International Research Journal of Applied Finance ISSN 2229 – 6891

Vol. VIII Issue – 1 January, 2017 www.irjaf.com

Emerging Markets Stock Exchange Linkages: Evidence from BRICS and

MIST Countrıes

Bengü Vuran* Kaan Türkay

Abstract
Due to the fast acceleration of globalization, the existing relationship between world
economies acquired different dimensions recently. These dimensions refer to elimination of
barriers based on protectionist policies and the increase of liberalization process. As the
barriers between economies are removed, markets establish closer and more intense
relationship with one another. While this situation offers economies new opportunities it also
exposes the economies to certain risks.

This study is designed to investigate the relationship between stock exchanges of emerging
markets. Stock price indices of BRICS ( Brazil, Russia, India, China and South Africa) and
MIST (Mexico, Indonesia, South Korea and Turkey) countries are selected as the scope of the
research. Engle-Granger and Johansen Cointegration Tests and Granger Causality Tests are
performed for the daily data from 2010 to 2016 using E-Views. Engle-Granger test results
show that BIST 100 (Turkey) cointegrates individually with IBOVESPA (Brasil), MICEX
(Russia), NIFTY100 (India), SSE Comp. (China), JSE All Shr.(South Africa), IPC (Mexico),
JSE Comp. (Indonesia) ve KOSPI (South Korea) implying that investors do not have
diversification benefits between the markets. Where as Johansen test finds no cointegration
among all indices for the time period investigated. According to Granger Causility test results
there are evidences for granger causility from BIST 100 and other stock indices.

Keywords: BRICS, MIST, BIST 100, Johansen Cointegration Test, Emerging Markets,
Stock Indices, Engle-Granger Cointegration Test.

I. Introduction
The interrelationship among the world stock markets have gained much popularity in the era
of globalization. Both investors and academic scholars have been interested in examining the
world equity market linkages. Most studies examined stock market comovements of
developed countries. Few have concentrated on the linkages between emerging stock
markets. Emerging markets are developing economies, many of which are experiencing rapid
growth and industrialization. In recent years, emerging markets have attracted considerable
attention from investors due to their rising shares of global economic output and stock market
capitalization. One of the many attractions of emerging markets is that the risks are very
different from those in developed countries. Investors can actually reduce the risk of their
overall portfolio and increase long-term return by making investments with a range of
different risks that then cancel each other out making their returns uncorrelated.

In 2001, the acronym BRIC was offered by U.S. investment bank Goldman Sachs, to
designate the four major fast-growing economies – Brazil, Russia, India, and China. In 2011,
because of its robust economic growth, South Africa joined this group and the acronym was
expanded to BRICS.

In 2011, the growth of the BRIC countries had already begun to slow down, and investors
withdrew their funds from the BRIC countries. 2012 a new grup of emerging economies was
proposed: Mexico, Indonesia, South Korea, and Turkey, or MIST.

1
International Research Journal of Applied Finance ISSN 2229 – 6891
Vol. VIII Issue – 1 January, 2017 www.irjaf.com

The purpose of this study is to contribute to literature on emerging stock market integration
through whether stock market indices of BRICS and MIST countries are cointegrated with
each other. This study also aims to extend the Turkish literature by including BRICS and
MIST countries.

The remainder of this paper is organized as follows: the next section reviews the literature,
the third section describes data and presents the results of the emprical study and final section
summurizes the results and concludes the paper.

II. Literature Review

There is a broad literature on cointegration of emerging market. In this part of the paper, an
attempt has been made to provide information about the recent research works undertaken in
the area of emerging markets.

Khan (2011 ) examines the long-run relationship of the United States and 22 other developed
and developing countries stock indices using daily data for the period from 1999 to 2010.
Johansen (1988) and the Gregory and Hansen (1996) test to show that stock markets of most
countries have become cointegrated by 2010. But short-run diversification opportunities
across the countries by comparing their daily returns to the daily returns of the global index
(S&P 1200) have been detected. China, Malaysia and Austria are found to be the countries
with highly favorable diversification opportunities as they are not cointegrated with the US
and are insensitive to the global index.

