An Investigation of the Weak Form of the Efficient Markets Hypothesis for the Kuwait Stock Exchange

Abstract

This article investigates the weak form of the efficient market hypothesis (EMH) for the Kuwait Stock Exchange (KSE). In particular, it tests whether share returns on the KSE exhibit patterns which may be used to predict future share price changes. Ten filter rules are tested on weekly data for 42 firms over the period 1998–2011. The results suggest that the KSE was not weak-form efficient because patterns and trends were present in security prices. In addition, the results are consistent with the substantive literature which has argued that emerging stock markets are informationally inefficient, such as Fifield, Power and Sinclair (2005, 2008) and Xu (2010) and particularly those early studies of Al-Shamali (1989) and Al-Loughani and Moosa (1999) that looked at trading rules for the KSE.

Keywords

Accounting information market efficiency trading rules filter rules capital market disclosures Kuwaiti

Introduction

This article examines the weak form of the efficient market hypothesis (EMH) for the Kuwait Stock Exchange (KSE). According to this hypothesis, a market is efficient ‘if it fully and correctly reflects all relevant information in determining security prices’ (Malkiel, 1992). The weak form of this hypothesis focuses on historic information and suggests that no trends or patterns are present in share return data. An investor, therefore, cannot achieve abnormal returns by trading on past information since the historical news is already impounded into share returns, and price changes follow a random walk process.

When investigating the weak form of the EMH, studies have typically concentrated on statistical tests by looking at the correlation between current and historic returns and conducting runs tests which examine for patterns in the sign of share price changes. In addition, another strand of the literature has focused on the performance of trading strategies; studies in this area normally investigate whether an investor who trades on the basis of trends in past security prices can outperform a passive investment approach. One of the most common strategies which have been studied in the literature is filter rates. The filter rule recommends the purchase (sale) of a share if its price has risen (falls) by a certain percentage from a previous low (high) value.

To date, a lot of studies which have tested the weak form of the EMH have concentrated on developed markets such as the USA and the UK (Fama, 1965; Fama & Blume, 1966; Hudson, Dempsey, & Keasey, 1996; Sweeney, 1988). In general, the findings from such studies suggest that markets are efficient with respect to historic information.¹

Correlation coefficients between current and past returns are typically close to zero (Fama, 1965), the sign of price changes exhibit no specific patterns (Fama, 1965) and trading strategies cannot outperform a passive investment approach where an investor buys a portfolio of diversified securities and holds them over a specific time horizon (Fama & Blume, 1966; Sweeney, 1988). However, a growing number of investigations indicate that stock exchanges in emerging market countries may not be weak-form efficient. The current study adds to research in this area.

Specifically, the current article tests a comprehensive set of filter rule strategies on weekly data for a sample of 42 Kuwaiti shares to examine whether the KSE is weak-form efficient. A number of previous studies have examined this issue (Abraham, Seyyed, & Alsakran, 2002; Al-Khazali, Ding, & Pyun, 2007; Al-Loughani, 1995; Al-Loughani & Moosa, 1999; Al-Mudhaf, 1983; Al-Shamali, 1989; Butler & Malaikah, 1992; Elango & Hussein, 2008; Gandhi, Saunders, & Woodward, 1980; Hassan, Al-Sultan, & Al-Saleem, 2003; Smith, 2007). However, a majority of these concentrate on statistical tests, analyse relatively old data and focus on indices rather than on individual security returns. The current article, therefore, contributes to the literature in this area by examining the weak form of the EMH for the KSE using individual company share price information over a recent period to test the profitability of a large number of filter rule trading strategies.

There are a number of reasons why the weak form of the EMH should be reconsidered for the KSE. First, a lot of the existing literature has arrived at mixed conclusions about whether or not the KSE is weak-form efficient; a comprehensive study using data for a relatively large sample of securities might help to clarify any confusion in the literature. Second, the KSE has changed since the early 1990s when data for a number of previous studies were used. For instance, several regulations were introduced² in order to improve the transparency and disclosure levels of companies listed on the KSE (Al-Yaqout, 2006). In addition, restrictions were lifted on the ownership of shares in Kuwaiti companies by non-nationals; since 2000, foreign investors are allowed to own up to 100 per cent of the equity of Kuwaiti firms listed on the KSE. The influx of foreign investors may have altered the analysis of Kuwaiti equities undertaken and improved the efficiency of the KSE. Further, the trading system of the KSE has been updated in recent years while a privatisation programme by the Kuwaiti government has more than doubled the number of companies listed over the period 1998–2011. As the supply of shares has increased, the appetite of investors for purchasing equities has grown (Almujamed, 2011) and the liquidity of the market may have improved.

Currently, the KSE has 230 listed companies from eight sectors with a market capitalisation of around $169 billion (Central Bank of Kuwait, 2011). It provides an interesting research site for examining the weak form of the EMH since it is a liquid market with a T+1 settlement system where the value of shares transacted and trading volume have grown by 971.0 per cent since 2000 (Central Bank of Kuwait, 2011). Further, it is a well-regulated market where transaction costs (at 0.1%) are minimal. Thus, unlike a lot of other emerging markets, any excess returns from following a filter rule investment strategy that are documented for the KSE may be achievable by practitioners.

Literature Review

Empirical studies from stock markets of the Gulf Cooperation Council (GCC) countries are relatively sparse when compared with investigations from other regions in the world.³ Investigations about the stock market of Kuwait are even less common; to date, only 11 investigations about the efficiency of the KSE have been published. Within these published articles, there is a lack of consensus about the efficiency of the KSE. For example, three of the studies have suggested that the KSE is weak-form efficient (Al-Khazali et al., 2007; Al-Mudhaf, 1983; Butler & Malaikah, 1992) while another eight papers have argued that returns in the market are predictable (Abraham et al., 2002; Al-Loughani, 1995; Al-Loughani & Moosa, 1999; Elango & Hussein, 2008; Gandhi et al., 1980; Hassan et al., 2003; Smith, 2007). A majority of the empirical studies that have been conducted over the last three decades, therefore, have suggested that the KSE is inefficient.

