The Effect of Implicit Market Barriers on Stock Trading and Liquidity

Transcription

1 The Effect of Implicit Market Barriers on Stock Trading and Liquidity Asem Alhomaidi Department of Economics and Finance University of New Orleans New Orleans, LA M. Kabir Hassan Professor of Finance Hibernia Professor of Economics and Finance Department of Economics and Finance University of New Orleans New Orleans, LA William J. Hippler (Corresponding Author) Assistant Professor of Finance College of Business and Public Management University of La Verne 1950 Third Street La Verne, CA

2 The Effect of Implicit Market Barriers on Stock Trading and Liquidity Abstract In this study, we compare the performance of Islamic and conventional stock returns in Saudi Arabia in order to determine whether the Saudi market exhibits characteristics that are consistent with segmented market and investor recognition effects. We sample the daily stock returns of all Saudi firms from September 2002 to 2015 and calculate important measures, including stock return, market integration (Pukthuanthong and Roll, 2009), turnover, liquidity (Amihud, 2002), and idiosyncratic volatility (Ang et al, 2006). Integration tests report that Islamic stocks are more sensitive to changes in key Saudi macroeconomic variables than conventional stocks, supporting the hypothesis that the Islamic and conventional stock markets are segmented in Saudi Arabia. In addition, our results show that Islamic stocks are more liquid, have lower idiosyncratic risk, and exhibit higher systematic turnover than conventional stocks, which supports the investor recognition hypothesis. Our results provide new evidence on asset pricing in emerging markets and the evolving Islamic financial markets. Keywords: segmented markets, Islamic finance, emerging markets, asset pricing, investor recognition

3 1. Introduction and Motivation Emerging markets have always been an important source of economic and financial market growth, and Islamic financial markets have begun to have a major impact in the global economy as well. In particular, much research has focused on asset-pricing issues in light of emerging and Islamic financial markets, as these markets can provide new insights into the efficiency of the global financial markets. Recently, in the context of Islamic finance, Merdad, Hassan, and Hippler (2015) investigate the Islamic-effect in a cross-sectional stock return context by looking at the difference in stock returns between Islamic and conventional firms in Saudi Arabia, and they find that, by complying with Sharia law, firms have a lower stock return on average than non-compliant, or conventional, firms. We extend their analysis by examining whether the return difference between the two classes of stocks are due to market segmentation, caused by religious-based restrictions on the portfolios of Islamic investors, or from liquidity effects caused by investor recognition, or both. It is important to examine these issues, because the existence of market segmentation and investor recognition effects has important implications as to the efficiency of emerging and Islamic financial markets. We examine the stock market of Saudi Arabia for several important reasons. First, the majority of traders in the Saudi stock market are individual traders, which gives us the opportunity to test the effect of individual recognition on trading behavior. Secondly, Saudi Arabia has a majority Muslim population, and it is

4 known for its strong adherence to Sharia law, which gives us the opportunity to test the effect of religious beliefs on investor investment decisions and portfolio construction. Third, in Saudi Arabia, as a way of increasing the public awareness of Sharia roles, clerics and Islamic finance scholars voluntarily screen stocks and financial instruments for their Sharia compliance and attempt to disseminate this information to the public through different media channels. This enables us to study the effect of such information on individual trading activity and determine whether this type of stock classification acts to increase the recognition of Sharia compliant stocks. Finally, Saudi banks, especially Islamic banks, provide the stock market with individual investors through their Murabaha personal loan product, which is an Islamic loan, whereby a bank sells stock on credit to a customer. The customer has the discretion to choose among a portfolio of stocks, but the stocks must be Sharia-compliant. The Murabaha loan is very common in Saudi banking, because of the ease of ownership transfer between the bank and potential borrowers. Therefore, the use of Islamic banking products has led to an increase in the involvement of relatively inexperienced individual investors in the stock market. Consequently, potential borrowers that may initially have no interest in investing in the stock market are incentivized to gather information pertaining to Sharia compliant stocks in order to maximize their profit. In addition, in order to benefit the public, the Saudi Capital Market Authority has adapted the strategy of allocating the majority new Saudi IPOs to individual investors. This allocation process, along with high post-ipo gains, has also caused large numbers of

5 individual investors to engage in IPO purchases and has increased the popularity of Saudi IPOs. As a consequence of these trends in the Saudi market, individual investors have become very active in trading Saudi stocks, and, as documented by Tadawul, 90 percent of Saudi stocks are traded by individual investors 1. Therefore, using the sample of Saudi stocks makes it possible to examine the impact of shariacompliance on measurable market segmentation and investor recognition. As a result of increased interest in pertinent stock information from relatively inexperienced Saudi investors, there also began an increased interest by Islamic scholars in educating investors about Islamic finance. As a part of efforts by Islamic scholars in providing important information to potential investors, Islamic scholars began screening stocks in order to classify them as Sharia compliant (Islamic), or non-sharia compliant (conventional). Depending on each investor s level adherence to their Islamic beliefs, Islamic investors typically refrain from trading stocks that are not compliant with Sharia law (conventional stocks) and will only trade Islamic, or Sharia compliant, stocks. On the other hand, non-muslim investors, who care less about Islamic religious restrictions, will trade stocks regardless of whether they are Sharia compliant (Islamic) or not. Based on this system, Islamic stocks will be accepted by a wider base of both Islamic and conventional investors, while conventional stocks will be traded only by investors that do not place an emphasis on Sharia compliance. We expect that, since the majority of stock trades in Saudi Arabia are conducted by individual investors, a stock s Sharia classification will affect its 1 From Tadawul periodical publications at