Bhattacharjee and Swaminathan (2016) conduct an analysis of the stock market integration of
India and a few selected countries of the globe. The long run relationship of India with some
of the selected countries across the globe is analyzed over three different phases. It is
observed that cointegration of India with other stock markets is increasingly improving over
the years with financial liberalization. The study finds that the Indian stock market is more
responsive to the other Asian stock markets during the recession phase than in any other sub-
sample periods.

Horvath and Petrovski (2013) examines international stock market comovements between
Western Europe (Czech Republic, Hungary and Poland) and South Eastern Europe (Croatia,
Macedonia and Serbia) using multivariate GARCH models for the period of 2006–2011.
They compared these two groups and find that the degree of comovements is much higher for
Central Europe. The correlation of South Eastern European stock markets with developed
markets is found to be zero.

Lingaraja, Selvam and Vasanth (2014) conduct an analysis on inter-linkages, comovement

and causal relationship among emerging stock market indices returns and developed stock
market index in Asia. Using daily time series data for a period of 12 years, from January 1st,
2002 to December 31st ,2013, they find that in the long run, three countries namely India,
Taiwan and Philippines exercise the greatest influence on Singapore. The developed Asian
market (Singapore) also exercises influence on India during whole study period. While India
enjoys highly interlinked comovements with Singapore, out of remaining seven markets, only
Philippines and Taiwan record one way bidirectional causality relationship with Singapore.
The other five emerging Asian markets, China, Indonesia, Malaysia, Korea and Thailand do
not have Inter-linkages and comovements with Singapore during the study period. Also, the
five emerging markets China, Indonesia, Korea, Malaysia and Thailand record higher risk
than India, Taiwan and Philippines.

2
International Research Journal of Applied Finance ISSN 2229 – 6891
Vol. VIII Issue – 1 January, 2017 www.irjaf.com

Yarovaya and Lau (2016) investigate stock market comovements for UK, BRICS and MIST
countries around global financial crisis. In their study, the application of conventional and
regime-switch cointegration techniques suggests an absence of diversification benefits.
Further evidence from application of a multivariate time-varying asymmetric model suggests
that conditional correlation among the stock markets exhibits higher dependency when it is
driven by negative shocks to the market. The results indicate that the Chinese stock market is
the most attractive option for the UK investor.

Fahami (2011) aims to examine the structure of linkages and causal relationship between
leading indices of BRICS economies and the United States (US), the United Kingdom and
Japan. The period of analysis is divided into three sample periods as pre-crisis (January 10th,
2005 to July 22nd, 2007), crisis period (July 29th, 2007 to January 10th, 2010) and post crisis
period (January 11th, 2010 to July 21st, 2011). Johansen Juselius (JJ) cointegration test and
Granger causality test are applied and the results show that all the stock markets under study
are cointegrated in pre-crisis, during crisis and post crisis period. China was the most
influential stock market before the crisis period, whereas United States influenced most of the
major equity markets during the period of turbulence. These indicate that what happens to US
stock market brings impact to the other equity markets worldwide despite the emergence of
BRIC countries.

Chittedi (2009) conducts a study which examines the integration of the stock market among
the BRIC (Brazil, Russia, India and China) economies in general and their integration with
the developed countries stock markets such as US, UK and Japan, using the Granger
causality, Johansen cointegration and error correction mechanism methodology based on
daily data for the period January 1998-August 2009. The results indicate that BRIC countries
and USA, UK and Japan stock markets are highly cointegrated. The US and Japan market
factors are influencing Indian stock market. It might be because of maximum international
trade commercial activities between these countries. Indian stock market is not influenced by
UK, Brazil, Russia and China markets. But Brazil and Russia markets are influenced by
Indian stock market.