A number of reasons may explain why some studies of the KSE reject the weak form of the EMH while others do not. A detailed investigation of the literature reveals that there have been variations in the time periods studied and the datasets used to examine the efficiency of the KSE. For example, 8 out of the 11 papers’ datasets were relatively old (Abraham et al., 2002; Al-Loughani, 1995; Al-Loughani & Moosa, 1999; Al-Mudhaf, 1983; Al-Shamali, 1989; Butler & Malaikah, 1992; Gandhi et al., 1980; Hassan et al., 2003) with share price information being studied from 1975 to 1999. In addition, only a minority (3 out of 11) used data for individual shares to test the weak form of the EMH for the KSE (Al-Mudhaf, 1983; Al-Shamali, 1989; Butler & Malaikah, 1992); eight focused on index values when conducting their tests. Further, a majority of studies about the KSE have concentrated on statistical tests to investigate the weak form of the EMH; most have employed the serial correlation test, runs test and/or the variance ratio test (10 out of 11 papers). By contrast, only two studies have investigated the performance of trading rules such as filter and moving averages rules (Al-Loughani & Moosa, 1999; Al-Shamali, 1989).

Studies by Al-Mudhaf (1983) and Butler and Malaikah (1992) have documented that the KSE is weak-form efficient. Both used data for individual shares and employed statistical tests such as serial correlation analysis and runs tests. For example, Butler and Malaikah (1992) used daily data for the most liquid shares listed on the KSE and the SSM (36 shares for the KSE and 25 for the SSM).⁴ The results indicated that the returns for 60.0 per cent of the sample of the most liquid Kuwaiti shares followed a random walk process. Other papers have documented similar results but employed different methods (Al-Khazali et al., 2007); they used a non-parametric version of the variance ratio test and the runs test after correcting for infrequent trading.⁵

By contrast, several investigations have used statistical analysis and found that KSE is not weak-form efficient. One of the early studies was conducted by Gandhi et al. (1980). They used serial correlation and run tests with monthly data for the All Share and Industrial indices from December 1975 to May 1978. They found that simple linear regressions of current returns on lagged returns suggested that significant patterns were present in the data. Al-Loughani (1995) also investigated the weak form of the EMH using a variance ration test but a different time period and an alternative market index called the Al-Shals Composite Index; he rejected the null hypothesis that share price changes followed a random walk.

The most relevant investigations for the current article are those that studied the performance of trading rules; a detailed review of the literature shows that only Al-Shamali (1989) and Al-Loughani and Moosa (1999) investigated the weak form of the EMH using such trading rules. Al-Shamali (1989) investigated the predictability of filter rules using daily and weekly share prices for 42 securities listed on the KSE over a 5-year period from January 1983 to December 1987. By contrast, Al-Loughani and Moosa (1999) examined the performance of moving average rules using weekly data for the KSE index over two sample periods from 27 August 1986 to 12 March 1997. Both studies suggested that the KSE was not weak-form efficient. For example, Al-Shamali (1989) used four filter rules (4%, 6%, 8% and 10%) and discovered patterns were present in the returns of most shares analysed;⁶ the returns from most of the filter rules outperformed a buy-and-hold strategy. For instance, an analysis of daily prices revealed that the 10.0 per cent filter rule outperformed the buy-and-hold strategy by 0.066 per cent (0.140 versus 0.074), while for the weekly data the naïve strategy underperformed by 0.037 per cent (0.098 versus 0.061). Further, he noted that the returns generated from filter rules appeared to decrease as the filter size increased suggesting that only small filter sizes should be employed.

A decade after the Al-Shamali (1989) study, Al-Loughani and Moosa (1999) analysed the performance of eight long-run moving average periods (5, 10, 15, 20, 25, 30, 35 and 40 weeks) compared to a buy-and-hold strategy. According to their investigation, a buy (sell) signal occurred when the current price was higher (lower) than a moving average of path prices; thus, their short-run moving average period was only one week. They assumed that KD1000 was invested to make the trade realistic. In addition, transaction costs of 0.1 per cent were taken into account. Further, the findings from their first sub-period indicated that trading rules outperformed the buy-and-hold strategy significantly when 5 and 10 week moving average periods were used; the buy-and-hold strategy only outperformed the trading rules when moving average periods of 35 and 40 weeks were employed. Moreover, the findings from the second sub-period were consistent with the results of the first sub-period; for instance, the 5-week moving average rule outperformed the naïve strategy by 0.0519 per cent (93.09% versus 87.9%). The researchers concluded that any developments which had occurred in Kuwait during its recent post-liberation period had not improved the efficiency of the KSE.

Data and Method

The empirical work in this article is based on weekly closing share prices of 42 companies from 5 January 1998 to 10 January 2011. The start date was selected because a sizeable number of firms were listed in 1997;⁷ an earlier start date would have reduced the sample size by a significant amount.⁸ The data were obtained from both KSE and Datastream; one source was used to check on the information supplied by the other.⁹ Both datasets were not adjusted for distributions such as cash dividends, share issues and capital reductions. Further, the KSE did not have information about any share distributions prior to 2001; information for the years 1997–2000 was obtained from two organisations: the Orient Consulting Center and the Gulf Investment House. Datastream only stored share price information from April 2001; thus, the data prior to 2001 were collected from the KSE on an Excel work sheet.

Table 1 reports details about the final sample of 42 firms that was used for this research. A visual inspection of the table reveals that the sample firms were drawn from seven different industries and ranged in size from a high of KD 429.7 m (TEL) to a low of KD 8.5 m (RRI).¹⁰ Thus, a good mix of firms were included in the analysis, although a majority were drawn from the banking, investment and industrial sectors; only two companies from the food industry were included in the final sample and none from the insurance sector were included.¹¹ Further analysis shows that banks are all classified among those large capitalisations. An analysis of the final column of Table 1 indicates that most of the firms were profitable in 2011. Only five firms made losses in that year.

Table 1.

