4.84% in 2001 to a high 127% in 2002); in the
Bulgarian and Croatian stock markets, higher than 290% and 200%, respectively (ranging from 10.6%
44 T. Syriopoulos / J. of Multi. Fin. Manag. 21 (2011) 40–54

Table 3
Equity Market Descriptive Statistics.

Levels (logs) S&P500 DAX GR-GI CYP-GI ISE-100 BET-C CROBEX SOFIX

Mean 7.0863 8.5189 8.0199 5.0100 9.5711 7.1618 7.1534 5.9301

Median 7.0984 8.5477 8.0162 4.7379 9.5184 7.0123 7.0400 6.1365
Maximum 7.3476 8.9987 8.7533 6.7444 10.9264 8.9132 8.5359 7.3945
Minimum 6.6853 7.7845 7.2911 4.2504 7.5456 6.0544 6.0663 4.2564
Std. Deviation 0.1485 0.2712 0.3317 0.6553 0.7857 0.9171 0.5477 0.9536
Skewness −0.5053 −0.4076 0.0517 0.8150 −0.3798 0.3894 0.8264 −0.2726
Kurtosis 2.7135 2.5623 2.0582 2.4527 2.6284 1.6286 3.0291 1.6080
Jarque- 22.578 17.515 18.361 60.490 14.629 50.888 55.911 33.709
Bera (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
(prob.)
Returns S&P500 DAX GR-GI CYP-GI ISE-100 BET-C CROBEX SOFIX

Mean 0.0005 0.0007 0.0014 0.0027 0.0050 0.0038 0.0033 0.0077

Median 0.0015 0.0042 0.0035 0.0001 0.0066 0.0055 0.0032 0.0048
Maximum 0.0749 0.1220 0.1573 0.3720 0.2857 0.1665 0.1808 0.2956
Minimum −0.1170 −0.1407 −0.1952 −0.2395 −0.3686 −0.1513 −0.2701 −0.1291
Std. Deviation 0.0234 0.0329 0.0430 0.0512 0.0666 0.0384 0.0375 0.0360
Skewness −0.4875 −0.3400 −0.4583 1.4604 −0.2338 −0.2618 −0.8276 2.0789
Kurtosis 5.7059 4.5360 6.2064 13.4768 6.1578 5.9294 11.0514 17.7435
Jarque- 168.899 57.613 227.062 2415.206 208.058 180.802 1379.454 3529.672
Bera (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
(prob.)

Notes: The indices above correspond to the following stock markets: S&P500: US; DAX: Germany; GR-GI: Greece; CYP-GI:
Cyprus; ISE-100: Turkey; SOFIX: Bulgaria; BET-C: Romania; CROBEX: Croatia.

in 2001 to 92% in 2003 in the former; and, from 0.72% in 2003 to 63% in 2007 in the latter); whereas,
in Cyprus and Greece, comparatively modest at 78% and 45%, respectively. In contrast, for that same
period, cumulative returns in the US and the German stock markets were standing at 14.5% and 26.8%,
respectively.
We employ weekly stock index closing prices of the sample markets in order to investigate the
issues of interest. The dataset is obtained from the Datastream base and was enriched with information
collected from the official stock market bureaus. The study period expands from April 27, 1998 to
September 10, 2007 (with the exception of the Bulgarian stock price index that starts on October
16, 2000). The stock market indices included in the study to represent emerging Balkan and mature
equity markets are: the BET-C of Romania; SOFIX of Bulgaria; CROBEX of Croatia; ISE-100 of Turkey;
GR-GI of Greece; CYP-GI of Cyprus; S&P500 of the US; and, DAX of Germany (Fig. 1). These stock
market indices are expressed in local currency terms (preliminary estimations with stock indices
converted into euros produced basically similar empirical outcomes). The stock price indices in local
currencies depict domestic market reactions to information from foreign markets (external shocks)
from the perspective of the domestic investors (Choudhry, 1994; Voronkova, 2004). Expressing stock
price indices in domestic currencies restricts their changes solely to stock price movements and avoids
potential distortions induced by exchange rate devaluations (as, for instance, in the case of the floating
Turkish lira in 2001, resulting to a 36% decline in two days; or, the more than 50% Romanian lei
devaluation against the US dollar in 1999).

2.3. Descriptive statistics

A number of descriptive statistics and tests are initially produced for the sample equity markets in
levels and returns (Table 3). The mean excess return is found to be higher for Bulgaria (SOFIX index)
and Turkey (ISE-100 index). As anticipated, the mean excess return is lower in the mature equity
markets (S&P500, DAX). The emerging Balkan equity markets exhibit higher volatility, as indicated by
larger standard deviation values. Most of the equity index series are negatively skewed (except from
Cyprus and Bulgaria), whereas kurtosis statistics indicate fatter tails for all series compared to normal
distribution. The Jarque–Bera test rejects normality in all cases.
 T. Syriopoulos / J. of Multi. Fin. Manag. 21 (2011) 40–54 45