Gözbaşı (2010) examines the interaction between Istanbul Stock Exchange 100 index and the
stock markets of 7 developing countries (namely Argentina, Brazil, Mexico, India, Malaysia,
Hungary and Egypt) employing cointegration and causality analyses using weekly data
between December 1995 and December 2008. The obtained results demonstrate the long-
term relationship between ISE 100 and Brazilian, Indian and Egyptian stock markets, while it
also indicates the short-term interaction of the ISE 100 with Mexican and Hungarian stock
markets.

III. Data and Methodologyy

The aim of the study is to investigate the long run relationship between the BRICS (Brasil,
Russia, India, China and South Africa ) and MIST (Mexico, Indonesia, South Korea and
Turkey) Countries’ major stock market indices. The indices used in the study are Bovespa
(Brazil), MICEX (Russia), Shanghai Stock Exchange Composite Index (China), Nifty 100
(India), Jakarta Composite Index (Indonesia), Kospi Composite Index (South Korea),
Meksico IPC Index ( Meksico ), Johannesburg Stock Exchange All Share Index ( South
Africa ) and BIST 100 Index ( Turkey ). The list of indices are presented in Table 1.

Refer Table 1

3
International Research Journal of Applied Finance ISSN 2229 – 6891
Vol. VIII Issue – 1 January, 2017 www.irjaf.com

The data used in this study consist of the daily closing values of the stock market index of the
major exchange in each country. The daily adjusted closing prices of each of the sample
indices were collected from Yahoo Finance database and website of national stock exchange
of the countries and they are all expressed in terms of local currencies. The data set covers the
period January,11th, 2010 through July,30th 2016; 1200 observation are used for investigation.
In this period, the synchronized trading days have been taken into consideration. The daily
closing prices are transformed in natural logarithmic values. Engle-Granger and Johansen
Cointegration analysis and Granger Causility Test have been applied by using E-Views 8.

While graph 1 represents the individual series based on original daily closing data, graph 2
shows the natural logarithmic series of daily data.

Refer Graph 1
Refer Graph 2

Table 2 demonstrates the descriptive statistics of stock indice series with natural logarithmic
values. Jarque-Bera statictic shows that the null hypothesis of normality is rejected for all the
markets at 1 % significance level. This means that all series are not normally distributed.

Refer Table 2

In Table 3, the correlation coefficients of logarithmic values of stock indices are given. It can
be seen from the table that BIST 100 index is 46 % negatively correlated with IBOVESPA
(Brazil) whereas 18 % positively correlated with SSE Composite Index (China), 82 %
positively correlated with Jakarta Composite Index (Indonesia ), 82 % positively correlated
with JSE ALSI (South Africa) , 79 % positively correlated with NIFTY 100 Index (India), 58
% KOSPI Composite Index ( South Korea) , 34 % with MICEX (Russia) and 83 % with
Meksico IPC. Also, it can be concluded that there exists 90 % positive correlation between
Meksico and Jakarta, 87 % positive correlation between India and South Africa indices. BIST
100 index has very high correlations with Indonesia, India, Meksico and South Africa stock
indices with 82 %, 79 %, 83 % and 82 % respectively.

Refer Table 3

To implement Johansen approach to cointegration, the order of the integration of the indices
have to be established. To ascertain the stationarity properties of each of the stock indices
Augmented Dickey Fuller (ADF) and Philips and Perron (PP) unit root tests are applied. The
results for the Augmented Dickey-Fuller (ADF) and PP unit root tests in Table 4 suggest that
all the stock indices in their natural logarithm level are non-stationary series. In first
difference form, however, these stock price indices are stationary. Thus, all those indices are
first-order integrated series, or I(1) processes.

Table IV displays the results of unit root tests. The natural logarithims of stock prices follow
a non-stationary process in the level form but their first differences are stationary. This means
the null hypothesis for unit root for the I (1) series is rejected.

