Managerial Finance 33:7 MENA Financial Markets 9781846634895, 9781846634888

Citation preview

mf cover (i).qxd

02/08/2007

12:07

Page 1

ISSN 0307-4358

Volume 33 Number 7 2007

Managerial Finance MENA financial markets Guest Editor: Dr Aktham Maghyereh

www.emeraldinsight.com

Managerial Finance

ISSN 0307-4358 Volume 33 Number 7 2007

MENA financial markets Guest Editor Dr Aktham Maghyereh

Access this journal online _________________________

447

Editorial advisory board___________________________

448

Oil prices and stock markets in GCC countries: new evidence from nonlinear cointegration analysis Aktham Maghyereh and Ahmad Al-Kandari ________________________

449

Financial integration, regulation and competitiveness in Middle East and North Africa countries Anastassios Gentzoglanis ________________________________________

461

The determinants of stock market development in the Middle-Eastern and North African region Samy Ben Naceur, Samir Ghazouani and Mohamed Omran ___________

477

Market efficiency, time-varying volatility and the asymmetric effect in Amman stock exchange Haitham Al-Zoubi and Bashir Kh.Al-Zu’bi __________________________

Access this journal electronically The current and past volumes of this journal are available at:

www.emeraldinsight.com/0307-4358.htm You can also search more than 150 additional Emerald journals in Emerald Management Xtra (www.emeraldinsight.com) See page following contents for full details of what your access includes.

490

CONTENTS

CONTENTS

Empirical testing of the loss provisions of banks in the GCC region

continued

Taisier A. Zoubi and Osamah Al-Khazali ___________________________

500

Call for papers ___________________________________

512

www.emeraldinsight.com/mf.htm As a subscriber to this journal, you can benefit from instant, electronic access to this title via Emerald Management Xtra. Your access includes a variety of features that increase the value of your journal subscription.

Structured abstracts Emerald structured abstracts provide consistent, clear and informative summaries of the content of the articles, allowing faster evaluation of papers.

How to access this journal electronically

Additional complimentary services available

To benefit from electronic access to this journal, please contact [email protected] A set of login details will then be provided to you. Should you wish to access via IP, please provide these details in your e-mail. Once registration is completed, your institution will have instant access to all articles through the journal’s Table of Contents page at www.emeraldinsight.com/0307-4358.htm More information about the journal is also available at www.emeraldinsight.com/ mf.htm

Your access includes a variety of features that add to the functionality and value of your journal subscription:

Our liberal institution-wide licence allows everyone within your institution to access your journal electronically, making your subscription more cost-effective. Our web site has been designed to provide you with a comprehensive, simple system that needs only minimum administration. Access is available via IP authentication or username and password.

E-mail alert services These services allow you to be kept up to date with the latest additions to the journal via e-mail, as soon as new material enters the database. Further information about the services available can be found at www.emeraldinsight.com/alerts

Emerald online training services Visit www.emeraldinsight.com/training and take an Emerald online tour to help you get the most from your subscription.

Key features of Emerald electronic journals Automatic permission to make up to 25 copies of individual articles This facility can be used for training purposes, course notes, seminars etc. This only applies to articles of which Emerald owns copyright. For further details visit www.emeraldinsight.com/ copyright Online publishing and archiving As well as current volumes of the journal, you can also gain access to past volumes on the internet via Emerald Management Xtra. You can browse or search these databases for relevant articles. Key readings This feature provides abstracts of related articles chosen by the journal editor, selected to provide readers with current awareness of interesting articles from other publications in the field. Non-article content Material in our journals such as product information, industry trends, company news, conferences, etc. is available online and can be accessed by users. Reference linking Direct links from the journal article references to abstracts of the most influential articles cited. Where possible, this link is to the full text of the article. E-mail an article Allows users to e-mail links to relevant and interesting articles to another computer for later use, reference or printing purposes.

Xtra resources and collections When you register your journal subscription online, you will gain access to Xtra resources for Librarians, Faculty, Authors, Researchers, Deans and Managers. In addition you can access Emerald Collections, which include case studies, book reviews, guru interviews and literature reviews.

Emerald Research Connections An online meeting place for the research community where researchers present their own work and interests and seek other researchers for future projects. Register yourself or search our database of researchers at www.emeraldinsight.com/ connections

Choice of access Electronic access to this journal is available via a number of channels. Our web site www.emeraldinsight.com is the recommended means of electronic access, as it provides fully searchable and value added access to the complete content of the journal. However, you can also access and search the article content of this journal through the following journal delivery services: EBSCOHost Electronic Journals Service ejournals.ebsco.com Informatics J-Gate www.j-gate.informindia.co.in Ingenta www.ingenta.com Minerva Electronic Online Services www.minerva.at OCLC FirstSearch www.oclc.org/firstsearch SilverLinker www.ovid.com SwetsWise www.swetswise.com

Emerald Customer Support For customer support and technical help contact: E-mail [email protected] Web www.emeraldinsight.com/customercharter Tel +44 (0) 1274 785278 Fax +44 (0) 1274 785201

MF 33,7

448

EDITORIAL ADVISORY BOARD

Professor Kofi A. Amoateng NC Central University, Durham, USA

Professor Moses L. Pava Yeshiva University, New York, USA

Professor Felix Ayadi Jesse E. Jones School of Business, Texas Southern University, USA Professor Mohamed E. Bayou The University of Michigan-Dearborn, USA Dr Andre de Korvin University of Houston-Downtown, USA Dr Colin J. Dodds Saint Mary’s University, Halifax, Nova Scotia, Canada Professor John Doukas Old Dominion University, Norfolk, Virginia, USA

Professor George C. Philipatos The University of Tennessee, Knoxville, Tennessee, USA Professor David Rayome Northern Michigan University, USA Professor Alan Reinstein Wayne State University, Detroit, Michigan, USA Professor Ahmed Riahi-Belkaoui The University of Illinois at Chicago, USA Professor Mauricio Rodriguez Texas Christian University, Fort Worth, Texas, USA

Professor Uric Dufrene Indiana University Southeast, New Albany, Indiana, USA

Professor Salil K. Sarkar Henderson State University, Arkadelphia, Arkansas, USA

Professor Ali M. Fatemi De Paul University, Chicago, Illinois, USA

Professor Atul A. Saxena Mercer University, Georgia, USA

Professor Iftekhar Hasan New Jersey Institute of Technology, USA

Professor Philip H. Siegel Monmouth University, New Jersey, USA

Professor Suk H. Kim University of Detroit Mercy, Detroit, USA

Professor Kevin J. Sigler The University of North Carolina at Wilmington, USA

Professor John Leavins University of Houston-Downtown, USA Professor R. Charles Moyer Wake Forest University, Winston-Salem, North Carolina, USA Dr Khursheed Omer University of Houston-Downtown, USA

Managerial Finance Vol. 33 No. 7, 2007 p. 448 # Emerald Group Publishing Limited 0307-4358

Professor Gordon Wills International Management Centres, UK Professor Stephen A. Zeff Rice University, Texas, USA

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0307-4358.htm

Oil prices and stock markets in GCC countries: new evidence from nonlinear cointegration analysis

Oil prices and stock markets in GCC countries 449

Aktham Maghyereh Department of Economics and Finance, United Arab Emirates University, Al-Ain, UAE, and

Ahmad Al-Kandari Department of Insurance and Banking, College of Business Studies, Kuwait Abstract Purpose – The purpose of this research is to examine the linkages between oil prices and stock market in Gulf Cooperation Council (GCC) countries. Prior work argues that oil prices and the GCC stock markets are not related. This conclusion could be due to the fact that only linear linkages have been examined. Design/methodology/approach – This study employs newly developed techniques of rank tests of nonlinear cointegration analysis proposed by Breitung and Gourieroux and Breitung. The Breitung’s method is selected in this study due its potential superiority at detecting cointegration when the error-correction mechanism is nonlinear. Findings – The empirical analysis of the paper supports that oil price impact the stock price indices in GCC countries in a nonlinear fashion. Thus, the statistical analysis in this paper obviously supports a nonlinear modeling of the relationship between oil and the economy. Research limitations/implications – The paper contains the normal limitations associated with the econometric method including statistical bias. Practical implications – The implication of this paper findings is that policy makers at GCC countries should keep an eye on the effects of changes in oil price levels on their own economies and stock markets. For individual and institutional investors, the nonlinear relationship between oil and stock markets imply predictability in the GCC stock markets. Originality/value – The paper presents new findings on the relationships between oil prices and the stock market in GCC countries. These findings should be of interest to researchers, regulators, and market participants. Keywords Oil industry, Stock markets, Persian Gulf States Paper type Research paper

1. Introduction The oil represents one of the most important macroeconomic factors in the world economy. Not surprising because the crude oil market is the largest commodity market in the world. What makes oil price changes even more interesting is not only their direct impact on economic activity, but also the changes in oil prices might reflect or even predict changes in international stability (Leigh et al., 2003). This means that oil price changes might not only have a direct effect on consumption and production, but that oil price changes can also proxy for changing risk aversion in the economy. As well documented, increases in oil demand without offsetting increases in supply lead to higher oil prices. Higher oil prices perform like an inflation tax on consumer and producers by reducing the amount of disposable income that consumers have left to spend on other goods and services and raising the costs of non-oil producing

Managerial Finance Vol. 33 No. 7, 2007 pp. 449-460 # Emerald Group Publishing Limited 0307-4358 DOI 10.1108/03074350710753735

MF 33,7

450

companies and, in the absence of fully passing these costs on the consumers, reducing profits and dividends, which are key drives of stock prices. In addition, oil prices respond to geopolitics, institutional arrangements (OPEC), and the dynamic of the future market (Sadorsky, 2004). Unanticipated changes in any of these factors can create volatility, and hence risk, in oil prices. Oil price volatility increases uncertainty, which may effect negatively on wealth and investment. While a large body of the empirical research has focus on the relationship between economic activity and oil price changes[1], few studies have been conducted on the relationship between financial markets and oil prices and those mainly for a few developed economies such as the USA and UK ( Jones and Kaul, 1992; Jones and Kaul, 1996; Huang et al., 1996; Gjerde and Saettem, 1999). Jones and Kaul (1992) examine the effect of oil prices on stock prices in the USA. They find an effect of oil prices on aggregate real stock returns, including a lagged effect, in the period 1947 to 1991. In a more recent study, Jones and Kaul (1996) test whether the reaction of international stock markets to oil prices can be justified by current and future changes in real cash flows and/or changes in expected returns. They find that in the postwar period, the reaction of US and Canadian stock prices to oil shocks can be completely accounted for by the impact of these shocks on real cash flows. In contrast, the results for both the UK and Japan are not strong. In an important study, Huang et al. (1996) examine the link between daily oil future returns and daily US stock returns. The evidence suggests that oil futures returns do lead some individual oil company stock returns but do not have much impact on general market indices. Gjerde and Saettem (1999) demonstrate that stock returns have a positive and delayed response to changes in industrial production and that the stock market responds rationally to oil price changes in the Norwegian market. In contrast to the work done on developed markets, relatively limited studies have focused on the relationship between oil prices and emerging stock markets (Papapetrou, 2001; Maghyereh, 2004; Hammoudeh and Eleisa, 2004). Papapetrou (2001) uses a multivariate vector regression model to study the dynamic interaction between oil prices, stock prices, and real economic activity in Greece. His empirical results show that changes in oil prices influence real activity and employment. Maghyereh (2004) uses a vector autoregression model to study the relationship between oil prices changes and stock returns in 22 emerging markets. He finds that oil price changes have no significant role in affecting stock returns. Hammoudeh and Eleisa (2004, HE henceforth) use the vector error-correction model to study the relationship between oil prices and stock prices for five members (Bahrain, Kuwait, Oman, Suadi Arabia, and the United Arab Emirates) of the Gulf Cooperation Council (GCC). They find that only the Saudi Arabia stock market has a relationship between oil prices and stock prices. Based on these results, one can conclude that the much-touted influence of oil price shocks on the economic activity is more of a legend than reality in the emerging markets. However, given the structure of the GCC countries, the findings of the HE represent a puzzle. These countries collectively account for about 16 per cent of the world a-day production in 2003, and possess 47 per cent of the world oil reserves. For these countries, oil exports largely determine their foreign earnings and their governments’ budget revenues and expenditures. Thus, they are the primary determinant of aggregate demand. The aggregate demand effect influences corporate output and domestic price levels, which eventually impact corporate earnings and stock market prices. This demand effect can also indirectly impact stock prices

through its influence on expect inflation, which in turn affects the expected discount rate. Oil prices also have a more direct effect on the GCC domestic interest rates through their effects on the availability of liquidity. Thus, one can expect that the oil prices change play a key role in determining the movements of the GCC stock markets. The conclusion of the HE could be due to the fact that their tests relied on are not powerful to detect nonlinear linkages. In their study, the Johansen cointegration test has been employed. This technique is built on the basis of linear autoregressive model and, implicitly, assumes that underlying dynamics are in linear form or can be made linear by a simple transformation. However, number of authors, such as Mork (1989), Mork et al. (1994), and Hamilton (1996, 2000), argue that there is a nonlinear relationship between the oil and economy. Specifically, these studies show that oil price increases are much more influential than oil price decreases, creating an asymmetric relationship between oil prices and economic activity. Furthermore, Bansal et al. (1993) demonstrate the economic significance of nonlinearities in asset pricing. Also, Hiemstra and Kramer (1997) argue that linear asset pricing models omit potentially useful aspects of the relationship between macroeconomic variables and stock returns. Furthermore, in a recent study, Ciner (2001) examines the linear and nonlinear casualty between oil futures returns and S&P 500 stock index and finds that oil price shocks impact the stock index returns in a nonlinear fashion. This suggests that there could be uncovered nonlinear relationships between oil prices and the stock market in GCC countries that was ignored by HE. Motivated by the above consideration, this study employs newly developed techniques of rank tests of nonlinear cointegration analysis proposed by Breitung and Gourieroux (1997) and Breitung (2001) to examine the relationship between oil prices and stock prices for the GCC countries. The Breitung’s method is selected in this study due its potential superiority at detecting cointegration when the error-correction mechanism is nonlinear. The study proceeds as follows. Section 2 provides an overview of the methodological issues. Section 3 presents the empirical evidence. Section 4 provides some concluding remarks. 2. Methodology 2.1 Rank unit root test Taking into account the previous argument, this paper implements rank test for unit root developed in Breitung and Gourieroux (1997). Consider ranked unit root rests that are invariant with respect to the error distribution and depend only on the ranks of observed differences. Let xt denote an observed time series, and of the observed series such as yt ¼ f ðxt Þ be a monotonic nonlinear transformation p f ðxt Þ ¼ x3t ; f ðxt Þ ¼ lnðxt Þ if xt > 0, and f ðxt Þ ¼ ðxx Þ if xt
0. Suppose the true data generating process is nonlinear; namely, it is yt rather than xt that are directly involved in the data generating process. In many cases, the functional form f ðÞ is unknown. To the particular form of transformation requires the estimation of the function f ðÞ from the data. According to Breitung and Gourieroux (1997), the estimation techniques are not yet available. So it is desirable to conduct unit root tests, which are unaffected by the choice of f ðÞ. To serve this purpose, Breitung and Gourieroux (1997) developed a tests procedure based on the ranks of the observed differences. Consider Rt as the uniform ranks of the differenced observations, such as Rt ¼ ðRank of
x1 ;
x2 ; . . . ;
xT Þ  ððT  1Þ=2Þ where T is the sample