S&P500 DAX
7.4
7.3 9
7.2 8.8
7.1 8.6
7 8.4
6.9 8.2
6.8 8
6.7 7.8
6.6 7.6
99 00 01 02 03 04 05 06 07 99 00 01 02 03 04 05 06 07

GR - GI CYP - GI
9 7
8.8
6.5
8.6
8.4 6
8.2
5.5
8
7.8 5
7.6
4.5
7.4
7.2 4
99 00 01 02 03 04 05 06 07 99 00 01 02 03 04 05 06 07

ISE - 100 BET - C

11
9
10.5
8.5
10
8
9.5
7.5
9
8.5 7

8 6.5

7.5 6

99 00 01 02 03 04 05 06 07 99 00 01 02 03 04 05 06 07

CROBEX SOFIX
9
7.5
8.5
7
8
6.5
7.5 6
7 5.5

6.5 5
4.5
6
4
99 00 01 02 03 04 05 06 07 01 02 03 04 05 06 07

Fig. 1. Equity market (log) plots.

Unconditional contemporaneous correlation coefficients in the sample equity markets indicate

strong pairwise correlations for the mature stock markets in some cases (e.g. US–Germany; US–Greece;
US–Cyprus; Germany–Greece; and, Germany–Cyprus). The Balkan equity markets exhibit relatively
modest pairwise correlations with the leading mature equity markets but robust ones with peer Balkan
markets (e.g. Greece–Cyprus; Romania–Turkey; Romania–Croatia; Turkey–Croatia; Bulgaria–Turkey;
Bulgaria–Romania; and, Bulgaria–Croatia). As an initial reaction, (static) correlation coefficients indi-
46 T. Syriopoulos / J. of Multi. Fin. Manag. 21 (2011) 40–54

Table 4
Unconditional correlation matrix (Returns).

S&P500 DAX GR-GI CYP-GI ISE-100 BET-C CROBEX SOFIX

S&P500 1.0000
DAX 0.7461 1.0000
GR-GI −0.0416 0.0307 1.0000
CYP-GI −0.0827 −0.0762 0.3144 1.0000
ISE-100 0.0342 0.0954 0.1970 0.1843 1.0000
BET − C −0.0496 −0.0494 0.1336 0.1821 0.1404 1.0000
CROBEX 0.0825 0.0697 0.2465 0.0537 0.2218 0.1227 1.0000
SOFIX 0.0344 −0.0261 0.0611 0.0190 0.0804 0.0691 0.1293 1.0000

Notes: Correspondence of equity markets and indices in Table 3.

cate that the Balkan equity markets, especially the late-comers in the EU, exhibit weak short-term
contemporaneous interactions with their mature counterparts (Table 4).

3. Empirical methodology and findings

Two distinct types of links between equity markets can be seen; the first one is typically related to
the ‘volatile behavior’ of stock indices and the second is related to the ‘trending behavior’ of the series.
In the first case, the models are based on estimating time-varying volatilities to investigate stock price
linkages and the transmission of shocks from one market to another. In the latter case, emphasis is
placed on joint tests of market integration and comovements in the long-run. Interrelations and link-
ages among emerging Balkan equity markets and mature counterparts are studied here by estimation
and testing for the presence and number of cointegrating vectors. Cointegration relationships allow for
the description of stable long-run stationary relationships among integrated variables and are defined
as independent linear combinations of these non-stationary variables achieving stationarity. When
a meaningful interpretation can be attached to this linear combination, it implies that the series do
not drift apart and are moving together by some long-run equilibrium relationship. Individual non-
stationary time series in a multivariate system are driven by a reduced number of common stochastic
trends. A common stochastic trend in a system of equity markets can be interpreted to mean that
the stochastic trend in one individual equity market is related to the stochastic trend in some other
individual equity market.

3.1. Unit root tests

In order to test for the presence of stochastic non-stationarity in the data, the integration order of the
individual time series is investigated employing standard unit root tests, the Augmented Dickey–Fuller
(ADF) test, the Phillips and Perron (PP) test and the Kwiatiwski–Phillips–Schmidt–Shin (KPSS) test. The
ADF test (including a drift plus a time trend) is described as:


k

yt = ˛ + t + yt−1 + i yt−i + εt , εt ∼i.i.d.(0,  2 ) (1)

i=1

where  indicates first differences of the series, the term t represents a trend variable and εt is the
white noise term. The null hypothesis implies that  = 0 for the yt sequence to contain a unit root and
rejection of the null hypothesis implies a stationary series.
The PP test is an ADF test modification and is described as:

yt = ˛0 + ˛yt−1 + ut (2)

where ut is white noise. This test is robust to a wide range of serial correlation and time-dependent
heteroscedasticity and allows for weak dependence and heterogeneity in the error term.
The KPSS test differs from the previous tests, since the yt series is assumed to be stationary under
the null hypothesis; rejection of the null hypothesis indicates a non-stationary time-series. The KPSS
 T. Syriopoulos / J. of Multi. Fin. Manag. 21 (2011) 40–54 47

test uses a frequency zero spectrum estimation of the residuals in:

yt = ˛0 + ˇt + εt (3)

The Lagrange Multiplier (LM) statistic (2 distribution) is defined as:


T t 2
( r=1 εt )
LM = (4)
T 2 f0
t=1

where f0 is the estimator of the residual spectrum at frequency zero.