Samples

Sector	No.	Name	Code	Paid-up Capital Size (KD in Million)	EPS
Banking (1)	1	National Bank of Kuwait	NBK	359.8	0.077
Banking (1)	2	Gulf Bank of Kuwait	GBK	250.8	0.011
	3	Commercial Bank of Kuwait	CBK	127.2	0.034
	4	Al-Ahli Bank of Kuwait	ABK	144.1	0.038
	5	Kuwait International Bank	KIB	103.7	0.019
	6	Burgan Bank	BUR	140.1	0.010
	7	Kuwait Finance House	KFI	248.9	0.031
Investment (2)	8	Kuwait Investment Company	KIV	55.1	–0.010
Investment (2)	9	Commercial Facilities Company	FAC	53.7	0.028
	10	International Finance Advisors	IFI	72.0	–0.020
	11	National Investment Company	NIV	87.6	–0.001
	12	KIIPCO Asset Management	KPR	121.2	0.038
	13	Al-Ahlia Holding Company	AIN	82.8	–0.026
	14	The Security House Company	SEC	68.0	–0.080
	15	Industrial Investment Company	IIC	44.9	–0.019
	16	Securities Group Company	SGC	25.5	–0.001
	17	Kuwait Financial Centre	KFC	50.6	0.007
Real Estate (3)	18	Kuwait Real Estate	KRE	90.6	–0.022
	19	United Real Estate	URE	118.7	0.004
	20	National Real Estate	NRE	81.4	–0.049
	21	Pearl of Kuwait Real Estate	PEA	25.1	–0.006
	22	Tamdeen Real Estate	TAM	37.3	0.012
Industrial (4)	23	National Industrial Group	NIN	129.5	–0.017
Industrial (4)	24	Pipes Industrial & Oil Service	PIP	22.5	–0.029
	25	Kuwait Cement Company	KCE	60.7	0.021
	26	Refrigeration Industries	RRI	8.5	0.011
	27	Heavy Engineering and Ship building	SHP	16.3	0.035
	28	United Industrial Company	UIC	49.5	0.005
	29	Boubyan Petrochemical Company	BPC	48.5	0.047
Service (5)	30	Agility Public Warehouse	WAR	104.6	0.015
	31	Zain Telecommunication	TEL	429.7	0.252
	32	Independent Petroleum Group	IPG	15.2	0.021
	33	National Cleaning Company	CLE	10.2	0.013
	34	Sultan Centre Food	SUL	57.8	–0.069
	35	Arabia Holding Group	AGH	13.5	0.006
Food (6)	36	Cattle Livestock Transport & Trading Co.	CAT	21.6	–0.022
	37	Danah Alsafat Foodstuff	DAN	28.8	–0.004
Non-Kuwaitis (7)	38	Sharjah Cement & Industrial Development	SCE	42.2	0.002
	39	Gulf Cement Company	GCE	62.7	–0.001
	40	Fujairah Cement Industries Company	FCE	27.2	0.002
	41	Ras Al Khaimah Company	RKW	35.7	0.013
	42	Arab Insurance Group	ARG	61.7	0.022

Source: www.Aljomant.net, 2011.

Note: This table provides details about the companies’ market values on 10 January 2011 (the end of the sample period), their earnings per share (EPS) and their price/earnings ratio (P/E). N/A = not available.

The tests in the current research are based on both the share prices as well as the natural logarithms of the security returns which were calculated according to the following equation:

R_i,tP_i,tD_i,tP_i,t_–1

where R_i,t is the return for share i in week t, P_i,t is the price of share i in week t, D_i,t is the dividend for share i and P_i,t_–1 is the price of the share in the previous week.

Table 2 highlights the descriptive statistics for the weekly returns of the sample companies. An analysis of the table highlights a number of points. First, the means of the average weekly returns among the sample firms were very small; they varied from a low of –0.36 per cent for PEA to a high of 0.31 per cent for NBK. Second, a majority (76.2%) of the average returns for the sample firms were positive which implies an upward trend in the share prices of most firms during the 13-year period studied. The standard deviation figures associated with these returns were relatively large; they varied from a high of 7.54 per cent for AGH to a low of 3.34 per cent for Commercial Bank of Kuwait (CBK). Also, it appears that there is no strong link between the mean and the standard deviation of returns for the firms. For example, the second best performing share (WAR) and the worst performing security (PEA) had standard deviations of 6.21 and 6.95 per cent, respectively. This image of volatile returns among the different shares is corroborated by an analysis of minimum and maximum values. The gap between the figures is sizeable for most firms; for instance, PEA recorded a drop of –51.38 per cent in one week and an increase of 37.20 per cent in another week during the 13-year period studied. Such a finding is not surprising since returns for emerging market securities typically exhibit a high level of volatility (Harvey, 1995). In addition, Table 2 highlights that sector influences affect the return series for the firms in the sample. For example, the banking shares, on average, were among the best performing firms with relatively low standard deviation values being recorded. By contrast, real estate shares were more volatile and had a higher risk on average.

Third, a visual inspection of the skewness and kurtosis for the firms being investigated reveals that, in most cases, the distributions of share returns were not normal; 33 of the 42 firms had skewness statistics that were statistically different from 0. Of these significant skewness statistics, 14 were negative and 19 were positive, suggesting that most of the firms’ returns series had a large tail of positive values. The kurtosis values are even more emphatic in confirming that the distribution of the returns series are non-normal; the values of this statistic for all the shares in the sample were more than twice their standard errors. This finding suggests that statistical tests, which are based upon the assumption of normally distributed data, may not be appropriate. Thus, care must be exercised when examining the outcomes of parametric tests and greater emphasis given to the trading rules results.

Table 2.

Descriptive Statistics

Code	Sector	Mean	St. Dev.	Minimum	Maximum	Skewness	Kurtosis
NBK	Banking (1)	0.0031	0.0402	–0.315	0.305	0.08*	17.11*
GBK		0.0016	0.0432	–0.725	0.254	–6.75*	119.53*
CBK		0.0029	0.0334	–0.165	0.218	0.58*	5.57*
ABK		0.0027	0.0386	–0.178	0.228	0.80*	4.70*
KIB		0.0009	0.0514	–0.246	0.228	–0.23*	3.84*
BUR		0.0021	0.0483	–0.262	0.249	–0.39*	6.09*
KFI		0.0026	0.0442	–0.357	0.322	–0.43*	14.96*
KIV	Investment (2)	0.0003	0.0543	–0.307	0.216	–0.23*	4.08*
FAC		0.0013	0.0378	–0.223	0.211	–0.29*	7.59*
IFI		0.0004	0.0743	–0.293	0.405	0.95*	5.18*
NIV		0.0023	0.0608	–0.316	0.296	–0.20*	5.13*
KPR		0.0021	0.0562	–0.3254	0.254	–0.05	3.90*
AIN		–0.0028	0.0735	–0.468	0.405	–0.67*	8.39*
SEC		–0.0005	0.0432	–0.237	0.2	–0.57*	5.99*
IIC		–0.0020	0.0560	–0.223	0.223	0.05	3.04*
SGC		0.0011	0.0459	–0.245	0.305	0.76*	7.58*
KFC		0.0005	0.0624	–0.223	0.336	0.57*	4.54*
KRE	Real Estate (3)	–0.0016	0.0646	–0.336	0.336	0.52*	5.58*
URE		–0.0005	0.0558	–0.357	0.357	0.46*	12.93*
NRE		0.0005	0.0684	–0.283	0.336	0.28*	3.82*
PEA		–0.0036	0.0695	–0.5138	0.372	–1.14*	12.76*
TAM		0.0008	0.0548	–0.521	0.325	–0.80*	15.66*
NIN	Industrial (4)	0.0006	0.0591	–0.31	0.278	–0.25*	4.69*
PIP		0.0000	0.0648	–0.357	0.268	0.12	3.98*
KCE		0.0018	0.0552	–0.357	0.288	–0.07	7.16*
RRI		0.0000	0.0487	–0.268	0.288	0.29*	7.09*
SHP		0.0011	0.0574	–0.189	0.336	1.00*	5.35*
UIC		–0.0007	0.0717	–0.405	0.388	0.07	5.20*
BPC		0.0012	0.0510	–0.268	0.238	–0.32*	5.23*
WAR	Service (5)	0.0029	0.0621	–0.293	0.262	0.05	3.48*
TEL		0.0028	0.0489	–0.511	0.223	–2.05*	21.49*
IPG		0.0008	0.0463	–0.179	0.231	0.24*	2.45*
CLE		–0.0013	0.0725	–0.365	0.388	0.60*	5.48*
SUL		0.0008	0.0671	–0.254	0.336	0.53*	4.28*
AGH		–0.0005	0.0754	–0.288	0.336	0.54*	3.39*
CAT	Food (6)	0.0004	0.0439	–0.305	0.305	0.71*	9.32*
DAN		0.0007	0.0714	–0.405	0.405	0.28*	5.53*
SCE	Non-Kuwaitis (7)	0.0015	0.0582	–0.223	0.326	0.67*	5.57*
GCE		0.0014	0.0557	–0.254	0.258	0.14	3.05*
FCE		0.0003	0.0581	–0.282	0.223	0.18	3.67*
RKW		0.0011	0.0658	–0.329	0.299	0.46*	5.31*
ARG		–0.0033	0.0508	–0.23	0.31	0.11	6.72*