Oil prices and stock markets in GCC countries 451

MF 33,7

size, the uniform version of the score-type rank test statistic is then computed as follows: !2 T t X X 1
T ðuniÞ ¼ 12 T Rs s ð1Þ t¼1

452

s¼1

However, the Breitung and Gourieroux (1997) version, which is based on inverse normal scores (INS) transformation of ranks, is given by !2 T t X X 1 0
T ðINSÞ ¼ 12 T Rs s ð2Þ t¼1

s¼1

where Rt0 is the INS transformation of Rt (such as Rr0 ¼ 1 ðT 1 Rt þ 0:5Þ)[2]. While the above two tests test the difference stationary model against the trendstationary model, they are directly based on the observed series xt rather than on the transformed series yt. This means that as long as that yt is a monotonic transformation of xt inspective of what functional forms the transformation might take, the rank test results using xt can carry over to yt. Breitung and Gourieroux (1997) conducted Monte Carlo simulation on the small sample properties and compared between the two traditional unit root tests (DF and SP)[3] and the two ranks tests statistics ð
ðuniÞ and
ðINSÞÞ and found that the two formal statistics suffer considerable size distortion from the misspecification of the process as a linear random walk. This is reflected in the fact that the formal two statistics have much higher rejection frequencies than the normal ones, while the latter two statistics have rejection frequencies close to the normal ones. Furthermore, Breitung and Gourieroux (1997) also argue that the rank tests are invariant with respect to the error distribution and are also robust against outliers and structural changes. 2.2 Rank tests for cointegration As a next step, we conduct the rank tests for cointegration developed by Breitung (2001). For a bivariate model, nonlinear cointegration may be tested using the -type and -type statistics. Suppose that we have two variables y1t ¼ f1 ðx1t Þ and y2t ¼ f2 ðx2t Þ are both I ð1Þ, where x1t and x2t are observed while f1 ðÞ and f2 ðÞ are monotonically increasing functions but are unknown. If the difference between y1t and y2t is integrated of order zero i.e. ut = y1t  y2t ~ I(0), then the a nonlinear cointegration relationship exist between x1t and x2t . Since the sequence of ranks is invariant to a monotonic transformation of the original data, the problem of unknown f1 ðÞ and f2 ðÞ may be circumvented by replacing them with the rank series, RT ðy1t Þ ¼ RT ½ f1 ðx1t Þ ¼ RT ðx1t Þ and RT ðy1t Þ ¼ RT ½f2 ðx2t Þ ¼ RT ðx2t Þ. Breitung considered two ‘‘distance measures’’: T ¼ T 1

supt jdt j ^
d

ð3Þ

and T

¼T

3

PT

2 T¼1 dt 2
d

ð4Þ

_

where dt ¼ RT ðx1t Þ  RT ðx2t Þ;  2
u is the variance of
e uRt , and supt jdt j is the maximum value of jdt j over t ¼ 1; 2; . . . ; T. The null hypothesis to be tested is no (nonlinear) cointegration. Breitung demonstrated with Monte Carlo simulations that theses tests perform reasonably well for small values of the correlation coefficient TR [4] between the ranked series RT ðx1t Þ and RT ðx2t Þ. The underlying assumption of the statistics expressed in equations (3) and (4) is that ranked series RT ðx1t Þ and RT ðx2t Þ have small correlation. To take into account the possibility that the correlation coefficient may be close to one, Breitung suggests the following correction (if 5 per cent is chosen to be the significance level): T ¼ ð5Þ  T 1  0:147ðTR Þ2 and T ¼

T 1  0:462TT

ð6Þ

The tests discussed above are designed to detect cointegration, i.e. to reject the null hypothesis of no cointegration when the residuals ut are I(0). Cointegration, if it exist, may be linear form or of nonlinear form. The Monte Carlo experiments in Breitung demonstrated that the rank tests have good power properties not only in the nonlinear case but also in the linear case. To decide whether a cointegrating relation is linear or non-linear, Breitung proposed a score statistic based on the rank transformation of the time series. This test is applied if the rank tests indicate cointegration. 2.3. Score statistic for nonlinear cointegration Consider the following nonlinear relationship between two time series: xit ¼ 0 þ 1 x2t þ f  ðx2t Þ þ ut , where 0 þ 1 x2t is the linear part. Under the null hypothesis, f  ðx2t Þ ¼ 0 and the ut are I(0) so that there is linear cointegration. Under the alternative hypothesis, f  ðx2t Þ 6¼ 0 and the ut are I(0) so that there is nonlinear cointegration. The score test statistic is given by TR2 from a least squares regression of _ u t on c0 þ c1 x2t þ c2 RT ðx2t Þ þ et . The ut are the residuals under the null hypothesis, possibly corrected for serial correlation and endogoneity using dynamic ordinary least squares method (DOLS) of Stock and Watson (1993). We apply the Schwarz-Bayesian criterion (SBC) to select appropriate leads and lags for DOLS. Under the null hypothesis, the test statistic is distributed as 2 with one degree of freedom. 3. Data and empirical analysis 3.1 Data The stock market data in this paper is drawn from value-weighted equity market indices of four GCC markets; namely, Bahrain, Kuwait, Oman, and Saudi Arabia. All data were collected piece by piece directly form individual exchanges and encompasses the period from 1 January 1996 to 31 December 2003. The crude oil market is the largest commodity market in the world. Total world consumption equals around 80 million barrels a day in 2003. Prices of three types of oil – Brent, West Texas Intermediate and Dubai – serve as a benchmark for other types of crude oil. Of total world oil consumption of about 80 million barrels a day in 2003, Brent oil serves as a benchmark for about 50 million barrels a day, West Texas Intermediate for about 15 million barrels a day, and Dubai for about 15 million barrels a day.

Oil prices and stock markets in GCC countries 453

MF 33,7

454

Even though price differences do exist, crude oil prices tend to move very closely together. Since Brent oil serves as a benchmark in the crude oil market, daily closing prices of crude oil Brent are used as our primary proxy for the world price of crude oil[5]. The daily closing prices for crude oil Brent for the period from 1 January 1996 to 31 December 2003 are obtained from the US Energy Information Administration. To assess the distributional properties of the data, Table I report summary statistical for the stock market returns and oil return. As we can see, all of the four series are skewed to the right. The kurtosis of each of the series is also higher compared to the normal distribution, which has a kurtosis of 3. This means that the empirical distribution has more weight in the tails and thus leptokurtic. A formal test for normally of the series can also be done by using the Jarque–Bera (1987) ( JB) test, which combines the skewness and kurtosis results. The results show that the JB test rejects the null hypothesis of normality for all series at the 1 per cent. 3.2 Results from traditional, linear cointegration analysis To better motivate the use of the rank test procedure, we first conduct traditional, linear cointegration analysis. Traditionally, the series transformed to logarithms before unit root and cointegration tests are conducted. Such a transformation is a special case of nonlinear ones; a log-linear generating process is still nonlinear with respect to the original data. To distinguish between strictly linear models and nonlinear models, we apply (augmented) Dickey–Fuller (ADF) and Phillips–Peron (PP) unit root tests to the original data without any transformations. To determine the appropriate lag length of the tests, we used the information criteria provided by the Akaike (AIC), SBC and the Hannan–Quinn (HQC) statistics. In the case of different recommendation provided by the criteria, greater weight is offered to the SBC and HQC statistics in view of the tendency of AIC statistic to overestimate the lag (Dickey and Fuller, 1981). The equation implied by the selection lag is then tested for autocorrelation in the residuals, and in no case, the hypothesis of non-autocorrelated residuals is rejected. The ADF and PP tests are reported in Table II. As expected, the typical nonrejection of a unit root in all stock price indices as well as oil price is upheld. Specifically, the both tests cannot reject the unit root null in any of the series. At their first differences, the both tests reject the unit root in all series. Thus, ADF and PP tests both confirm that all series under consideration are unit root processes. Our next task is to check whether the series are cointegrated. Specifically, having established the presence of a unit root in the first-difference of each variable, we need to test whether stock market and oil price has different unit roots (non-cointegrated) or shares the same unit root (cointegrated). Cointegrated variables, if disturbed, with not

Table I. Descriptive statistics for the stock market and oil returns

Bahrain Kuwait Oman Saudi Arabia Oil price

Mean

Maximum

Minimum

Standard deviation

Skewness

Kurtosis

JB-Statistics

0.0207 0.0107 0.0171 0.0622 0.1007

2.871 3.846 8.041 6.403 12.048

3.487 3.817 4.867 3.506 22.364

0.518 0.632 0.672 0.771 2.703

0.703 0.488 1.666 0.632 0.737

12.347 8.915 19.406 12.288 9.590

3,473.4* 1,108.4* 10,895.4* 3,416.1* 1,738.7*

Note: * Indicates significance at the 1 per cent levels

Level

Bahrain

Test statistics P 0.21 PP(1) 1.54 PP(2) 5.26 ADF(1) 1.6093 ADF(2) 1.0255 ADF(3) 0.9476 First difference P 29.2038** PP(1) 24.2269** PP(2) 29.2150** ADF(1) 10.9631** ADF(2) 9.9721** ADF(3) 9.9000**

Kuwait

Oman

4.04 8.03 9.50 0.7430 2.3160 3.3315*

0.28 3.80 8.81 0.7197 1.4483 1.6623

66.9563** 66.4869** 67.0217** 40.5947** 41.5720** 41.6173**

24.7441** 23.7378** 24.7532** 11.4241** 10.4473** 11.4171**

Saudi Arabia

0.46 2.21 10.71 1.6907 1.5874 2.2894 27.7662** 25.7610** 26.6786** 11.8908** 10.8983** 11.7453**

Oil price

Critical regions 5% 10%

0.1983 2.5813 2.4409 0.1903 1.4424 2.3494

8.29 14.51 21.78 1.93 2.89 3.40

5.88 11.65 18.42 1.60 2.58 3.13

29.757** 29.766** 30.7681** 13.4289** 14.4466** 16.4627**

8.29 14.51 21.78 1.93 2.89 3.40

5.88 11.65 18.42 1.60 2.58 3.13

Notes: P, Philips tests with zero-mean stationarity; PP, Phillips–Perron test with (1) stationarity and (2) trend stationarity; ADF, augmented Dickey–Fuller test, with (1) zero-mean stationarity, (2) stationarity, and (3) trend stationarity; KPSS, Kwiatkowski, Phillips, Schmidt, and Shin tests, with (1) stationarity and (2) trend stationarity. * and ** denotes rejection of the null hypothesis at the 10 and 5 per cent significant level, respectively

drift apart from each other and thus posses a long-run equilibrium relationship. The existence of cointegration would imply that the two series would not drift too far apart. A none-stationary variable, by definition, tends to wander extensively over time, but a pair of non-stationary variables may have the property that a particular linear combination would keep them together, that is, they do not drift far apart. Under this scenario, the two variables are said to be cointegrated or possess a long-run linear stable relationship. We test the linear cointegration between each of GCC stock market and the oil price with the methodology suggested by Johansen (1988, 1991, 1995) and Johansen and Juselius (1990). In testing the long-run linear relationship between each of stock market price index and oil price, the null hypothesis states that there is no cointegration relation. The results of the Johansen cointegration test are provided in Table III. The results suggest that none of stock market under consideration has linear long-run relation with the oil price. These results are highly consistent with the conclusion of Hammoudeh and Eleisa (2004). If one stopped here, it would be concluded that the significant of oil price for the aggregate economy in these countries is greatly exaggerated. However, the above results should be treated with caution for the following reasons. First, it has been well documented that there is a nonlinear relationship between oil and the economy. This suggests that there could be uncovered nonlinear linkages between oil prices and the stock market. Second, while the most unit root tests are based on the assumption of normality distributed errors (and so is the Johansen cointegration test), the distribution of the series under consideration are found to be leptokurtic (i.e. fat tailed). That explains why the unit root and cointegration tests results in Tables II and III all suffer from failure to pass the normality test, which is also the case in the previous

