Printer Friendly

Cointegration and causality between the GCC stock indices and gold indices.

1. Introduction

The gold is considered as a strategic resource that is used by many countries as a national security and a hedging investment instrument and it has been considered as the most precious metal since ancient times (Bilal, Talib, Haq, Khan, & Naveed, 2013). The gold is considered as a conservative investment, security collateral, and risk diversifier and inflation hedger. Uncertainty of economic environment turns the attention to investing in gold as a safe haven.

Throughout history, gold maintained a unique function of both exchange and a store of value. Baur & McDermott (2010) found that gold was a strong risk management tool for major European and US stock markets during the recent global financial crisis. For GCC countries and many other both gold and oil are with no doubt considered as strategic resources for national security and various economic activities. Furthermore, there is strong evidence in the literature that equity portfolio investors uses gold to hedge the risk associated with their investment (Chkili, 2016; Ciner, Gurdgiev, & Lucey, 2013). Whenever there is a financial or economic crisis, investors tend to move out from risky assets, such as, stocks and invest in gold (Narang & Singh, 2012). The price fluctuation of gold and market volatility will impose some effects on market participants. Hence, stock market investors, producers, consumers, governments and business will face some challenges due to the changes in gold prices. The recent statistics for the official gold holding for GCC are as follows: (1) Bahrain holds 4.7 tons, (2) Kuwait holds 79 tons, (3) Qatar holds 26.6 tons, (4) Saudi Arabia holds 322.9 tons, (5) UAE holds 7.5 tons and Oman with no holding of gold (World Gold Council, 2017). As gold have emerged as an important investment class, market participants are especially interested in the interaction across gold and the stock markets, given the increased price volatility, which affects market stabilization policies and impose more challenges to all market participants including the government.

The objective of this study is to investigate, establish and determine the causality and cointegration between GCC stock indices including Bahrain Stock Exchange (BSE), Kuwait Stock Exchange (KSE), Qatar Stock Exchange (QSE), Saudi Stock Exchange (SSE), Muscat Securities Market (MSM), Dubai Stock Exchange (DSE) and Abu Dhabi Stock Exchange (ADSE), and gold indices. Our sample period is from January 2010 to December 2016. This allows us to provide a recent view on the causality and cointegration between GCC stock indices and gold, while most of the prior studies has only accounted for the periods before and not for all GCC markets. Compared to prior studies, this study therefore provides insights from the post 2008 global financial crisis and increased market volatility due to the changes in oil price, which considered the main sources of income for all GCC countries. In fact, during this period, the price of an ounce of gold had peaked to around $1900 in September 2011 from a bottom of $1050 in December 2015. As for the GCC markets, it has experienced a major changes in market index in which most of them changed by more than 50% during the period of January 2010 to December 2016.

The rest of the paper is organized as follows: Section 2, discuss the previous studies in the field and reviews the related literature. Section 3, explains the data and methodology including the econometrics model. Section 4 makes the analysis and the implications, and finally section 5 concludes.

2. Literature review

Many of the past literature had concentrated on studying the relationship between GCC stock markets and foreign exchange markets. GCC stock markets are not only influenced by the changes in foreign exchange markets but also the changes in gold, oil and other commodities.

Different timeframe and econometrics techniques have been used to model the linkages between gold and stock market indices. There is no clear rejection or confirmation for the relationships between gold and stock markets movements. Smith (2002) discovered a short-run correlation between gold price and stock price indices but small and negative for Japan and European markets. Dirk & Baur (2010) concluded that most developed countries during the financial crisis used gold as safe haven. Horng & Huang (2013) investigated the relationship of both Malaysian Stock Market and Thailand Stock Market with gold market. They applied the bivariate asymmetric integrated generalized autoregressive conditional heteroscedasticity (IGARCH) model and found that gold market affects both risk and return of Malaysian Stock Market and Thailand Stock Market. In GCC stock markets, Al Janabi, Hatemi, & Irandoust (2010) found that these markets are informationally efficient with both gold and oil price indexes. They used daily price stock indexes for the period 2006 to 2008. In Turkey, Buyuksalvarci (2010) concluded that gold price had no significant effects on the Istanbul Stock Exchange-100 index return.

Mishra, Das, & Mishra (2010) proved that stock market returns Granger-causes the gold prices and gold prices Granger-causes stock market returns. The finding are consistent with previous theory that consider the gold as "safe haven".