Source: Various publications from KSE and Datastream.

Note: This table shows descriptive statistics for the shares for the sample firms. The ‘Mean’ is the average return over the period while St. Dev. is the standard deviation of the values around the mean. Maximum refer to the minimum and maximum weekly return over the sample period respectively. Skewness is a measure of the symmetry of the distribution, while Kurtosis examines whether the data are peaked or flat relative to a normal distribution. An * indicates significantly greater than 0 at the 5% level; the standard error for skewness was 0.94 and 0.187 for kurtosis.

This research investigates the profitability of filter rules and compares the trading rule profits relative to a buy-and-hold strategy. The tests attempt to discover whether various filter rules can outperform a buy-and-hold strategy. If any filter strategy based on past information can generate excess returns relative to the naive buy-and-hold strategy then the weak form of the EMH is rejected and the market is inefficient. However, the weak form of the EMH is accepted if the returns from filter strategies are equal, or less than, those of the buy-and-hold strategy.

Ten different filter rules were investigated to test the weak form of the EMH for the KSE. These include filter sizes of 1.0, 2.0, 4.0, 6.0, 8.0, 10.0, 12.0, 14.0, 16.0 and 18.0 per cent. The filter sizes were chosen because (a) they have been used by other researchers who have investigated this topic in different countries (Fama & Blume, 1966; Fifield et al., 2005; Xu, 2010), (b) the small increments of either 1.0 or 2.0 per cent will be able to detect whether any patterns exist based on the magnitude of previous price changes in the data and (c) relatively few trades might be generated for filter sizes that are larger than 18.0 per cent.¹² The filter rule strategy suggests that a buy signal emerges when a share’s price increases by X per cent from the previous low. It recommends that the share is held until its price declines by X per cent from a subsequent high. Any price changes of less than X per cent are ignored (Fifield et al., 2005).

In implementing the filter rules, the assumptions proposed in Fifield et al. (2005, 2008) are followed. First, it is assumed that an investor always begins with a buy position; after a buy signal, the investor holds the security until a sell signal is generated. Following the sell signal, the investor sells the share and remains out of the market until a subsequent buy signal emerges. This process is repeated over the 13-year period analysed. Thus, the returns generated are calculated for all buy–sell transactions and compared with the profits from a corresponding buy-and-hold strategy which assumes that the investor buys the security on the first day and holds it in their portfolio until the last day when the investor sells the share. The returns from both filter rules and buy-and-hold strategies are calculated taking into consideration transaction costs of 0.1 per cent in the KSE. Second, each investor has a limited amount of funds so that all cash is invested at each buy transaction. Thus, no borrowing is allowed to purchase or sell securities. Third, profits generated from the rules are not assumed to be reinvested. Fourth, no short selling is allowed since this is not permitted in the KSE. In addition, multiple buys (sells) are not permitted; the purchase of security has to be followed by a sale before another purchase can be made.¹³ Finally, any interest earned, when an investor is out of the market, is not considered in the analysis. Such assumptions make the study more realistic and bias the results against finding evidence of trading rule profitability.

Results of Filter Rules

Table 3 and Figure 1 present the performances of the 10 filter rules and compare these with the profitability of the buy-and-hold strategy.¹⁴ The profit figures for each filter rule and buy-and-hold strategy are reported after transaction costs of 0.1 per cent have been deducted. This commission fee was added to inject a measure of realism into the analysis and to overcome any criticisms of other studies in the substantive literature which have ignored the impact of transaction costs (Huang, 1995; Sweeney, 1988).

From Table 3, it appears that the buy-and-hold strategy achieved an average return of 0.20 per cent for all 42 firms in the sample; only 13 of the 42 shares achieved positive returns although the market witnessed an upward rise over the 13-year period studied. The findings from the buy-and-hold strategy reveal that large-size firms outperformed small- and mid-sized companies. For example, a passive investment in large firms such as NBK, CBK, ABK, KFI and TEL generated returns of over 180.0 per cent while small- and mid-sized companies reported gains which ranged from 3.5 per cent for SGC to 46.2 per cent for FAC; the majority (80.0%) of the large profitable firms were in the banking sector (NBK, CBK, ABK and KFI), which confirms the finding from the descriptive statistics that the banking industry was the most profitable among the various sectors in the KSE over the period investigated. Indeed, the vast majority of Kuwaiti banks (six out of the seven) achieved returns which ranged from a high of 398.3 per cent for the CBK to a low of 19.3 per cent for Gulf Bank of Kuwait (GBK). Moreover, an analysis of the results for other sectors shows that all firms in the real estate, the food and non-Kuwaiti sectors as well as most of the industrial and service sector companies recorded losses for the buy-and-hold strategy.

An analysis of the number of trades in Table 3 shows that, on average, the number of trades declines significantly as the filter size increases. For example, as the filter size rises from 1.0 to 18.0 per cent, the average number of trades generated by the filter rules decreases from 223.7 to 23.9. Therefore, an investor who followed the 18.0 per cent filter rule over the 11-year period studied would only have transacted about 24 times for each share. A more detailed investigation of the number of trades for the individual firms in the sample reveals that the 1.0 per cent filter for three firms (TEL, IPG and AGH) generated the largest number of transactions (262), while GBK was associated with the smallest number of transactions (4) when the 18.0 per cent filter rule was tested.

