EXAMINATION OF THE INTERRELATIONSHIP BETWEEN ARAB MARKETS AND ECONOMIC INDICATORS


Said Elfakhani, Ph.D.
Harvey R. Wickes Chair in International Business
College of Business and Management
Saginaw Valley State University
7400 Bay Rd., C307
University Center, MI 48710Cell: 989-971-9735
Office: 989-964-6076
Fax: 989-964-4699
Email: said.elfakhani@svsu.edu

Nabil El Meslmani
Ph.D. Finance Candidate
Concordia University
Email: n_elmes@jmsb.concordia.ca

Abstract

The stock markets in the MENA region, and particularly in the Gulf, have experienced significant growth and earned exceptional returns in the past decade. This observation deserves further investigation aimed at understanding the dynamics within which these markets have developed. Toward this end, this paper explores whether the sampled seven MENA Arab markets have long-term relationships with specific global economic indicators such as movements in oil prices, gold prices, S&P 500 index, and MSCI index. Such a relationship, if confirmed, could help in assessing whether these markets are mispriced, and the extent to which these markets movements can be predicted. Using cointegration analysis, our tests confirmed the existence of long-term bivariate relationships between oil prices and the markets of KSA, Kuwait, Egypt, Oman, and Tunisia. Tunisia, Oman, and Egypt also exhibited long-term relationships with gold prices. With regard to other international indexes, long-term relationships were found among KSA and Oman with S&P 500 index; whereas KSA, Morocco, and Oman markets were individually integrated with MSCI. We conclude that investors who are bullish about oil prices can find markets cointegrated with oil to offer good investment opportunities. On the other hand, the lack of integration of some markets with the international capital markets could be considered a sign that these markets need more time to mature.

Key words: Arab Markets, MENA, Cointegration, Economic Indicators, Investment Opportunities

I. Introduction
The capital markets in the Middle East North Africa (MENA) region experienced a period of high returns throughout most of the past decade. The various economic stimulants designed by MENA governments are attracting waves of local and foreign investors. The potential gains in these markets raise the question of what factors are driving their success.

This paper examines the existence of long-term relationships between some MENA stock markets on the one hand, and economic and global indicators, namely oil prices, gold prices, S&P 500 index, and Morgan Stanley Capital Index (MSCI) on the other hand.

Long-term relationships, if confirmed, have several implications for regional and international markets. A long-term relationship with oil prices suggests that MENA stock markets can suffer from declining prices when oil prices decrease, if no proper protective measurements are taken by concerned governments. A long-term relationship with indexes such as S&P 500 and MSCI reflects that these markets are interconnected with the US and world markets, which would imply that they can be subjected to the same trends of stagnation, recession, or growth experienced in the rest of the world. A long-term relationship with gold prices indicates that these markets could be interlinked with gold prices on a long-term basis. Moreover, they indicate that gold can be a good alternative investment for people in the MENA region who do not want to invest in local markets.

Prior literature related to capital markets in the MENA region has relied on simple regression techniques. While these techniques may be sufficient, there is inadequate statistical evidence showing that the results of these regression studies were not spurious.

The major contribution of this research is that it addresses the question of what type of long-term relationships dominate stock markets in the MENA region. For that, we use proper econometric techniques adopted commonly by researchers to test long-term relationships with global economic factors and international stock markets. Findings from this study will shed light on the particular forces that drive MENA markets on the long-term. In addition, it can be an important reference for future research and analysis on short-term relationships and causality issues.

Toward this end, we collected monthly data for the stock markets of the Kingdom of Saudi Arabia (KSA), Kuwait, Jordan, Egypt, Oman, Morocco, Tunisia, S&P 500 and MSCI indexes, and oil and gold prices. Using the Augmented Dickey-Fuller (ADF) technique (1979), we test for the existence of unit roots. We use Johansen's Multivariate Approach (JMA) (1991) to test for cointegration.

The results of the bivariate cointegration analysis indicated that the markets of KSA, Kuwait, Egypt, Oman, and Tunisia each have a long-term relationship with oil prices, whereas Morocco and Jordan did not. The stock markets of Tunisia, Oman, and Egypt had a long-term relationship with gold prices, whereas the markets of KSA, Kuwait, Jordan, and Morocco did not. KSA and Oman each had a long-term relationship with S&P 500, whereas Kuwait, Jordan, Morocco, Egypt, and Tunisia did not. KSA, Morocco, and Oman markets had a long-term relationship with MSCI, whereas Kuwait, Jordan, Egypt, and Tunisia did not. Multivariate cointegration results were consistent with bivariate results, with the exception of KSA and Kuwait, which needed further investigation. We conclude that the integration of some markets with oil prices can enhance investment opportunities for those who are bullish about oil prices, particularly markets which are not yet integrated with international markets. On the other hand, the lack of integration of some markets with the international capital markets could be considered a sign that these markets still need more time to mature.