Table 5 summarizes the results produced by unit root tests in logarithms and first differences. The
ADF, PP and KPSS tests are estimated with and without trend on the first differences of (log) equity
market indices. According to the null hypothesis in the ADF and PP test, each of the time series contains
a unit root (i.e. testing the series as I(1) against I(0)); this should be rejected if the first differences of
(log) equity indices are stationary. The null hypothesis of a unit root is rejected at the 5% significance
level by both (ADF and PP) tests for each equity index series. These findings are also supported by the
KPSS test. The sample equity index returns are found to be stationary and the equity markets under
study are integrated of order I(1), as no unit root was detected in the first differences of the series.

3.2. Cointegrating vectors

It is widely established in empirical finance that if two or more variables (time series) are coin-
tegrated, then stationary linear combinations of these variables may exist even though the variables
themselves are individually non-stationary. The absence of cointegration suggests that the time series
involved have no long-run interdependence and can drift arbitrarily away from each other. As the null
hypothesis of unit roots cannot be rejected, multivariate models can be built to enable investigation
of short- and long-run linkages in the sample equity markets. Testing for the presence and number
of cointegrating vectors in the Balkan and mature equity markets under study jointly is based on a
vector error correction model (VECM), applying the procedure advanced by Johansen (Johansen, 1991;
Johansen and Juselius, 1990).
Defining a vector zt of n potentially endogenous non-stationary variables, it is possible to specify
the following data generating process and model zt as an unrestricted vector autoregression (VAR)
involving up to k-lags of z:

zt = A1 zt−1 + A2 zt−2 + · · · + Ak zt−k + ut ut ∼IN (0, ˙) (5)

where zt is a (n × 1) matrix and each of Ai is a (n × n) matrix of parameters. Eq. (5) can be reformulated
into a VECM form:
zt = 1 zt−1 + 2 zt−2 + · · · + k−1 zt−k+1 + ˘zt−k + ut or

k−1
(6)
zt = i zt−i + ˘zt−k + ut
i=1

where  i = −(I − A1 − · · · − Ai ), (i = 1,· · ·,k − 1),  i are interim multipliers and ˘ = −(I − A1 − · · · − Ak ). If
the coefficient matrix ˘ has reduced rank r < n, there exist (n × r) matrices ˛ and ˇ each with rank r
such that ˘ = ˛ˇ and ˇ zt is stationary. Testing for cointegration hence is related to the consideration
of the rank of ˘, that is finding the number of r linearly independent columns in ˘ (cointegrating
vectors). The hypothesis of the existence of r cointegrating vectors can be tested by the ‘trace’ test, i.e.
the LR test statistic that there are at most r distinct cointegrating vectors against a general alternative:


n

trace (r) = −2 log (Q ) = −T log (1 − li∧ ) (7)
i=r+1

where i = r + 1,. . ., n, are the (n − r) smallest squared canonical correlations, r = 0, 1, 2,. . ., n − 1 and
trace
(r) = 0, when all
i = 0. Alternatively, the ‘maximum eigenvalue’ test can be used to compare the null
 48
Table 5

T. Syriopoulos / J. of Multi. Fin. Manag. 21 (2011) 40–54

Unit root tests.

ADF PP KPSS

zt zt zt zt zt zt

(1) (2) (1) (2) (1) (2) (1) (2) (1) (2) (1) (2)

S&P500 −1.493 −1.537 −23.499 −23.485 −1.374 −1.418 −23.593 −23.582 0.452 0.455 0.148 0.103
DAX −0.940 −0.902 −22.147 −22.176 −0.940 −0.893 −22.148 −22.180 0.477 0.464 0.250 0.129
GR-GI −1.230 −1.265 −24.512 −24.502 −1.093 −1.129 −24.561 −24.555 0.466 0.468 0.203 0.174
CYP-GI −1.041 −1.012 −9.531 −9.528 −0.954 −0.914 −22.073 −22.062 0.401 0.345 0.262 0.232
ISE-100 −0.996 −2.238 −22.897 −22.874 −1.066 −2.458 −22.937 −22.915 2.311 0.162 0.043 0.043
BET − C 1.300 −4.272 −20.396 −20.723 0.915 −4.116 −20.763 −20.923 2.622 0.537 0.687 0.161
CROBEX 1.011 −2.342 −22.154 −22.324 0.895 −2.462 −22.213 −22.341 2.443 0.465 0.362 0.035
SOFIX −0.060 −1.935 −17.523 −17.503 −0.084 −1.978 −17.492 −17.472 2.303 0.344 0.119 0.123

Notes: ADF: Augmented Dickey–Fuller test; PP: Phillips–Perron test; KPSS: Kwiatkowski–Phillips–Schmidt–Shin test. zt : variables in logarithms; zt : variables in first differences. (1):
without trend; (2) with trend. ADF-PP critical values – without trend: −3.443 (1% level); −2.867 (5% level); −2.569 (10% level); ADF-PP critical values – with trend: −3.976 (1% level);
−3.419 (5% level); −3.132 (10% level); KPSS critical values – without trend: 0.739 (1% level); 0.463 (5% level); 0.347 (10% level); KPSS critical values – with trend: 0.216 (1% level); 0.146
(5% level); 0.119 (10% level).
 T. Syriopoulos / J. of Multi. Fin. Manag. 21 (2011) 40–54 49

Table 6
Tests for the presence of cointegrating vectors.