Table 3.

Filter Rule Results

	1.0%			2.0%		4.0%		6.0%		8.0%
Share	B&H Profits	No. T.	Diff %	No. T.	Diff %	No. T.	Diff %	No. T.	Diff %	No. T.	Diff %
Sector 1
NBK	365.3	236	–247.4	144	–254.0	70	–191.7	44	–158.7	34	–155.8
GBK	19.3	226	95.9	142	121.8	74	153.7	50	145.7	32	179.7
CBK	398.3	258	–435.1	168	–355.8	86	–278.3	40	–183.7	24	–140.9
ABK	272.4	242	–291.9	178	–240.1	106	–216.2	56	–147.1	38	–117.7
KIB	–25.3	236	151.4	188	191.5	112	175.9	82	113.6	60	119.6
BUR	89.1	232	21.4	188	22.5	118	–19.3	80	5.6	54	–4.5
KFI	201.2	230	–34.2	150	–60.5	90	–35.8	68	–100.2	46	–73
Sector 2
KIV	–55.0	224	24.4	200	10.3	136	42.4	88	130.2	58	138.5
FAC	46.2	250	–104.3	168	–111.1	86	–89.2	62	–51.6	42	–32.4
IFI	–78.6	228	310.6	180	327.6	128	422.3	94	402.0	86	453.3
NIV	37.0	244	132.3	180	165.1	140	123.2	98	173.9	74	188.8
KPR	43.4	244	71.4	208	89.0	146	102.6	98	152.7	70	210.2
AIN	–98.2	200	–47.0	186	–63.7	144	–92.1	108	–70.5	84	–27.5
SEC	63.3	220	–11.0	136	42.7	80	137.9	54	176.1	42	206.1
IIC	–91.7	180	–149.5	164	–142.9	140	–123.5	104	–177.3	92	–160.2
SGC	3.5	196	–17.0	172	–24.6	104	27.9	74	–11.5	46	46.5
KFC	–61.3	222	–7.6	198	–11.8	150	18.4	106	101.2	72	145.6
Sector 3
KRE	–91.8	172	–47.7	158	–23.7	126	–0.5	88	90.8	68	96
URE	–74.9	180	–213.7	160	–195.7	124	–174.4	78	–88.9	64	–99.8
NRE	–72.2	260	45.8	226	3.4	160	114.5	106	211.8	76	255.3
PEA	–98.8	196	–119.4	180	–91.6	148	–103.7	114	–112.8	82	–70.4
TAM	–43.6	216	17.5	188	–10.4	144	–62.2	96	73.1	58	212.3
Sector 4
NIN	–55.3	258	258	208	282.4	126	328.2	80	348.4	64	346.1
PIP	–76.3	250	154.8	226	164.2	148	225	96	294.4	74	260.2
KCE	17.1	246	34.1	196	–15.1	136	–28.3	92	33.4	54	131.3
RRI	–55.5	220	1.0	166	–8.3	106	60	78	36.5	54	73.2
SHP	–28.4	254	–98.7	232	–101.4	156	19.9	98	89.3	74	32.6
UIC	–89.5	234	–47.1	208	–44.7	160	–74.8	114	–34.6	96	–85.2
BPC	–5.4	260	42.6	186	91.6	118	131.7	82	123.8	56	120.7
Sector 5
WAR	93.2	236	219.1	196	236.8	132	322.7	102	251.2	82	209.6
TEL	181	262	–92.4	154	10.1	106	–28.6	62	53.0	48	6.6
IPG	–18.6	262	–114.3	220	–112.4	132	–7.6	90	11.3	66	38
CLE	–93.1	230	178.4	184	210.1	134	229.1	98	166.4	80	250.8
SUL	–61.2	244	145.8	196	211.7	140	298.9	102	209.3	84	206.9
AGH	–89.3	262	–55.4	230	–63.8	170	–1.6	136	12.7	108	–14.8
Sector 6
CAT	–30.6	200	15.2	176	12.7	98	121.5	56	163.7	48	95.9
DAN	–71.5	216	55.6	194	71.9	158	59.3	130	30.7	92	125.1
		1.0%		2.0%		4.0%		6.0%		8.0%
Share	B&H Profits	No. T.	Diff %	No. T.	Diff %	No. T.	Diff %	No. T.	Diff %	No. T.	Diff %
Sector 7
SCE	–11.9	190	64.3	156	88.5	94	132	70	181.0	62	161.7
GCE	–9.7	182	142.9	158	115.5	120	132.7	94	170.9	76	118.1
FCE	–61.1	206	–161.9	184	–143.3	130	–131.2	94	–114.1	84	–100.4
RKW	–52.6	162	–215.7	142	–194.6	126	–200.5	98	–175.5	76	–175.9
ARG	–95.7	130	35.8	114	23.3	96	26	62	85.3	54	2.1
AVG	0.2	223.7	–7.0	180.7	5.3	123.8	36.8	86.2	62.2	65.1	75.5