The following section reviews scholarly literature which employed cointegration techniques to study long-term relationships of capital markets. We present the markets they have analysed and the results they have obtained. This same section also lays the foundation of our methodology and the choice of variables considered as global economic indicators. Section III presents a market overview of the countries selected for our analysis. Section IV provides a description of the data and its sources along with the methodology as well as the results of both unit roots tests and cointegration tests. Section V presents the final conclusions and recommendations.

II. Literature Review and Testing Techniques
The finance literature has examined the relationship between stock market performance and a wide range of economic factors. Regression analysis has been the most commonly used
tool for testing possible relationships. Other researchers applied cointegration with time series data as regression techniques may not be adequate for this type of data.

International studies have focused on markets integration in the western world and the group of eight industrialised countries (e.g. Taylor and Tonks, 1989; Kasa, 1992; Arshanapalli and Doukas, 1993; Masih and Masih, 1997a and 1999; Dickinson, 2000; Sheng and Tu, 2000; Lamba and Otchere, 2001; and Bessler and Yang, 2003). Other studies were concerned with integration in the South East Asia and Japan (e.g. Chan et. al., 1992; Ghosh et al., 1999; and Yang et. al., 2003). Some studies explored such relationships between East and West (e.g. Masih and Masih, 1997b and 2001; and Phylaktis and Ravazallo, 2005), whereas Defusco et al (1996) studied three groups of international markets (namely, Latin America, the Pacific Basin and the Mediterranean), and Muradoglu et al (2001) for evidence on Turkish versus international markets.

Unfortunately, little research has been conducted on MENA capital markets, and on the factors affecting returns on stock prices in that region. In their test of the market efficiency hypothesis (EMH) and the random walk hypothesis (RWH), which considered three emerging Gulf markets (Kuwait, Saudi Arabia, and Bahrain), Abraham, Seyyed and Alsakran (2002) corrected for infrequent trading characterising the GCC markets and found that price changes are independent for all three markets. The RWH could not be rejected in the Saudi Arabian and Bahraini markets. Correcting for infrequent trading problem as well as using the Granger causality test, Ratanapakron and Sharma (2002) found that no long-term relations existed among the Middle East, U.S., Europe, Asia, Latin America, and Eastern Europe stock indexes during the pre-Asian crisis period.

Neaime (2004) examined the integration between some Arab markets and the US, UK and France. His results indicated that the stock markets of Egypt, Turkey, Jordan, and Morocco have matured and are cointegrated with the world financial markets. However, the Gulf Cooperation Council (GCC) stock markets appear to be segregated from the rest of the world. Consequently, they can offer diversification potentials to international and regional investors.

Hassan and Jung-Suk (2004) tested the stock markets of Bahrain, Oman, Saudi Arabia, Jordan, Egypt, Morocco, Turkey, US, UK, France, and MSCI. They found no significant cointegration between MENA stock markets and US stock markets. Only when they included UK and French stock markets within developed stock markets did they find a long-term relationship between GCC and developed stock markets. GCC countries appeared to cointegrate with developed stock markets including US, UK, and France, while MENA stock markets generally appeared segmented from US markets.

Elfakhani et al (2008) tested the long-term relationship among Arab stock markets, US and emerging markets indexes during the 1997-2002 sub-period. They show that Jordan and Kuwait, Jordan and Tunisia, Kuwait and Tunisia, and Kuwait and Saudi Arabia are cointegrated, while the others are not, thus creating good diversification opportunities. On the other hand, only Jordan, Kuwait, and Morocco were cointegrated with the US general market index, offering a viable substitute for those investing in the US markets.

Finally, Alsakran and Al-Shaikh (1998) tried to identify the effects of inflation, interest rates, and oil prices on the Saudi stock market. However, they did not provide sufficient evidence that their regression results were not spurious.

Hence, the above studies fall short of finding any economic intuition for the reported cointegration. This paper differs from previous research in that we do not simply rely on regression analysis in defining relationships of capital markets with their major determinants; rather, we have gone further by utilising an econometric test that enables us to determine the existence of a long-term relationship between these markets and some major determinants.
In this study, we have considered four major determinants: oil prices, gold prices, and S&P 500 and MSCI indexes. Oil prices were selected because many of the MENA markets that were analysed are big oil producer economies that depend heavily on oil production. The S&P 500 index was selected as a representative of the U.S. capital market. Being the world's biggest economy and the strongest political player worldwide, most countries have economies highly associated with the U.S. Hence, it is important to check whether markets in the MENA region are affected by the U.S.. Economic globalisation has caused most markets worldwide to be linked to each other. Some economists and financial analysts have argued that global diversification is becoming less valuable in portfolio management as most countries are becoming increasingly interconnected. MSCI is considered representative of the world capital market, and we therefore use it to measure the degree to which MENA countries are integrated in world markets. The last determinant considered is gold. Gold is a favourable alternative investment for investors who want to invest their money in alternatives to the capital markets. We were interested in observing if gold is used as an alternative investment option for individuals who live in the MENA region.