6 trading and liquidity through two channels. First, Sharia classification creates an implicit barrier between two classes of stocks by restricting Islamic investors to trade only stocks that are classified as Islamic, or Sharia compliant. We call this the market segmentation hypothesis, or the impediment-to-trade hypothesis. To illustrate the segmentation hypothesis, consider a two-country model of partial segmentation similar to that of Errunza and Losq (1985) in which investment barriers are asymmetric: country 1 s investors can invest in country 2 s securities, but country 2 s investors are prohibited from investing in country 1 s securities. Errunza and Losq (1985) show that country 2 s (eligible) securities are priced as if markets were completely integrated, but country 1 s (ineligible) securities command a super risk premium. If a company from country 1 cross-lists its shares in country 2, comparative statics show that the super risk premium disappears, the share price increases, and the expected return decreases. A Similar situation exists in segmented markets, where firms issue restricted shares that only local citizens can hold and unrestricted shares that can be held by both local and foreign investors. Studies uniformly find that unrestricted shares trade at premium prices, relative to those of restricted shares. Bailey, Chung and Kang (1999) study the impact of barriers to international capital flows using stock price data from eleven countries, and they find that unrestricted shares have large price premiums, compared to the restricted shares, except for the case of Chinese stocks 2. Therefore, 2 Bailey (1994) suggests that the lack of alternative investments to low-yielding bank accounts drives domestic Chinese savings into stock investments and pushes prices beyond what foreigners are willing to pay. He concludes that local Chinese demand for any available investment vehicle is even greater than the foreign demand to invest in China. Lee (2009) provides another explanation for the price premium of the restricted shares by proposing that they provide better market liquidity than the unrestricted counterparts.

7 an explanation of what occurs in the Islamic markets is that there are two types of investors, Islamic and conventional investors. Islamic investors trade only in Sharia-compliant stocks and cannot trade in conventional stocks, whereas conventional investors can trade in both Islamic and conventional stocks. Secondly, Sharia classification actively promotes Islamic stocks by increasing the base of potential investors for Islamic stocks and, therefore, makes Islamic stocks recognized by a greater number of investors than conventional stocks. The wider base of potential investors for Islamic stocks facilitates the dissemination of new information related to Sharia compliant stocks and, therefore, increases their liquidity and trading activity. We call this hypothesis the investor recognition hypothesis. The idea that neglected stocks earn a return premium over recognized stocks has been in existence for many years (e.g., Arbel et al., 1983). Merton (1987) develops an asset pricing model that explains this apparent pricing anomaly. The key difference between Merton s model and standard asset pricing models such as the CAPM is that Merton s model assumes that investors only know about a subset of available securities and that these subsets differ across investors. A number of studies examine the investor recognition hypothesis. Loughran and Schultz (2005) compare Islamic and conventional firms and investigate whether the visibility of Islamic firms affect liquidity. Particularly, they show how investor access to information and familiarity with the firm affect stock liquidity. Their

8 results indicate that, after adjusting for size and other factors, conventional firm (firms with low visibility) stocks trade much less, are covered by fewer analysts, and are owned by fewer institutions than Islamic firms (firms with high visibility). Additionally, using an event-study methodology, Amihud, Mendelson and Uno (1999) find that a reduction in the minimum trading unit of a stock facilitates liquidity, increases the firm s shareholder base, and is associated with a significant increase in stock price. In addition, Kadlec and McConnell (1994) find that firms that announce a listing on the New York Stock Exchange (NYSE) earn abnormal stock returns of 5 percent in response to the listing announcement and experience an increase in the number of shareholders, followed by a decrease in bid-ask spreads. Moreover, Foerster and Karolyi (1999) find that non-u.s. firms that cross-list on U.S. exchanges earn cumulative abnormal stock returns of 19 percent during the year before cross-listing, but incur a loss of 14 percent during the year following crosslisting. Also, the cross-listing decision is associated with an increase in the shareholder base and a subsequent fall in returns. Therefore, Foerster and Karolyi (1999) conclude that there is evidence in favor of the market segmentation and investor recognition hypotheses. Finally, Chen, Noronha and Singal (2004) find that inclusion into the S&P 500 index leads to a permanent positive stock price increase, but there is no decline for firms that de-list. We examine the sample of daily Saudi Arabian-listed stock returns from 2002 to 2015, and our results support the segmentation and impediment-to-trade