Mulyadi & Anwar (2012) concluded that there is inverse relationship between the gold and the performance of the stock markets. Generally, gold is considered as a safer asset relative to stocks. Kumar (2014) examined the relationship between the gold prices and Indian sectoral indices and concluded that there is no evidence of a volatility spillover running from gold prices. Sreekanth & Veni (2014) study revealed the existence of long-run cointegration between gold prices and stock index (S&P CNXNIFTY) of the Indian National Stock Exchange. They have used different econometrics tool including Dickey-Fuller (ADF) test, Granger causality test (GCT), Vector Error Correction Model, Wald's coefficient diagnosis, residual analysis and Johansen cointegration test for the period 2005 to 2013. Similar negative relationship between gold and stock prices found by Shahzadi & Chohan (2012) using the KSE-100 index, the index of Pakistan's Karachi Stock Exchange in Pakistan. They used ADF test, GCT and Johansen cointegration test for five years data from December 2005 to December 2010.

Choudhry, Hassan, & Shabi (2015) examined the co-movements between stock market returns and gold returns during the recent global financial crisis for the S&P500 in United States and FTSE100 in United Kingdom. They found that gold may not perform well as "safe haven" during the financial crisis but it can be used as a hedge against stock market returns and volatility. Tuysuz (2013), Emmrich & Mcgroarty (2013), Miyazaki & Hamori (2013), and Chen & Lin (2014) provided empirical evidence that gold used as a "safe haven" during the financial crisis.

Considerable amount of research is done to investigate the relationship between gold and stock prices but literature gives mixed views about this relationship. The reasons for these differences in results could be due to the time series modeling techniques and time period of study used in the studies. To address the literature gap, this study employs to determine long and short time cointegration and causality between GCC stock indices including Bahrain Stock Exchange, Kuwait Stock Exchange, Qatar Stock Exchange, Saudi Stock Exchange, Muscat Securities Market, Dubai Stock Exchange and Abu Dhabi Stock Exchange to study the relationship with the gold prices during a recent timeframe after the 2008 global financial crisis. This will help institutional investors and private portfolio managers to make informed investment decisions and diversification of portfolios.

3. Data and methodology

In this study we used the international gold prices and six daily stock price indices as follow; Bahrain Stock Exchange (BSE), Kuwait Stock Exchange (KSE), Qatar Stock Exchange (QSE), Saudi Stock Exchange (SSE), Muscat Securities Market (MSM), Dubai Stock Exchange (DSE) and Abu Dhabi Stock Exchange (ADSE) from 2010 to 2016. These series' data consist of daily closing prices from January 1, 2010 to December 31, 2016. There are 2557 observations within each series.

Cointegration can be calculated with different methods in the literature. In this study, we have selected Johansen Test to calculate cointegration. The Johansen test and its estimation strategy depend on maximum likelihood. The test can be used to estimate all cointegration when there are more than two variables. Furthermore, the test helps to estimate all cointegration vectors with numerous variables that may include at least two cointegration vectors (Acikalin & Basci, 2016).

When studying the time series data, it is necessary to ensure that all data are stationary to avoid econometric problems. We used Augmented Dickey-Fuller (ADF) test to determine all series whether stationary or non-stationary using Equation 1, which is given below.

[DELTA][y.sub.t] = [phi][y.sub.t-1] + [p.summation over t=1] [[alpha].sub.i][DELTA][y.sub.t-i] + [] (1)

According to Equation (1), null hypothesis means non-stationary or it exist a unit root with [phi] = 0. Contrarily, alternative hypothesis does not include a unit root and it is stationary with [phi] < 0.

When all series are stationary, then we can use using Johansen's Methodology to test the cointegration. The Johansen Cointegration Test starting a point with Vector Autoregression (VAR) equation. In the VAR equation, it includes p order and it can be shown in Equation (2).

[y.sub.t] = [mu] + [A.sub.1][y.sub.t-1] + ... + [A.sub.p][y.sub.t-p] + [[epsilon].sub.t] (2)

Where [y.sub.t], is an nx1 vector of variables which are integrated of one order, and it is denoted I(1). If vectors are more than 1, then Equation (2) can be re-written as Equation (3) below,

[DELTA][y.sub.t] = [mu] + [PI][y.sub.t-1] + [p-1.summation over i=1] [[gamma].sub.i][DELTA][y.sub.t-i] + [[epsilon].sub.t] (3)


[PI] = [p.summation over i=1][A.sub.i] - I and [[gamma].sub.i] = -[p.summation over j=i+1][A.sub.j]

In Equation (3) [PI] means coefficient matrix and if reduced rank lower than n then it has matrices like n x r. This matrices has [alpha] and [beta] within rank in r and if all variables are stationary, then r is the number of cointegration relations. Thus, A is the parameter in the Vector Error Correction Model act as an adjustment parameter. And [beta] is a cointegration vector of each column of the model (Johansen, 1995).