In order to test the long-term relationship between MENA stock markets and the proposed possible determinants, we use the cointegration technique, in lieu of the traditional regression model. Our use of cointegration relates to the nature of certain time series. Regression techniques may fail to represents the relationship among them. This issue deserves further justification and will be discussed next.

Using standard regression techniques, instead of cointegration, with non-stationary data, can lead to the problem of spurious regressions involving invalid inferences (Harris 1995). In particular, the results obtained from analysing non-stationary data suggest that there are statistically significant relationships between the variables in the regression model.

Cointegration is the statistical approach that tests for the existence of a long-term equilibrium relationship among non-stationary variables. In order to test whether non-stationary time series are cointegrated, it is enough to test whether or not the error term resulting from the regression among these time series is stationary. If the error is stationary, then the variables do not cointegrate, and there runs the possibility that the regression obtained is spurious, hence the use of the ADF to determine data stationarity (Harris 1995, Ch. 3). The ADF tests the null hypothesis that a series does contain a unit root (non-stationary) against the alternative of stationarity. A series is called Integrated of order 'd', I (D), if it contains 'd' unit roots. It is important to emphasise that too few lags may result in over-rejecting the null of unit roots when it is actually true. On the other hand, too many lags may reduce the power of the test. We have adopted the optimal lag as suggested by EasyReg software. JMA Technique is applied in testing for the existence of cointegration in our time series.

III. An Overview of MENA Markets
Seven Arab markets were sampled, four of which (Saudi Arabia, Kuwait, Oman, and to a lower degree Egypt) are oil-producing, and the remaining three are non-oil producing (Jordan, Morocco and Tunisia). Understanding the major characteristics of the above sampled Arab countries will allow us to draw better insights from the cointegration results.

The economic performance of each of the four aforementioned oil-producing countries has been strengthened by high oil prices. High oil prices have generated relatively huge revenues and have led to increased domestic liquidity, which combined with deregulation of some key sectors, increased confidence in the equity market.

In Saudi Arabia, high domestic liquidity and low interest rates have caused the stock market index to more than double since early 2004, before the sharp decline in 2006. The state's monopoly on mobile telecommunications has been broken. Confidence in the equity market has improved after the appointment of a special board to the Capital Market Authority (CMA) in July 2004. Several IPOs were launched in the past decade and the outcome has been encouraging to date.

Kuwait is the most oil-dependent economy among GCC countries. Hence, higher oil prices produced exceptionally large current account and fiscal surplus, leading to increased government spending. The fall of the Iraqi regime of Saddam Hussein, which posed a continuous threat to Kuwaiti nationals, increased domestic confidence and spending abilities. The Kuwait Stock Market Exchange (KSE) performed strongly in 2003 and 2004. Daily trading volumes and amounts reached a level acceptable enough to avoid infrequent stock trading syndrome. However, the Kuwait stock market was perceived as overvalued and stock prices had taken a plunge in 2006.

Oman's oil production decreased in 2004. However, this decline in oil production was offset by an increase in oil prices worldwide, which still reflected positively on Oman's economy. In addition, the rising prices of Liquefied Natural Gas (LNG) were also helping. The Omani government has also been trying to diversify its economy by taking positive steps toward developing hydrocarbon industries, and nurturing its tourism sector. A free trade agreement with the U.S. is also underway, and is expected to increase Omani exports to the U.S.

The privatisation of OmanTel was the biggest IPO in Omani history, and future privatisation plans are expected. The Muscat security market has also performed well in the last few years. The Omani market increased in 2004 by almost 70 percent over the previous year, sending positive signs of the future of the Omani capital market to investors. Trading volume continued to increase reflecting rising interest in the market and eliminating the problem of infrequent trading. This trend in the Omani economy, combined with strengthened investor confidence in the Omani capital market, conveys an optimistic view of the Omani Capital Market.

In addition to limited oil production, Egypt started to rely on natural gas production after discovering a new reserve estimated to last for 120 years. The construction of two large LNG
plants, and the government strict plan to focus on gas production, has made Egypt the world's sixth largest producer of LNG. Egypt has an ongoing plan to link its natural gas production to the European markets. The increase in oil and gas prices has boosted Egypt's fortunes. However, oil production is progressively declining due to the depletion of Egypt's oil reserves.