Refer Table 4

4
International Research Journal of Applied Finance ISSN 2229 – 6891
Vol. VIII Issue – 1 January, 2017 www.irjaf.com

Having established that each of the stock indices are I(1), the presence of unit root in the error
terms is tested. If the stock indices are integrated in the same order, the equation below is
estimated.
∆et = aet-1 + ∑ q+ ¢∆ej-t + vt (1)

The presence of unit root in error terms is tested by ADF test. If the error terms of the
regression is stationary, I(0), it can be concluded that the stock indices are cointegrated.

Refer Table 5

Table V gives the results of ADF test. It is understood that the error terms has no unit root,
I(0), so the stock indices have a long term equilibrium point implying a long-run equilibrium
relationship among variables.

Refer Table 6

Having established that all prerequisites have been provided to conduct the cointegration
analysis, to determine the lag lenght, the VAR analysis which is proposed by Hall (1991) has
been applied. The results are displayed in Table 6. The appropriate lag structure of the model
is determined by using Akaike Information Criterion (AIC) as 2.

The results of the cointegration analysis can be found in Table 7 and Table 8. Firstly, test for
bivariate cointegration between each pair of markets (Engle-Granger Methodology ) and then
multivariate cointegration ( Johansen Cointegration ) behavior of group of markets have been
conducted.

According to Engle-Granger analysis; where the dependent variable is BIST 100 indice and
the independent variables are choosen among aforementioned indices one by one, if the
coefficient of is positive and significant, it can be concluded that there is a short term
cointegration between those indices. If the coefficient of one day lagged value of error terms
of the model (ec_ln_index t-1) is between 1- and 0 (-1< conficient error correction < 0) and is
significant, it can be said that there is a long term relationship between BIST 100 and relevant
indices.

In Engel Granger analysis, since Brazil and Mexico are six hours back to Turkey, one day
lagged value of those indices are taken into consideration.

Refer Table 7

As seen in Table VII, BIST 100 index is cointegrated with IBOVESPA, MICEX, NIFTY
100, Shangai Composite, Johanessburg All Share, IPC, Jakarta and KOSPI Indices both in
the short and long run at 10 % significance level.

Based on the these results the next step is to investigate the significance of the cointegration
vector(s) by means of Johansen cointegration test. Table VIII shows that the null of no
cointegration among indices are accepted for 5 % and 10 % significance level meaning that
when all indices are observed together there is no evidence of moving comovement.

Refer Table 8
Refer Table 9

5
International Research Journal of Applied Finance ISSN 2229 – 6891
Vol. VIII Issue – 1 January, 2017 www.irjaf.com

Table 9 shows the results of granger causility tests. Since all the indices have 5 trading days
in a week, the lag lenght has been choosen as 5. It appears that Granger causality runs one-
way from IBOVESPA to BIST and IPC to BIST 100 meaning that information containing in
these indices is statistically significant up to 5 days. And according to the results, BIST100
index is found to be the granger cause of SSComp., JKSE and KOSPI indices.

It is also found that there is bidirectional causility between BIST 100 and JKSE Composite
index both in the short run and the long run. On the other hand, no causility has been
observed between BIST100 and MICEX, NIFT100 and JSE All Share indices both in the
short and long run.

IV. Results
With the increase of financial integration, while the investors who find great opportunity of
mobility in global financial system, are trying to increase their return on one hand, on the
other hand, are targeting investment strategies that will allow them to manage their risks. One
of these strategies is to diverse their risks by invesing in different markets which have no
correlations instead of investing in the markets with high integration level in the meanwhile.
Cointegration analysis, in a sense, can be helpful in determining whether a statistically
significant relationship between the different markets. Also the determination of short-or
long-term relationship even the causility test can help them in creating strategies for risk
diversification.