Oil prices and stock markets in GCC countries 455

Table II. Unit root and stationarity test results

MF 33,7 Bahrain Kuwait

456

Oman Saudi Arabia

Table III. Johansen’s cointegration tests

H0: rank ¼ p

Model with trend max Trace

Model without trend max Trace

p¼0 p0 p¼0 p0 p¼0 p0 p¼0 p0

14.475 4.561 11.954 3.873 10.847 3.421 17.745 2.832

13.258 4.594 11.935 4.982 14.824 3.493 17.754 2.036

19.874 4.561 21.484 3.873 16.955 3.421 18.952 2.832

21.281 4.594 22.921 4.982 19.549 3.493 21.838 2.036

Notes: max is the likelihood ratio test based on maximal eigenvalue of the stochastic matrix; Trace is the likelihood ratio test based on trace of the stochastic matrix. The truncation lag length m is so determined that the computed vector portmanteau statistic for the chosen s ¼ 20 lags (¼40 lags for each pair) is smaller than 2n2(s – m), where n is the number of dependent variables. This ensures no residual autocorrelation. All models fail normality tests (not reported)

study. Ignoring such a problem of non-Gaussianity of series may invalidate the test results. To sum up, the absence of normal distribution of series and the non-existence of the linear way in which stock market price indices response to oil price render the test results in this section dubious. Taking into account the Non-Gaussianity and nonlinearity in the data, we then employ newly developed techniques of rank tests of nonlinear cointegration analysis proposed by Breitung and Gourieroux (1997) and Breitung (2001). 3.3 Results from rank tests Although the rank unit root tests are applied directly to the observed series, it is the order of integration of their monotonic transformations that is to be determined. Table IV shows that the null hypothesis of difference-stationarity cannot be rejected for all series, but can be rejected at a higher than 1 per cent level for the firstdifferences of all series. This implies that their transformed series are all characterized as difference-stationary rather than trend-stationary, and thus the observed four stock indices as well as oil price are likely to be nonlinear random walk process (i.e. their monotonically transformed series are random walk processes). Level

Bahrain

Kuwait

Oman

Saudi Arabia

Oil price

Test statistics T(uni) T(INS)

0.35679 0.08638

0.31421 0.19542

0.36576 0.07774

0.28381 0.08722

0.24812 0.09625

0.00000* 0.00000*

0.00059* 0.00025*

0.00271* 0.00013*

0.00182* 0.00053*

First difference T(uni) 0.00118* T(INS) 0.00022* Table IV. Rank tests for a unit root

Note: * indicate significance at the 5 per cent level, using value with T ¼ 1 in Table 6 of Breitung and Gourieroux (1997)

Giving that the monotonic transformation of all series are I(1) processes, as explained in section 2, the rank rests for possible nonlinear cointegration in the relationship between the GCC market and oil price can be conducted. The results of cointegration rank test statistics between each of GCC stock market and oil price are presented in Table V. From the table, it turns out that a cointegration relationship exist between each of the GCC market and oil price, since both the -type and the -type rank test statistics unanimously reject the no-cointegration null at a higher than 5 per cent significant level. Comparing Table III with Table V where linear cointegration test statistics are all insignificant, it may claim that the cointegration relationship between the stock markets in GCC countries and oil price as detected by the rank tests is nonlinear. To further support to the above evidence that there are nonlinear long-run relationship between stock market price indices and oil price, we perform score test statistics for nonlinear cointegration. As shown in the last column in Table V, the nonlinear-score tests turns out to be significantly different from zero at the 1 per cent level. The same holds true at the 5 per cent level, except for the Oman that shows some week evidence in favor of linearity. We conclude that our findings support a nonlinear cointegration relation between the GCC stock market and oil price. Overall the statistical analysis puts forward main empirical findings. There is a nonlinear relation between oil and stock markets in GCC countries. This is, in general, consistent with the previous studies which giving an unique place for oil in the economy and also with Huang et al. (1996), Gjerde and Saettem (1999), and Ciner (2001), who argue that oil shocks affect stock returns using different data and estimation methods. A recent study of Cochrane (2001) in his treatment of asset pricing model offers an explanation for the nonlinear relation between the oil and stock market. His analysis suggests that returns on oil and the stock market are linked via a stochastic (random) discount factor, that is frequent to all assets. If the association between oil and the stock market depends on the properties of the discount factor, it is questionable that the inflation plays an important role in the transmission process. It is well known that Fisher’s hypothesis argue that the risk-free component of the discount factor is comprised of expected inflation rate and a real rate of return. For this reason, shocks to the inflation rate are likely to affect both oil and stock returns, creating a tie.

Bahrain Kuwait Oman Saudi Arabia

T*

T*

T**

T**

TR

Nonlinear score

0.2921* 0.2358* 0.1975* 0.2510*

0.0059* 0.0062* 0.0019* 0.0028*

0.2365* 0.2384* 0.2977* 0.2580*

0.0059* 0.0068* 0.0023* 0.0027*

0.058 0.138 0.117 0.036

4.817* 6.419** 3.283*** 9.712**

Notes: The critical values of T*, T**, T*, and T** (not reported) are obtained from Monte Carlo simulation with 10,000 replications and T ¼ 1,883. Comparing the obtained critical values with Table 1 in Breitung (2001), we find that, as the sample size changes, the critical values not change so much. The nonlinear-score test follows a 2 distribution with one degree of freedom. * indicate significance at 5 per cent level, ** indicate significance at 1 per cent level and *** indicate significance at 10 per cent level

Oil prices and stock markets in GCC countries 457

Table V. Rank tests for nonlinear cointegration

MF 33,7

458

4. Conclusion The purpose of this paper is to contribute to the literature on stock markets and oil prices by studying the nonlinear impact of the oil prices on the GCC stock market price indices. This is an important and interesting topic to study because, for GCC countries, oil exports largely determine their foreign earnings and their governments’ budget revenues and expenditures. The risk from price changes, thus, play a key role in the development of these countries and their financial markets. In an important study, HE detect no relationship between oil prices and the GCC stock market returns, which is against the importance of the oil prices on the economy of these countries. This study argues that the conclusion of HE is due to the fact that they focus solely on linear dependences. In this paper, we consider an application of rank tests for a nonlinear cointegration relationship between oil price and the stock markets in GCC countries. Our empirical analysis supports that oil price impacts the stock price indices in GCC countries in a nonlinear fashion. Thus, the statistical analysis in this paper obviously supports a nonlinear modeling of the relationship between oil and the economy, which is consistent with some authors, such as Mork (1989), Mork et al. (1994), and Hamilton (1996, 2000). The implication of our findings is that policy makers at GCC countries should keep an eye on the effects of changes in oil price levels on their own economies and stock markets. For individual and institutional investors, the nonlinear relationship between oil and stock markets implies predictability in the GCC stock markets. Notes 1. The early studies include Pierce and Jared (1974), Rasche and Tatom (1977), and Darby (1982), all of which documented and explained the inverse relationship between oil price increases and aggregate economic activity. Later empirical studies – such as Hickman et al. (1987), Jones and Leiby (1996), Hooker (1999), Hammes and Wills (2003), and Leigh et al. (2003) – confirm the inverse relationship between oil prices and aggregate economic activity. 2. The critical values of
T ðuniÞ and
T ðINSÞ are available in Appendix B of Breitung and Gourieroux (1997). 3. DF and SP denotes to Dickey-Fuller and Schmidt and Phillips unit root tests. 4. TR is obtained by the following way: PT T t¼2
RT ðxit Þ
RT ðx2t Þ q ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi R ¼ P PT ð t¼2
RT ðx1t Þ2 Þð Tt¼2
RT ðx2t Þ2 Þ 5. We also estimated the results using daily for Arab light, Arab Medium, Dubai, and West Texas as alternatives for the world price of oil and found these measures did not substantively affect our results. References Bansal, R.D., Hsieh, A. and Viswanathan, S. (1993), ‘‘A new approach to international arbitrage pricing’’, Journal of Finance, Vol. 48, pp. 1719-48. Breitung, J. (2001), ‘‘Rank tests for nonlinear cointegration’’, Journal of Business and Economic Statistics, Vol. 19, pp. 331-40. Breitung, J. and Gourieroux, C. (1997), ‘‘Rank tests for unit roots’’, Journal of Econometrics, Vol. 81, pp. 7-27.

Ciner, C. (2001), ‘‘Energy shocks and financial markets: nonlinear linkages’’, Studies in Nonlinear Dynamics and Econometrics, Vol. 5, pp. 203-12. Cochrane, J.H. (2001), Asset Pricing, Princeton University Press, Princeton, NJ. Darby, M.R. (1982), ‘‘The price of oil and world inflation and recession’’, American Economic Review, Vol. 72, pp. 738-51. Dickey, D. and Fuller, W.A. (1981), ‘‘Likelihood ratio statistics for autoregressive time series with a unit root’’, Econometrica, Vol. 49, pp. 1057-72. Gjerde, O. and Saettem, F. (1999), ‘‘Causal relationships among stock returns and macroeconomic variables in small, open economy’’, Journal of International Financial Markets, Vol. 9, pp. 61-74. Hamilton, J.D. (1996), ‘‘This is what happened to the oil price–macroeconomy relationship’’, Journal of Monetary Economics, Vol. 38, pp. 215-20. Hamilton, J.D. (2000), ‘‘What is an oil shock’’, working paper, No. W7755, MBER. Hammes, D. and Wills, D. (2003), ‘‘Black gold: the end of Bretton Woods and the oil price shocks of the 1970s’’, working paper, University of Hawaii Hilo. Hammoudeh, S. and Eleisa, L. (2004), ‘‘Dynamic relationships among GCC stock markets and NYMEX oil futures’’, Contemporary Economic Policy, Vol. 22, pp. 250-69. Hickman, B., Huntington, H. and Sweeney, J. (1987), Macroeconomic Impacts of Energy Shocks, North-Holland, Amsterdam. Hiemstra, C. and Kramer, C. (1997), ‘‘Nonlinearity and endogeneity in macro-asset pricing’’, Studies in Nonlinear Dynamics and Econometrics, Vol. 2, pp. 61-76. Hooker, M. (1999), ‘‘Are oil shocks inflationary? Asymmetric and nonlinear specifications versus changes in regime’’, working paper, Federal Reserve Board of Governors. Huang, R.D., Masulis, R.W. and Stoll, H.R. (1996), ‘‘Energy shocks and financial markets’’, The Journal of Future Markets, Vol. 16, pp. 1-25. Jarque, C.M. and Bera, A.K. (1987), ‘‘A test for normality of observations and regression residuals’’, International Statistical Review, Vol. 55, pp. 163-72. Johansen, S. (1988), ‘‘Statistical analysis of cointegration vectors’’, Journal of Economic Dynamics and Control, Vol. 12, pp. 231-54. Johansen, S. (1991), ‘‘Estimation and hypothesis of cointegration vectors in Gaussian vector autoregressive models’’, Econometrica, Vol. 59, pp. 1551-81. Johansen, S. (1995), Likelihood Based Inference in Cointegrated Vector Autoregressive Models Oxford University Press, Oxford. Johansen, S. and Juselius, K. (1990), ‘‘Maximum likelihood estimation and inference on cointegration with applications to the demand for money’’, Oxford Bulletin of Economics and Statistics, Vol. 52, pp. 169-210. Jones, C.M. and Kaul, G. (1992), ‘‘Oil and stock markets’’, Journal of Finance, Vol. 51, pp. 463-91. Jones, C. M. and Kaul, G. (1996), ‘‘Oil and stock markets’’, Journal of Finance, Vol. 51, pp. 463-91. Jones, D.W. and Leiby, P. (1996), ‘‘The macroeconomic impacts of oil price shocks: a review of the literature and issues’’, working paper, Oak Ridge National Laboratory, Oak Ridge, TN. Leigh, A., Wolfers, J. and Zitzewitz, E. (2003), ‘‘What do financial markets think about the war of Iraq?’’ working paper, Stanford Graduate School of Business, Stanford, CA. Maghyereh, A. (2004), ‘‘Oil price shocks and emerging stock markets: a generalized VAR approach’’, International Journal of Applied Econometrics and Quantitative Studies, Vol. 1 No. 2, pp. 27-40. Mork, K.A. (1989), ‘‘Oil and macroeconomy when prices go up and down: an extension of Hamilton’s results’’, Journal of Political Economy, Vol. 91, pp. 740-4.

Oil prices and stock markets in GCC countries 459

MF 33,7

460

Mork, K.A., Olsen, O. and Mysen, H.T. (1994), ‘‘Macroeconomic responses to oil price increases and decreases in seven OECD countries’’, Energy Journal, Vol. 15, pp. 19-35. Papapetrou, E. (2001), ‘‘Oil price shocks, stock markets, economic activity and employment in Greece’’, Energy Economics, Vol. 23, pp. 511-32. Pierce, J.L. and Jared, J.E. (1974), ‘‘The effects of external inflationary shocks’’, Brooking Papers on Economic Activity, Vol. 1, pp. 13-61. Rasche, R.H. and Tatom, J.A. (1977), ‘‘The effect of the new energy regime on economic capacity, production and prices’’, Economic Review, Vol. 59, pp. 2-12. Sadorsky, P. (2004), ‘‘Stock markets and energy prices’’, Encyclopedia of Energy, Vol. 5, pp. 707-17. Stock, J.H. and Watson, M.W. (1993), ‘‘A simple estimator of cointegration vectors in high order integrated systems’’, Econometrica, Vol. 61, pp. 783-820. Further reading Sadorsky, P. (1999), ‘‘Oil price shocks and stock market activity’’, Energy Economics, Vol. 21, pp. 449-69. Corresponding author Aktham Maghyereh can be contacted at: [email protected]

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0307-4358.htm

Financial integration, regulation and competitiveness in Middle East and North Africa countries Anastassios Gentzoglanis

Financial integration, regulation 461

University of Sherbrooke CEREF (Centre for the Study of Regulatory Economics and Finance) Faculty of Business Administration, Department of Finance, Sherbrooke, Quebec, Canada Abstract Purpose – The purpose of this article is to examine is to the link between stock markets and economic growth in advanced and emerging economies in the Middle East and North Africa (mena) region. Design/methodology/approach – Indices measuring the degree of financial openness and market development are constructed and used to perform various Granger causality tests to identify predictors of current growth rates. Findings – It is found that the link exists only in the group of high income countries but this relationship is rather weak for the low income MENA economies. Privatization alone, although necessary, is not enough to spur economic growth. The establishment of sound institutions and welldefined regulatory policies are needed to protect investors’ rights and entice them to invest in real and financial assets in the MENA region. Originality/value – The paper offers insights into financial integration, regulation and competitiveness in MENA countries. Keywords Regulation, Privatization, Stock markets, Economic development, Middle East, North Africa Paper type Research paper

1. Introduction There is a growing theoretical and empirical literature dealing with the effects of privatization on a country’s financial system (Kogut and Spicer, 2001; Morgenstern, 1995). The growth of the latter is associated with the country’s economic development (Filer et al., 1999; Ross and Zervos, 1996; Klaus and Kugler, 1998). Regulatory reforms and corporate governance are deemed necessary to the creation of an investment environment that attracts foreign capital and local investment (Gentzoglanis, 2000, 2001). Middle East and North Africa (MENA) countries privatized a number of firms during the 1990s and in the past few years privatized large-scale infrastructure companies such as telecommunications, energy and transport. As a result of these privatizations, stock market capitalization has dramatically increased, and with it the financial bond markets. Nevertheless, stock markets in these countries are not particularly active and without any further reforms of the banking and financial system most of the stock exchanges will eventually decline and/or shut down altogether. This may affect their future economic growth. Apparently, countries with better-developed financial systems tend to grow faster than countries with poor and relatively less liquid stock exchanges and weak financial and banking sectors (Levine, 2003). Nevertheless, empirical studies (Levine and Zervos, 1998) have demonstrated that the mere size of the stock market is not enough to foster This paper was written while the author was visiting the University of Crete.