The depreciation of the pound since mid 2003 has directly increased Egyptian exports and increased the availability of foreign currencies, which itself enhanced domestic confidence and GDP. The government has also adopted significant customs reforms that have added to the confidence of local and foreign investors. Tourism started to grow recently after the September 11 attacks and the Iraq war. Moreover, tourism was boosted by the depreciation of the Egyptian pound, continued infrastructural development, and successful worldwide advertisement campaigns. Revenues from the Suez Canal have also risen due to an increase in shipment activities in the region. On another front, the stock market has performed very well, boosted mainly by export oriented firms. This improvement was steady despite the fact that listed companies were no longer partially exempt from their tax, pushing some companies to deregister from the stock market. Finally, privatisation in Egypt was moving at a relatively slow pace, but the government is expected to bring privatisation to acceptable levels soon.
The non-oil producing countries in our sample (i.e., Jordan, Morocco and Tunisia) have also experienced significant developments. The war in Iraq has negatively affected Jordan's economy, but there have been other improvements since then. For instance, the manufacturing sector in Jordan is growing, and tourism is progressively flourishing. Demand for housing has increased, leading to a construction boom. Textile exports, mostly to the U.S., have increased the country's foreign currency reserve. The stock market's rally upward is probably causing some stocks to be overvalued. In fact, almost ten companies dominate the entire stock market, making it less diversified and more sensitive to shocks and bad news. Finally, more privatisation projects have been undertaken by the Jordanian government.

The Moroccan economy, which is mostly driven by its agricultural sector, has performed well in the past few years. Its economy is thus especially influenced by weather changes and drought conditions. The lack of natural resources such as oil and gas will put the Moroccan economy under pressure if oil prices continue their upward trend. Nevertheless, tourism and emigrant remittance play an important role in the economy. Also, the government is trying to expand the role of the private sector in the economy and to attract foreign investment. In this context, tariffs have been reduced significantly. Consequently, a privatisation plan for several public sectors is under way. Moreover, the capital market in Morocco has been performing well. Privatisation is proceeding at a good pace. Trading volume in the stock market has also increased, helping the market rid itself of an infrequent trading problem.

The agricultural sector constitutes 12 percent of the Tunisian economy, rendering it dependent on weather conditions, similar to Morocco. Structural reforms are being implemented by the government to enhance the investment climate and reduce country risk. For instance, there are plans to remove tariffs with the European Union (EU), and this is expected to increase Tunisian trade with the EU. Also, the tourism sector in Tunisia has grown recently to the credit of careful government planning. The Tunisia Stock Exchange has performed well and has enhanced investors' confidence, however it is still relatively small compared to other markets in the MENA region. This is due to several problems such as the small size of the market compared to the country's GDP, the lack of liquidity, and the dominance of financial stock companies. In addition, regulation requirements tended to keep foreign investors away from the Tunisian market. Nevertheless, privatisation is progressing slowly but gradually in Tunisia. The privatisation of the government stake in the GSM Telecom's Operators was the biggest in recent history.

IV. Empirical Tests and Results
Monthly data for the Indices, oil and gold prices, and MSCI were gathered from Reuters from October 1998 until August 2005. This sampling period constitutes eighty-three months. We opted to limit our study to data from 1998 to 2005 for several reasons. Although some Arab markets started half a century ago, they were small and unorganised. Most of these markets matured post to the Asian crisis in 1997-98. Our sample ended in 2005 because it marked the start of a new era of a sharp rise in oil prices (almost four-fold), disturbances in global stock markets, the tumultuous political events that have afflicted the US starting with the Iraq war, the recent US subprime mortgage crisis, the depreciation of the US dollar in relation to strong competitive currencies (e.g., Euro and GBP). At the regional level, there has been accelerated talk since 2005 about starting a free trade Arab market zone, and a new wave of accelerated rising stock prices followed by sharp declines in 2008. All of these factors and events can influence the long-term relationship between Arab markets and global markets, and render the 2005-2009 period noisy, thus confounding our results, so we decided to restrict our sampling period to 2005.
Table 1 presents major statistics for the seven Arab markets over the 1998-2003 sampling period, as available.(1) Panel A of Table 1 shows the annual market capitalisation and dollar trading volume, while panel B of the same table shows the number of shares traded and the number of listed companies.

Recently, equity markets have surged in almost all MENA countries, with an average increase in 2003 of 124 percent over the1998 level (Table 1, Panel A). However, this growth was mainly driven by the 186 percent increase in oil-producing countries compared to a 45 percent decrease in non-oil producing Arab countries. A similar pattern to equity capitalisation existed for the total dollar value of traded stocks from 2000 until 2003. The overall growth in their values was 31 percent, while it was 34 percent and -37 percent for oil producing versus non-oil (1) producing countries, respectively.