This study investigates the linkages between the main equity indices of BRICS (Brazil,
Russia, India, China and South Africa) and MIST (Mexico, Indonesia, South Korea and
Turkey) countries. According to the results of Engle-Granger Analysis; BIST 100 index is
found to be cointegrated with IBOVESPA, MICEX, NIFTY 100, Shangai Composite,
Johanessburg All Share, IPC, JKSE Composite and KOSPI Indices both in the short and long
run. This implies that for an investment strategy covering the bivariately selected stock
market indices which is created on the purpose of international risk diversification will be
ineffective.

As for the results of Johansen Cointegration Tests, which examine the existance of long-run
relationship in a multivariate framework, it is found that there is no long-run relationship
between the selected stock market indices for the given period. It can be concluded that the
opportunity for long-run diversification within that group of markets is possible and it gives
an opportunity of risk diversification for international investors.

According to Granger Causility test results, BIST100 index is found to be the granger cause
of Shangai Stock Exchange Composite Index and KOSPI Composite Index. There is one-way
causility which runs from IBOVESPA to BIST and IPC to BIST 100. It is also found that
there is bidirectional causility between BIST 100 and JKSE Composite index.

References
Bhattacharjee, Sudarshan; Swaminathan, A. M.. Stock Market Integration of India with Rest
of the World : An Empirical Study. Indian Journal of Finance, p. 22-32, May. 2016, ISSN
0973-8711.
Chittedi, K. R. Global stock markets development and integration: with special reference to
BRIC countries” MRPA Paper, 2009, no. 18602
Engle,R.F. and Granger, C.W.J. 1987 “Cointegration and Error Correction Representation:
Estimamtion and Testing”, Econometrica, Vol.55, pp.251-76.

6
International Research Journal of Applied Finance ISSN 2229 – 6891
Vol. VIII Issue – 1 January, 2017 www.irjaf.com

Fahami, N.A., “The structure of linkages and causal relationships between BRIC and
developed equity markets”, 2011 International Conference on Information and Finance
IPEDR vol.21 (2011) © (2011)IACSIT Press, Singapore.
Hall,S.G.1991, “The Effect of Varying Lenght VAR Models on the Maksimum Likelihood
Estimates of Cointegrating Vectors”, Scottish Journal of Political Economy,Vol.38, pp.
317-23.
Hovarth, R. and Petrovski D., International stock market integration: Central and South
Eastern Europe compared, Economic Sysytems, Vol.37, Issue 1, March 2013, p. 81-91.
Gözbaşı, O., IMKB İle Gelişmek Olan Ülke Hisse Senedi Piyasalarının Etkileşimi:
Eşbütünleşme Ve Nedensellik Yaklaşımı, Erciyes Ü. İİBF Dergisi Sayı:35 Ocak-Temmuz
2010,s:98-118.
Johansen, S.1988, “Statistical Analysis of Cointegrating Vectors”, Journal of Economic
Dynamics and Control, Vol.12,pp.231-54.
Khan, Taimur A. (2011) "Cointegration of International Stock Markets: An Investigation of
Diversification Opportunities," Undergraduate Economic Review, Vol.8, Issue1, Article 7.
Lingaraja,K., Selvam, M., Vasanth,V. , “Co Movements and Inter-Linkages among Emerging
and Developed Stock Markets in Asia with Reference to Singapore Stock Exchange”,
International Research Journal of Finance and Economics, ISSN 1450-2887 Issue 122,
May, 2014,p.102-120.
Yarovaya, L., Lau,M.C.K, “Stock market comovements around the Global Financial Crisis:
Evidence from the UK, BRICS and MIST Markets”, Research in International Business
and Finance, Vol. 37,May 2016, p. 605–619

Table 1: List of the indices

COUNTRY INDEX COUNTRY INDEX
Brasil IBOVESPA Mexico IPC Index
Russia MICEX Indonesia Jakarta Composite Index
India NIFTY 100 South Korea KOSPI Composite Index
China Shangai SE Composite Index Turkey BIST 100
South Africa Johannesburg SE All Share Index