There are two tests related to Johansen Cointegration Test that are used in canonical correlations. Significance of the correlation is used in [PI] matrix for reduced rank (Hjalmarsson & Osterholm, 2007).

[J.sub.trace] = -T [n.summation over i=r+1] ln(1 - [[??].sub.i]) [J.sub.max] = -T ln(1 - [[??].sub.r+1]) (4)

In Equation (4), T means sample size and [??] largest canonical correlation of i. The null hypothesis of r cointegration vectors can be tested by the trace test that tests against the n cointegration vectors, which is the alternative hypothesis.

The causality is a concept related to prediction. In literature, Granger Causality is the most known statistical concept of causality. With the Granger Causality, if there are two variables and first variable's past values then it should include information, which helps predict the second variable. However if future information include past values of second variables, then we can conclude that there are Granger Causality between both variables (Granger, 2001). Granger Causality can be written as bivariate linear autoregressive model with two variables using Equation (5) below,

[Y.sub.1](t) = [p.summation over 1][A.sub.11], j[Y.sub.1] (t-j) + [p.summation over 1][A.sub.12], j[Y.sub.2] (t-j) + [[epsilon].sub.1](t) [Y.sub.2](t) = [p.summation over 1][A.sub.21], j[Y.sub.1] (t-j) + [p.summation over 1][A.sub.22], j[Y.sub.2] (t-j) + [[epsilon].sub.2](t) (5)

Where, p means model order, A represents coefficient of the model and [epsilon] represents residuals, which is the prediction errors for all series in the model. If variance of [[epsilon].sub.1] in first equation in the Equation (5) contains the [Y.sub.2] reduced version, then it can be concluded that [Y.sub.2] Granger Cause of [Y.sub.1]. However, if the coefficients in the model accordingly and significantly different from zero, it can be concluded that [Y.sub.2] is Granger Cause of [Y.sub.1].

4. Results and implications

We have started to analyze all series log version with Unit Root test. According to the results of ADF test statistics, we need to generate all series with first difference to having stationary series.

Figure 1 shows that all series in the model with daily closing price and Table 1 shows their descriptive statistics before log version.

Figure 2 shows the log series of each variable in the model including the seven GCC stock markets and gold prices.

According to Table 2, Johansen Cointegration Test result shows that there is a 1-cointegration equation in the model between series. It means that at least one cointegration relationship in long period.

In Table 3, there are so many Granger Causality relations between variables. According the Granger Causality Test results, only statistically significant result has been showed in Table 3. Relationships and way of causality can be understand from probability and Chi-Sq. test statistics from the Table. Granger Causality can be either in one direction or in both directions. Figure 3 shows the VAR Granger Causality results for the stock markets. There five relationship with two direction effects: (1) Abu Dhabi Stock Exchange with Saudi Stock Exchange, (2) Abu Dhabi Stock Exchange with Dubai Stock Exchange, (3) Kuwait Stock Exchange with Muscat Securities Market, (4) Qatar Stock Exchange with Saudi Stock Exchange and (5) Qatar Stock Exchange with Muscat Securities Market. Furthermore, there are another nine relationship with one direction effect between the seven stock markets.

5. Conclusion

The literature provides abundant evidence on the importance of gold for hedging and safe haven purposes that is used by governments, institutions and individuals. In this study, we used Johansen Cointegration and VAR based Granger Causality to examine the causality and cointegration between GCC stock indices and gold indices from January 2010 to December 2016. The main finding highlights that there is a strong Granger Causality between the gold indices and GCC stock indices according the statistically significant level. Furthermore, there five relationships with two direction effects another nine relationships with one direction effect between the seven stock markets. This means that changes in one market is followed by later changes in another market during the period of study. Hence, we can conclude that stock market performance in all seven GCC markets provides useful information in predicting the changes in other markets within GCC. Our results is particularly useful for policy makers who have been using gold as indicator and hedging instrument for portfolio investment in stock markets. This is important as a risk management tool for institutional investors and multinational corporations. Further research should focus on whether the relationship with oil and stock indices will be move in a different direction of gold or not.


Acikalin, S., & Basci, E. S. (2016). Cointegration and causality relationship between BIST 100 and BIST gold indices. Yonetim ve Ekonomi [Management and Economics], 23(2), 565-574 [in Turkish].

Bilal, A. R., Talib, B. A., Haq, I. U., Khan, M. N. A., & Naveed, M. (2013). How gold prices correspond to stock index: A comparative analysis of Karachi Stock Exchange and Bombay Stock Exchange. World Applied Sciences Journal, 21(4), 485-491.