Null Eigenvalues
trace test Critical values at 95%

Model 1 Model 2 Model 1 Model 2 Model 1 Model 2

r=0 0.169914 0.171255 212.1689 236.8555 156.00 182.82

r≤1 0.109543 0.124906 145.5000 169.6079 124.24 146.76
r≤2 0.091990 0.099412 103.9647 121.8421 94.15 114.90
r≤3 0.069177 0.076744 69.41760 84.35687 68.52 87.31
r≤4 0.051755 0.054385 43.75413 55.77094 47.21 62.99
r≤5 0.039965 0.040020 24.72915 35.75182 29.68 42.44
r≤6 0.027893 0.030494 10.12790 21.13019 15.41 25.32
r≤7 0.000007 0.027664 0.000262 10.04336 3.76 12.25

Notes: H1 (r) against H1 (n). Model 1: model with unrestricted constant; Model 2: model with a linear trend in the cointegration
vector; critical values from Ostenwald-Lenum (1992).

hypothesis of r cointegrating vectors against the alternative of (r + 1) cointegrating vectors. The LR test
statistic for this hypothesis is given by:
∧

max (r, r + 1) = −2 log (Q ) = −T log(1 − lr+1 ) (8)

where r = 0, 1, 2,. . ., n − 1.
Two alternative models are compared and contrasted for the Balkan and developed equity markets
of interest: a model with unrestricted constant term (model 1); and, a model with a linear trend
in the cointegration vector (model 2). The fact that all equity markets are integrated of order I(1)
supports the application of the Johansen cointegration methodology in order to examine the long-run
relationship between the stock market indices. The results of the
trace and
max tests for the two
models are presented in Tables 6 and 7, respectively. According to the
trace test, the null hypothesis
that the equity markets are not cointegrated (r = 0) is rejected for both models estimated. However,
the
trace test indicates that there are at most two cointegrating vectors in model 1, while at most
three cointegrating vectors in model 2. The results obtained from the
max test indicate that there is
one cointegrating vector in both models. In case these two tests do not produce similar findings, it
has been proposed that the indications of the
trace test should be preferred (Johansen and Juselius,
1990). The
trace test has been argued to provide a more power result than the
max test, as it takes
into account all n − r values of the smallest eigenvalues (Kasa, 1992; Serletis and King, 1997). Overall,
the larger the number of cointegration vectors the higher the probability that a long-run relationship
is present in the equity markets under study.
Since our empirical findings indicate that there are more than one cointegration vectors, comove-
ments and linkages are expected for the Balkan and mature equity markets. Although these markets
may follow a diverging path in the short-run, a robust long-term (statistical) relationship is estab-
lished. For an international investor, these findings may indicate that potential long-term benefits

Table 7
Tests for the number of cointegrating vectors.

n−r
max test Critical values at 95%

Model 1 Model 2 Model 1 Model 2

r=0 66.668 67.247 51.42 55.50

r=1 41.535 47.765 45.28 49.42
r=2 34.547 37.485 39.37 43.97
r=3 25.663 28.585 33.46 37.52
r=4 19.024 20.019 27.07 31.46
r=5 14.601 14.621 20.97 25.54
r=6 10.127 11.086 14.07 18.96
r=7 0.003 10.043 3.76 12.25

Notes: H1 (r) against H1 (r + 1). Model 1: model with unrestricted constant; Model 2: model with a linear trend in the cointegration
vector; critical values from Ostenwald-Lenum (1992).
50 T. Syriopoulos / J. of Multi. Fin. Manag. 21 (2011) 40–54

associated with portfolio diversification to the Balkan equity markets may be rather limited. Never-
theless, potential short-term gains should not be ruled out. Taking into account the two cointegration
vectors indentified in the sample markets, we can, furthermore, examine the influence of the two
mature markets (Germany and the US) on the Balkan equity markets. Table 8 summarizes the two
normalized cointegration vectors obtained from models 1 and 2. The cointegration vectors indicate
that the US equity market exerts a long-run impact on Greece and Croatia, while the German equity
market affects Greece, Cyprus, Turkey, Romania and Bulgaria. These findings appear to be reasonable
given the leading role of the US and Germany in world equity markets, related to market size and
depth, trading volumes and liquidity, and are in line with past empirical findings (Ratanapakorn and
Sharma, 2002; Syriopoulos, 2006, 2007).