9 hypothesis. Conventional stocks are isolated and segmented from Islamic stocks in the Saudi market, which means that there exists an implicit barrier to trade between them. By applying the Pukthuanthong and Roll (2009) measure of market integration, we find that Saudi macroeconomic factors better explain Islamic stock returns than conventional stock returns. In some regressions, the percentage of macroeconomic factors explaining Islamic stock returns is twice that explaining conventional stock returns. One important implication of this result is that, since conventional securities cannot be traded by Islamic investors, investors who invest in conventional securities must hold undiversified portfolios and, thus, require a return premium for bearing some idiosyncratic risk. To test this implication, we follow Ang et al. (2006) and estimate firm idiosyncratic volatility as the standard deviation of abnormal stock return, relative to the CAPM and Fama and French (1993) three-factor models. The results show that conventional stocks have higher estimated idiosyncratic risk than Islamic stocks. We also test the investor recognition hypothesis and the results indicate that the differences in liquidity and trading activity between Islamic and conventional stock classes comes mainly from stock visibility and individual investor recognition of these stocks. In our study, this finding means that conventional stocks are known to relatively few investors in the Saudi markets, due to the fact that they do not meet Sharia requirements. Merton (1987) shows that expected returns decrease with the size of a firm s investor base, which he characterizes as the degree of investor recognition.

10 Our empirical results indicate that Islamic stocks enjoy higher liquidity and more trading activity than conventional stocks in Saudi markets. We find that Islamic stock turnover is 20 percent higher than conventional stock turnover in the Saudi market. In addition, according to Amihud s (2002) liquidity measure, Islamic stocks are 24 percent more liquid than their conventional counterparts in the Saudi market. Our results also imply that the greater turnover of Islamic stocks comes completely from systematic turnover. We apply a methodology similar to Loughran and Schultz (2005), and the results show that, when investors trade stocks for market-wide reasons, they trade Islamic stocks. Our findings imply that familiarity is an important criterion in determining trading. If investors trade in response to information that concerns a particular stock, they have to trade that stock. On the other hand, if investors can trade a number of different stocks in response to information with market-wide ramifications, they choose to trade the stocks that they are familiar with Islamic stocks. Our paper contributes to multiple strands of the literature. To our knowledge, ours is the first study to examine the liquidity and trading activity differences between Islamic and conventional stocks in the context of the market segmentation and investor recognition hypotheses, and this represents a significant contribution to the literature on the subject. Our results link market segmentation and stock cross-listing to stock liquidity and trading activity by showing that the market segmentation effect can result not only from explicit, but also from implicit barriers, like religious beliefs. We also contribute to the literature on investor

11 behavior by showing that individual preferences and religious beliefs can impact investor financial decisions and portfolio construction. The results link investor recognition to stock liquidity by showing that stock Sharia classification acts as a source of information dissemination to outside investors. Finally, this paper contributes to the growing literature on Islamic finance by showing that stock screening and Sharia classification increases investor awareness of Islamic finance and the role of Sharia compliance in the Saudi market. In addition, results imply that Islamic loans can benefit the economy as a whole by increasing trading in stock markets, improving the circulation of money across different economic sectors, and lowering the external costs of funds for Sharia compliant firms through increased liquidity and recognition. The remainder of this paper is organized as follows. In section 2, we review the relevant literature on Islamic finance. Section 3 explains the data and methodology used in our analyses. Section 4 presents and interprets the results. Section 5 summarizes the findings and gives a conclusion. 2. Literature review of Islamic stock performance Recently, much research has been conducted in the field of Islamic finance, due to the increasing interest in Islamic investors and their strong demand to invest in products that are only combatable with Sharia law. This has led to the development of different Sharia-compliant products, like Islamic stocks, mutual funds, market indices, and bonds (sukuk). The increased interest in Islamic finance has grown globally, and, consequently, many large western financial institutions

12 have established their own Islamic subsidiaries and offer Islamic financial instruments targeted directly at their Islamic clientele. One measure of Islamiccompliant stock market is the Dow Jones Islamic market index (DJIMI), which tracks the stocks of US corporations whose business and activities are compatible with Islamic law. The index is a basket of stocks consisting of firms that satisfy Islamic investing principles (avoiding unethical or highly-indebted firms, or engaged in gambling, alcohol sales, and other prohibited (non-halal) activities). As Islamic financial products grow in demand, several empirical studies investigate whether these Islamic products are associated with higher or lower returns, compared to their conventional counterpart products. Hakim & Rashidian (2004) study the DJIMI and compare it with the broad market index (Wilshire 5000) and the three-month Treasury bill. They conclude that Muslim investors are not penalized by investing in the Dow Jones Islamic Market Index (DJIMI), because investors who invest in the DJIMI have diversification benefits, since the Islamic index is influenced by factors independent from the broad market or interest rates. In another study, Hakim & Rashidian (2004) analyze the DJIMI and investigate how the Sharia selection restriction affects index performance in order to determine the competitive risk-return tradeoff of a Sharia-compliant index. In addition, they examine the extent to which a Sharia-compliant index correlates with the broad stock market. They find that the DJIMI return is competitive, relative to the world stock market index, DJW, but underperforms when compared to other restricted, non-islamic indices, like the