Two factors contributed to the higher capitalisation trend: the shift of investors' focus from international markets (which have been in decline since 2000) to regional markets, and the low interest rates on domestic and international deposits. Capital that has been repatriated from overseas rose in the past five years, adding to the stock of excess liquidity in the region. The latest data from the Bank of International Settlement show that Saudi and other Gulf citizens have been moving part of their funds out of the international banking centres. Also, starting mid-2001, Arab investors have been diversifying their international portfolios in response to falling interest rates, tumbling equity markets, and threats facing the security of their investments abroad. For example, total deposits at Saudi banks rose 7.5 percent, from $85.3 billion at the end of 2002 to $91.7 billion by August 2003. Liquidity proxied by the number of shares sold has also increased significantly in Arab markets (Table 1, Panel B). The overall growth in traded shares was 270 percent, from 2000 until 2003, mostly driven by oil-producing countries in our sample (275 percent and 72 percent for oil versus non-oil producing countries, respectively). However, the number of listed companies has dropped over the same period (-4 percent, -1 percent and -24 percent for the overall sample, oil and non-oil producing countries, respectively).

Table 2 illustrates the Indices selected and their Reuters Tick Code. The data are rearranged based on initial formats and corresponding values expressed in natural logarithm. Unit roots tests were conducted on every time series to determine whether or not it is stationary.

Table 2. Data Sources and Codes. Sampled MENA Arab countries names, variable name and variable tick code are shown here. All data are sourced from REUTERS.

A. Methods Followed in Conducting Unit Roots Test

The EasyReg software tests the following unit root null hypothesis:

H0: Z(t) is not a trend stationary process

Against the following alternative hypothesis:

H1: Z(t) is a trend stationary process

Failing to reject the null hypothesis means that the time series is not stationary and that it contains at least one unit roots. On the other hand, rejecting the null hypothesis means that the data is stationary. All sampled countries time series except Jordan were found to be nonstationary. Jordan was found to have a stationary time series data based on critical values of 10 percent confidence.

The main interest in this approach is that if two series are stationary, then a simple regression between them would not result in spurious results and the need for cointegration would be eliminated. However if data series are non-stationary, the cointegration test is needed to determine a long-term relationship among them. We have to note that if two series do not have the same order of integration, then they cannot be cointegrated. However, if the number of series exceeds two, then a combination of these series could be cointegrated even though they do not have the same order of integration. In our analysis we did not investigate the exact order of each time series. Instead, we focused on determining whether series are stationary or not. The justification for this is that if two time series contain different integration order, and we are testing them for cointegration, then the results would indicate no cointegration vector to start with.

The results of cointegration reported in Table 3 are divided into seven panels, one for each of the seven sampled Arab countries.(2) For each country, the first test is set to determine whether a multivariate cointegration exists between the country stock index and the set of four time series variables, namely oil prices, gold prices, S&P 500 index, and MSCI. Also, bivariate cointegration was conducted between a country index and each of the four time series separately. The reported results list the variables that were tested, the lag length that was used in conducting the tests, the resulting number of cointegrating vectors obtained, and the cointegrating vectors that have resulted from the test if they do exist.

Our results for Saudi Arabia are reported in Panel A. The test between the KSA index and all four variables confirms the existence of two cointegrating vectors that link the four variables together on a long-term basis. The first vector presents a minimal inverse contribution of oil prices in the long-term. To maintain the long-term equilibrium drawn from the first vector, KSA stock markets will move in the same direction with gold prices and S&P 500 index, whereas it will move in an opposite direction with respect to MSCI. The second vector presents an insignificant positive contribution in equilibrium to the long-term relationship of both the Saudi stock market and oil prices.

Results from bivariate tests have generated one cointegration vector that determines a long-term relationship between the KSA market and oil prices. When oil prices increase, the Saudi market will increase on a long-term basis by a factor equal to almost '0.45' in order to preserve the long-term equilibrium. The Saudi Market tended to decrease by a factor of '1' when S&P 500 increase by a factor of '0.80' in order to preserve the long-term equilibrium, and tends to decrease by a factor of '0.83' when MSCI increases by a factor of '1' in order to preserve the long-term equilibrium. On the other hand, there is no long-term cointegration between the KSA market and gold prices.

The findings of bivariate tests for the Saudi market are consistent with the reported development in Saudi Arabia. The regulatory openness toward foreign investments, increased liquidity, increased number of listed companies, membership in the World Trade Organization (WTO), and the privatisation of some public sector institutions, all supported by steady and growing strategic oil commodity production and revenues, allowed the Saudi market to prosper on a long-term basis. On the other hand, lower global interest rates as well as uncertainty about international investments provide a plausible explanation for the inverse cointegration with S&P 500 and MSCI.