	10.0%		12.0%		14.0%		16.0%		18.0%
Share	No. T.	Diff %	No. T.	Diff %	No. T.	Diff %	No. T.	Diff %	No. T.	Diff %
Sector 1
NBK	22	–47.5	20	–71.9	16	–124.3	14	–143.0	8	85.2
GBK	26	167.1	20	157.2	14	150.4	6	297.1	4	460.7
CBK	20	–168.7	12	–81.3	10	123.5	8	133.1	6	194.0
ABK	24	–91.1	24	–147.0	20	–153.4	10	–78.9	8	–66.5
KIB	52	83.4	42	80.9	36	71.1	28	94.9	26	60.4
BUR	42	–4.3	38	–15.5	22	152.2	18	122.0	14	152.1
KFI	42	–110.5	28	–10.2	28	–64.8	20	–52.2	16	–0.6
Sector 2
KIV	46	122.7	38	99.9	38	46.5	30	48.9	28	45.2
FAC	30	–16.8	22	0.5	16	23.6	16	−23.4	14	–21.2
IFI	62	519.2	54	486.7	48	443.2	34	478.1	28	880.3
NIV	52	201.8	38	235.6	36	223.3	34	166.5	30	138.3
KPR	48	274.0	32	328.7	30	232.5	28	177.1	26	156.8
AIN	70	–3.2	60	–0.2	56	–75.8	46	–63.7	42	–78.8
SEC	28	234.7	24	212.7	18	209.8	18	117.7	16	107.9
IIC	64	–84.1	50	–54.0	34	16.8	24	59.5	24	16.4
SGC	36	69.2	26	105.8	22	54.8	20	15.5	16	19.3
KFC	46	141.7	42	117.0	36	83.5	32	118.0	30	105.8
Sector 3
KRE	60	88.4	56	–4.9	48	22	28	158.6	26	141.3
URE	38	16.6	32	–10.5	24	–6	24	–40.0	24	–59.0
NRE	66	251.5	50	292.4	46	173.3	34	222.3	32	207.7
PEA	58	–31.4	42	43.9	36	24.8	34	–8.8	30	1.7
TAM	42	230.8	30	271.7	28	189.4	24	201.0	24	143.9
Sector 4
NIN	44	474.5	38	413.8	36	268	36	196.6	30	139.9
PIP	60	229.4	52	234.1	38	253.4	32	247.1	30	197.1
KCE	44	111.2	34	95.4	26	155.1	24	95.6	16	154.8
RRI	38	137.9	30	95.6	24	114.6	22	61.3	16	120.2
SHP	62	79.7	40	184.5	34	143.9	28	115.8	22	131.0
UIC	80	–24.1	54	52.7	50	21.8	40	17.9	40	–35.2
BPC	36	186.8	32	124.0	32	57.4	28	85.5	24	70.8
	10.0%		12.0%		14.0%		16.0%		18.0%
Share	No. T.	Diff %	No. T.	Diff %	No. T.	Diff %	No. T.	Diff %	No. T.	Diff %
Sector 5
WAR	66	269.1	52	330.4	48	461.5	34	575.3	26	557.5
TEL	38	41.3	26	112.6	24	39.8	20	82.1	18	51.5
IPG	42	99.9	30	90.8	26	74.3	22	60.7	22	–0.5
CLE	66	164.7	58	148.4	48	163.9	40	175.7	34	207.3
SUL	68	203.1	54	288.9	48	278.6	36	290.1	32	311.8
AGH	88	7.8	72	–1.3	54	45.3	42	99.8	38	80.1
Sector 6
CAT	40	73.7	40	10.4	30	26.3	28	12.0	20	42.9
DAN	66	120.5	62	73.2	46	131	38	182.5	32	185.4
Sector 7
SCE	46	126.8	42	110.2	38	105.4	36	106.9	36	65.2
GCE	60	417.3	42	488.2	32	446.2	26	646.1	26	605.5
FCE	70	–110.2	44	78.1	38	111.8	24	166.6	20	185.1
RKW	70	–190.8	58	–136.9	44	–79.5	34	–57.4	28	–44.3
ARG	46	–12.6	38	10.1	28	49	24	57.0	22	16.3
AVG	50.1	101.2	40.0	115.3	33.5	111.5	27.2	124.2	23.9	136.5

Source: Authors’ analysis using SPSS & Minitab.

Note: The table presents the findings of 10 filter rules ranging from 1.0% to 18.0%. Specifically, it reports the number of trades (No. T.) produced by each filter rule strategy. In addition, it highlights, in percentage terms, the profits from the buy-and-hold strategy and the difference between the rule and the buy-and-hold profits (Diffs). AVG is the average of the number of trades, of the buy-and-hold strategy and of the difference between the rule and corresponding buy-and-hold profits across all 42 firms in the sample. All profits for the filter rules and the buy-and-hold strategy are reported net of transaction costs of the KSE. Sector 1 denotes banking while 2 to 7 represent investment, real estate, industrial, service, food and non-Kuwaiti sectors, respectively.

A visual inspection of Table 3 shows that, when the difference between the rule profits and the buy-and-hold gains are compared, the differences are positive for 9 of the 10 filter rules. Indeed, all filter strategies except the small filter sizes (1.0%) outperformed their buy-and-hold counterparts in the KSE.¹⁵ Such findings were different from the results reported by Al-Shamali (1989) and Fama and Blume (1966). Al-Shamali documented that profits from filter rules for Kuwaiti firms declined as the filter size increased. In the USA, Fama and Blume suggested that only small filters of 0.5 per cent outperformed their corresponding buy-and-hold strategies; possible reasons for this difference between the results of previous papers and the findings of the current study could be: (a) the market witnessed an upward trend over the 13-year period investigated and (b) large-sized filters generated higher returns relative to small-sized filters over the bullish market for the time period analysed.

Figure 1 provides a complete picture of the profitability of filter rules and the buy-and-hold strategy; the box plots of this figure show the distribution of profits (losses) of all filter rule sizes and the buy-and-hold strategy; it summarises the results of the filter rules. Overall, the results are similar to those in Table 3. However, a visual inspection of Figure 1 shows that none of the medians across the filter rules underperformed the buy-and-hold strategy which suggests that filter rules are more profitable than their buy-and-hold alternatives. Moreover, it appears that some outliers are present in the results which suggests that some shares generate enormous profits for most of the filter rules. For example, when the 18.0 per cent is employed, a firm such as IFI achieves profit of 801.7 per cent. Therefore, investors could have achieved incredible profits over the period 1998–2011, if they had implemented this filter rule.

A more detailed analysis of Table 3 shows that, on average, larger sized filters (12.0%, 14.0%, 16.0% and 18.0%) outperformed both medium-sized filters (6.0%, 8.0% and 10.0%) and small-sized filters (1.0%, 2.0% and 4.0%). Indeed, on average, the 18.0 per cent filter rule achieved the largest mean profit of 136.3 per cent while the 1.0 per cent filter recorded the largest loss of –7.0 per cent. In addition, the analysis illustrates that the number of firms that outperformed the buy-and-hold strategy increased significantly as the filter size rose except in the case of the 2.0, 8.0, 16.0 and 18.0 per cent filters. For instance, in the case of 22 firms, the 1.0 per cent filter outperformed the buy-and-hold strategy, in the case of 29 firms the 6.0 per cent filter was more profitable and for 36 of the 42 firms studied the 14.0 per cent filter outperformed the buy-and-hold alternative. Overall, in 68.3 per cent (287 out of 420 instances)¹⁶ of the cases investigated, the filter rules outperformed the buy-and-hold approach, which suggests that patterns exist in the KSE. As a result, the filter tests indicate that the KSE may not be weak-form efficient since a trading strategy based on historic data outperformed the corresponding naïve trading tactic.

Figure 1.

Source: Authors’ analysis using SPSS & Minitab.