7
International Research Journal of Applied Finance ISSN 2229 – 6891
Vol. VIII Issue – 1 January, 2017 www.irjaf.com

Graph 1: Daily closing data of indices ( January 2010 – June 2016 )

BIST100 IBOVESPA MICEX
100,000 80,000 2,000

90,000
70,000 1,800

80,000
60,000 1,600
70,000
50,000 1,400
60,000

40,000 1,200
50,000

40,000 30,000 1,000

250 500 750 1000 250 500 750 1000 250 500 750 1000

NIFTY100 SSCOMP_ JSEALLSHR

10,000 6,000 60,000

9,000
5,000
50,000
8,000
4,000
7,000 40,000
3,000
6,000
30,000
2,000
5,000

4,000 1,000 20,000

250 500 750 1000 250 500 750 1000 250 500 750 1000

IPC JSECOMP_ KOSPI

48,000 6,000 2,400

44,000 2,200
5,000

40,000 2,000
4,000
36,000 1,800

3,000
32,000 1,600

28,000 2,000 1,400

250 500 750 1000 250 500 750 1000 250 500 750 1000

Graph 2: Indice Series with Logarithmic Value ( January 2010 – June 2016 )
12

11

10

7
250 500 750 1000

LN_BIST100 LN_IBOVESPA
LN_MICEX LN_NIFTY100
Ln_SSEComp. LnJSEAllShrIndx.
LN_IPC LnJKSEComp.
LN_KOSPI

8
 International Research Journal of Applied Finance ISSN 2229 – 6891
Vol. VIII Issue – 1 January, 2017 www.irjaf.com

Table 2: Descriptive Statistics of Logarithmic Series

ln_BIST100 ln_IBOVESPA ln_MICEX ln_NIFTY100 ln_SSECOMP ln_JSEALLSHR ln_IPC ln_JKSECOMP ln_KOSPI
Mean 11.14524 10.93588 7.327480 8.732213 7.862643 10.57791 10.58597 8.338996 7.567182
Median 11.16393 10.92883 7.304788 8.670240 7.815584 10.59255 10.61378 8.372050 7.579648
Maximum 11.44228 11.19351 7.585733 9.104024 8.543227 10.91851 10.74319 8.616729 7.700001
Minimum 10.79424 10.53596 7.087899 8.391267 7.575590 10.15786 10.32115 7.814227 7.347809
Std. Dev. 0.151053 0.133820 0.098961 0.197188 0.202936 0.228867 0.110196 0.184211 0.063712
Skewness -0.230528 -0.146875 0.477792 0.470959 0.851962 -0.104819 -0.602954 -0.869877 -0.943137
Kurtosis 1.974633 2.573048 2.380231 1.757538 3.362354 1.540053 2.278875 3.183318 3.856377

Jarque-Bera 63.19747 13.42887 64.86278 121.5461 151.7330 108.7697 98.71166 153.0174 214.5705
Probability 0.000000 0.001213 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000

Sum 13374.29 13123.05 8792.976 10478.66 9435.171 12693.50 12703.16 10006.80 9080.618
Sum Sq. Dev. 27.35768 21.47145 11.74220 46.62087 49.37845 62.80392 14.55976 40.68631 4.867029

Observations 1200 1200 1200 1200 1200 1200 1200 1200 1200

Table 3: Correlation Matrix of Logarithmic Series

CORRELATION ln_BIST100 ln_IBOVESPA ln_MICEX ln_NIFTY100 ln_SSECOMP ln_JSEALLSHR ln_IPC ln_JKSECOMP ln_KOSPI
ln_BIST100 1.000000
ln_IBOVESPA -0.463210 1.000000
ln_MICEX 0.343940 -0.148483 1.000000
ln_NIFTY100 0.795284 -0.549890 0.492883 1.000000
ln_SSECOMP 0.189752 -0.013532 0.607011 0.487221 1.000000
ln_JSEALLSHR 0.820574 -0.737474 0.360147 0.872915 0.173755 1.000000
ln_IPC 0.839872 -0.593849 0.365480 0.733976 0.067818 0.907467 1.000000
ln_JKSECOMP 0.827993 -0.653222 0.270946 0.720776 0.009003 0.888367 0.900882 1.000000
ln_KOSPI 0.586948 -0.161482 0.549015 0.468321 0.140267 0.554521 0.635995 0.681932 1.000000

Table 4: The Results of Stationarity Test (Engle-Granger Analysis)

ADF PP ADF PP
Test for unit root in Test for unit root in Test for unit root in Test for unit root in
I(0) I(0) I(I) I(I)
AIC None None None None
t-stat. Prob. t-stat. Prob. t-stat. Prob. t-stat. Prob.
ln_BIST100 0.561843 0.8375 0.564256 0.8380 -34.7921 0.0000 -34.7921 0.0000
ln_IBOVESPA -0.57838 0.4668 -0.58065 0.4659 -34.2047 0.0000 -34.2023 0.0000
ln_MICEX 0.466124 0.8152 0.470344 0.8163 -35.1373 0.0000 -35.1373 0.0000
ln_NIFTY100 1.073817 0.9266 1.08861 0.9285 -32.2243 0.0000 -32.2243 0.0000
ln_SSComp. -0.07100 0.6589 -0.19011 0.6177 -7.84835 0.0000 -33.7085 0.0000
ln_JSEALLShr. 1.829828 0.9843 1.68373 0.9780 -26.7475 0.0000 -35.6654 0.0000
ln_IPC 1.049219 0.9234 1.03306 0.9213 -19.2573 0.0000 -34.1296 0.0000
ln_JKSEComp. 1.591121 0.9731 1.556768 0.9711 -22.9747 0.0000 -34.6047 0.0000
ln_KOSPI 0.617685 0.8497 0.377872 0.7932 -8.52973 0.0000 -33.3146 0.0000

Table 5: Unit root test for Residual of Regression (Johansen Analysis)

t-statistic Prob.*
Augmented Dickey-Fuller test statistic
-3.867794 0.0001
1% level -2.566894
Test critical values: 5% level -1.941088
10% level -1.616521

9
International Research Journal of Applied Finance ISSN 2229 – 6891
Vol. VIII Issue – 1 January, 2017 www.irjaf.com

Table 6: Lag Length Selection (Johansen Analysis)

Lag LogL LR FPE AIC SC HQ

0 12117 NA 1.17E-20 -20.34958 -20.31115 -20.3351

1 32475.97 40375.77 1.86E-35 -54.43021 -54.04587* -54.28537
2 32639.58 322.0002 1.62e-35* -54.56905* -53.83882 -54.29387*
3 32703.38 124.5817 1.66E-35 -54.54013 -53.464 -54.13459
4 32756.36 102.6654 1.74E-35 -54.49304 -53.07101 -53.95715
5 32800.67 85.20302 1.86E-35 -54.43138 -52.66346 -53.76514
6 32847.37 89.07373 1.97E-35 -54.37372 -52.2599 -53.57714
7 32879.41 60.64065 2.14E-35 -54.29145 -51.83173 -53.36451
8 32926.79 88.93847 2.26E-35 -54.23493 -51.42932 -53.17764
9 32964.15 69.57947 2.44E-35 -54.1616 -51.01008 -52.97396
10 33021.13 105.2511* 2.54E-35 -54.12123 -50.62382 -52.80324

Table 7: Engle-Granger Cointegration Analysis Results

Short Period Linking Long Period Linking
CONFICIENT PROBALITY CONFICIENT PROBALITY
Criteria %10 Criteria %10
d(ln_IBOVESPA) (-1) 0.13526 0.0000 ec_ln_IBOVESPA(-1) -0.00601 0.0824
d(ln_MICEX) 0.48501 0.0000 ec_ln_MICEX(-1) -0.00690 0.0288
d(ln_NIFTY100) 0.49735 0.0000 ec_ln_NIFTY100(-1) -0.01609 0.0016
d(ln_SSComp.) 0.13275 0.0000 ec_ln_SSComp.(-1) -0.00791 0.0177
d(ln_JSEALLShr.) 0.72604 0.0000 ec_ln_JSEALLShr.(-1) -0.01567 0.0019
d(ln_IPC) (-1) 0.08137 0.0605 ec_ln_IPC(-1) -0.01981 0.0004
d(ln_JKSEComp.) 0.44151 0.0000 ec_ln_JKSEComp.(-1) -0.01814 0.0011
d(ln_KOSPI) 0.45858 0.0000 ec_ln_KOSPI(-1) -0.00763 0.0506

Table 8: Results of Johansen Cointegration Test

Hypothesized Trace 0.05 Critical

No. of CE(s) Eigenvalue Statistic Value Prob.**

None 0.037409 167.0456 197.3709 0.5551

At most 1 0.02464 121.4076 159.5297 0.8255
At most 2 0.017193 91.54443 125.6154 0.8351
At most 3 0.016436 70.78577 95.75366 0.6961
At most 4 0.013327 50.94838 69.81889 0.5961
At most 5 0.011437 34.88866 47.85613 0.4539
At most 6 0.009283 21.11965 29.79707 0.3502
At most 7 0.005971 9.956595 15.49471 0.2841
At most 8 0.002326 2.787821 3.841466 0.0950

10
International Research Journal of Applied Finance ISSN 2229 – 6891
Vol. VIII Issue – 1 January, 2017 www.irjaf.com

Table 9: The Results of Granger Causility Tests

Lag 5 Long Period (ln series) Short Period (d_ln series)
F-Statistic Prob. (%5) Result F-Statistic Prob. (%5) Result
ln_IBOVESPA 5.444613779 0.000058366 H0 reject! 6.282033926 0.000009200 H0 reject!
ln_BIST100 1.598962416 0.157450885 H0 accept! 0.485896331 0.786988497 H0 accept!

ln_MICEX 0.73280141 0.598882512 H0 accept! 0.222831474 0.952763409 H0 accept!

ln_BIST100 0.349630419 0.882660122 H0 accept! 0.175400126 0.971808911 H0 accept!

ln_NIFTY100 1.847975027 0.100765451 H0 accept! 1.618885411 0.152038547 H0 accept!

ln_BIST100 1.797318265 0.110509205 H0 accept! 1.841264207 0.102010123 H0 accept!

ln_SSComp. 0.344685695 0.885817162 H0 accept! 0.128788831 0.985871032 H0 accept!

ln_BIST100 2.912198269 0.012776372 H0 reject! 2.061217434 0.067822009 H0 accept!

ln_JSEALLShr. 1.027973052 0.399600240 H0 accept! 0.292979919 0.916976917 H0 accept!

ln_BIST100 0.693341461 0.628534707 H0 accept! 1.450690587 0.203361967 H0 accept!

ln_IPC 2.901357973 0.013059986 H0 reject! 2.017678365 0.073598954 H0 accept!

ln_BIST100 1.582588291 0.162027157 H0 accept! 1.611153565 0.154119332 H0 accept!

ln_JKSEComp. 2.678069724 0.020458133 H0 reject! 2.292128323 0.043661261 H0 reject!

ln_BIST100 8.39388985 0.000000082 H0 reject! 8.469252784 0.000000069 H0 reject!

ln_KOSPI 1.034692445 0.395637702 H0 accept! 0.764397329 0.575514602 H0 accept!

ln_BIST100 9.829548894 0.000000003 H0 reject! 9.797579004 0.000000003 H0 reject!
(H0 : Indice X does not granger cause of Indice Y.)

Authors

Bengü Vuran*
Associate Professor, PhD, Istanbul University, School of Business, Finance Department.
[email protected]

Kaan Türkay
MS Finance, Istanbul University, Institute of Social Sciences, Finance Programme.

* corresponding Author

11