Managerial Finance Vol. 33 No. 7, 2007 pp. 461-476 # Emerald Group Publishing Limited 0307-4358 DOI 10.1108/03074350710753744

MF 33,7

462

economic growth. Simply listing privatized companies on the stock exchanges is not enough to spur further real economic activity. It is rather liquidity of stock exchanges that fosters resource allocation and growth. Developed economies such as the USA and France, have a turnover ratio (an index of stock exchange’s liquidity) of approximately 0.5 (or 50 per cent), while less developed economies such as MENA countries (Egypt, Morocco, Turkey, etc.) have a turnover ratio hardly equal to 0.06 (or 6 per cent). While the turnover ratio (domestic stock transactions on a country’s national stock exchanges) may be less important for large firms in developed economies, it is particularly relevant for MENA countries mostly dominated by SMEs which heavily rely on local markets for their financial needs for expansion (Kmeir, 2003). As economies become more financially integrated and large firms list and issue stocks on foreign exchanges, the development and the location of the stock market are important for the provision of liquidity. MENA countries cannot rely on foreign exchanges to satisfy the liquidity needs of their vulnerable industrial firms. A fundamental question then arises. If economic growth heavily depends on a country’s financial development, what are the policies for MENA countries to develop a well-functioning financial system? What are the legal, regulatory and policy changes that would foster the emergence of growth-enhancing financial markets and intermediaries in MENA region? This paper tackles these issues for a selected number of MENA countries by adopting an empirical approach. The remainder of the paper proceeds as follows. The next section discusses the theoretical underpinnings of the empirical relation that exists between financial development and economic growth. The third section develops the empirical model to be examined for a sample of MENA countries, the data and the econometric techniques to be used. The fourth section presents the results and reports on their quality for MENA countries. Lastly, the final section concludes and offers some policy recommendations. 2. Theoretical underpinnings of the financial effects of liberalization and the cost of capital for MENA countries 2.1 The theoretical model Theoretical work has demonstrated the link between a country’s financial liberalization and the cost of capital. When a MENA country liberalizes its equity market, its cost of capital will be reduced. Firms’ in MENA countries can thus get greater financing opportunities at better terms increasing thereby their international competitiveness. To demonstrate the link, we use Stulz’s (1999) partial equilibrium model. Let us suppose a small MENA economy whose equity market is isolated from the world equity markets (before liberalization of its local financial market). Suppose as well that investors care about the expected return and variance of their investment.
M Þ represent the equilibrium required rate of return on the aggregate Letting EðR domestic stock market before liberalization and rf the domestic risk-free interest rate. Following traditional mean–variance model formulation, the price of risk is the
M Þ
rf] divided by the variance VARðR
M Þ of the aggregate risk premium [EðR aggregate return on the market. The price of risk in a MENA country before liberalization is a constant T and:
M Þ ¼ rf þ T VARðR
MÞ EðR

ð1Þ

This required rate of return will change once the domestic financial market is liberalized and the small MENA economy allows its residents to invest abroad and foreigners are allowed to invest locally. Let us suppose that the mean and variance of domestic dividends remain the same after liberalization and the required rate of return
 Þ, while the required rate of return on the on the market after liberalization is EðR M
world equity market is EðRW Þ. With liberalized equity markets, the relevant risk-free interest rate is the world risk-free rate rf . With liberalized markets, two factors will determine the risk premium on the
W Þ
r  ] and (b) the beta of domestic stock markets: (a) the world risk premium [EðR f the domestic stock market with the world stock market (
MW Þ, i.e.
 Þ ¼ r  þ
MW ½EðR
W Þ
r EðR M f f

Financial integration, regulation 463

ð2Þ

Given the negligible size of the local stock market, the risk premium of the world’s portfolio is barely affected by the addition of the stock market of a MENA economy to
W Þ and after some
W Þ
r  ] ¼ TVARðR the world’s stock market. Then, [EðR f mathematical manipulations, we can establish that the required rate of return on the domestic stock market after liberalization is: 


M; R
WÞ
Þ ¼ r  þ TCOVðR EðR M f

ð3Þ

Subtracting equation (1) from equation (3) we get:
M; R
W Þ
VARðR
MÞ
M Þ ¼ ðr 
rf Þ þ TCOVðR EðR f

ð4Þ

Given the differences in capital endowments between rich and poor countries, it is reasonable to expect (rf > rf ), therefore, the first term of the equation is negative. The change in the risk premium is negative as well since the covariance of the local market
W Þ, is necessarily less than the variance of the local
M; R with the world market, COVðR
market, VARðRM Þ. The net result is a reduction in the cost of capital after financial liberalization. Thus, from a theoretical point of view, MENA countries can enjoy a reduction in the cost of capital and local firms can get better financing once their capital markets are liberalized. This is the main thrust behind regulatory reforms and restructuring programmes. This relationship can be captured schematically (Figure 1).

Figure 1. The conceptual framework

MF 33,7

464

Privatization increases the deepening of the stock exchanges while sound financial regulation contributes to the increase of the efficiency of the whole financial system. This results in an efficient capital allocation, higher productivity and finally faster economic growth. Privatization and regulatory reforms create efficient domestic capital markets which can funnel savings to private sector investments (Bekaert et al., 2002) decreasing thereby the cost of borrowed capital given that there is abundance of it for equity issues. It is then expected that, debt finance will decrease. As a result, operational efficiency, competitiveness and solvency are bound to improve. The issue whether the financial activity has any impact on real economic variables is heavily debated among economic theorists and practitioners (Lucas, 1988; Miller, 1999). Recent empirical research tries to resolve this quadroon by employing a number of econometric techniques and financial data for developed and developing countries (Levine, 2001, 2002, 2003; Gentzoglanis, 2003). In spite of the large number of empirical studies on this issue, there is no unanimity on this debate (Henry, 2003). 2.2 Financial markets and economic growth – some empirical facts Emerging and developed stock markets have experienced rapid growth rates during the last decade. Market capitalization in emerging countries has risen from $488 billion in 1988 to $2,439 billion by 1999, while the annual trading on their exchanges rose from $411 billion in 1988 to $2,439 billion by mid-1999 (IFC, 2002). It is alleged that important structural, technological and institutional changes have triggered the exceptional growth rates in stock markets and in economic development of both developed and emerging economies during this decade. It is important, however, to make a distinction in the way the above changes have affected differently both groups of countries. Emerging economies, especially the ones in MENA region, have departed from a different economic environment than the industrialized economies. Some MENA countries have adopted liberalization and privatization policies that dramatically affected the structure and performance of their whole economies. Financial liberalization is one policy aiming at promoting market development, which further results in a reduction of the cost of capital for large and small firms (Henry, 2000b; Bekaert and Harvey, 2000). Liberalization policy, although necessary, it is not sufficient to increase growth. Other policies are necessary to achieve this goal. For instance, building institutions and developing the necessary enforcement mechanisms such as competition authorities and regulatory agencies are deemed necessary for confidence building and the creation of markets that will attract capital to assure sustained levels of high economic growth and the development of capital markets. There is some evidence that privatization and financial liberalization result in a gradual financial integration because higher returns of local capital markets attract new investments that enhance economic growth. Henry (2000b), reports that the excess returns on the announcement of liberalization policies are around 6 per cent in the first month and 26 per cent in the following eight months but for some countries these excess returns last longer. Nevertheless, some countries fail to attract investment despite the privatization policies and the liberalization of their financial markets. It is hypothesized in the theoretical and empirical literature that the failure of privatizations to create sustained levels of high economic growth is due to lack of strong institutional and regulatory mechanisms that protect investors. Market reforms cannot create viable financial markets and economic growth. Only the creation of institutions and regulations can.

It is true that the main goal behind the reforms is to spur economic growth. Nevertheless, there is an enduring theoretical debate whether financial developments are the cause of economic growth or the consequence of increased economic activity. There is a scant empirical literature establishing the link between financial development and growth in general and stock markets in particular. These studies regress average growth rates in per capita output across countries on a set of variables controlling for initial conditions and country characteristics as well as measures of financial development (Levine and Zervos, 1996, 1998; Harris, 1997 and King and Levine, 1993a, b). These studies are beset with the notorious problems of causality and unmeasured cross-country heterogeneity in factors such as savings rates and financial sector differences (Caselli et al., 1996). Admittedly, industrialized countries have more efficient capital markets than MENA and other emerging economies. In an efficient equity market stocks are forward-looking in a sense that current prices reflect the present discounted value of future profits. Future growth rates are thus reflected in initial stock prices. The empirical examination of the link between stock prices and market development should take into account this causality problem. The development of indicators of market development that are independent of stock prices is indicative in this case. Some recent empirical studies (Rajan and Zingales, 1998) tried to circumvent the causality problem by using a number of alternative indicators such as the turnover velocity (the ratio of turnover to market capitalization) and financial deepening (the percentage increase in the number of listed companies). Their findings are consistent with the hypothesis that active capital markets are important engines of growth in emerging economies and the creation and fostering of such markets has an important impact on long-term growth and development. This article hypothesizes that privatization increases capital deepening which may further lead to economic growth. Privatization alone is not enough though to guarantee such a development. Efficient regulations and the establishment of a playing level field for domestic and foreign business are necessary for the development of active capital markets. Some MENA countries have gone to a process of massive privatizations which contributed to the deepening of their capital markets significantly. Privatization opened up the market to competition and attracted foreign direct investment in many industrial sectors in general and in the finance and in the banking industries in particular. Yet the performance of these countries varies significantly despite their privatization policies. We argue that the differences in performance is mainly due to the different regulatory regimes adopted by each country for their banking and financial markets. An empirical comparison of the difference in performance is done for two groups of countries, industrialized and MENA economies. Although differences may exist among countries in the same group, the classification is made according to World Bank’s 2001 classification index. It is then possible to have low performers and high performers within the same group of countries. 3. Capital markets and regulatory reforms in MENA countries MENA countries have initiated their restructuring and privatization policies since the early 1990s. After the timid policies of privatization of some of their manufacturing industries, they went ahead with the privatization of the financial sector and the creation of capital markets borrowing a lot from the European and the USA models. Privatization of major infrastructure industries, such as telecommunications, electricity, gas, water and transport, is a more recent phenomenon.

Financial integration, regulation 465

MF 33,7

466

Table I compares the performance of some countries in terms of GDP growth in the past few years. MENA countries as a group has managed to realize some important gains in growth but other regions and countries have managed better results at the same period. Not all countries in MENA have had equal success. Jordan is probably the single MENA country in the group that has managed to do better than the others. Economic growth can be sustained not only by reforming capital markets but by also restructuring the real economy, by promoting SMEs and by attracting more FDI. Tunisia, Egypt and Morocco are lagging behind in restructuring of the real economy and, together with Turkey, have very low levels of FDI despite the almost finalized privatization of their banking sector. As a result, large banks in these countries have now excess liquidity and difficulties to lend to creditworthy domestic customers. There is a credit crunch which seriously inhibits economic growth. The importance of institutional mechanisms of government regulation and trust in the creation of well functioning markets is becoming evident from the experiences of the two largest economies, Egypt and Turkey. Not only have these countries failed to experience any respectable positive growth, they have even lost roughly half of their pre-restructuring GDP. They have weak prudent governance with limited competition and regulation and supervision, although formally advanced, remain poorly enforced. MENA countries have gone through the process of modernization of their stock markets and the establishment of regulatory institutions. Some of the reformers, mainly Jordan, managed to increase the depth of their financial markets and reduce the cost of equity at competitive levels, enough to attract international investments. Table II compares the financial market openness of selected MENA countries. Egypt The liberalization and deregulation programmes greatly managed to reform Egypt’s financial sector especially after the 1994 privatization of major banks and their listing to Cairo’s Stock Exchange. Despite these reforms, the financial market is still Percentage growth

Table I. GDP growth rates for regions and countries (1999-2003)

World production Industrialized countries USA Japan European union Germany UK France Italy Emerging countries Emerging Asia Latin America Central and East Europe Africa Middle East (including Turkey) World trade

1999

2000

2001

2002

2003 (forecast)

3.6 3.4 4.1 0.7 2.6 1.8 2.1 3 1.6 3.9 6.1 0.2 3.6 2.6 1.2 5.3

4.7 3.8 4.1 2.2 3.4 3 2.9 3.6 2.9 5.7 6.7 4 6.6 3 6.1 12.6

2 0.7 0.3
0.1 1.5 0.6 1.9 1.8 1.8 3.6 5.2 0.1 4.5 2.4 1.5
0.1

2.4 1.4 2.3
0.4 1.0 0.5 1.6 1.3 0.8 3.8 5.6
1.4 3.2 3 3 2.1

3.5 2.5 3 1 2 1.9 2.9 2.5 2.3 4.8 6 2.8 3.7 3.1 3.5 6.1

Source: www.trading-safely.com/sitecwp/ceen.nsf/vwNL/E8B9CB5DB1B811EAC1256C4E004CB698

dominated by four biggest state banks[1] which control more than half of market activity. In 2001, the privatization of the financial sector was hit by a serious setback when the government announced a banning for joint-venture banks to hold shares in the four biggest banks. Four biggest state-owned insurance companies also dominate the insurance sector[2]. They control much of the market. The 1998 law allows a 100 per cent foreign ownership of insurance companies. Foreign operations are allowed in Egypt’s stock exchange but share performance has been very weak especially after the 11 September attacks and the world economic slowdown. Foreign brokers’ transactions fell to a mere 6 per cent in March 2002 from a total 25 per cent. The new Cairo and Alexandria Stock Exchange (CASE) is very narrowly based with just only 100 truly active stocks of the 1,129 listed. The Hermes share index is at its lowest level for eight years.