When testing the Kuwaiti market, the results in Panel B confirmed the existence of long-term equilibrium with the Kuwaiti market moving in the same direction as gold prices, S&P 500, and MSCI indexes. A weaker and negative relationship existed between Kuwaiti markets with oil prices in the presence of the other variables. At the bivariate level, as oil prices increase, the Kuwaiti Stock Market tended to increase on a long-term basis by a factor equal to almost '0.71' in order to preserve the long-term equilibrium; while no independent long-term relationship existed between the Kuwaiti market and gold prices, S&P 500, and MSCI indexes.

The Kuwaiti findings imply that the Kuwait stock market is still not internationally integrated. However, this lack of integration in the long run also suggests that the Kuwaiti markets offer good diversification opportunities for international portfolio managers. The high association with oil prices leaves it subject to the fluctuations of oil prices. This volatility, however, can be offset by the presence of rich oil reserves and strong foreign reserves that can protect the market from a crash by injecting liquidity in the system whenever necessary.

With regard to the Omani stock market (Panel C), we find that in order to preserve the long-term equilibrium established, the Omani market had to move in the same direction as MSCI and in the opposite direction to S&P 500, while oil and gold prices do not contribute significantly in the multivariate cointegrating vector.

At the individual level, we found one cointegration vector that determined a positive long-term relationship between the Omani stock market and oil prices by a factor equal to almost '0.86' in order to preserve the long-term equilibrium. Similarly, the Omani stock market would increase by a factor of almost '0.36' whenever gold prices increase by a factor equal to '1', would increase by a factor of '0.49' whenever the S&P 500 increases by a factor of '1', and would increase by a factor equal to '0.62' whenever MSCI increases by a factor of '1', in order to maintain the long-term equilibrium.

Distinct from other Gulf markets, the Omani market seems to be well-integrated with the international capital market. As the country with the least oil production capacity, Oman experienced more pressure than other countries to finance its spending using non-oil resources, which made the country more affected by international shocks, and rendered the Omani stock market cointegrated with international markets, in addition to being cointegrated with oil prices.
In panel D, we found that Egypt did not cointegrate significantly in the first vector with the four global variables. For the second vector, in order to preserve the long-term equilibrium established, the Egyptian market had to move in the same direction with oil prices, gold prices, and S&P 500, and in the opposite direction to MSCI. When testing the index with oil prices only, as oil prices increased the Egyptian stock market tended to increase on a long-term basis by a factor equal to almost '0.90', in order to preserve the long-term equilibrium. Also, as gold prices increased, the Egyptian market index tended to increase on a long-term basis by a factor equal to almost '0.36', while no cointegration between the Egyptian stock market and S&P 500 and MSCI, respectively.

Our tests conveyed that the lack of cointegration between the Egyptian market and international markets may be due to hesitation about confidence in economical and political reforms in Egypt. The Egyptian capital market, however, has more upside potential that has not yet been reflected in the stock market values, especially with the new gas reserves being used to boost the economy, tourism, and Suez Canal revenues. This suggests that the Egyptian economy is the most diversified compared to its neighbouring countries in the MENA region. For example, taking all these factors into consideration, the Egyptian market should be less affected by fluctuations in oil prices than Saudi Arabia and Kuwait.

The last three panels in Table 3 report the results for non-oil producing countries. Panel E summarises the findings for the Jordan stock market index relationship with oil prices, gold prices, and S&P 500 and MSCI indexes. There existed two cointegrating vectors linking Jordan and the four variables together on long-term basis. Jordan, however, is the least contributor in the first vector. For the second vector, in order to preserve the long-term equilibrium established, the Jordanian market had to move in the same direction as MSCI and gold prices, and in an opposite direction to S&P 500 and oil prices. With regard to paired cointegration tests between the Jordanian index and each of the four variables, we could not perform any of them as the Jordanian market was found to be non-stationary at the five percent significance level, but stationary at the 10 percent significance level. The stationarity at the 10 percent significance level renders the result of bilateral cointegration tests with non-stationary time series inconclusive. In brief, the Jordanian market still needs to mature. The Iraqi war, economic problems, and the dominance of ten companies on the stock markets make investment opportunities in the Jordanian market limited and non-beneficial for diversification purposes. However, investment in the Jordanian market can serve as a good hedge tool for international investors.

The test results on whether there is cointegration among the Moroccan stock market index with oil prices, gold prices, S&P 500, and MSCI indexes are reported in the first row of Panel F of Table 3. We noticed that gold prices do not contribute significantly in the cointegrating vector, while in order to preserve the long-term equilibrium, the Moroccan market had to move in the same direction with MSCI, and in the opposite direction to S&P 500 index and oil prices. At the level of paired cointegration tests, we find no cointegration between the Moroccan stock market and each of oil prices, gold prices, and S&P 500 index. Yet, there is a long-term relationship between the Moroccan stock market and MSCI. The Moroccan stock market will increase by a factor of '1' whenever the MSCI increases by a factor equal to almost '0.81', in order to preserve the long-term equilibrium.