13 socially responsible, or Green, index. Overall, they conclude that investing in the Muslim index does not coincide with a discernible cost for complying with the Sharia restrictions. Another index that tracks Islamic stocks is The Financial Times Stock Exchange Global Islamic Index, or Sharia Global Equity Index (FTSE Sharia Global index). Hussein (2004) compares this index with the FTSE All-World Index and tests whether the returns earned by investors who purchase shares in one of these two indices are significantly different from the other. He divides the sample period into two sub-periods, the bull period (July 1996 March 2000) and the bear period (April August 2003). A comparison of the raw and risk-adjusted performance measures show that the Islamic index performs as well as the FTSE All-World index over the entire period. He also finds that the Islamic index yields statistically significant positive abnormal returns in the bull market period, although it underperforms the counterpart index in the bear market period. He concludes that the application of ethical screening does not have an adverse effect on FTSE Global Islamic Index performance. Similar results are presented in Hussein (2007), where he adds the Dow Jones Islamic Index and Dow Jones world Index to the comparison. Similarly, Girard and Hassan (2008, 2010) find no convincing performance differences between Islamic and non-islamic indices. In their study, they use five Islamic indices and five corresponding FTSE indices that do not explicitly claim to use Islamic screening (FTSE Global Index, FTSE Asia Pacific Index, FTSE

14 Americas Index, FTSE Europe Index, and FTSE South Africa Index). Their results indicate that similar reward to risk and diversification benefits exist for both Islamic and non-islamic indices. After controlling for market risk, size, book-tomarket, momentum, and local and global factors, they conclude that the difference in returns between Islamic and conventional indices is not significant. Their findings suggest that the difference in performance of Islamic indices, as compared with conventional indices, is attributed to style differences between the two types of series. In a mutual fund context, Merdad, Hassan, and Alhenawi (2010) examine 28 mutual funds (both Islamic and conventional) that operate in Saudi Arabia and investigate their performance during different economic conditions to assess any differences in their corresponding returns. On a non-risk adjusted basis, they find that there is no significant performance difference between Islamic and conventional funds. However, both funds, Islamic and conventional, significantly underperform the GCC Islamic Index and Saudi Stock market index during the bull periods. However, on a risk-adjusted basis, conventional funds outperform Islamic funds during the overall and bull periods. Moreover, Hoepner, Rammal, and Rezec (2011) analyze the financial performance and investment style of 265 Islamic equity funds from twenty countries. The results from examining the fund performances did not allow them to conclude whether Islamic funds generally under or outperform equity markets. Instead, they attribute the heterogeneity in Islamic fund performance to national characteristics. Islamic funds from the six largest Islamic

15 financial centers in the study (the GCC countries and Malaysia) perform competitively with, or even outperform, international equity market benchmarks. In contrast, Islamic fund portfolios from most other nations with less developed Islamic financial services significantly underperform their benchmarks. In addition, they find that Islamic funds from non-islamic economies underperform their benchmarks. Hoepner, Rammal, and Rezec (2011) suggest that Islamic funds benefit from proximity to Islamic financial centers, as this allows them to develop more expertise. They also suggest that Islamic funds appear to have a financial advantage in Muslim economies, which could be a result of the higher utility consumers and agents in those economies receive from the Sharia compliance of investments. They argue that governments can increase the attractiveness of Islamic funds in their jurisdiction by developing Islamic financial markets. While some research indicates that Islamic indices do not significantly underperform conventional indices, Hayat and Kraeussl (2011) find that Islamic equity funds, in fact, do underperform in many cases. They attribute the bad performance of Islamic equity funds to the bad timing of the equity fund managers, who try to time the market, but, in doing so, reduce the return, rather than increase it. By analyzing the risk and return characteristics of a sample of 145 IEFs over nine years, their results indicate that, on average, IEFs substantially underperform both their Islamic and conventional benchmarks. They argue that some risks should be taken into account when assessing IEFs as an investment alternative that are

16 usually not present in conventional investments. These risks include: changing Sharia (Islamic law) rules; the lack of a sufficient track record; high exposure to companies that might be sub-optimally leveraged; and the persistence firms with low working capital. When considering the current research on Islamic stock portfolio returns, there is mixed evidence as to whether Islamic investments outperform or underperform their conventional counterparts. One explanation of such conflicting results is that current comparative analyses may be inappropriate. In fact, Derigs and Marzban (2008) analyze the impact of applying alternative Sharia screens on the universe of halal assets and show that Sharia screening procedures currently used in practice are inconsistent with respect to discriminating between halal and haram firms. Their analysis reveals that the asset universes are significantly different in size and consistency, i.e. for every index, there is a substantial number of assets which are specified as Sharia compliant or noncompliant, but classified in the opposing categories for other indices. They argue that, so far, there is no universal or generally accepted understanding of how to transform the descriptive Sharia rules into a system of verifiable investment guidelines. The Saudi Arabian stock market is an example of an Islamic financial system having many alternative Sharia screening procedures. There are more than five different Sharia screening procedures. Merdad, Hassan, and Hippler (2015) use one of the strictest Sharia screening procedures, the classifications of Dr. Al-Osaimi, and their justification for using such a rigorous screening process is to ensure