Financial integration, regulation 467

Morocco Despite the privatization programme, Morocco’s stock exchange performs poorly. Such a weak performance is attributed to a number of factors; the government’s failure to float some capital of privatized companies on the national stock exchange; the lack of investor interest in emerging markets; the poor results of listed companies; Morocco’s dependence on traditional bank funding. The launch of new indexes[3] in January 2002, failed to increase interest in stocks and the stock market failed to revive[4]. In 2001, and despite the government efforts to offer tax incentives, the Casablanca stock exchange remained sluggish. The main share index (IGB) fell 7.4 per cent, after a 15.3 per cent decline in 2000, with 31 per cent fewer transactions (60 per cent fewer in 2000). Morocco’s determination to boost growth through structural reforms is reflected in its efforts to increase transparency in public services by simplifying procedures, obliging government bodies to give account of their decisions, establishing an ombudsman’s office, Al Wassit (the mediator) and abolishing the privileges of top government officials. Despite the progress, a lot remains to be done[5].

Country

Liberalization date

1st ADR introduction

Description of liberalization of the capital markets in MENA

Egypt

1992

November 1996

Jordan Morocco

December 1995 June 1988

December 1997 April 1996

Tunisia

June 1995

Turkey

August 1989

February 1998 July 1990

No restrictions on foreign investment in the stock exchange. Foreign investors have full access to Egypt’s capital markets (Capital Market Law 95) Foreign investors are allowed to buy stocks without government approval Foreign investors were allowed to subscribe to two Treasury bond issues (June 1988); unrestricted repatriation of capital and income from investment Partial privatization of inward portfolio investments Foreign investors can purchase listed securities without restrictions and pay no taxes on capital gains as long as they are register with the Capital Market Board and the Treasury

Source: www.worldbank.org/data

Table II. Openness of equity markets in selected MENA countries

MF 33,7

468

Tunisia The structural adjustment programme dates back to January 1987 but progress has been slow and the reforms in the financial system are still going on. The banking sector is dominated by many players[6] and given the small number of savers and investors in the country, business opportunities and banks’ profitability is low. However, the new strategy of many secondary development and commercial banks to find market niches and increase their specialization brought some interesting results increasing thereby their holdings of all financial assets in 63.6 per cent in 2000 (in contrast, the development banks held less than 4 per cent in 2000). The commercial banking industry is highly concentrated and the two main state-owned commercial banks (the Banque Nationale Agricole and the Socie´te´ Tunisienne de Banque) control more than 40 per cent of all deposits. The Banque Tunisienne de Solidarite´ is the most recent arrival in the sector specializing in micro-credit. The insurance sector was also reformed by a number of laws in 1992, 1994 and 1997. These reforms aimed at encouraging households to increase their savings rate but these non-bank financial institutions did not grow much and currently they do not play a big part in the country’s financial system. The stock exchange activity is very limited and approximately 90 per cent of issues on the primary market are government stocks and debentures of large firms. Lack of interest of foreign investors, the general lack of local interest in financial products and the predominance of small family run firms in Tunisia are some factors that may explain its relatively weak market activity. Table III presents the capital market characteristics for MENA countries. Jordan Among emerging stock markets, Jordan’s is one of the most open to foreign investors[7] and the most sophisticated among the Arab countries. The Amman Financial Market (AFM) is the fastest growing stock market in the region with a market capitalization close to US $5 billion, representing a market capitalization to GDP close to 77 per cent. The growth rate of market capitalization has been rising by 158 per cent over the last five years. The banking and finance sector leads the market with 54.4 per cent of total market capitalization while the industrial sector is second with 33 per cent of capitalization, with the service and insurance sectors representing 10.3 per cent and 1.5 per cent, respectively. The performance of the stock market is greatly attributed to bald measures taken by the government in June 1997. Jordan introduced a modern Securities Law by which it separated the regulatory function from the technical side of the market. It created a regulatory body, the Jordan Securities and Exchange Commission (JSEC) to organize,

Capital market characteristics

Table III. Capital markets development in selected MENA countries in 2001

Developing countries

Market capitalization of listed companies (% of GDP) Stocks traded, total value (% of GDP) S&P/IFC investable index (annual % change) Source: www.worldbank.org/data/

Egypt

Jordan

38

28.9

58.4

33.5

11.2
45.6

–

Morocco

Tunisia

Turkey

32.7

14.5

35

4.9

3.3

3.2


24.5


19.1

9

89.9
51.2

develop and monitor the securities market according to internationally accepted and proven standards. This led to an increase in investors’ confidence as well as in stocks’ activity. The maintenance of a transparent flow of information among market institutions, participants and investors and the creation of sophisticated, professional and efficient organizational and administrative functions of market institutions helped to boost Jordan’s financial sector in the region. To further enhance the competitiveness of Jordan’s financial sector, the government created four other entities. The Jordan Stock Exchange (JSE), the Jordan Stock Depository (JSD), an institute to provide proper training concerning dealing in securities and an association to represent private sector participants in the securities industry in their dealings with the JSEC[8]. 4. Variables, data and empirical methodology 4.1 Data and variables The data set used in this study consists of monthly time series observations spanning from 1996 to 2002. The data on stock markets and information on economic and institutional variables were taken from various sources (International Finance Corporation, International Financial Statistics (IMF), National Stock Exchanges, Federation of European Stock Exchanges and World Federation of Exchanges. The Fraser Institute. Stock market development is measured by three variables: (1) market capitalization over GDP, (2) the variation of listed domestic shares and (3) turnover velocity. We employ indicators of market development that are not dependent on stock prices. Indicators such as turnover velocity (the ratio of turnover to market capitalization), and financial deepening, i.e. the annual percentage increase in the number of listed companies, are better indicators than market capitalization over GDP for these purposes. Following Morck et al. (2000) and Durnev et al., (2001), we develop synchronicity indices which measure the correlation between individual share return variation with market return variation. Synchronicity indices are constructed for each of five MENA economies (Egypt, Jordan, Morocco, Tunisia and Turkey) and three developed economies (Spain, Italy, Sweden, Canada and France). Financial markets are less vulnerable to ‘‘policy capture’’ when they are based on confidence and trust. Regulations that safeguard the interests of investors favour informed risk arbitrage in stocks and lead to asynchronous stock price movements. Synchronicity, a characteristic of less efficient capital markets, may be attributed to factors such as, the low degree of regulation of financial markets, the thinness of stock exchanges (the number of firms listed on the exchange) and institutional environment that poorly protects private property and the rights of individual and institutional investors. Synchronicity is associated with market inefficiencies and low rates of economic growth. Given that the impact of stock market development on growth is likely to vary across levels of development, we divide our sample of countries into two groups according to per capita GDP (World Bank’s 2001 classification). Thus, in the high income group there are the three industrialized countries, Italy, Spain, Canada, France and Sweden; while in the low-middle income class are the five MENA countries. Based on the indices on Economic freedom developed by the Fraser Institute and the Heritage Foundation/Wall Street Journal (2003), we calculate an index of financial market openness appropriate for examining the impact of regulation and privatization policies on stock market development and growth. The latter takes into account an array of institutional factors determining economic freedom and assigns a score to each of its ten categories (trade, fiscal burden, government intervention, monetary policy, foreign investment, banking and finance, wages/prices, regulation and black

Financial integration, regulation 469

MF 33,7

470

market). Although we recognize that all these categories are important and have an impact on a country’s classification as more open or less open, we only take into account the three sub-categories (regulation, foreign investment and banking) to construct an index of a country’s openness of its financial markets. Thus, countries with an average of below 2.5 are classified as having a more open financial market while the ones with an average between 2.5 and 3.5 as medium financial openness and those above 3.5 are classified as having a less open financial market[9]. Under these criteria, Jordan falls into the second category, but for the purposes of our analysis, it is classified into the low-middle income class of countries. Table IV presents the sample statistics for the key variables for the full sample and the subgroups according to per capita GDP and financial market openness. Over the time period under examination, countries with high financial openness grew less than the ones with low financial openness. Furthermore, high income countries have a higher turnover to market capitalization ratio and market capitalization over GDP ratio. 4.2 Empirical methodology To test for the direction of causality (stock market development and economic growth) we apply Granger’s (1969) prototype causality model where the existence of causal ordering in Granger’s sense implies predictability and exogeneity. The following causality analysis is used to detect the direction of information flow among the variables. Suppose X Granger-causes Y but Y does not Granger-cause X, then past values of X should be able to help predict future values of Y, but past values of Y should not be helpful in forecasting X. More specifically, let Y be the economic growth variable and X an indicator of stock market development, while the subscripts t and t
i be the current and lagged values. Yt ¼  0 þ

k1 X

i þ Yt
i þ

i¼1

k2 X


i Xt
i þ "t

ð5Þ

i Xt
i þ t

ð6Þ

i¼1

and Xt ¼ 0 þ

k3 X

i þ Yt
i þ

i¼1

k4 X i¼1

where and k is the order of the lag for X and Y[10] and "t is white noise vector (k  1) of random shocks which arePindependently, identically and normally distributed with mean zero and covariance . The null hypothesis that X does not Granger-cause Y is that the coefficients
i ¼ 0 for i ¼ 1, 2, . . . , k, which can be determined by a standard F-test.

Statistics

Table IV. Sample characteristics

All countries High income Low-middle income High financial openness Low financial openness

Mean Mean Mean Mean Mean

SE SE SE SE SE

GDP growth 2.26 2.44 1.65 2.05 3.49

(2.45) (1.94) (1.61) (3.6) (3.51)

Mkt. Turnover/Mkt. capitalization/ GDP capitalization 45.68 67.45 31.24 54.64 24.53

(57.12) (65.75) (30.67) (59.09) (21.12)

0.49 0.95 0.31 0.39 0.24

(0.37) (0.74) (0.39) (0.26) (0.16)

Change in No. of companies
53.14
24.79
113.37
28.67 7.99

(246.21) (132.67) (176.35) (254.19) (24.36)

In this paper, the causality terminology is used to indicate the direction of the causal relation among variables. For example, if market capitalization is found to be Granger causing economic growth then it is interpreted as the ability of the market capitalization to contain information on the future course of that variable. In this analysis, we do not claim that Granger type causality should be necessarily interpreted as evidence for a structural causality from the stock market variables to the economic growth. 5. Results Table V reports the results of the empirical estimations. As it might be expected, lagged growth rates are, in general, significant predictors of current growth rates. This is particularly true for high income economies but for low-middle income economies this effect is rather weak. This may be explained by the fact that the low-middle income economies in our sample have experienced unstable macroeconomic conditions and uncertain political outcomes. These countries have difficulties in attracting foreign and local investors alike. As far as the link between financial variables and growth is concerned, there is a positive relationship between the ratio of market capitalization/GDP and future economic growth. This is what it is expected to happen especially in high income countries which have more efficient capital markets and stocks incorporate future earnings into current prices and therefore market capitalization. The link between turnover velocity and future economic growth is present for highincome countries but not for low-middle income countries. Thus, a higher turnover velocity Granger-causes growth but the location of the effect is not the same for low and high income economies. For the latter category, the effect is within countries while for the first group this effect is between countries. This is particularly important for this group of countries suggesting that a more active stock market is associated with growth rates far more significant compared to the presence of less active stock markets. Low income countries can thus increase their growth rates by creating active stock markets. The latter is more likely to happen when these countries abide to strong regulatory and institutional frameworks and judicial systems. It is thus important that the institutional changes initiated by these countries take root and continue to work towards the adoption of more market oriented mechanisms and adopt policies favouring regional and international integration of their markets. It is time to turn to the link between market deepening as it is measured by the number of listed domestic companies on the stock exchange and future economic growth. As Table V reports, there is no evidence for such a link. If privatizations increase the number of listed firms on stock exchanges, the latter may not have an effect on growth. As it was mentioned earlier, privatization is a necessary but not a sufficient condition to spur economic growth. Institutional and sound regulatory changes are some other important factors to entice investors to actively participate in real and capital markets. Table VI reports the results when the variables representing the openness of financial markets are taken into account. Active stock markets in countries with less open financial markets, do not contribute to economic growth. On the contrary, active stock markets in these countries are used for entrenchment purposes given that corruption is quite prevalent. Stock markets can have a positive effect on growth when they function normally and efficiently. The latter can be safeguarded when these countries can enact sound regulations and institutions. These results are confirmed by

Financial integration, regulation 471

Table V. Tests of Granger causality running from financial variables to growth (countries classified by income)

Total Between Within

Total Between Within

High income

Low-middle income

0.759* (0.066) 0.892 (0.692) 0.777*** (0.088)

0.573* (0.058) 1.218* (0.322) 0.778* (0.091)

0.738* (0.035) 0.869* (0.258) 0.163* (0.064)

0.009 (0.045)
0.009 (0.043) 0.054 (0.012)

0.008* (0.003)
0.009 (0.652) 0.011* (0.005)

0.004** (0.001) 0.009 (0.002) 0.008* (0.002)

0.564* (0.059)
0.355 (0.542) 0.554*** (0.073)

0.857* (0.057) 1.337* (0.39) 0.649* (0.056)

0.963* (0.014) 0.977* (0.067) 0.75* (0.049)

4.716* (0.686) 6.936* (1.576)
0.987 (1.943)

0.825** (0.297)
0.767 (0.651) 1.897* (0.482)

0.956* (0.489) 1.96* (0.362) 1.96** (0.658)

X ¼ Mkt. capitalization/GDP Lagged Y Lagged X

0.744* (0.067) 0.852* (0.0582) 0.7310* (0.0721)

0.588* (0.076) 0.947* (0.058) 0.934* (0.068)

0.597* (0.027) 0.786* (0.054) 0.584* (0.048)


0.0006* (0.0005)
0.0007 (0.009)
0.00005 (0.00001)


0.0008 (0.002) 0.009 (0.004) 0.0008 (0.004)


0.008** (0.0001)
0.003 (0.004) 0.00005 (0.00001)

X ¼ Change in No. of companies Lagged Y Lagged X

Notes: *Significant at the 1 per cent confidence level; **significant at the 5 per cent confidence level; ***significant at the 10 per cent confidence level

Total Between Within

All countries

X ¼ Turnover/Mkt. capitalization Lagged Y Lagged X

472

Group of countries

MF 33,7

0.621* (0.035) 0.893* (0.351) 0.1953* (0.059) 0.538* (0.061) 0.853* (0.069) 0.674* (0.087) 0.5497 (0.837) 1.854** (0.768) 0.0739 (0.867)

Total Between Within

Total Between Within

Total Between Within


0.097 (0.067)
0.032 (0.095) 0.054 (0.039)

0.007** (0.002) 0.010 (0.0005) 0.009* (0.003)

0.010** (0.005) 0.008 (0.005) 0.008* (0.002)

X ¼ Turnover/Mkt. capitalization Lagged Y Lagged X

0.486*** (0.096) 1.598** (0.642) 0.008 (0.341)

0.672* (0.097) 0.879* (0.007) 0.639* (0.069)

0.675* (0.041) 0.893* (0.074) 0.675* (0.072)

1.956 (2.976)
2.763 (2.964)
3.881*** (1.759)

0.874** (0.956) 1.793 (0.759) 1.446* (0.761)

1.246* (0.548) 1.891* (0.872) 1.932** (0.638)

X ¼ Mkt. capitalization/GDP Lagged Y Lagged X

0.673** (0.052) 0.879* (0.056)
0.074 (0.237)

0.697* (0.095) 0.926* (0.034) 0.656* (0.091)

0.875* (0.077) 0.979* (0.042) 0.679* (0.037)


0.0083 (0.074) 0.0058 (0.035) 0.0039 (0.004)


0.007 (0.0004)
0.006 (0.003)
0.008 (0.003)


0.007** (0.0002)
0.008 (0.003) 0.000002 (0.00001)

X ¼ Change in No. of companies Lagged Y Lagged X

Notes: *Significant at the 1 per cent confidence level; **significant at the 5 per cent confidence level; ***significant at the 10 per cent confidence level

Low openness

Highly opened financial markets

All countries

Group of countries

Financial integration, regulation 473

Tests of Granger causality running from financial openness to growth (countries classified by financial openness)

Table VI.

MF 33,7

474

a similar study (Filler et al., 1999) realized to examine whether stock markets promote economic growth in emerging and advanced economies but do not agree with the results found in a study by Harris (1997) which finds that the link between stock market and growth exist only for advanced economies. 6. Conclusions This article examines the link between stock markets and economic growth in advanced and emerging economies in the MENA region. Indices measuring the degree of financial openness and market development are constructed and used to perform various Granger causality tests to identify predictors of current growth rates. It is found that the link exists only in the group of high income countries but this relationship is rather weak for the low income economies. Privatization alone, although necessary, is not enough to spur economic growth. The establishment of sound institutions and well-defined regulatory policies are needed to protect investors’ rights and entice them to invest in real and financial assets in the MENA region. It is possible that in countries with lower openness of their financial system, active stock markets may even thwart growth. Although the sample examined is rather small, the results are interesting and informative. The conclusions are of great interest to policy makers and civil servants. Capital markets must be assorted with institutions, sound regulations and enforcement mechanisms to provide confidence and trust to individual and institutional investors. Privatization alone, although increases the number of listed firms on the stock exchange, is not enough to provide incentives in investing in financial and real, productive assets of the economy. Economic growth is a more complex phenomenon to be explained with simple empirical models and more data and better empirical techniques are needed in order to be able to shed more light on this question. Nevertheless, this study is in the right direction. Notes 1. MISR, the National Bank of Egypt (NBE), the Bank of Cairo and the Bank of Alexandria. 2. There are 11 in total. 3. Two new indexes were launched in January 2002; the Moroccan All Share Index (MASI) of 55 quoted companies and the Moroccan Most Active Share Index (MADEX). 4. The new indexes, MASI and MADEX fell by 15 per cent and 24 per cent respectively, between January 2002 and the end of September. In the first half of 2002, trading volume fell 33 per cent (year-on-year basis). 5. On its first national conference on administrative reform entitled The Moroccan public sector and the challenges of 2010 in May 2002, officials stressed the degree of corruption and disorganization that still plagues public bodies. 6. There are 14 deposit (or commercial) banks, six development (or investment) banks, eight offshore and two merchant banks and about 20 finance companies. 7. About 39 per cent of market value is owned by non-Jordanians while the government of Jordan, through the Jordan Investment Corporation ( JIC), owns approximately 18 per cent of total market capitalization. 8. A private sector depository and stock exchange. 9. As it must be expected, industrialized countries fall in the first category while the MENA countries into the second category.

10. The optimal length of the lag structure is tested by applying various information criteria (Akaike, 1969; Hannan and Quinn, 1979) as suggested by Hsiao (1981), but we found that more than one lag in either Y or X was never optimal. References Akaike, H. (1969), ‘‘Fitting autoregressive models for predictions’’, Annals of Mathematical Statistics, Vol. 40, pp. 243-47. Bekaert, G. and Harvey, C. (2000), ‘‘Foreign speculators and emerging equity markets’’, Journal of Finance, Vol. 55 No. 2, pp. 565-613. Bekaert, G., Harvey, C. and Lundblad, C. (2002), ‘‘Does financial liberalization spur economic growth?’’, working paper No. 8245, NBER. Caselli, F., Esquivel, G. and Lefort, F. (1996), ‘‘Reopening the convergence debate: a new look at cross-country growth empirics’’, Journal of Economic Growth, Vol. 1, pp. 363-89. Durnev, A., Morck, R. and Yeung, B. (2001), ‘‘Value enhancing capital budgeting and firm-specific stock returns variation’’, paper presented at the NBER Behavioral Finance Conference, 10 November. Gentzoglanis, A. (2001), ‘‘Privatization and regulatory regimes in telecommunications and energy sectors in CEEC’’, working paper. Gentzoglanis, A. (2003), ‘‘Capital market reforms and financial integration of CEEC’’, draft paper. Granger, C.J. (1969), ‘‘Investigating causal relationships by econometrics models and cross spectral methods’’, Econometrica, Vol. 37, pp. 425-35. Hannan, E. and Quinn, B. (1979), ‘‘The determination of the order of an autoregression’’, Journal of Royal Statistical Society, Vol. 41, pp. 190-5. Harris, R. (1997), ‘‘Stock markets and development: a re-assessment’’, European Economic Review, Vol. 41, pp. 139-46. Henry, P.B. (2000b), ‘‘Do stock market liberalizations cause investment booms?’’, Journal of Financial Economics, Vol. 58 No. 1-2, pp. 301-34. Henry, P.B. (2003), ‘‘Commentary on equity market liberalization in emerging markets’’, Federal Reserve Bank of St. Louis Review, pp. 75-80. Hsiao, C. (1981), ‘‘Autoregressive modeling and money-income causality’’, Journal of Monetary Economics., Vol. 7, pp. 85-106. King, R.G. and Levine, R. (1993a), ‘‘Finance and growth: schumpeter might be right’’, Quarterly Journal of Economics, Vol. 108, pp. 715-35. King, R.G. and Levine, R. (1993b), ‘‘Finance, entrepreneurship and growth’’, Journal of Monetary Economics, Vol. 32, pp. 513-42. Levine, R. (2001), ‘‘International financial liberalization and economic growth’’, Review of International Economics, Vol. 9 No. 4, pp. 688-704. Levine, R. (2002), ‘‘Finance and growth’’, unpublished manuscript, University of Minnesota. Levine, R. (2003), ‘‘More on finance and growth: more finance, more growth?’’, Federal Reserve Bank of St. Louis Review, pp. 31-46. Levine, R. and Zervos, S. (1996), ‘‘Stock market development and long-run growth’’, World Bank Economic Review, Vol. 10 No. 2, pp. 323-39. Levine, R. and Zervos, S. (1998), ‘‘Capital control liberalization and stock market development’’, World Development, Vol. 26 No. 7, pp. 1169-83. Lucas, R.E., Jr. (1988), ‘‘On the mechanics of economic development’’, Journal of Monetary Economics, Vol. 22 No. 1, pp. 3-42.

Financial integration, regulation 475

MF 33,7

476

Miller, D. (1999), ‘‘The market reaction to international cross-listings: evidence from depository receipts’’, Journal of Financial Economics, pp. 103-23. Morck, R., Yeung, B. and Yu, W. (2000), ‘‘The information content of stock markets: why do emerging markets have synchronous stock price movements?’’, Journal of Financial Economics, Vol. 59 No. 1 and 2, October. Morgenstern, C. (1995), ‘‘Capital market development and financing russia’s transformation’’, in Lieberman, I. and Nellis, J. (Eds), Russia: Creating Private Enterprise and Efficient Markets, The World Bank, Washington, DC. Rajan, R.G. and Zingales, L. (1998), ‘‘Financial dependence and growth’’, American Economic Review, Vol. 88, pp. 559-86. Stulz, R.M. (1999), ‘‘Globalization and the cost of equity capital’’, working paper No. 99-02, The New York Stock Exchange, New York, NY. Further reading Bekaert, G., Harvey, C. and Lundblad, C. (2000), ‘‘Emerging equity markets and economic development’’, Journal of Development Economics, Vol. 66, pp. 465-504. Gentzoglanis, A. (2002), ‘‘Privatization and investment in telecommunications industry’’, background paper, ERF. Granger, C.W.J. (1986), ‘‘Developments in the study of co-integrated economic variables’’, Oxford Bulletin of Economics and Statistics, Vol. 48, pp. 213-28. Granger, C.W.J. and Newbold, P. (1974), ‘‘Spurious regressions in econometrics’’, Journal of Econometrics, pp. 111-20. Harrinson, A., Love, I. and McMillan, M. (2001), ‘‘Global capital flows and financing constraints’’, working paper No. 2782, World Bank. Henry, P.B. (2000a), ‘‘Stock market liberalization, economic reform, and emerging market equity prices’’, Journal of Finance, Vol. 55 No. 2, pp. 529-64. International Finance Corporation (various years), Emerging Stock Markets Factbook, International Finance Corporation, Washington, DC. International Monetary Fund (various years), International Financial Statistics, International Monetary Fund, Washington, DC. Levine, R. (1997), ‘‘Financial development and economic growth: views and agenda’’, Journal of Economic Literature, Vol. 35, pp. 688-726. Rajan, R. and Zingales, L. (2002), ‘‘The great reversals: the politics of financial development in the 20th century’’, Journal of Financial Economics. United Nations, Statistical Division (various years), ‘‘Industrial statistics yearbook’’, General Statistics, Vol. I, United Nations, New York, NY United Nations Industrial Development Organization (UNIDO) (various years), International Yearbook of Industrial Statistics, UNIDO, Vienna. World Bank (various years), World Development Indicators, World Bank, Washington, DC. Corresponding author Anastassios Gentzoglanis can be contacted at: [email protected], anastassios. [email protected]

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0307-4358.htm

The determinants of stock market development in the Middle-Eastern and North African region

Stock market development

477

Samy Ben Naceur Faculty of Economics and Business, Universite´ Libre de Tunis, Tunis, Tunisia

Samir Ghazouani Institut Supe´rieur de Comptabilite´ & d’Administration des Entreprises (ISCAE), Campus Universitaire de Manouba, Manouba, Tunisia, and

Mohamed Omran College of Management and Technology, Arab Academy of Science and Technology, Alexandria, Egypt Abstract Purpose – The purpose of this study is to investigate the role of stock markets in economic growth and to shed some light on the macroeconomic determinants which must have an important influence on stock markets development. Design/methodology/approach – The empirical study is conducted using an unbalanced panel data from 12 Middle Eastern and North African (MENA) region countries. Econometric issues are based on estimation of some fixed and random effects specifications. Findings – It is found that saving rate, financial intermediary, stock market liquidity and the stabilization variable are the important determinants of stock market development. In addition, it is found that financial intermediaries and stock markets are complements rather than substitutes in the growth process. Practical implications – This paper has some policy implications to MENA region countries. In order to promote stock market development in the region, it is important to encourage savings by appropriate incentives, to improve stock market liquidity, to develop financial intermediaries and to control inflation. Originality/value – Since it is unclear whether emerging markets in the MENA region respond, similarly, to economic and political shocks like other emerging markets and/or developed markets. This paper fills this gap by making an in-depth analysis of 12 MENA capital markets in order to assess how they can improve their capital markets, and hence, benefit the global investor. Keywords Stock markets, Capital markets, Middle East, North Africa Paper type Research paper

1. Introduction Even though most Middle-Eastern and North African (MENA) countries have embarked on economic reform and structural adjustment programs, the Asian crisis of 1997 shifted the focus of such programs to financial markets. However, the lack of institutional development is a powerful obstacle to an increased access to MENA capital markets. Additionally, the region witnessed, and still, wars, political turmoil and economic instability. Consequently, MENA countries have not yet emerged as economic powers, which might explain the lack of academic research on MENA capital markets. In fact, it is unclear whether emerging markets in the MENA region respond,

Managerial Finance Vol. 33 No. 7, 2007 pp. 477-489 # Emerald Group Publishing Limited 0307-4358 DOI 10.1108/03074350710753753 .

MF 33,7

478

similarly, to economic and political shocks like other emerging markets and/or developed markets. Hence, the purpose of this paper is to fill this void in the literature and make an in-depth analysis of 12 MENA capital markets in order to assess how they can improve their capital markets, and hence, benefit the global investor. More precisely, the paper addresses the issue of the impact of macroeconomic variables on stock market development in 12 MENA markets. The measurement of stock market development is important because it is the precept for predicting economic growth and therefore, the principle for country selection by foreign investors. Thus, the results of this paper are expected to contribute to the paradigm of global asset allocation strategies in emerging markets. The paper is organized as follows. In section 2, the outlines which characterize the evolution of the financial systems in the MENA region countries, especially those considered in the empirical study, are presented. Section 3 Identifies organic relationships between stock market development and economic growth by means of some main macroeconomic determinants. Section 4 outlines the data and the adopted econometric methodology. The findings are reported in section 5 and the paper concludes with a summary and policy implications. 2. Evolution of financial systems in the MENA region Most of the Middle-Eastern countries have engaged in implementing economic reform and structural adjustment programs in recent decades. The core of the abovementioned programs is the reform in the financial sector that have enabled most of MENA countries to establish or resurrect their stock markets. As a result, stock exchanges in these countries and other emerging economies became very important to the world’s economy and their role in the international financial system witnessed a significant increase. As seen from Table I, Panel A, over the nine years between 1990 and 1999, the market capitalization of emerging stock markets rose from $604 billion to $3,074 billion. In the meantime, the value of trade represented by the emerging markets also increased significantly from $613 billion in 1990 to $2,867 billion in 1999, a more than fourfold Years 1990

1999

Panel A: Developed and emerging markets performance between 1990 and 1999 Stock market indicators Value Percentage Value DM EM DM EM DM EM Market capitalization (billion of US$) 8,795 604 93.5 6.5 32,956 3,074 Value traded (billion of US$) 4,406 613 88 12 28,154 2,867 Number of listed companies 16,504 8,920 65 35 23,326 26,314

Percentage DM EM 91.5 8.5 91.8 9.2 47 53

Panel B: Weight of Middle-Eastern markets in the S&P/IFCGCI between 1990 and 1999 Latin America 15.8% 21.3% Asia 73.3% 50.7% Europe 9.7% 14% Middle East and North Africa 1.2% 14% Table I. Some indicators for emerging markets over the period [1990-1999]

Notes: DM, Developed Markets; EM, Emerging Markets; S&P/IFCGCI, Standard and Poor and International Finance Corporation Global Composite Index, respectively Source: International Finance Corporation (2000)

increase in their level of activities. In addition, the emerging stock markets’ participation in the number of listed companies throughout the world jumped from only 35 per cent in 1990 to 53 per cent in 1999, suggesting that these markets are seen to grow much faster than developed markets. However, because of major changes in the economic and financial environment in the MENA countries, emerging equity markets in this region have been the focus of much attention recently from international investors. As a result, the weighting of these markets in the Standard and Poor/International Finance Corporation Global Composite Index surged substantially from below 2 per cent in 1990 to around 14 per cent in 1999 (Table I, Panel B). Nevertheless, the latest development in the stock market in MENA countries varies across countries as seen in Table II. Jordan, Kuwait and Qatar seem to outperform other countries as they achieved very strong progress between 1999 and 2002. On the other hand, Morocco, Turkey and Saudi Arabia come at the end of the list. On average, it is clear from the table that the whole sample countries achieved high level of performance between 1999 and 2002. The only negative indicator is the market capitalization that decreased by 21.6 per cent, but the main reason here is Turkey because of the huge depreciation in its currency[1].

1999

Bahrain Egypt Iran Jordan Kuwait Lebanon Morocco Oman Qatar Saudi Arabia Tunisia Turkey Total

2000

2001

% change 2002 [1999-2002]

1999

Market capitalization (billions of US$) 7.16 6.62 6.60 7.72 7.76 41 33.04 30.79 24.31 26.34
20.28 1,033 21.86 7.54 9.70 14.34
34.38 295 5.83 4.94 6.31 7.09 21.46 152 19.60 19.85 26.66 35.10 79.09 85 1.92 1.58 1.25 1.40
27.38 13 13.70 10.88 9.03 8.56
37.49 54 4.30 3.52 2.63 5.27 22.42 140 5.50 5.17 7.30 10.57 92.13 21 60.95 67.17 73.20 74.85 22.80 72 2.64 2.81 2.23 2.13
19.43 44 114.27 69.51 47.69 34.40
69.89 285 290.78 230.37 216.92 227.76
21.67 2,235

Value traded (billions of Bahrain 0.44 0.25 0.25 0.21 Egypt 9.73 11.80 5.91 6.44 Iran 2.27 1.08 1.10 2.21 Jordan 0.55 0.41 0.93 1.33 Kuwait 6.00 4.21 11.71 22.12 Lebanon 0.09 0.12 0.05 0.11 Morocco 2.52 1.21 0.84 1.44 Oman 0.71 0.55 0.42 0.58 Qatar 0.34 0.24 0.41 0.88 Saudi Arabia 15.09 17.31 22.22 30.97 Tunisia 0.46 0.69 0.34 0.25 Turkey 84.03 181.93 80.40 70.76 Total 122.24 219.79 124.60 137.31

US$)
53.59
33.75
2.82 142.96 268.66 26.93
42.95
18.57 159.73 105.31
46.10
15.80 12.33

6.21 29.44 10.39 9.41 30.62 4.71 18.43 16.60 6.18 24.75 17.33 73.54 42.04

2000

2001

2002

Stock market development

479

% change [1999-2002]

Number of listed companies 41 42 40 1,071 1,110 1,150 304 316 327 163 161 158 86 88 95 13 14 13 54 55 55 131 96 140 22 22 25 75 76 68 44 45 46 315 310 288 2,319 2,335 2,405


2.44 11.33 10.85 3.95 11.76 0.00 1.85 0.00 19.05
5.56 4.55 1.05 7.61

Turnover ratio 3.71 3.79 2.67 38.32 24.32 24.46 14.30 11.33 15.39 8.21 14.80 18.83 21.20 43.93 63.03 7.45 4.24 8.24 11.13 9.31 16.82 15.67 15.94 11.04 4.64 5.67 8.36 25.78 30.36 41.38 24.45 15.34 11.60 261.75 168.59 205.68 95.41 57.44 60.29


56.93
16.89 48.09 100.03 105.86 74.80
8.72
33.49 35.18 67.19
33.10 179.69 43.41

Table II. Selected indicators of stock market performance [1999-2002]

MF 33,7

480

On the other hand, most of our sample countries embarked on institutional setting and regulations such as establishing security market regulation, investors protections, trading rules based on shared regulatory responsibility, etc. Additionally, except for Gulf countries and Iran, foreign investors are allowed to participate in stock market activities with nearly no limitation[2]. Given the importance of these markets to the global financial system, this paper empirically investigates the impact of several macroeconomic indicators on stock market development. 3. Stock market development, economic growth and macroeconomic factors The financial system is essential to an economy because it is responsible for resource allocation. Well-working financial intermediaries may affect positively the economic development through four main channels which consist in reducing information and transactions costs, improving the allocation of resources (through fund pooling, risk diversification, liquidity management, screening and monitoring), increasing saving rates and promoting the development of markets and instruments that enable risk sharing organic relationships between stock market development and economic growth are identified[3]. Conversely, the financial crisis literature points to the destabilizing effect of financial liberalization as it leads to over-lending. Over-lending would appear through several channels, including a limited monitoring capacity of regulatory agencies, the inability of banks to discriminate good projects during investment booms, and the existence of an explicit or implicit insurance against banking failures (Aghion et al., 1999). Empirical investigations provide evidence of a positive relationship between finance and economic growth. The first evidence that financial sector development promotes growth was reported by Goldsmith (1969) in a paper covering 35 countries over the period (1860-1963). However, his work did neither control for initial conditions and country characteristics, nor did it permit any conclusions on causality or the relative strength of the transmission channels. Recent years have witnessed a vivid interest in empirical research on the relationship between financial sector development and economic growth. The relationship was found to be bi-directional, that is financial development is crucial for economic growth and economic growth can also promote financial development. The evidence deals with this causal relationship along three lines: .

Financial development accelerates economic growth or is conducive to growth slowdowns. The financial sector increases savings and allocates them to more productive investments. Thereby financial development can spur economic growth. For instance, recent findings in King and Levine (1993), Rousseau and Wachtel (1998), Levine and Zervos (1998), Levine et al. (2000), Beck et al. (2000a, b), and Rousseau and Sylla (2001). Conversely, the banking and currency crisis literature find that monetary aggregates, such as domestic credit, are among the best predictors for crises (e.g. Demirgu¨c¸-Kunt and Degatriache, 1998, 2000; Gourinchas et al., 1999). Since banking crises usually lead to recessions, an expansion of domestic credit would then be associated to growth slowdowns.

.

Economic growth promotes financial development. According to this hypothesis, financial development appears as a consequence of the overall economic

expansion. It has a passive role and adapts itself to the financing needs of the real sector (Gurley and Shaw, 1967; Goldsmith, 1969). .

The reciprocal relationship. Economic growth makes the development of financial system profitable and the establishment of an efficient financial sector contributes to stimulate economic development. Luintel and Khan (1999) reveal evidence for bi-directional causality from a sample of ten developing countries. Shan et al. (2001) confirm this finding from a sample of nine OECD countries.

Most of the evidence quoted above uses banks-based measures of financial development such as total lending by non-bank public per capita, bank credit to GDP (Shan et al., 2001) and broad money to GDP (Rousseau and Sylla, 2001). Banks dominate financing in many places and even in the most developed countries. Stock markets are only a small part of the overall financial system[4]. However, the existence of an equity market is important because it provides investors with an exit mechanism, it attracts foreign capital inflows, it provides important information that improves the efficiency of the financial system and it provides the valuation of companies. Not until recently has the focus increasingly shifted to stock market indicators, due to the increasing contribution of financial markets in economies. Ajte and Jonanovic (1993) show that trading volume (stock market development indicator) has a strong incidence on economic growth while bank credit does not. Similarly, Levine and Zervos (1996, 1998) and Singh (1997) find evidence on a positive relationship between stock market development and long-run economic growth. In addition, Levine and Zervos (1996) show that stock market liquidity is positively and robustly associated with long-run economic growth after controlling for economic and political variables. In the same vein, Rousseau and Wachtel (2000) use two measures of stock market development, that is the ratio of market capitalization to GDP and the ratio of total value traded to GDP. Both have a positive coefficient, but only the latter is significant. The results show that the development of a liquid and highly capitalized equity market accelerates growth. They also make an important contribution to the literature by using panel techniques with annual data. Finally, Garcia and Liu (1999) empirically explored the determinants of stock market development, particularly market capitalization. They also examined the association between financial intermediary development and stock market development using a sample of fifteen industrial and developing countries from 1980 to 1995. They concluded that real income level, saving rate, financial intermediary development, and stock market liquidity are important predictors of market capitalization, while macroeconomic stability does not have any explaining power. They confirmed that banks and markets are complement instead of substitutes. 4. Data and econometric modeling 4.1 Data sources Data were extracted from various sources. Arab Monetary Fund was a main source for data on Arab countries. We consult the capital market unit database to collect stock market indicators from 1994, and the economic and technical department database for other economic data series. As for the stock market data prior to 1994, we collect them from world development indicators and local stock markets. With regard to Iran and Turkey, world development indicators was the main source for both economic and stock market data.

Stock market development

481

MF 33,7

482

Our original intention was to cover all countries in the MENA region, but given that some countries have not yet established stock markets (for example, Iraq, Libya, Sudan, Syria and Yemen), and other countries established stock markets only in the past couple of years (United Arab Emirates), the sample countries included only 12 countries, in which ten countries are from the Arab world. Of course, data were not available for a uniform period for each country, and many countries have no stock market until recently. Consequently, it is expected that the number of observations varies across our sample countries leading to conduct estimations over an incomplete panel data. The number of time observations ranges from three annual observations for Qatar to 21 observations for Jordan. For the most other countries, the periods of observations cover mainly the 1980s and 1990s (see Appendix 1). This paper focuses on the determinants of stock market capitalization – defined as the total market value of all listed shares divided by GDP – as a proxy for financial market growth[5]. Besides, we use the following indicators as explanatory variables (Garcia and Liu, 1999): .

Income. We use real GDP in US dollars to measure the income level. As income increases, its cyclical component should have a positive incidence on the size of the stock market. In addition, higher income means better education, better business environment and wealthy citizens. We expect to have a positive impact on stock market development.

.

Saving rate. The saving rate is calculated as the ratio of gross saving to gross disposable income. Like banks, stock markets convey saving to investment projects. Usually, the larger the saving rate, the higher the flow of capital to stock markets. We expect a positive effect of the saving rate on the stock market size.

.

Investment rate. The investment rate is calculated as the ratio of gross fixed capital to gross disposable income. As investment rate depends on saving rate, we expect investment to be important determinants of stock market capitalization.

.

Credit to private sector. We use the domestic credit to the private sector divided by GDP to account for financial intermediary development. Since both banks and stock markets intermediate savings towards investment projects, they can be either complements or substitutes. Boyd and Smith (1996) suggest that banks and stock markets may behave as complements rather than as substitutes. Empirically, Demirgu¨c¸-Kunt and Levine (1996a) show that the degree of stock market development is positively related to bank development. Conversely, Garcia (1986) finds that central banks may generate a negative correlation between bank growth and stock market development.

.

M3. Another indicator for bank development is the ratio of broad money supply M3 to GDP. This ratio is a measure of the size of the banking sector in relation to the economy as a whole whereas credit to private sector measures the role of financial intermediaries in provision of longer-run financing of investment projects by private corporations.

.

.

Stock market liquidity. We measure the stock market liquidity using two indicators. The first variable is the value traded which is the ratio of total value traded to GDP and it measures the value of stock transactions relative to the size of the economy. The second variable is the Turnover ratio calculated as the ratio of the total value traded by stock market capitalization (It often measures the value of equity transactions relative to the size of the stock exchange). Liquid stock market enables investors to modify their portfolios quickly and cheaply. It facilitates investment projects and make them less risky (Levine, 1991; Bencivenga et al., 1996). Therefore, we expect liquidity to have a positive impact on stock market capitalization because larger amount of savings are channeled through stock markets. Macroeconomic instability. To measure the incidence of macroeconomic instability on stock market development, we use inflation change. We expect that the higher the volatility of the economy (inflation change) the less incentive companies and investors would have to put their money in the stock exchange (stock market development).