Buyuksalvarci, A. (2010). The effects of macroeconomics variables on stock returns: Evidence from Turkey. European Journal of Social Sciences, 14(3), 404-416.

Chen, A. S., & Lin, J. W. (2014). The relation between gold and stocks: An analysis of severe bear markets. Applied Economics Letters, 21(3), 158-170.

Chkili, W. (2016). Dynamic correlations and hedging effectiveness between gold and stock markets: evidence for BRICS countries. Research in International Business and Finance. 38, 22-34.

Choudhry, T., Hassan, S. S., & Shabi, S. (2015). Relationship between gold and stock markets during the global financial crisis: Evidence from nonlinear causality tests. International Review of Financial Analysis, 41, 247-256.

Ciner, C., Gurdgiev, C., & Lucey, B. M. (2013). Hedges and safe havens: an examination of stocks, bonds, gold, oil and exchange rates. International Review of Financial Analysis, 29, 202-211.

Dirk, G., & Baur, T. K. (2010). Is gold a safe haven? International evidence. Journal of Banking and Finance, 34(8), 1886-1898.

Emmrich, O., & Mcgroarty, F. J. (2013). Should gold be included in institutional investment portfolios? Applied Financial Economics, 23(19), 1553-1565.

Granger, C. W. J. (2001) Essays in econometrics: The collected papers of Clive W.J. Granger. Cambridge: Cambridge University Press.

Hjalmarsson, E., & Osterholm, P. (2007). Testing for cointegration using the Johansen methodology when variables are near integrated (IMF Working Paper, WP/07/141). International Monetary Fund, June.

Al Janabi, M. A. M., Hatemi-J., A., & Irandoust, M. (2010), International Review of Financial Analysis, 19(1), 47-54.

Johansen, S. (1995). Likelihood-based inference in cointegrated vector autoregressive models. New York: Oxford University Press.

Horng, W., & Huang, M. (2013). Threshold model of gold price market on the two stock market returns' influence: Empirical study of Thailand and Malaysian markets. International Review of Management and Business Research, 2(4), 1057-1064.

Kumar, K. S. (2011). Study on dynamic relationship among gold price, oil price, exchange rate and stock market returns. International Journal of Applied Business and Economic Research, 9(2), 145-165.

Mishra P. K., Das J. R., & Mishra, S.K. (2010). Gold price volatility and stock market returns in India. American Journal of Scientific Research, 9, 47-55.

Miyazaki, T., & Hamori, S. (2013). Testing for causality between the gold return and stock market performance: Evidence for 'gold investment in case of emergency'. Applied Financial Economics, 23(1), 27-40.

Mulyadi, M. S., & Anwar, Y. (2012). Gold versus stock investment: An econometric analysis. International Journal of Development and Sustainability, 1(1), 1-7.

Narang, S. P., & Singh, R. (2012). Causal relationship between gold price and Sensex: A study in Indian context. Vivekananda Journal of Research, 1(1), 33-37.

Shahzadi, H., & Chohan, M. N. (2012). Impact of gold prices on stock exchange: A case study of Pakistan (Working Paper Series). Karachi Stock Exchange, 10(2), 1-12.

Smith, G. (2002). London gold prices and stock price indices in Europe and Japan. Unpublished paper. February 2002, pp. 1-31.

Sreekanth, D., & Veni, L.K. (2014). Causal relationship between gold price and S&P CNX NIFTY - An empirical study in Indian context. Asian Journal of Research in Banking and Finance, 4(5), 253-265.

Tuysuz, S. (2013). Conditional correlations between Stock Index, investment grade yield, high yield and commodities (gold and oil) during stable and crisis periods. International Journal of Economics and Finance, 5(9), 28.

Sami Al Kharusi, (1) Esref Savas Basci (2)

(1) Economics and Finance Department, Sultan Qaboos University, Oman

(2) Faculty of Economics and Administrative Sciences, Hitit University, Corum, Turkey

corresponding e-mail: ksami[at]squ(dot)edu{d}om

address: Sultan Qaboos University, Economics and Finance Department, P.o. Box 20; Al Khodh, Postal Code 123, Sultanate of Oman


                  BAHRAIN          KUWAIT            OMAN

Mean              1278.318         6490.286         6203.646
Median            1254.130         6364.600         6201.160
Maximum           1605.980         8430.730         7547.650
Minimum           1035.300         4936.510         4867.000
Std. Dev.          145.5795         802.7583         552.4852
Skewness             0.157839         0.289803         0.244415
Kurtosis             1.598340         2.039601         2.122258
Jarque-Bera        219.9342         134.0623         107.5417
Probability          0.000000         0.000000         0.000000
Sum            3268659.        16595662         15862723
Sum Sq. Dev.  54170320                1.65E+09         7.80E+08
Observations      2557             2557             2557