17 Sharia compliance. According to the classification they use, any stock that does not meet the classification criteria will be considered a non-sharia compliant, or conventional, stock; however, stocks considered conventional under this methodology could be classified as Sharia compliant using another classification methods. Accordingly, we utilize the Sharia classification of Dr. Al-Osaimi in our analyses in order to minimize the possibility of the results being impacted by ineffective or inconsistent Islamic screening. 3. Data and Methodology In order to compare the returns between Islamic and conventional stocks, we use stock market data for the sample of Saudi Arabian firms. The data are collected from the Global Compustat database, and the sample consists of daily stock returns that span from May 2002 to April 2015 and is subject to availability. From these daily data, monthly returns for all listed firms are calculated. We also calculate the monthly market value (size) for each firm by multiplying the number of shares outstanding by the stock price at the end of each month. Saudi macroeconomic factors are collected from the Saudi Arabian Monetary Agency and the Saudi Central Department of Statistics & Information. Crude oil production data are collected from the Organization of the Petroleum Exporting Countries (OPIC) 3. The riyal value of traded shares is calculated by multiplying the daily stock price by the volume of shares traded in a day for each listed firm; the monthly riyal value of traded shares is the sum of the daily riyal values of the traded shares in each month. As a proxy for stock liquidity, we calculate the percentage of the firm s 3 some of these data are collected from a third party website which is

18 shares traded each month. This variable is calculated by dividing the volume of shares traded during the month by the total shares outstanding for the firm. In order to classify a firm as either Islamic (Sharia compliant) or conventional (non-sharia compliant), we use the most common and market available classifications. In Saudi Arabia, there exist more than one classification, and these classifications differ in their screening processes and criteria for selecting Islamic firms. The classifications share some criteria, like prohibited activities, which include: (1) activities that involve in any form of usury or interest rates (riba). For example, borrowing or investing in interest bearing or fixed-income instruments; (2) Activities that involve excessive risk, uncertainty, ambiguity, or deception (gharar); (3) Activities that are related by any means to gambling, lottery, or games of chance (maysir); (4) Activities that are related to non-halal businesses, such as those that deal with pork, adult entertainment, tobacco, non-medical alcohol; and all other unethical businesses. For many firms, it is quite difficult not to engage in some of these activities, especially in activities that involve dealing with any form of interest rates. As a consequence, some Sharia scholars add certain exceptions to the filtering process and the adherence to Sharia law. For example, if a firm is engaged in an impermissible activity, the ratio of income generated by that impermissible activity to total income must be less than a certain percentage in order to classify the firm as an Islamic firm, otherwise it is considered a non-sharia compliant, or conventional, firm. The classifications often differ in the percentage of income

19 generated by impermissible activities the firm may undertake in order to be considered an Islamic firm. Another heterogeneity between the classification methodologies is that some classifications consider firms that operate in the insurance industry as conventional firms, since they believe an insurance company must be a non-profit in order to be considered Islamic. In our study, we rely on Dr. Al-Osaimi s classification 4, which is the most common classification. By using his classification, we are sure that firms are appropriately classified as Shariacompliant, because this classification is a sub-set of all other classification methods that exist in the market. 4. Results and interpretation Before we compare liquidity and trading activity among conventional and Islamic stocks, we analyze the differences in firm characteristics and performance between the two classes. Table 1 shows the summary statistics of firms during the sample period, which spans from May 2002 to April We separate firms into two groups, Islamic and conventional firms, based on Al-Osaimi s classification. Conventional firms tend to be larger than Islamic firms in terms of average and median totals assets, the market value of assets, and equity capitalization. Conventional firms have an average of 23,153 million Saudi riyals in total assets, which is almost three times the average total assets of Islamic firms of 8,590 million Saudi riyals. Conventional firms also have higher market values of equity, with an average of 14,120 million Saudi riyals, which is almost twice the average of the Islamic firm s average market value of equity. The median market value of equity, 4 Dr. Al-Osiami s classification list and criteria could be found on the following website:

20 however, is higher for Islamic firms than conventional firms, and this suggests that Islamic firms have extreme market values of equity that result in depressing the average market value of equity for the full sample of Islamic firms. As referenced by the reported market-to-book ratios, Islamic firms tend to have higher market-tobook ratios than conventional firms, since both the mean and median values are higher for Islamic firms. Higher market-to-book ratios suggest that Islamic firms have higher growth opportunities than conventional firms and, thus, have higher stock returns than their conventional counterparts. [Insert Table 1 here] As the reported by debt ratios, Islamic firms have lower average and median debt-to-asset and debt-to-equity ratios than conventional firms. Islamic firms have an average (median) debt-to-asset ratio of 17.8 percent (0.0 percent), whereas conventional firms have an average ratio of 71.6 percent (40.6 percent), which is four times the average debt-to-asset ratio of Islamic firms. Higher debt ratios suggest that conventional firms should have higher stock returns to compensate investors for the risk associated with higher leverage. In terms of profitability, conventional firms have higher average, but lower median, operating income than Islamic firms. In addition, the average and median earnings per share are higher for Islamic firms than for conventional firms. The average price to earnings ratio, however, is higher for conventional firms than Islamic firms. In terms of payout, Islamic firms have higher average and median dividend yields than conventional

21 firms. In conclusion, the results in table 1 suggest that Islamic stocks should be priced higher than conventional stocks. In order to analyze the difference between Islamic and conventional stocks in terms of performance, trading, and liquidity, we divide the sample into yearly subsamples. Table 2 shows the average monthly non-risk adjusted returns for value and equally weighted portfolios for each class of stocks every year from 2002 to For value weighted portfolios, Islamic stocks outperform their conventional counterparts in six out of the fourteen years, but they have approximately the same return over the whole sample. The returns of Islamic stocks tend to vary relatively less between years. From 2002 to 2006, Islamic stocks underperform conventional stocks, whereas, between 2007 and 2015, they outperform their conventional counterparts. The results also show that Islamic portfolios outperform conventional portfolios during bearish trends. This result is consistent with Merdad, Hassan and Alhenawi (2010), who find that conventional funds underperform Islamic funds during bearish periods. [Insert Table 2 here] In comparing Islamic and conventional firms in terms of liquidity or activity, we divide the cumulative number of shares traded every month by the total number of shares outstanding. A higher percentage indicates higher activity, or higher liquidity, for a stock. Table 3 shows the trading activity of Islamic and conventional portfolios in each sector of the Saudi stock market. From the results, it is clear that Islamic portfolios outperform conventional portfolios in terms of trading activity in

22 Saudi Arabia. In every sector where we have both Islamic and conventional firms, Islamic firms have a higher percentage of shares traded than conventional firms. This result supports the segmented market hypothesis, where we have Islamic investors seeking only Islamic stocks and conventional investors who can invest in both conventional and Islamic stocks. Therefore, Islamic stocks will have a higher liquidity and will be more actively traded than conventional stocks, since they are traded by all type of investors. In addition, Amihud s (2002) illiquidity measure indicates that Islamic stocks are less liquid than conventional stocks, but this illiquidity comes mainly from the period before From 2007 to 2015, Islamic stocks are more liquid than conventional stocks, according to Amihud s (2002) illiquidity measure. [Insert Table 3 here] In Table 3, we investigate the performance, liquidity, and trading activity of Islamic and conventional portfolios across different market sectors. The Islamic firms are absent from three sectors, as indicated earlier. Of the remaining 12 sectors, where both Islamic and conventional firms exist, Islamic portfolios outperform conventional portfolios in 6 sectors and underperform in the remaining 6. Also from table 3, we show that, according to Dr. Al-Osaimi s classification, there exist no Islamic firms in the media and publishing, insurance, and multiinvestment sectors, because this classification system follows a very strict screening and filtering process and, as a result, only 58 of 166 listed firms are classified as Islamic in the Saudi market.

23 Table 4 compares the yearly performance between Islamic and conventional portfolios in the Banks & Financial services sector, and the Islamic portfolios outperform the conventional portfolios in 9 of the 13 years examined. The outperformance of Islamic stocks may be related to the fact that, in this sector, firms will be rewarded if they are classified as Islamic firms. The stocks of Islamic banks are very attractive to Saudi public, since they claim Sharia compliance in all their business. [Insert Table 4 here] 4.1 The Segmented Markets Hypothesis The rational agent framework offers two explanations for the cross-sectional return differential between the two classes of stocks (Islamic and conventional). If the classification effect represents an arbitrage opportunity, it can only persist if large impediments prevent rational agents from trading. Particularly, Islamic investors are prohibited from trading in non-islamic stocks. We call this the impediments-to-trade or market segmentation hypothesis. Alternatively, the return differential may not reflect a mispricing, but a fair compensation for risks not captured by standard factors. We examine these two explanations in turn. In testing the impediments-to-trade hypothesis, we examine whether there exists a market segmentation between the two stock classes, or whether these two classes are integrated. The most widely used measure of integration is the crossclass correlations of the stock portfolio returns of each class. However, as suggested by Pukthuanthong and Roll (2009), this quantitative measure of integration suffers

24 from a fundamental flaw. Pukthuanthong and Roll (2009) provide an alternative measure of integration, which is the R-squared computed from regressing stock returns on multiple potential factors that should explain stock return variation. If the impediments-to-trade hypothesis holds, we expect to see a clear segmentation between the two stock classes. Table 5 shows the results of regressing each group s stock returns on a number of macroeconomic factors that are related to the economy of Saudi Arabia. The macroeconomic factors include: the consumer price index; the central bank balance sheet; money supply (M3); current account; capital flows; loan growth rate; monthly oil production; and Brent crude oil spot price. We regress stock returns on the five most influential macroeconomic factors and then add some additional macroeconomic factors into two regressions. The results show that Islamic stocks are more sensitive to Saudi Arabian macroeconomic factors than their conventional counterparts. By examining the R-squared of the three regressions, we confirm that these macroeconomic factors better explain Islamic stocks than conventional stocks. [Insert Table 5 here] Table 5 shows that the first regression explains percent of Islamic stock returns, but it only explains 7.35 percent of conventional stock returns. This suggests that Islamic stocks are more integrated with Saudi macroeconomic factors than conventional stocks. In addition, the coefficients of the first regression are statically significant in the case of Islamic stocks; however, the only significant coefficient is the one related to the change in the consumer price index for

25 conventional stocks. We add additional macroeconomic factors for the second and third regressions and show that these factors can improve the predictability of stock returns, and the explained percentage of stock returns is higher for Islamic stocks than it is for conventional stocks. The heterogeneity of R-squared differences between the two classes suggests that there exists market segmentation between Islamic and conventional stocks, which support our hypothesis of market segmentation and impediments-to-trade in the Saudi stock market. Not all investors can trade both classes, because Islamic investors can only trade stocks that comply with their religious principles, which are classified as Islamic stocks. 4.2 The Investor Recognition Hypothesis Alternatively, Merton (1987) offers another explanation to the classification effect within the rational agent paradigm. He models informationally incomplete markets in which investors only know about a subset of the available stocks. In such markets, stocks that are recognized by fewer investors need to offer higher returns to compensate their holders for being imperfectly diversified. This hypothesis, known as the investor recognition hypothesis, has particular relevance to the classification effect in the context of Islamic stocks. Since Islamic stocks are traded by both Islamic and conventional investors, they have a broader clientele and, thus, have higher media and analyst coverage, which leads to a higher degree of investor recognition. In addition, it is important to note that, in Islamic countries, a major mechanism for Islamic banks to provide Islamic or Sharia-compliant loans is through the purchase and sale of Islamic

26 stocks. The use of Islamic compliant stocks to facilitate Islamic bank loans can have a major effect on Islamic stocks through two channels. First, it increases the liquidity of Islamic stocks. Secondly, it increases the Islamic stock investor base by introducing new investors, bank borrowers, to the Islamic stock market. If Islamic stock classification improves investor recognition, then its effect should be stronger among stocks that would otherwise have a lower degree of recognition. In Merton s (1987) framework, firm idiosyncratic risk is priced, because of the imperfect diversification that stems from a lack of investor recognition. Firms with higher idiosyncratic volatility should offer a return premium to compensate shareholders for the undiversified risk they impose. In this way, idiosyncratic volatility measures the amount of idiosyncratic risk borne by investors due to imperfect diversification. Hence, there should be a clear difference in the idiosyncratic risk between the two classes of stocks. The intuition behind this is that, since conventional stocks cannot be traded by all the market participants, investors hold undiversified portfolios and idiosyncratic risk should be priced in the Saudi market. Following Ang et al. (2006), we estimate firm idiosyncratic volatility as the standard deviation of the daily abnormal stock returns, relative to the CAPM and Fama and French (1993) three-factor models. If stock classification increases investor recognition and improves diversification, its effect should be stronger among firms with higher idiosyncratic volatility. Table 6 shows the difference in idiosyncratic risk between the two stock classes. Islamic stocks report lower

27 idiosyncratic volatility than conventional stocks using both pricing models. Using the CAPM model, the idiosyncratic risk for Islamic stocks is 1.38 percent, which is lower than their conventional counterpart of 1.64 percent. Islamic stocks also have lower idiosyncratic risk than conventional stocks when using the Fama and French (1993) three-factor model. Islamic stocks have idiosyncratic risk of 1.24 percent, compared to 1.48 percent for conventional stocks. The difference in idiosyncratic risk suggests that conventional stocks suffer from isolation and supports our hypothesis of market segmentation. conventional stocks are traded by fewer investors, because they are restricted from being traded by Islamic investors, which leads to conventional stocks being segmented and isolated from Islamic stocks. Therefore, conventional stocks have to compensate their investors in the form of an additional risk premium for being undiversified. [Insert Table 6 here] We now examine whether Islamic stocks are more liquid than conventional stocks, and we follow Loughran and Schultz (2005) by regressing stock turnover on an Islamic firm dummy variable and several control variables to test the difference in liquidity between Islamic and conventional stocks. Using a measure of stock turnover and Amihud s (2002) illiquidity measure, Table 7 shows the difference between Islamic and conventional stocks in liquidity within each sector and across the whole market. The results show that Islamic stocks are more liquid than conventional stocks across the whole market. In terms of turnover, the Islamic dummy variables have a positive and statistically significant coefficient value of

28 0.20, which means that Islamic stock turnover is 20 percent higher than conventional stock turnover. In regards to Amihud s (2002) illiquidity measure, the Islamic dummy variable has a positive and statistically significant value of 0.24, which means that Islamic stocks are 24 percent more liquid than conventional stocks according to Amihud s (2002) illiquidity measure. [Insert Table 7 here] Table 7 also shows the turnover difference between Islamic and conventional stocks across market sectors. The Banks and Financial sector is a special case, because all classifications agree on classifying Banks into Islamic and conventional banks. Banks also declare their Sharia compliance to the public. Therefore, investors know whether a particular bank is Sharia compliant, or not. Interestingly, the results show that Islamic banks are more than two times more liquid than conventional banks in the banking and finance sector. Specifically, Islamic banks are 236 percent and 246 percent more liquid than conventional banks in terms of stock turnover and Amihud s (2002) illiquidity measure, respectively. Islamic stocks also dominate conventional stocks in terms of liquidity in the transport, energy and utilities, and hotel and tourism sectors. Since our results indicate that turnover is significantly lower for conventional stocks than Islamic stocks, we now examine whether conventional stocks trade less in response to firm-specific or market-wide factors. If the systematic component of turnover is smaller for conventional firms, this would suggest that investors trade Islamic stocks in response to market-wide or industry-wide information. This could

29 mean that Islamic firms are more visible, or are known to more investors, and, hence, are more heavily traded when information that is relevant to numerous firms is revealed. Following Loughran and Schultz (2005), for every trading day over the 2002 to 2015 period, we estimate market turnover for the Saudi stock market. The Saudi stock market turnover is calculated by dividing the total number of Saudi shares traded on that day by the total number of shares outstanding for all Saudi-listed firms. Then, we run a regression for every stock each calendar year of the form: turnover i,t = α 0 + α 1 turnover m,t+1 + α 2 turnover m,t + α 3 turnover m,t 1 + ε i,t, (1) where turnover i,t is the turnover of stock i on day t, and turnover m,t is the market turnover on day t. To obtain the systematic turnover for stock i on day t, we multiply the coefficients by the average contemporaneous and lagged market turnovers. The unsystematic turnover is, therefore, measured by the intercept coefficient. We then average the turnover components across all Islamic and conventional stocks for a year. Table 8 reports the mean systematic and unsystematic turnovers across all years. The average systematic turnover for conventional stocks is only 1.59 percent of the outstanding shares per day, while the systematic turnover averages 2.24 percent of outstanding shares each day for Islamic stocks. Using the standard deviation of the difference in systematic turnover across the 14 years, we calculate a t-statistic for the difference between the two classes of 2.73, which, with 13 degrees of freedom, is significant at the 5 percent level. For unsystematic turnover,

30 conventional stocks have an average of 0.94 percent of outstanding shares per day, but it is only 0.81 percent per day for Islamic stocks. [Insert Table 8 here] The results in Table 8 imply that the greater turnover of Islamic stocks comes from systematic turnover. That is, when investors trade stocks for marketwide reasons, they trade Islamic stocks. This finding seems to confirm that familiarity is an important criterion in determining trading. If investors trade in response to information that concerns a particular stock, they are inclined to trade that stock. On the other hand, if investors can trade a number of different stocks in response to information with market-wide ramifications, they trade the stocks that they are familiar with, and more Saudi investors are familiar with Islamic stocks. Table 8 also reports the mean coefficients of stock turnover on contemporaneous market turnover, as well as market turnovers for the previous and following days. For conventional stocks, the mean coefficient for contemporaneous market turnover is 2.66, and for Islamic stocks it is Additionally, the mean difference between them is 0.5, which is statically significant at the 5 percent level. Also, conventional stock turnover does not lag Islamic stock turnover; The coefficient for lagged turnover is 0.14 for conventional stocks and 0.03 for Islamic stocks. 5. Conclusion The objective of this paper is to compare the performance of Islamic and conventional stocks in terms of liquidity and trading activity and to examine

31 whether the differences are attributed to investor recognition and to the implicit trading barrier created between the two classes of stock by Islamic classification. Due to the religious nature of the population, the degree to which individuals participate in the stock market, and the recent growth in the listed firms available for purchase in the stock market, the Saudi Arabian stock market provides an ideal sample for our analysis. Accordingly, we sample the daily stock returns of all Saudi firms from September 2002 to 2015 and calculate important measures pertaining to performance, liquidity, and integration. For example, we compute a measure of market integration along the lines of Pukthuanthong and Roll (2009), a measure of liquidity similar to Amihud (2002), and a measure of idiosyncratic volatility like that described in Ang et al (2006). Our results support the segmentation and impediment-to-trade hypothesis. Conventional stocks are isolated and segmented from Islamic stocks, which means that there exists an implicit barrier of trade between them. By applying a methodology similar to Pukthuanthong and Roll (2009) to measure market integration, we find that Saudi macroeconomic factors can better explain Islamic stock returns than conventional stock returns. In some regressions, the percentage of macroeconomic factors explaining Islamic stock returns is twice as high as that for conventional stock returns. One important implication is that conventional securities cannot be traded by Islamic investors, therefore, investors who invest in conventional securities must hold undiversified portfolios and, thus, require a return premium for bearing idiosyncratic risk. To test this implication, we follow