Despite the fact that the Moroccan market is not directly linked to oil prices, it remains a good hedge for investors of those regions where the markets are integrated with oil prices on a long-term basis. In addition, the fact that the Moroccan stock market is not really associated with S&P 500 whereas it is linked to MSCI can be explained by the observation that the Moroccan economy relies more on the European Union than on the U.S..

The last Panel G in Table 3 shows the findings for the Tunisian market. The cointegration test confirmed the existence of two cointegrating vectors that link the Tunisian stock market with the four time series variables together on a long-term basis. However, Tunisia does not contribute significantly to both the first and second vectors. Consequently we cannot build on these two vectors to make any inferences.

While the bivariate cointegration tests indicate no long-term relationship between the Tunisian stock market and both S&P 500 index and MSCI, we find that as oil prices increased by a factor equal to almost '0.16', the Tunisian stock market index tended to increase on a long-term basis by a factor equal to '1', in order to preserve the long-term equilibrium. Similarly, as gold prices increased by a factor equal to almost '0.18', the Tunisian stock market tended to increase on a long-term basis by a factor of '1'.

The lack of cointegration among the Tunisian market and the four global variables, and the finding that the long-term relationship of the Tunisian market index with oil and gold prices is relatively low, indicate that the nature of the Tunisian market as still being in the early developing phases, in addition to other problems, has made the integration of the Tunisian market with the rest of the world relatively slower than other markets in the region.

V. Conclusions and Recommendations
In this paper, we have tried to determine whether a long-term relationship exists between seven MENA Arab capital markets and some relevant economic indicators including oil prices, gold prices, S&P 500 and MSCI indexes. For this purpose, monthly data were collected on the stock market indexes of the involved markets for the period extending from October 1998 until August 2005.

Cointegration tests were conducted on the stock markets of each of the seven selected Arab countries with the four candidate determinants mentioned above. Results differ from one Arab country to another. All four oil-producing countries (i.e., KSA, Kuwait, Oman, and to a lesser degree Egypt) as well as Tunisia appeared to be especially integrated with oil prices. In addition, the markets of Oman, Egypt and Tunisia were also integrated with the other strategic product gold. On the other hand, Oman was cointegrated positively and Saudi Arabia negatively with international markets proxied by MSCI, while Morocco was cointegrated with MSCI.

We conclude that most MENA markets still need time to mature and become integrated with world markets, although they are expected to do well as long as oil prices remain high for oil-producing countries, and the trend for economic diversification and more international trade engagement continues for non-oil producing countries. Jordan and Morocco seem to be shielded from fluctuations in oil prices, and hence they can serve as a good investment opportunity to Gulf and international investors who want to hedge their exposure to oil prices.

This paper is not without limitations. The lack of monthly samples for some Arab Indexes prevented us from including them in our analysis. These markets, if included, might have added more insight on the long-term behaviour of the remaining MENA capital markets.

Finally, we suggest a number of directions for future research that can extend the present work and lead to independent, but related, research:

1. Some markets that were missed in this study are presently highly active (e.g. UAE and Bahrain). They are attracting large investments and are expected to become more mature with time.
2. Future research should also conduct short term analysis and causality tests.
3. New variables can be useful in understanding our results, such as the monthly trading volumes, natural gas prices, size of oil reserves, interest rates levels, inter-bank rates, credit spread, S&P100 as a representative of large cap companies, and Russell 2000 as a representative of small cap companies.
4. It would be useful to test whether the chosen economic determinants have long-term relationships with one another, and consequently, whether they are affecting each other.
5. In order to avoid possible bias when testing long-term relationships between the selected Arab countries and the international index, we recommend the use of another international index that does not include MENA region in its composition.
6. In our analysis, some multivariate cointegration tests generated two cointegrating vectors (two long-term relationships). In such cases, it would be useful if future research tested which of these two vectors is good for hedging.

Footnotes:
1. Sources for Table 1 include central banks, ministries of finance, stock exchanges, and Arab Monetary Fund
websites, 2003, ESCWA surveys of Economic and Social Developments in the ESCWA Region, 2001 and 2002.
2. 2 Numbers for 2003 in Morocco were not available; this fact understates the total for all values in Table 3.