                   QATAR        SAUDI ARABIA       UAE_ADX

Mean              9742.219         7402.912        3585.843
Median            9275.560         6922.960        3630.370
Maximum          14350.50         11149.36         5253.410
Minimum           6502.930         5323.270        2293.090
Std. Dev.         1837.019         1312.713         953.5949
Skewness             0.515199         0.964351        0.133788
Kurtosis             2.348611         2.874351        1.335882
Jarque-Bera        158.3237         398.0053        302.6725
Probability          0.000000         0.000000        0.000000
Sum           24910854         18929247         9169001.
Sum Sq. Dev.         8.63E+09         4.40E+09        2.32E+09
Observations      2557             2557            2557


Mean             2667.365
Median           2382.810
Maximum          5374.110
Minimum          1301.240
Std. Dev.        1151.349
Skewness            0.477532
Kurtosis            1.874896
Jarque-Bera       232.0483
Probability         0.000000
Sum           6820453.
Sum Sq. Dev.        3.39E+09
Observations     2557


Hypothesized  Eigenvalue  Trace statistic  0.05 Critical  Prob. (**)
No. of CE(s)                                   value

  None (*)     0.015545      129.2455        125.6154       0.0295

Note: Trace test indicates 1 cointegrating eqn(s) at the 0.05 level;
(*) denotes rejection of the hypothesis at the 0.05 level; (**)
MacKinnon-Haug-Michelis (1999) p-values



Dubai Stock        Abu Dhabi Stock Exchange (ADSE)  17.22152   0.0018
Exchange (DSE)     Saudi Stock Exchange (SSE)       18.69425   0.0009
                   Muscat Securities Market (MSM)    9.303134  0.0540
Abu Dhabi Stock    Dubai Stock Exchange (DSE)       11.79596   0.0189
Exchange (ADSE)    Saudi Stock Exchange (SSE)       13.62397   0.0086
                   Qatar Stock Exchange (QSE)        7.835716  0.0978
Saudi Stock        Abu Dhabi Stock Exchange (ADSE)  15.21007   0.0043
Exchange (SSE)     Qatar Stock Exchange (QSE)        8.768569  0.0672
Qatar Stock        Saudi Stock Exchange (SSE)       36.27102   0.0000
Exchange (QSE)     Muscat Securities Market (MSM)   19.60416   0.0006
Muscat Securities  Saudi Stock Exchange (SSE)       64.92152   0.0000
Market (MSM)       Qatar Stock Exchange (QSE)       11.62942   0.0203
                   Kuwait Stock Exchange (KSE)       8.660808  0.0702
                   Saudi Stock Exchange (SSE)       21.62660   0.0002
Kuwait Stock       Qatar Stock Exchange (QSE)       13.39750   0.0095
Exchange (KSE)     Muscat Securities Market (MSM)   10.53437   0.0323
                   Bahrain Stock Exchange (BSE)     18.86820   0.0008
Bahrain Stock      Qatar Stock Exchange (QSE)       11.86250   0.0184
Exchange (BSE)


Dubai Stock           GRANGER CAUSE
Exchange (DSE)        GRANGER CAUSE
                      GRANGER CAUSE
Abu Dhabi Stock       GRANGER CAUSE
Exchange (ADSE)       GRANGER CAUSE
                      GRANGER CAUSE
Saudi Stock           GRANGER CAUSE
Exchange (SSE)        GRANGER CAUSE
Qatar Stock           GRANGER CAUSE
Exchange (QSE)        GRANGER CAUSE
Muscat Securities     GRANGER CAUSE
Market (MSM)          GRANGER CAUSE
                      GRANGER CAUSE
                      GRANGER CAUSE
Kuwait Stock          GRANGER CAUSE
Exchange (KSE)        GRANGER CAUSE
                      GRANGER CAUSE
Bahrain Stock         GRANGER CAUSE
Exchange (BSE)
COPYRIGHT 2019 Prague Development Center
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2019 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Kharusi, Sami Al; Basci, Esref Savas
Publication:Business and Economic Horizons
Geographic Code:7OMAN
Date:Jan 1, 2019
Previous Article:Dividend policy as a multi-purpose mechanism; the case of conventional and Islamic banks before and after the 2008 crisis.
Next Article:Influence of regulatory capital requirements on the self-financing capacity of a banking company.

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters |