Co-integration between Pakistan equity market and emerging ...



DYNAMIC RELATIONSHIP BETWEEN PAKISTAN EQUITY MARKET AND ASIAN EMERGING EQUITY MARKETS AND DEVELOPED WORLD

Abstract

The study examines the linkage of Pakistan Stock Exchange[1] (PSX) with emerging and developed equity markets over the time period of 1997 to 2014. The Johansen’s multivariate co-integration tests both maximum eigenvalue statistics and trace statistics suggest that the PSX has long run relation with the emerging and developed market based. However, the bivariate co-integration trace and maximal-eigenvalues conclude that the PSX individually have no linkage with other markets. The study also concluded that PSX is highly volatile equity market range from 44% to -24%, which risk takers may prefer.

INTRODUCTION

The determination of co-integration among World's equity markets becomes very important due to the globalization of an economy. Today the movement of investment funds across the boundary is facilitated by the improved technology, advance trading system and innovations of financial products in financial markets. It influences both the practitioners and academicians interest in the study of integration of the World's equity markets. The relationship among the world's equity markets got considerable momentum after the October 1987 global stock market crash and even more in the Asian financial crisis of 1997. Many researchers in past have examined this relationship among the world’s developing and developed equity markets.

Due to the globalization, the behaviors of investors are rapidly changing in the equity markets of the world; that may be resulted in growing of economies smoothly. It may lead weaker economies towards integration among emerging economies. Because of the globalization the investments trend into the financial assets increased extremely. Today lot of equity markets in the world are moving towards integration like MENA, EU, NAFTA, ASEAN and Scandinavian countries. It reduces the risk of financial crisis in any equity market and that is the positive sign for the world economic growth. Karachi Stock Exchange (Pakistan equity market) is considered an emerging equity market that their financial infrastructure and industries are in a growing stage. Asjeet S. Lamba (2005) examined the short and long term relationship among the famous equity markets of South Asia that is India, Sri Lanka and Pakistan against developed stock markets of the world like UK, US and Japan. The study concluded that the given Asian markets are relatively isolated from the major developed ones and also found that the South Asian equity markets are becoming more integrated with each other but at a relatively slow pace. Most of the world's equity markets are emerging. Due to similarities in emerging equity markets, there are more chances for fund managers and other institutional investors to build portfolio diversification. To maximize benefits and minimize cost of portfolio diversification for intended investors it is necessary to determine the Co-integration between the developed stock markets and emerging equity markets of the World. The scope of Co-integration among the Asian emerging stock markets increases after Asian financial crisis in 1997. To examine Co-integration of Pakistan Stock Exchange (PSX) with other emerging equity markets is interesting because PSX became noticeably popular market since 2002. This study focuses on examining the Co-integration of Pakistan equity market with the five Asian emerging equity markets i.e. India, Malaysia, Indonesia, Hong Kong and China and the six developed equity markets i.e. Japan, US, UK, Germany, France and Australia.

LITERATURE REVIEWS

Kashif. H and Arshad H (2011) examined the causal and dynamic relation of KSE-100 index with emerging equity markets of China, Malaysia, Hong Kong, Turkey, Indonesia, Thailand, Brazil, and India as well as with the developed stock markets of France, UK, USA, and Japan for the time period of 1998 to 2008. By using Johansen co-integration technique the study concluded that KSE 100 index has long run relationship with Indonesia (JCI) and Brazil (BCI) equity markets and short run relations with the China (SCI) equity markets. Further the study indicated that stock markets of JCI, Nikkei-225, KLSE, SCI, ISE, BSE, BCI and SET granger causes to KSE-100, while KSE 100 index granger cause to JCI., KLSE, SET and HSI. A. Iqbal, N. K and S. Rafiq (2011) examined the dynamic linkage between emerging stock market of Pakistan, India and the developed equity market of USA for the time period of 2003 to 2009 by using daily data. The study applied Johansen co-integration technique and found that there is no co-integration between stock markets of Pakistan, India and USA. While results from granger causality test showed USA granger cause to India and Pakistan equity markets. Worthing , A .C and Higgs, H (2006) examined the weak form of market efficiency between ten Asian emerging stock markets, that are India equity market, China equity market, Korea equity market, Indonesia equity market, Pakistan equity market, Malaysia equity market, Sri Lanka equity market, Philippines equity market, Thailand equity market and Taiwan equity market as well as five developed equity markets are Hong Kong equity market, Australia equity market, New Zealand equity market, Japan equity market, and Singapore equity market. The study tested daily data from 1987 to 2003 through variance ratio test. It indicated that all the selected emerging equity markets are characterized by random walks whereas some developed market like Hong Kong stock market, Japan stock market and New Zealand stock market exhibit random walk. Lim, K. P et. al (2007) re-examined the weak form of efficiency in 10 Asian markets China Shanghai SE composite, India B.S.E National , Indonesia Jakarta SE composite, Korea SE composite, Malaysia Kuala Lumpur composite, Pakistan KSE 100, Philippines SE composite, Sri Lanka Colombo all share, Taiwan SE weighted and Thailand Bangkok S.E.T. Bi-correlation test shows that these Asian indices have efficiency to interlink with each other for long period of time, only to combine with brief periods of strong nonlinear dependence. Asjeet S. Lamba (2005) conducted a study in the context to examine the long term and short term linkage among the equity markets of South Asia composed of India, Sri Lanka and Pakistan as well as three key well developed equity markets that is UK, US and Japan, from period from 1997 to 2003 by applying Johansen’s co-integration tests. The findings were that the Indian stock market is affected by the selected developed equity markets that are UK, US and Japan and the terrorist attacks on September 11, 2001 on United State are considered the reason of influence on Indian equity market. The study also concluded that Pakistan and Sri Lanka are comparatively isolated from the major developed markets whereas it is also found that the selected Asian stock markets are gradually coming close to each other. Abd. Majid, S. M et al (2007) focused on eight equity markets of OIC (Organization of the Islamic Conference) to examine the financial linkage between these markets. It consists of four Middle East and North Africa regions. From Middle East and North Africa regions Egypt, Turkey, Kuwait and Oman were included, and the other four understudy markets are from Asian region i.e; Malaysia, Bangladesh, Pakistan and Indonesia. The daily data was tested for the time period from January 2002 to December 2006 to determine integration among the identified variables. The study concluded that international integration is only found between the stock markets of Malaysia, Indonesia and developed markets and no integration is found between Bangladesh, Pakistan and developed markets. The study also concluded that there is no evidence of integration among OIC markets. Arshad H. et al (2008) examined the long run relationship among Pakistan (PSX) equity markets and developed world by using Johansen and Juselius (1990) cointegration test for the time period of 2000 to 2006. It was found that PSX is integrated with the developed markets. While pair wise cointegration test explained that the PSX is not integrated with UK, Canada, U.S.A, Italy, Australia and Germany equity markets. However PSX is integrated with France and Japan equity markets. Paresh N. R. S. and M. N (2004) observed the dynamic linkages between the stock markets of Bangladesh, Pakistan, Sri Lanka and India by applying granger causality and multivariate techniques. The study concluded that stock prices in India, Sri Lanka and Bangladesh granger cause Pakistan stock prices in long run. While in the short run there is unidirectional granger causality running from stock prices in Pakistan and Sri Lanka to India and from stock prices in Pakistan to Sri Lanka. Kadir K. and S. Ergun (2010) conducted a study to determine the financial integration in Balkans by applying multivariate co-integration among four emerging markets. They found evidences of co-integration between the Balkans equity market indexes.

DATA DESCRIPTION AND METHDOLOGY

As at end of March 2010, 660 companies were listed with a market capitalization of Rs. 894.2 billion. The study utilizes 206 monthly closing stock indices values of the 12 stock markets for the period of August 1997 to October 2014 that includes Pakistan stock exchange (P.S.E) along with five Asian emerging stock markets that is India (BSE), Malaysia (KLSE), Indonesia (JCI), Hong Kong (HSI), and China (CSE), and the six developed equity markets are also included in the analysis i.e. Japan (NIKKEI), US (NYSE), UK (FTSE-100), Germany (DAX), France (CAC-100) and Australia (AORD). The databases used in the study are Yahoo Finance and other web sources.

The natural log is used to measure the compound rate of return. To analyze the co-integration of Pakistan stock markets with the developing Asian stock markets and the developed equity markets, the study uses: Correlation test, Johansen co-integration test and Granger Causality test.

To test the stationarity among the time series unit root tests are applied. The Phillips-Perron Test (1988) “PP” and the Augmented Dickey Fuller Test (1981) “ADF” are used to test the stationarity. The mathematical representations of the ADF test with a constant and trend and with a constant and without trend are given as:

[pic] (1)

[pic] (2)

Equation 1 includes both a constant ([pic]) and a trend tern ([pic]2 ) and Equation 2 includes only a constant trend. In both equations, ∆ is a difference operator and µt is the error term with zero mean and constant variance.

The PP test is also estimated with constant & trend and with constant and without trend as:

[pic] (3)

[pic] (4)

Here Equation 3 includes both a constant ([pic]) and (α2) trend and Equation 4 includes only a constant term ([pic] ). The null hypothesis in Equation (3) and (4) is π = 0.

Johansen’s (1991) is a method for testing co-integration for more than few time series. It examines the existence of long run relation between the applied variables. This test permits more than one co-integrating relationship, so is more generally applicable than the Engle and Granger test which is based on the Dickey-Fuller (1979, 1981) unit root test in the residuals from a single co-integrating relationship.

The general specification of the JJ co-integration model is given as:

[pic]

These are Trace test and Maximal Eigenvalue test.

[pic]

For the purpose of selecting the lag length for Vector Autoregressive process Akasike Information Criterion (AIC) is used. Similarly to examine the bivariate co-integration between the two financial series, the model stated is exercised in an autoregressive manner among two financial time series. It is also used for the purpose of determining the long run association among the two time series. There are some deficiencies while using the Johansen multivariate co-integration to identify the lead lag association between the variables. Hence, the study apply granger causality test to examine the causality among the data series that will determine whether changes in a variable will have an impact on changes of other variables.

RESULTS ANALYSIS

The descriptive statistics shown in table 1 present the stock returns indexes of the 12 equity markets. It includes the distribution of mean, median, maximum, minimum and standard deviation. According to the table, the descriptive statistics of average returns of all the stock indices are offering negative returns for given time period except Japan. The average negative return of PSX is -1.02% at given level of total risk equal to 9.84%. Table 1 indicates that the average return of NIKKEI (Japan) equity market is the highest among the selected countries for the given time period that is equal to 0.4% at the given level of standard deviation 5.9%. The PSX is the most risky stock market with the return volatility ranges from 44.87% to -24.11%, while the Japan market returns volatility range from 27.2% to -12%. The conclusive result is that risk taker investors may prefer PSX as the maximum returns could be up to 44%.

Table 1: Descriptive Statistics

|Equity Markets |Country | Mean | Median |Maximum |Minimum | Std. Dev. |

|RET PSX |Pakistan |-0.0102 |-0.0177 |0.4488 |-0.2411 |0.09841 |

|RET BSE |India |-0.008 |-0.0114 |0.27299 |-0.2489 |0.07796 |

|RET KLSE |Malaysia |-0.0033 |-0.0094 |0.28463 |-0.2944 |0.07461 |

|RET JCI |Indonesia |-0.0116 |-0.0218 |0.3772 |-0.2502 |0.08877 |

|RET HIS |Hong Kong |-0.0015 |-0.0095 |0.34824 |-0.2532 |0.07841 |

|RET CSE |China |-0.0123 |-0.0114 |0.18417 |-0.2252 |0.07602 |

|RET NIKKEI |Japan |0.0043 |-0.0024 |0.27216 |-0.1209 |0.05925 |

|RET NYSE |USA |-0.0021 |-0.0087 |0.21738 |-0.102 |0.04833 |

|RET FTSE_100 |UK |-0.0007 |-0.0044 |0.13954 |-0.0849 |0.04439 |

|RET DAX |Germany |-0.0024 |-0.0122 |0.29333 |-0.1937 |0.06995 |

|RET CAC_100 |France |-0.0008 |-0.0123 |0.19225 |-0.1259 |0.05919 |

|RET AORD |Australia |-0.003 |-0.0122 |0.15088 |-0.0736 |0.03952 |

The correlation table indicates that PSX is highly significantly correlated with the developed equity markets and most of emerging markets. The table indicates that there exist 17.6%, 17.2%, 19.2%, 16.5% correlation between PSX and CSE, NIKKEI, NYSE, DAX respectively at 0.05 level while 29%, 21% and 20.5% with BSE, KLSE and HIS respectively. The correlation among the all developed equity markets exists specifically among European countries.

Table2: Correlation matrix of Emerging Asian and Developed Equity Markets.

|  |

|*. Correlation is significant at the 0.05 level (2-tailed). |

To test whether a series contains a unit root, the table 3 has been classified into two different panels that is panel A and B. The panel A shows the result of log levels while applying the ADF test and PP test. The panel A results suggests that all the variables contain unit root while applying both the ADF test and PP test. The panel B shows the result of the data series at the first difference by applying both the ADF test and PP test that suggests that all the variables are stationary. Thus the conclusive results are that all variables remained non-stationary at log level but the series became stationary at 1st difference at the 5% of the probability level, so all these indices can be define as integrated of order one I (1).

Table 3: Unit Root Analysis by applying ADF and PP Test

|Stock Index |A D F Tests |P P Tests |

|Panel A Log Levels |

|PSX |-0.5108 |-0.528 |

|BSE |-0.3601 |-0.5165 |

|KLSE |-1.5503 |-1.348 |

|JCI |-0.3375 |-0.0777 |

|HIS |-1.7552 |-1.6744 |

|CSE |0.9112 |0.63124 |

|NIKKEI |-1.5675 |-1.8264 |

|NYSE |-2.2944 |-2.4299 |

|FTSE-100 |-1.9196 |-2.1564 |

|DAX |-1.9026 |-2.1351 |

|CAC-100 |-1.9472 |-2.2812 |

|AORD |-1.4078 |-1.638 |

|Test critical values: |

|1% level |-3.469 |-3.469 |

|5% level |-2.8784 |-2.8784 |

|10% level |-2.5758 |-2.5758 |

|Panel B 1st Difference |

|PSX |-12.502 |-12.503 |

|Ln BSE |-12.193 |-12.252 |

|KLSE |-10.589 |-10.397 |

|JCI |-10.343 |-10.198 |

|HIS |-11.529 |-11.549 |

|CSE |-11.484 |-11.559 |

|NIKKEI |-11.709 |-11.783 |

|NYSE |-10.715 |-10.755 |

|FTSE-100 |-12.16 |-12.175 |

|DAX |-11.874 |-11.895 |

|CAC-100 |-11.24 |-11.327 |

|AORD |-11.283 |-11.487 |

As the ADF and PP test concluded that all the indices are integrated at order one I (1). The next step is to examine the Johansen and Juselius bi-variate and multivariate co-integration tests. Table 4 presents the results of the Johansen’s multivariate co-integration test. It follows up both the Maximum Eigenvalue statistics and Trace statistics. The trace statistics shows that there are three co-integrating vectors at 0.05 critical values and one integrating vector at 0.1 critical values for the given time period. The same result is followed by applying maximum Eigenvalue statistics that also indicates that there are three co-integrating vectors at 0.05 critical values and one integrating vector at 0.1 critical values. The conclusive result is that there exists long run relationship among these four equity markets. The Johansen’s co-integration multivariate test determines that the PSX has long run relation with the emerging Asian market and no relation with the developed market.

Table 4: Multi-variate Co-integration Analysis Trace Statistics and Maximum Eigenvalue Statistics.

|Hypothesis |Trace |Critical |P-value |Max-Eigen |Critical |P-value |

| |Statistics |Values |at 0.05 |Statistics |Values |at 0.05 |

| | |at 0.05 | | |at 0.05 | |

|None * |422.4482 |334.9837 |0 |88.08722 |76.57843 |0.0033 |

|At most 1 * |334.361 |285.1425 |0.0001 |72.36145 |70.53513 |0.0336 |

|At most 2 * |261.9995 |239.2354 |0.003 |68.56103 |64.50472 |0.0195 |

|At most 3 |193.4385 |197.3709 |0.0769 |56.46515 |58.43354 |0.0773 |

|At most 4 |136.9733 |159.5297 |0.4252 |45.5167 |52.36261 |0.2115 |

|At most 5 |91.45664 |125.6154 |0.837 |30.96898 |46.23142 |0.723 |

|At most 6 |60.48766 |95.75366 |0.9445 |19.1515 |40.07757 |0.9822 |

|At most 7 |41.33616 |69.81889 |0.9247 |13.45457 |33.87687 |0.9931 |

|At most 8 |27.88158 |47.85613 |0.8188 |11.99837 |27.58434 |0.9328 |

|At most 9 |15.88322 |29.79707 |0.7205 |7.486184 |21.13162 |0.9323 |

|At most 10 |8.397032 |15.49471 |0.424 |6.565072 |14.2646 |0.5418 |

|At most 11 |1.831961 |3.841466 |0.1759 |1.831961 |3.841466 |0.1759 |

Table 5 reports the positive and significant relationship of PSX with BSE, KLSE, NYSE and FTSE_100, while negative and significant relation with CAC_100, JCI and NIKKEI. It allows the investors to diversify the risk by investing in these integrating equity markets.

Table 5: Normalized co-integrating coefficients.

|Variables |Coefficient |S.E |t-stat |

|BSE |10.87467 |-1.6227 |-6.70159 |

|KLSE |8.702302 |-1.77373 |-4.90622 |

|JCI |-13.8761 |-1.54719 |8.968588 |

|HIS |-3.16293 |-2.09831 |1.507371 |

|CSE |-0.91498 |-0.71796 |1.274419 |

|NIKKEI |-5.51881 |-1.05311 |5.240485 |

|NYSE |7.27877 |-3.24271 |-2.24466 |

|FTSE_100 |16.04872 |-3.3907 |-4.73316 |

|DAX |0.331925 |-2.17213 |-0.15281 |

|CAC_100 |-14.3135 |-2.28871 |6.253951 |

|AORD |1.177194 |-2.19734 |-0.53574 |

In table 6, the study examines the pairwise co-integration properties of PSX index with other Asian emerging equity market and the selected developed equity market of the world in bi-variate form. The appropriate lag length was one, which is in accordance to the Akaike Information Criterion (AIC) and Schwarz Information Criterion (SIC). By applying Trace test and Maximal-Eigenvalues test at the 0.05 level, the study concludes that PSX has three co-integrating vectors at 0.05 levels with the given Asian equity markets and developed world. It may attract the investors of the selected countries as well as others for the purpose of the portfolio diversification. In other words, PSX individually have long run relationship with the Asian and developed equity markets.

Table 6: Bi-variate Co-integration Analysis Trace Statistics and Maximum Eigenvalue Statistics