It is apparent from Table 3 that when the 1.0, 2.0, 4.0, 6.0, 8.0, 10.0 and 12.0 per cent filters are tested on bank shares, the buy-and-hold strategy dominates; this suggests that the banking sector is more efficient than other industries in the sample. For example, when the 1.0 per cent filter rule was tested for the CBK, it underperformed the buy-and-hold strategy by −435.1 per cent. The results of filter rules from other sectors appear to be successful in most cases. Finally, Table 3 highlights that some shares such as IFI, GCE and WAR achieved incredible returns for most of the filters investigated. For instance, when the 18.0 per cent filter rule was implemented for IFI, GCE and WAR, the buy-and-hold strategy underperformed by 880.3, 605.5 and 557.5 per cent, respectively. Thus, sizeable profits were available to investors who followed this technical approach over the period 1998–2011.

Analysis of Variance

A general linear model (GLM) was fitted to the data in order to explain any variance between those profits earned by filter rules and the returns achieved by the corresponding buy-and-hold strategy (Diffs). The purpose of this testing is to investigate whether any difference in profits (Rule-B&H) varied systematically from one sector to another of across different filter sizes or levels. The GLM also seeks to uncover whether any difference in profits is related to the size of the firms (the market capitalisation) analysed in the sample. In explaining the variance of any difference between those profits earned by filter rules and the returns achieved by the corresponding buy-and-hold strategy, the model took the form:

_{j(s, r, m)}α_sβ_rαβ_srα_m_,

where Diffs _{j(s, r, m)} is the difference between the returns generated by filter rules and the profits earned by the corresponding buy-and-hold strategy for company j in sector s for market capitalisation m; µ is the overall mean for the difference between the filter rule profits and the buy-and-hold strategy for all firms across the rules, α_s is the main effect for sector and β_r is the main effect for the filter’s size. (αβ)_sr is the interaction effect for sectors and the rule’s size; this allows for the possibility that the profitability of rules may vary across sectors for different rule characteristics_. Finally, α_m is the regression coefficient for the market capitalisation.

Table 4 shows that the model was capable of explaining approximately 22.08 per cent (R²) of the variation in profit differences. A visual inspection of Table 4 shows that there is very strong evidence of differences in profitability between rules (F = 3.35, p = 0.001) and between sectors (F = 6.02, p < 0.0005); these main factors are statistically significant since the F-ratios are large and the p-values are less than 0.05. However, there is no evidence that the rules behave differently in different sectors; the interaction term had an F-ratio of only 0.26 with a p-value of 1.000. Thus, no single filter consistently outperformed the buy-and-hold strategy in different sectors.¹⁷

Table 4.

Analysis of Variance (ANOVA) of the Difference Between Filter Rule Returns and the Profits of the Buy-and-hold Strategy

Source of Variation	Degrees of Freedom	Sum of Squares	Mean Square	F-ratio	Sig. of F
Sector	6	877103	146184	6.02	0.000*
Rule	9	967340	81292	3.35	0.001*
Rule*Sector	54	342782	6348	0.26	1.000
Market Cap	1	215969	49954	2.06	0.152
Error	349	8478826	24295	–	–
Total	419	10882020	–	–	–
R-Sq	22.08%

Source: Authors’ analysis using SPSS & Minitab.

Note: The table shows the results of an ANOVA on the difference between the filter rule profits relative to the buy-and-hold strategy. The table tests whether any differences between sectors, rule percentages and firm sizes as measure by market capitalisation and for interactions between rules and sectors are significant. R-square represents the percentage of the variability in profit differences explained by the fitted model. Finally, Sig of F presents the significance of the F-test or P-value. An * significance at the 5.0% level. Data is based on information in Table 3.

The analysis suggests that selecting the filter size is extremely important for determining the overall profitability of the trading strategy. Not surprisingly, this result is consistent with the substantive literature which argues that the selection of an appropriate filter size influences the success of the filter strategy (Huang, 1995). For example, Huang found that, of the 24 filters tested, filters sizes, ranging from 4.5 to 18.0 per cent, outperformed the buy-and-hold strategy once transaction costs were considered, whereas filters below 4.5 per cent did not. Such a finding is consistent with the results in the fourth section which indicated that larger sized filters (12.0%, 14.0%, 16.0% and 18.0%) outperformed both medium- (6.0%, 8.0% and 10.0%) and small-sized filters (1.0%, 2.0% and 4.0%). Table 3 shows that the difference in profits achieved by the filter rules increased as the filter size rose. Further, all filters, except the 1.0 filter size, outperformed their buy-and-hold counterparts.¹⁸ Indeed, on average, the small-sized filters (1.0%, 2.0% and 4.0%) recorded the lowest returns of 11.7 per cent; medium-sized filters (6.0%, 8.0% and 10.0%) achieved returns of 79.6 per cent while the larger sized filters (12.0%, 14.0%, 16.0% and 18.0%) recorded the largest returns of 121.88 per cent. In addition, an inspection of the results shows that the banking industry is the only sector which consistently achieves losses when filter trading strategies are implemented. For example, this sector achieved large losses, on average, of approximately –15.6 per cent from the various filters tested. Such a conclusion corroborates the findings in the fourth section which suggested that shares in the banking sector were the most efficiently priced amongst the securities traded on the KSE. Finally, the investigation shows that there was no significant association between differences in filter rule profitability and firm size as measured by market capitalisation (F = 2.06, p = 0.152).

Conclusion

This study employs 10 filter rules to test the weak form of the EMH for the KSE over the 13-year period (1998–2011) for 42 firms listed in the KSE. The analysis showed that the KSE was not weak-form efficient because patterns and trends were present in the share prices; individuals who had followed filter strategies based on past price information could have made profits.

Further, the profitability of the technical trading rules shows that transaction costs of 0.1 per cent in Kuwait have no impact on the performance of filter rules. This finding is very different from the conclusion reached by studies such as Fama and Blume (1966) and Tijjani (2008) who believed that transaction costs eliminated a lot of gains achieved by trading rules. One explanation for this difference is that the level of transaction costs imposed by the KSE is very small. The findings from the sectoral analysis suggest that securities in the banking industry are the most efficiently priced amongst the shares traded on the KSE. One possible justification for this finding is that the level of transparency and disclosure is much higher for banks relative to other industries analysed in the sample.

Therefore, the results are consistent with the substantive literature which has argued that emerging stock markets are informationally inefficient, such as Fifield et al. (2005, 2008) and Xu (2010). In particular, the current findings also confirm the results from the early studies of Al-Shamali (1989) and Al-Loughani and Moosa (1999) that looked at trading rules for the KSE and found that trading strategies could exhibit patterns in share price changes. This study might assist policymakers in the state of Kuwait with understanding the context of the KSE; it might, therefore, lead them to introduce regulatory changes which could improve the level of the efficiency of the KSE.