3.3. Error correction and Granger causality

The presence of cointegrating vectors in the sample equity markets supports the employment of a
dynamic vector error correction (VEC) model to examine short- and long-tem behavior and linkages of
emerging Balkan and developed equity markets. The VECM depicts the feedback process and adjust-
ment speed of short-run deviations towards the long-run equilibrium path and reveals the short-run
(uni- or bi-directional) Granger causalities (causal flows) and lead-lag relationships in any Balkan
equity market relative to the other markets. For two cointegrated (equity market) series xt and yt , the
error correction mechanism (ECM) can take the form:


m1 
m2

xt = ˛1 + ˇ1i xt−i + ˇ2i yt−i + 1 zt−1 + u1t (9a)

i=1 i=1


m3

m4
yt = ˛2 + ˇ3i xt−i + ˇ4i yt−i + 2 zt−1 + u2t (9b)
i=1 i=1

The magnitude of the coefficients  1 and  2 determines the speed of adjustment back to the long-
run equilibrium following a market shock. When these coefficients are large, adjustment is quick, so z
will be highly stationary and reversion to the long-run equilibrium will be rapid. Table 9 summarizes
the results obtained from the VECM model.
The implementation of the VEC model allows for two channels of causality between the equity
markets. The first one is the lagged values of zS&P500 , zDAX , zGR-GI , zCYP-GI , zISE-100 , zBET-C ,
zCROBEX and zSOFIX and the other one is the error correction term (ECT) of the cointegration vector.
The joint significance of the equity market lags is tested by the F-test statistic, whereas the lagged error
correction term coefficient is tested by the relevant t-test statistic. The VECM lag length is chosen on the
basis of (Akaike and Schwarz) information criteria. We first examine the error correction term in order
to gain insight into the adjustment process of the equity markets towards equilibrium (Table 9, Panel
A). Since the ECT is not found to be statistically significant in the two developed markets (Germany
and the US), it can be concluded that these markets are statistically exogenous to the system and
follow an autonomous path. On the contrary, the Balkan equity markets can be potentially affected
by short-run effects in the developed markets, since the respective ECTs are found to be statistically
significant in all Balkan equity markets (excluding Bulgaria). The developed equity markets appear to
have a dominant role on the trends that the other markets will follow in the long-run. This implies that
short-run dynamics revert back the emerging equity markets that deviate from the common trend.
The larger ECT coefficient for the Greek equity market indicates a more rapid adjustment compared to
the other equity markets; this may be related to the fact that the Greek market has been upgraded from
an emerging to a newly developed market recently and is closely correlated to the mature markets.
The direction and impact of any causal relationship and lead-lag effects between the emerging
Balkan and developed equity markets can be investigated by Granger-causality tests (Table 9, Panel
B). The joint significance of the lagged cross-market returns in the eight-equation system is tested
with relevant F-test statistics. The core empirical findings indicate that the Balkans equity markets
are influenced by changes in both mature equity markets. These (statistically) causal relationships are
Table 8

T. Syriopoulos / J. of Multi. Fin. Manag. 21 (2011) 40–54

Normalized cointegrating vector.

S&P500 DAX GR-GI CYP-GI ISE-100 BET − C CROBEX SOFIX C TREND

Model 1
1.0000 0.0000 −0.6761 0.0094 0.0786 −0.0428 0.0986 −0.0073 −2.8656
(0.067) (0.032) (0.056) (0.048) (0.048) (0.037)
[−10.091] [0.293] [1.403] [−0.891] [2.054] [−0.197]
0.0000 1.0000 −0.7795 −0.0924 −0.3971 0.1223 0.0270 0.1588 0.0422
(0.098) (0.047) (0.083) (0.071) (0.070) (0.055)
[−7.954] [−1.966] [−4.784] [1.722] [0.385] [2.887]
Model 2
1.0000 0.0000 −9.1970 2.2974 8.7028 0.9913 1.3630 0.9958 −37.186 −0.0549
(1.688) (0.719) (1.369) (1.176) (1.478) (1.014) (0.014)
[−5.448] [3.195] [6.357] [0.842] [0.922] [0.982] [−3.921]
0.0000 1.0000 14.802 −4.2665 −16.149 −1.7683 −2.3386 −1.7146 63.045 0.1009
(3.090) (1.316) (2.507) (2.152) (2.706) (1.856) (0.027)
[4.790] [−3.242] [−6.441] [−0.821] [−0.864] [−0.923] [3.737]

Notes: Figures in () and [] are standard errors and t-statistic, respectively. Model 1: model with unrestricted constant; Model 2: model with a linear trend in the cointegration vector.