|Variables |Hypothesis |Trace |

| | |Statistics |

|BSE does not Granger Cause PSX |0.05516 |0.81461 |

|PSX does not Granger Cause BSE |2.96084 |0.08717 |

|KLSE does not Granger Cause PSX |5.92194 |0.01601 |

|PSX does not Granger Cause KLSE |5.46104 |0.02064 |

|JCI does not Granger Cause PSX |1.49288 |0.2235 |

|PSX does not Granger Cause JCI |3.29616 |0.07125 |

|HIS does not Granger Cause PSX |3.20154 |0.07539 |

|PSX does not Granger Cause HIS |3.03869 |0.08315 |

|CSE does not Granger Cause PSX |0.05007 |0.82322 |

|PSX does not Granger Cause CSE |0.02204 |0.88215 |

|NIKKEI does not Granger Cause PSX |3.80287 |0.05285 |

|PSX does not Granger Cause NIKKEI |6.33594 |0.01278 |

|NYSE does not Granger Cause PSX |2.31778 |0.12981 |

|PSX does not Granger Cause NYSE |5.10305 |0.02519 |

|FTSE_100 does not Granger Cause PSX |1.64276 |0.20173 |

|PSX does not Granger Cause FTSE_100 |2.37016 |0.12558 |

|DAX does not Granger Cause PSX |1.55913 |0.21355 |

|PSX does not Granger Cause DAX |1.36022 |0.24517 |

|CAC_100 does not Granger Cause PSX |1.3232 |0.25167 |

|PSX does not Granger Cause CAC_100 |1.7399 |0.18897 |

|AORD does not Granger Cause PSX |8.56155 |0.00391 |

|PSX does not Granger Cause AORD |3.41176 |0.06651 |

Conclusion

The innovation speeds up the movement of investment fund across the boundaries. It attracts the attention of both the practitioners and academicians stake in the study of integration of the World's equity markets. The globalization has influence in integration of financial markets like MENA, EU, NAFTA, ASEAN and Scandinavian countries.

The descriptive statistics shows that on average NIKKEI (Japan) equity market outperform among the selected equity markets, while the PSX (Pakistan) is the most risky market with the range of 44.87% to -24.11%. It might encourage the aggressive investors to earn the maximum upto 44.87%. The correlation matrix shows that European countries are highly correlated because of the free flow of fund in between the European Union. The co-integrating coefficients values determine the risk averse investors to design a portfolio by investing in negative related equity markets that is CAC_100, JCI and NIKKEI.

References

A. Iqbal, N. Khalid and S. Rafiq (2011), ‘Dynamic interrelationship among the stock markets of India, Pakistan and United states; World Academy of Science, Engineering and Technology, pp 98-104.

Asjeet S. Lamba (2005), “Analysis of the short- and long-run relationships between South Asian and developed equity markets”, International Journal of Business, 10(4), 2005 ISSN: 1083−4346.

Dickey, D.A. and Fuller W.F. (1979), “Distribution of estimates for autoregressive time series with a unit root” Journal of American Statistical Association, 74: 427-31.

Dickey, D.A. and Fuller W.F. (1981), “The likelihood ratio statistic for autoregressive time series with a unit root” Econometrica 49(4): 1057-1072.

Engle, R.F. and Granger (1987), “Co-integration and error correction regression estimation and testing” Econometrica. 52 (2): 251-76.

Johansen, S. (1991),”Estimation and hypothesis testing of co-integrating vectors in Gaussian vector autoregressive models”. Econometrica 59 (November):1551-1580.

Kadir Karagoz and Suzan Ergun (2010), ‘Stock Market Integration among Balkan Countries; MIBES Transactions, Vol 4, Issue 1, pp 49-59.

Kian-Ping Lim, Robert D. Brooks and Melvin J. Hinich (2007), "Nonlinear serial dependence and the weak-form efficiency of Asian emerging stock markets". Journal International Financial Markets, Institutions & Money, 18(1) INTFIN-323.

M. Shabri Abd. Majid, Rosylin Mohd. Yusof, and Ahmad Nazri Razal (2007), "Dynamic Financial Linkages among Selected OIC Countries". Journal of Economic Cooperation, 28 (2) 25-56.

Paresh Narayan, Russell Smyth and Mohan Nandha (2004), ‘Interdependence and dynamic linkages between the emerging stock markets of South Asia; Accounting and Finance 44, 419–439.

Phillips, P., and P. Perron (1988),”Testing for a unit root in time series regression”, Biometrica, 75 (June):335-346.

Worthing A .C and Higgs, H (2006), "Weak-Form market efficiency in Asian and developed equity markets comparative tests of random walk behaviour" Accounting Research Journal, 19 (1), 54-63.

-----------------------

[1] Pakistan Stock Exchange (PSX) was formerly named as Karachi Stock exchange (KSE). However the indices are still named same as previous i.e. KSE100, KSE 30 etc.

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download