Future research of the KSE could use new listed companies, focus on daily data and investigate the performance of a wider range of statistical and non-statistical tests. In order to facilitate future research in this area, a comprehensive database of high-frequency share price information is needed so that future researchers do not have to spend time gathering information about stock splits and share dividends before manually adjusting the returns earned by firms. Further, a methodical data capture system is needed which records the closing price of every share each day so that the academics might have confidence in the dataset being analysed.

Footnotes

Acknowledgements

This study was funded by the Public Authority for Applied Education & Training—Kuwait/Research Grant No. BS-12-07.

Notes

References

Abraham

Seyyed

F. J.

& Alsakran

S. A.

(2002). Testing the random walk behavior and efficiency of the Gulf stock markets. The Financial Review, 3(3), 469–480.

Al-Abdulqader

K. A.

Hannah

& Power

D. M.

(2007). A test of the weak form of the efficient markets hypothesis for the Saudi stock market. Journal of Emerging Market Finance, 2(6), 167–190.

Alexander

S. S.

(1964). Price movements in speculative markets: Trends or random walks. Industrial Management Review, 1, 25–46.

Aljoman Centre for Economic Consultancy (2011). Listed firm in the KSE. Retrieved 21 June 2011 , from http://aljoman.net/aljomanar/FreeZone/Fz_StartDate.aspx and http://www.aljoman.net/

Al-Khazali

O. M.

Ding

D. K.

& Pyun

C. S.

(2007). A new variance ratio test of random walk in emerging markets: A revisit. Financial Review, 2(2), 303–317.

Al-Loughani

N. E.

(1995). Random walk in thinly traded stock markets: The case of Kuwait. Arab Journal of Administrative Sciences, 34(1), 25–47.

Al-Loughani

N. E.

& Moosa

(1999). Testing the efficiency of an emerging stock market using trading rules: The case of Kuwait. Journal of the Gulf and Arabian Peninsula Studies, 95, 217–237.

Al-Mudhaf

J. M. S.

(1983). The Kuwait Stock Exchange: Its efficiency (Unpublished PhD thesis). Birmingham: Birmingham University.

Almujamed

(2011). Can the use of technical analysis in emerging markets be profitable?: The case of Kuwait. Saarbrücken: VDM Verlag Dr. Mülle.

10.

Al-Shamali

(1989). Weak form efficiency and factors leading to market efficiency in the Kuwait Stock Market (Unpublished PhD thesis). Loughborough: Loughborough University.

11.

Al-Yaqout

(2006). The usefulness of quarterly financial reports (Unpublished PhD thesis). Loughborough: Loughborough University.

12.

AMF (2012). Several bulletins from Arab Monetary Fund. Retrieved 21 April 2012 , from http://www.amf.org.ae/amdb

13.

Benjelloun

& Squalli

(2008). Do general indexes mask sectoral efficiency? A multiple variance ratio assessment of Middle Eastern equity markets. International Journal of Managerial Finance, 4(2), 136–151.

14.

Brock

Lakonishok

& Lebaron

(1992). Simple technical trading rules and the stochastic properties of stock returns. Journal of Finance, 47(5), 1731–1764.

15.

Butler

K. C.

& Malaikah

S. J.

(1992). Efficiency and inefficiency in thinly traded stock markets: Kuwait and Saudi Arabia. Journal of Banking and Finance, 16(1), 197–210.

16.

Central Bank of Kuwait (2011). Several statistical bulletins. Retrieved 5 June 2012 , from http://www.cbk.gov.kw/WWW/index.html

17.

Elango

& Hussein

M. I.

(2008). An empirical analysis on the weak-form efficiency of the GCC markets applying selected statistical tests. International Review of Business Research Papers, 4(1), 1–19.

18.

Fama

E. F.

(1965). The behavior of stock-market prices. The Journal of Business, 38(1), 34–105.

19.

Fama

E. F.

& Blume

M. E.

(1966). Filter rules and stock-market trading. Journal of Business, 39(1), 226–241.

20.

Fifield

S. G. M.

Power

D. M.

& Sinclair

C. D.

(2005). An analysis of trading strategies in eleven European stock markets. The European Journal of Finance, 11(6), 531–548.

21.

Fifield

S. G. M.

Power

D. M.

& Sinclair

C. D.

(2008). The performance of moving average rules in emerging stock markets. Applied Financial Economics, 18(19), 1515–1532.

22.

Gandhi

D. K.

Saunders

& Woodward

R. S.

(1980). Thin capital markets: A case study of the Kuwaiti sock market. Applied Economics, 12(2), 341–349.

23.

Gunasekarage

& Power

D. M.

(2001). The profitability of moving average trading rules in South Asian stock markets. Emerging Markets Review, 2(1), 17–33.

24.

Harvey

C. R.

(1995). Predictable risk and returns in emerging markets. Review of Financial Studies, 8(3), 773–816.

25.

Hassan

K. M.

Al-Sultan

W. S.

& Al-Saleem

J. A.

(2003). Stock market efficiency in the Gulf Cooperation Council Countries (GCC): The case of Kuwait Stock Exchange. Scientific Journal of Administrative Development, 1(1), 1–21.

26.

Huang

Y. S.

(1995). The trading performance of filter rules on the Taiwan Stock Exchange. Applied Financial Economics, 5(6), 391–395.

27.

Hudson

Dempsey

& Keasey

(1996). A note on the weak form efficiency of capital markets: The application of simple technical trading rules to UK stock prices—1935 to 1994. Journal of Banking and Finance, 20(6), 1121–1132.

28.

Malkiel

(1992). Efﬁcient market hypothesis. In Newman

Milgate

& Eatwell

(Eds), New Palgrave dictionary of money and finance. London: Macmillan.

29.

Moustafa

M. A.

(2004). Testing the weak-form efficiency of the United Arab Emirates stock market. International Journal of Business, 9(3), 309–325.

30.

Park

& Irwin

(2005). The profitability of technical trading rules in US futures markets: A data snooping free test (AGMAS Project Research Report 2004–05).

31.

Smith

(2007). Random walks in Middle Eastern stock markets. Applied Financial Economics, 17(7), 587–596.

32.

Sweeney

R. J.

(1988). Some new filter rule tests: Methods and results. Journal of Financial and Quantitative Analysis, 23(3), 285–300.

33.

Tijjani

(2008). An empirical investigation of the equity investment and share valuation process in Nigeria (Unpublished PhD thesis). Dundee: University of Dundee.

34.

(2010). An empirical investigation of weak-form efficiency in the Chinese Stock Market (Unpublished PhD thesis). Dundee: University of Dundee.