References
1. Abraham, A., F.J. Seyyed and S.A. Alsakran, 2002. Testing the random walk behaviour and efficiency of the Gulf Stock Markets, The Financial Review 37 (3), 469-480.
2. Alsakran, S. and S. Al-Shaikh, 1998. Macro-Economic Forces and the Saudi Stock Market. Proceedings of the Administrative Sciences Conference at KFUPM, March 1998.
3. Arshanapalli, Bala and John Doukas. "International Stock Market Linkages: Evidence from the Pre- and Post-October 1987 period." Journal of Banking and Finance 17 (1993): 193- 208. North-Holland.
4. Bessler, David A. and Jian Yang. "The Structure of Interdependence in International Stock Markets". Journal of International Money and Finance 22 (2003): 261 - 287.
5.Chan, Kam C., Benton E. Cup and Ming-Shiun Pan. "An Empirical Analysis of Stock Prices in Major Asian Markets and the United States." The Financial Review 27, 2 (May 1992): 289-307.
6. Defusco, Richard A., John M. Geppert, and George P. Tsetsekos. "Long-Run Diversification Potential in Emerging Stock Markets." The Financial Review 31, 2 (May 1996); ABI/INFORM Global.
7. Dickey, David A. and Wayne A. Fuller, 1979. Distribution of the estimators for autoregressive time series with a unit root, Journal of the American Statistical Association 74, 427-431.
8. Dickinson, David G. "Stock Market integration and Macroeconomic Fundamentals: An
Empirical Analysis, 1980-95." Applied financial Economics 10 (2000): 261-276.
9. Elfakhani, S., M. Arayssi, and H. Smahta, 2008. "Globalization and Investment Opportunities: A Cointegration Study of Arab, U.S., and Emerging Stock Markets." The Financial Review, Nov. 2008, Vol. 43, pp. 591-611.
10. Ghosh, Asim, Reza Saidi, and Keith H. Johnson. "Who Moves the Asia-Pacific Stock Markets-US or Japan? Empirical Evidence Based on the Theory of Cointegration." The Financial Review 34 (1999): 159-170.
11. Harris, Richard I.D. Using Cointegration Analysis in Econometric Modeling. New York: Prentice Hall, 1995.
12. Hassan, Kabir and Jung-Suk Yu. "Financial Integration and Volatility Spillovers on the MENA Stock Markets." Departments of Economics and Finance, College of Business Administration, University of New Orleans, 2004, (Unpublished).
13. Johansen, S., 1991. Estimation and hypothesis testing of cointegration vectors in Gaussian Vector Autoregressive Models, Econometrica 59, 1551-80.
14. Kasa, Kenneth. "Common Stochastic Trends in International Stock Markets". Journal of Monetary Economics 29 (1992): 95-124, North-Holland.
14. Lamba, Asjeet S. and Isaac Otchere. "An Analysis of the Dynamic Relationships between theSouth African Equity Market and Major World Equity Markets." Multinational Finance Journal 5, 3 (September 2001): 201; ABI/INFORM Global.
15. Masih, Rumi and Abul M.M. Masih. "Long and Short Term Dynamic Causal Transmission amongst International Stock Markets." Journal of International Money and Finance 20 (2001): 563-587.
16. Masih, Abul M.M. and Rumi Masih. "Are Asian Stock Market Fluctuations Due Mainly to Intra-Regional Contagion Effects? Evidence Based on Asian Emerging Stock Markets." Pacific-Basin Finance Journal 7 (1999): 251-282.
17. Masih, Abul M.M. and Rumi Masih. "A Comparative Analysis of the Propagation of Stock Market Fluctuations in Alternative Models of Dynamic Causal Linkages." Applied Financial Economics 7(1997a): 59-74.
18. Masih, Abul M.M. and Rumi Masih. "Dynamic Linkages and the Propagation Mechanism Driving Major International Stock Markets: An Analysis of the Pre- and Post- Crash Eras". The Quarterly Review of Economics and Finance 37, 4 (Fall 1997b): 859-885.
19. Muradoglu, Gulnur, Kivilcim Metin and Reha Argac. "Is There a Long Run Relationship between Stock Returns and Monetary Variables: Evidence from an Emerging Market"? Applied Financial Economics 11 (2001): 641-649.
20. Neaime, Simon. "Liberalization and Financial Integration of MENA Stock Markets." American University of Beirut/ Department of Economics, 2004, (Unpublished).
21. Phylaktis, Kate and Fabiola Ravazzolo. "Stock Market Linkages in Emerging Markets: Implications for International Portfolio Diversification." Int. Fin. Markets, Inst. And Money 15 (2005): 91-106.
22. Ratanapakron, O. and S.C. Sharma, 2002. Interrelationships among regional stock indices, Review of Financial Economics 11(2), 91-108.
23. Sheng, Hsiao-Ching and Anthony H. Tu. "A Study of Cointegration and Variance Decomposition among National Equity Indices before and during the Period of the Asian Financial Crisis." Journal of Multinational financial Management 10 (2000): 345-365
24.Taylor, Mark P. and Ian Tonks. "The Internationalization of Stock Markets and the Abolition of U.K. Exchange Control." The Review of Economics and Statistics 71, 2 (May, 1989): 332-336.
25.Yang, Jian, James W. Kolari and Insik Min. "Stock Market Integration and Financial Crises: the Case of Asia". Applied Financial Economics 13 (2003): 477-486.