51
52 T. Syriopoulos / J. of Multi. Fin. Manag. 21 (2011) 40–54

Table 9
VECM estimates.

zS&P500 zDAX zGR-GI zCYP-GI zISE-100 zBET-C zCROBEX zSOFIX

Panel A. Error correction term (ECT) estimates

ECT −0.0124 0.0010 0.0730 −0.0638 0.0691 0.0674 −0.0328 0.0232
(0.013) (0.020) (0.017) (0.024) (0.037) (0.021) (0.016) (0.023)
[−0.900] [0.050] [4.179] [−2.619] [1.851] [3.126] [−1.967] [1.006]

F-statistic F-statistic

Panel B. Granger causality test

S&P500 → DAX 0.97345 {0.404} DAX → S&P500 2.1134c {0.097}
S&P500 → GR-GI 45.6762a {0.000} GR-GI → S&P500 5.8098a {0.000}
S&P500 → CYP–GI 11.1137a {0.000} CYP–GI → S&P500 1.1231 {0.339}
S&P500 → ISE-100 14.2795a {0.000} ISE-100 → S&P500 0.5797 {0.628}
S&P500 → BET-C 3.0061b {0.030} BET-C → S&P500 1.0886 {0.353}
S&P500 → CROBEX 15.9810a {0.000} CROBEX → S&P500 2.1011c {0.099}
S&P500 → SOFIX 0.0473 {0.986} SOFIX → S&P500 1.2080 {0.306}
DAX → GR-GI 44.5898a {0.000} GR-GI → DAX 2.2066c {0.086}
DAX → CYP–GI 11.9441a {0.000} CYP–GI → DAX 1.3270 {0.264}
DAX → ISE-100 15.3069a {0.000} ISE-100 → DAX 1.6160 {0.184}
DAX → BET-C 4.1068a {0.006} BET-C → DAX 0.4535 {0.714}
DAX → CROBEX 12.3552a {0.000} CROBEX → DAX 1.3895 {0.245}
DAX → SOFIX 0.0890 {0.966} SOFIX → DAX 1.5249 {0.207}
GR–GI → CYP–GI 3.2629b {0.021} CYP–GI → GR–GI 5.4463a {0.001}
GR – GI → ISE-100 1.0833 {0.355} ISE-100 → GR–GI 1.2797 {0.280}
GR – GI → BET-C 4.0318a {0.007} BET-C → GR–GI 2.7958b {0.039}
GR – GI → CROBEX 0.7581 {0.517} CROBEX → GR–GI 2.3620c {0.070}
GR – GI → SOFIX 0.4918 {0.688} SOFIX → GR–GI 0.1970 {0.898}
CYP – GI → ISE-100 2.2779c {0.078} ISE-100 → CYP–GI 0.3802 {0.767}
CYP – GI → BET-C 0.5653 {0.638} BET-C → CYP–GI 0.5197 {0.668}
CYP – GI → CROBEX 0.2768 {0.842} CROBEX → CYP–GI 0.7671 {0.512}
CYP – GI → SOFIX 2.6225b {0.050} SOFIX → CYP–GI 0.2077 {0.891}
ISE-100 → BET-C 0.8353 {0.474} BET-C → ISE-100 0.2632 {0.851}
ISE-100 → CROBEX 3.1638b {0.024} CROBEX → ISE-100 0.6952 {0.555}
ISE-100 → SOFIX 3.0448b {0.028} SOFIX → ISE-100 1.0095 {0.388}
BET-C → CROBEX 2.8907b {0.035} CROBEX → BET-C 4.8994a {0.002}
BET-C → SOFIX 1.2038 {0.308} SOFIX → BET-C 1.5952 {0.190}
CROBEX → SOFIX 0.3308 {0.803} SOFIX → CROBEX 0.9894 {0.397}

Notes: Figures in (), [] and { } are standard errors, t-statistics and p-probability values, respectively; a, b, c indicate 1%, 5% and
10% significant level, respectively.

found to be unidirectional, since the Balkan equity markets are not seen to exert any significant impact
on the mature equity markets, except from Greece. The feedback from the Greek equity market can be
justified by this market’s increased level of internationalization, its leading role in the Balkan region
and the substantial portfolio fund flows from and to the other developed markets. Plausible causal
relationships are also indentified between the Balkans equity markets, as, for instance, in the case
of bidirectional short-run effects between Greece, Romania and Cyprus; the unidirectional effects
of Turkey on Bulgaria and Croatia; and, the interdependence of Romania and Croatia. Robust trade
transactions, direct foreign investments, portfolio asset allocation, cross-border corporate activity as
well as a rich historical and cultural background are reasons that can justify strong economic ties
between the Balkan markets.

3.4. Impulse response function

Impulse response analysis traces out the dynamic responsiveness of each equity market to shocks
that come from the other equity markets and indicates whether the impact of a shock is permanent
or transitory in each equity market. In the former case, the effect of an unanticipated shock to an
equity market does not die away quickly and, as a result, leads the group of the markets to a new
equilibrium (permanent shocks). In the latter case, the system returns to its previous equilibrium
 T. Syriopoulos / J. of Multi. Fin. Manag. 21 (2011) 40–54 53

(transitory effects). Hence, the pattern of the impulse response of each Balkan equity market to a
shock in the US and German stock markets is examined, in order to obtain additional insight into the
structure of the Balkan equity market linkages. For that, the impulse responses based on the simulated
responses of the estimated VAR system are produced. The plots of the time path of impulse responses of
each equity market to one standard deviation shock of the US and German stock indices are inspected
over an eight-week time horizon (impulse response graphs not shown).
As it becomes apparent, the US equity market responses fast to shocks in Germany, Greece and
Turkey, but the impact is rather limited. Shocks in the Balkan equity markets have only limited impact
in Germany. The smaller Balkan markets (in terms of capitalization) appear to be more sensitive to
shocks that come either from the developed markets or the other Balkan markets.

4. Conclusion

We have examined the dynamic interdependences, linkages and causality effects between major
Balkan equity markets, namely, Romania, Bulgaria, Croatia, Turkey, Cyprus and Greece and developed
equity markets, the US and Germany. The empirical findings enrich the thin body of the empirical
financial literature focusing on emerging economies and the Balkan equity markets in particular.
The motivation for this study arises from the need to investigate the short- and long-run dynamics
of equity markets that either have been or are prospective members of the European Union (EU).
Financial empirical literature in the emerging Balkan equity markets remains surprisingly thin.
Since a number of Balkan States have recently joined the EU, the examination of possible long-run
interdependencies and comovements with major international markets remains a crucial issue and
provides some indication of the overall economic convergence process. The empirical findings sup-
port the presence of cointegration vectors indicating a stationary long-run relationship between the
Balkan and mature equity markets. Both domestic and external forces affect the Balkan equity mar-
kets and shape their long-run equilibrium path. Most of the Balkan equity markets have a relatively
short active life and lack a substantial market depth (in terms of listed companies and capitaliza-
tion). However, there are growing inflows of international portfolio investments and trading activity
in the Balkan equity markets. The US, German and Greek equity markets appear more integrated, as
deviations from equilibrium are restored relatively fast. These markets also hold a leading role, since
they can induce strong movements in the Balkan markets but are not influenced by them. The pres-
ence of cointegrating relationships, however, has important implications for portfolio management.
Long-run comovements in the Balkan equity markets imply that potential risk diversification and
attainment of superior portfolio returns may be limited for international investors. However, there
are still interesting short-run opportunities for international asset allocation to the Balkan markets,
especially to the Romanian and Turkish equity markets that appear to have a smoother reaction to
external shocks. It appears then, that in the short-run, international investors may exploit arbitrage
mispricing opportunities, based on active portfolio management strategies.
To conclude, the Balkan markets follow a common path of growth and become gradually more
integrated with the international mature markets. These linkages are anticipated to strengthen in
the medium term, as the Balkan economies are newcomers in the EU. In order to attain economic
convergence, the Balkan States are bound to remove any remaining trade and investment barriers and
proceed to tighter policy coordination with the EU. These developments will eventually lead to new
challenges for the enlarged Euroland as a whole.

References

Alexakis, P., Apergis, N., Xanthakis, E., 1997. Integration of international capital markets: further evidence from EMS and non-EMS
membership. Journal of International Financial Markets, Institutions and Money 7, 277–287.
Alexander, C., 2001. Market Models: a Guide to Financial Data Analysis. John Wiley and Sons, New York.
Bessler, D.A., Yang, J., 2003. The structure of interdependence in international stock markets. Journal of International Money
and Finance 22, 261–287.
Booth, G.G., Martikainen, T., Tse, Y., 1997. Price and volatility spillovers in Scandinavian stock markets. Journal of Banking and
Finance 21, 811–823.
Chan, K.C., Gup, B.E., Pan, M.S., 1997. International stock market efficiency and integration: a study of eighteen nations. Journal
of Business Finance and Accounting 24, 803–813.
54 T. Syriopoulos / J. of Multi. Fin. Manag. 21 (2011) 40–54

Chaudhuri, K., Wu, Y., 2003. Random walk versus breaking trend in stock prices: Evidence from emerging markets. Journal of
Banking and Finance 27, 575–592.
Chen, G., Firth, M., Rui, O., 2002. Stock market linkages: evidence from Latin America. Journal of Banking and Finance 26,
1113–1141.
Choudhry, T., 1994. Stochastic trends and stock prices: an international inquiry. Applied Financial Economics 4, 383–390.
Clausing, K., Dorobantu, C., 2005. Re-entering Europe: does European Union candidacy boost foreign direct investment? Eco-
nomics of Transition 13, 77–103.
Corhay, A., Rad, A.T., Urbain, J.P., 1995. Long run behavior of Pacific-Basin stock prices. Applied Financial Economics 5, 11–18.
Defusco, R.A., Geppert, J.M., Tsetsekos, G.P., 1996. Long-run diversification potential in emerging stock markets. The Financial
Review 27, 289–295.
Dekker, A.K., Sen, K., Young, M., 2001. Equity market in the Asia Pacific region: a comparison of the orthogonalized and
generalized VAR approaches. Global Finance Journal 12, 1–33.
Dickinson, D.G., 2000. Stock market integration and macroeconomic fundamentals: an empirical analysis. Applied Financial
Economics 10, 261–276.
Ejara, D.D., 2001. Comovement of European equity markets after euro, Discussion Paper , 8th Annual Global Finance Conference,
Los Angeles, California.
Francis, B., Leachman, L., 1998. Superexogeneity and the dynamic linkages among international equity markets. Journal of
International Money and Finance 17, 475–492.
Gelos, G., Sahay, R., 2000. Financial market spillovers in transition economies. Working Paper, International Monetary Fund.
Gerrits, R.J., Yuce, A., 1999. Short- and long-term links among European and US stock markets. Applied Financial Economics 9,
1–9.
Huang, B.N., Yang, C.W., Hu, J.W.S., 2000. Causality and cointegration of stock markets among the United States, Japan and the
South China growth triangle. International Review of Financial Analysis 9, 281–297.
International Monetary Fund, 2005. Transition Economies: An IMF Perspective on Progress and Prospects.
www.imf.org/external/np/exr/ib/2000/110300.htm.
Janakiramanan, S, Lamba, A.S., 1998. An empirical examination of linkages between Pacific-Basin stock markets. Journal of
International Financial Markets, Institutions and Money 8, 155–173.
Johansen, S., 1991. Estimation and hypothesis testing of cointegration vectors in Gaussian vector autoregressive models. Econo-
metrica 59, 1551–1580.
Johansen, S., Juselius, K., 1990. Maximum likelihood estimation and inference on cointegration with application to the demand
for money. Oxford Bulletin of Economics and Statistics 52, 169–210.
Karolyi, G., Stulz, R., 1996. Why do markets move together? An investigation of US-Japan stock return comovements. Journal of
Finance 51, 951–986.
Kasa, K., 1992. Common stochastic trends in international stock markets. Journal of Monetary Economics 29, 95–124.
Kekic, L., 2005. Foreign direct investment in the Balkans: recent trends and prospects. Southeast European and Black Sea Studies
5, 171–190.
Koutmos, G., 1996. Modeling the dynamic interdependence of major European stock markets. Journal of Business Finance and
Accounting 23, 975–988.
Kwan, A.C., Sim, A.B., Cotsomitis, J.A., 1995. The causal relationships between equity indices on world exchanges. Applied
Economics 27, 33–37.
Masih, A.M., Masih, P., 1997. Dynamic linkages and the propagation mechanic driving major international markets: an analysis
of the pre- and post-crash era. Quarterly Review of Economics and Finance 37, 859–885.
Phylaktis, K., Ravazzolo, F., 2002. Stock market linkages in emerging markets: implications for international portfolio diversifi-
cation. Journal of International Financial Markets, Institutions and Money 15, 91–106.
Pynnonen, S., Knif, J., 1998. Common long-term and short-term price memory in two Scandinavian markets. Applied Financial
Economics 8, 257–265.
Ratanapakorn, O., Sharma, S.C., 2002. Interrelationships among regional stock indices. Review of Financial Economics 11, 91–108.
Samitas, A., Kenourgios, D., Paltalidis, N., 2006. Short and long run parametric dynamics in the Balkans stock markets. Interna-
tional Journal of Business, Management and Economics 2, 5–20.
Samitas, A., Kenourgios, D., Paltalidis, N., 2007. Integration and behavioural patterns in emerging Balkan stock markets. Working
Paper , EFMA Annual Conference, Vienna.
Serletis, A., King, M., 1997. Common stochastic trends and convergence of European Union stock markets. The Manchester
School 65, 44–57.
Sharma, S.C., Wongbangpo, P., 2002. Long term trends and cycles in ASEAN stock markets. Review of Financial Economics 111,
299–315.
Swanson, P.E., 2003. The interrelatedness of global equity markets, money markets, and foreign exchange markets. International
Review of Financial Analysis 12, 135–155.
Syriopoulos, T., 2004. International portfolio diversification to Central European stock markets. Applied Financial Economics
14, 1–16.
Syriopoulos, T., 2006. Risk and return implications from investing in emerging European stock markets. Journal of International
Financial Markets, Institutions and Money 16, 283–299.
Syriopoulos, T., 2007. Dynamic linkages between emerging European and developed stock markets: has the EMU any impact?
International Review of Financial Analysis 16, 41–60.
Syriopoulos, T., Athanassiou, E., Kollias, C., 2006. Dynamic volatility and external security related shocks: the case of the Athens
Stock Exchange. International Financial Markets, Institutions and Money 16, 411–424.
Voronkova, S, 2004. Equity market integration in Central European emerging markets: a cointegration analysis with shifting
regimes. International Review of Financial Analysis 13, 633–647.
Yang, J.J., Kolari, J.W., Min, I., 2003a. Stock market integration and financial crises: the case of Asia. Applied Financial Economics
13, 477–486.
Yang, J.J., Min, I., Li, Q., 2003b. European stock market integration: does EMU matter? Journal of Business Finance and Accounting
30, 1253–1276.