4.2 Econometric modeling According to the available data, the treatment of incomplete panels is imperative. Indeed, the available panel dataset for the twelve MENA region countries is unbalanced since each variable is observed over varying time period length. In this study, fixed effects as well as random effects models are considered. The fixed effects model is more simple to conduct and is defined according to the following regression model: yit ¼
i þ 0 Xit þ "it ;

i ¼ 1; . . . ; N ; t ¼ 1; . . . ; Ti

ð1Þ

yit indicates the dependent variable while Xit determines the vector of k explanatory variables.
i, i ¼ 1, . . . , N, are constant coefficients specific to each country. Their presence assumes that differences across the considered countries appear by means of differences in the constant term. These individual coefficients are estimated together with the vector of coefficients . In order to validate the fixed effects specification, the question is to prove, according to the empirical application, that the individual coefficients
i, i ¼ 1, . . . , N, are not all equal. This corresponds to the following joint null hypothesis: H0 :
1 ¼    ¼
N ¼


ð2Þ

It is rather the acceptation of the alternative hypothesis which is interesting if we want to differentiate between the situation in each country considered in the sample and confirm the existence of significant heterogeneity across countries. The appropriate statistic of the test is a Fisher distributed one with (N
1,
Ni¼1 Ti
N
k) degrees of freedom under the null hypothesis and is defined as follows: SSR0
SSR1 F¼ SSR1

PN

i¼1

Ti
N
k N
1

ð3Þ

Stock market development

483

MF 33,7

where SSR0 and SSR1 are, respectively, the sum of squared residuals provided by the estimation of the constrained model (under the null hypothesis that is no individual specific coefficients are considered) and the sum of squared residuals relative to the fixed effects model (equation (1)). In the random effects case, the model is defined as follows: yit ¼ 0 Xit þ "it ;

484

i ¼ 1; . . . ; N ; t ¼ 1; . . . ; Ti

ð4Þ

where "it ¼ i þ it reflect the error component disturbances. The individual specific effects are random and distributed normally (i ! IIN(0, 2 )). They are independent of the residual terms it which are also distributed normally (it ! IIN(0, 2 )). The estimation of the model is conducted by the feasible generalized least squares method. First, convergent estimates of the variances 2 and 2 are needed. They are obtained by the following formulae:  PN PT i
^i 2 ^it
 i¼1 t¼1  ^2 ¼ P ð5Þ N i¼1 Ti
N
k

^2

N  2 1  1 X i0  ^  yi
b Xi
^2 ¼ N
k i¼1 Ti

ð6Þ

^it are the residuals issued from the estimation of the fixed effects model (equation (1)) ^i are individual means of these residuals over each time period relative to each and  country. Next, the first term in equation (6) indicates the residuals issued from the estimation of the unit means regression where ^ ib are called the between estimators. The second stage consists in the estimation by ordinary least squares of the following transformed regression model: qffiffiffiffi  qffiffiffiffi   qffiffiffiffi   0 ^ ^ ^ i
1 yi ¼  Xit þ i
1 Xi þ "it þ i
1 "i ð7Þ yit þ with ^i ¼

^2

^2 ; þ Ti ^2

i ¼ 1; . . . ; N

ð8Þ

Finally, a Hausman specification test is conducted in order to compare the two categories of specifications. It is proven that, under the null hypothesis, the two estimates (equations (1) and (7)) could not differ systematically since they are both consistent. So, the test can be based on the difference. Under the null hypothesis, the Hausman statistic is asymptotically distributed as chi-square with k degrees of freedom and is written down as follows:  0     
1   ^ ^F
V ^ ^GLS V ^GLS
^F ð9Þ H ¼ ^GLS
^F where ^F and ^GLS are, respectively, the estimates of the fixed effects and ^ ðÞ are the corresponding variance–covariance matrices of random effects models. V these estimated coefficients.

5. Empirical results The estimation of fixed effects as well as random effects specifications was carried out using the econometric methodology presented in section 4[6]. First, the F-test (equation (3)) led to the validation of the fixed effects specification, that is the presence of individual effects which are not equal. So the heterogeneity across countries is confirmed. On the other hand, the Hausman test also led, for this application, to the acceptation of the null hypothesis according to which the estimates issued from each type of model are equivalent. For all the specifications, the estimated values for the statistics of the test (equation (9)) are below the corresponding critical values of chi-square at the 5 per cent level (see Appendix 2 for random effects results). Indeed and from a purely practical point of view, it is difficult to make a real distinction between fixed and random effects models in such situation because each one have some technical and conceptual advantages and drawbacks. Since we need to consider heterogeneity across countries, we adopt the fixed effects specifications. Table III shows the results of regressions on determinants of stock market capitalization from a sample of twelve MENA countries using the fixed effects specifications. The results from column (1) display that last year’s saving rate, domestic credit to private sector and last year’s value traded to GDP ratio have a positive and significant effect on stock market capitalization. Conversely, last year’s income has no significant impact on market capitalization meaning that high income growth does not promote development in the stock market. On the other hand, growth in revenues is conveyed to other vehicles, such as real investment, bank sector and foreign direct investment (specially Gulf countries), rather than stock markets.

Regressions Income Saving rate

(1) 0.00167 (1.205) 0.402 (2.377)

Investment rate Credit to private 0.846 sector (9.112) M3 Value traded

0.283 (3.434)

Turnover ratio

(2)

(3)

(4)

0.00163 0.00246 0.00197 (1.125) (1.379) (1.334) 0.174 0.471 (0.767) (2.686)
0.0684 (
0.382) 0.83 0.942 (8.407) (10.077) 0.268 (2.557) 0.31 0.543 (3.676) (5.575) 0.0406 (1.0896)

0.0012 (0.866) 0.424 (2.529)

(6) 0.00183 (1.0037)
0.341 (
1.555)

0.798 (8.388) 0.311 (3.758)

(7) 0.00213 (1.364)

134 0.925 75.442

134 0.885 48.032

134 0.922 69.66

485

(8) 0.00127 (0.873)


0.00883
0.0405 (
0.0469) (
0.226) 0.947 0.791 (9.47) (7.854)

0.254 (2.497) 0.557 (5.758)

0.335 (3.939) 0.041 (1.0582)


0.247 (
1.913)

Inflation change No. observations 134 R2 0.929 Statistic F 75.294

(5)

Stock market development

134 0.931 74.726

134 0.887 52.723

134 0.918 68.179


0.222 (
1.67)

Table III.

134 0.927 74.951

Fixed effects specifications of determinants of market capitalization

MF 33,7

486

To test the effect of the investment rate on stock development, column (2) includes last year’s investment rate instead of saving rate. The results reached in the first regression are confirmed. However, the investment rate is not a good predictor of market capitalization since its coefficient is insignificant. This confirms the void relationship between the real economy and the financial market since we found in the first regression no link between last year’s income and stock market capitalization. Such issue could be explained by the smallness of the stock markets in the MENA region. To test the incidence of another measure of financial intermediary development on stock market capitalization, column (3) contains M3 to GDP ratio instead of domestic credit to private sector. This confirms the positive impact of financial intermediary growth on development of stock market. Comparing regressions (1) and (3), we record that domestic credit to private sector seems to be a better measure of financial intermediary and a better predictor of stock market development. This is consist with our expectation. To test the incidence of another measure of stock market liquidity, regression (4) includes least year’s turnover ratio instead of the ratio of value traded to GDP. Comparing with regression (1), we notice that the value traded plays a more important role in explaining stock market capitalization since the coefficient of least year’s turnover ratio is not significant. In addition, inflation change is introduced to control for the macroeconomic stability. In regression (5), the coefficient of this variable have the expected negative sign and is significant which indicate that macroeconomic stability does play a considerable role in determining stock market capitalization. Finally, to test the hypothesis that saving rate is better than investment rate to predict stock market capitalization development, we compare regressions (3)-(5) with regressions (6)-(8), respectively. The results prove the predominance of saving rate as a good predictor stock market capitalization. 6. Conclusions Using a sample of twelve MENA region countries over a varying period, this study tries to identify the main macroeconomic determinants of stock market development. It also examines the impact of financial intermediary development on stock market capitalization. We find that saving rate, financial intermediary (specially credit to private sector), stock market liquidity (specially the ration of value traded to GDP) and the stabilization variable (inflation change) are the important determinants of stock market development, while income as well as investment do not prove to be significant. In addition, we find that financial intermediaries and stock markets are complements rather than substitutes in the growth process. This paper also has some policy implications to MENA region countries. In order to promote stock market development in the region, it is important to encourage savings by appropriate incentives, to improve stock market liquidity, to develop financial intermediaries and to control inflation. Notes 1. If we eliminate Turkey, the market capitalization indicator would be increased by 1 per cent instead. 2. Gulf countries in this paper are Bahrain, Kuwait, Oman, Qatar and Saudi Arabia. 3. For more detailed descriptions, see King and Levine (1993), Levine (1997), Pagano (1993) and Wachtel (2003).

4. The literature omits measures of stock market development because measures of stock market development for a 20-year period are only available for about 40 countries. 5. This definition of stock market development is used here rather than a composite index of stock market development because it is a good proxy for general development and individual measures and indexes of stock market development are strongly associated (Demirgu¨c¸-Kunt and Levine, 1996a). 6. An appropriate algorithm was written on TSP.43 software. References Aghion, P., Bacchetta, P. and Banerjee, A. (1999), ‘‘Capital markets and the instability of open economies’’, mimeo. Atje, R. and Jovanovic, B. (1993), ‘‘Stock markets development’’, European Economic Review, Vol. 37, pp. 632-40. Beck, T., Demirgu¨c¸-Kunt, A. and Levine, R. (2000a), ‘‘A new database on financial development and structure’’, World Bank Economic Review, Vol. 14, pp. 597-605. Beck, T., Levine, R. and Loayza, N. (2000b), ‘‘Finance and the sources of growth’’, Journal of Financial Economics, Vol. 58, pp. 261-300. Bencivenga, V., Smith, B. and Starr, R. (1996), ‘‘Equity markets, transaction costs, and capital accumulation: an illustration’’, World Bank Economic Review, Vol. 10, pp. 241-65. Boyd, J. and Smith, B. (1996), ‘‘The coevolution of the real and financial sectors in the growth process’’, World Bank Economic Review, Vol. 10, pp. 371-96. Demirgu¨c¸-Kunt, A. and Degatriache, G. (1998), ‘‘The determinants of banking crises in developing and developed countries’’, IMF Staff Papers, Vol. 45, pp. 81-109. Demirgu¨c¸-Kunt, A. and Degatriache, G. (2000), ‘‘Banking sector fragility: a multivariate logit approach’’, World Bank Economic Review, Vol. 14, pp. 287-307. Demirgu¨c¸-Kunt, A. and Levine, R. (1996a), ‘‘Stock markets, corporate finance, and economic growth: an overview’’, World Bank Economic Review, Vol. 10, pp. 223-39. Garcia, V.F. (1986), A Critical Inquiry Into Argentine Economic History [1946-1970], Garland Publishing Co., New York, NY. Garcia, V. and Liu, L. (1999), ‘‘Macroeconomic determinants of stock market development’’, Journal of Applied Economics, Vol. 2, pp. 29-59. Goldsmith, R.W. (1969), Financial Structure and Development, Yale University Press, New Haven, CT. Gourinchas, P.O., Landerretche, O. and Valdes, R. (1999), ‘‘Lending booms: some stylized facts’’, mimeo. Gurley, J. and Shaw, E. (1967), ‘‘Financial structure and economic development’’, Economic Development and Cultural Change, Vol. 34, pp. 333-46. International Finance Corporation (2000), Emerging Stock Market Fact Book 2000, Washington, DC. King, R.G. and Levine, R. (1993). ‘‘Finance and growth: Schumpeter might be right’’, Quarterly Journal of Economics, Vol. 108, pp. 717-38. Levine, R. (1991), ‘‘Stock markets, growth and tax policy’’, Journal of Finance, Vol. 46, pp. 1445-65. Levine, R. (1997), ‘‘Financial development and economic growth: views and agenda’’, Journal of Economic Literature, Vol. 35, pp. 688-726. Levine, R. and Zervos, S. (1996), ‘‘Stock markets, banks, and economic growth’’, working paper No. 1690, World Bank Policy Research. Levine, R. and Zervos, S. (1998), ‘‘Stock markets, banks and economic growth’’, American Economic Review, Vol. 88, pp. 537-58.

Stock market development

487

MF 33,7

488

Levine, R., Loayza, N. and Beck, T. (2000), ‘‘Financial intermediation and growth: causality and causes’’, Journal of Monetary Economics, Vol. 46, pp. 31-77. Luintel, and Khan, M. (1999), ‘‘A quantitative reassessment of the finance-growth nexus: evidence from a multivariate VAR’’, Journal of Development Economics, Vol. 60, pp. 381-405. Pagano, M. (1993), ‘‘Financial markets and growth: an overview’’, European Economic Review, Vol. 37, pp. 613-22. Rousseau, P.L. and Sylla, R. (2001), ‘‘Financial systems, economic growth and globalization’’, NBER working paper No. 8323. Rousseau, P.L. and Wachtel, P. (1998), ‘‘Financial intermediation and economic performance: Historical evidence from five industrial countries’’, Journal of Money, Credit, and Banking, Vol. 30, pp. 657-78. Rousseau, P.L. and Wachtel, P. (2000), ‘‘Equity markets and growth: cross country evidence on timing and outcomes’’, Journal of Banking and Finance, Vol. 24, pp. 1933-57. Shan, J.Z., Morris, A.G. and Sun, F. (2001), ‘‘Financial development and economic growth: an egg and chicken problem?’’, Review of International Economics, Vol. 9, pp. 443-54. Singh, A. (1997), ‘‘Financial liberalisation stock markets and economic development’’, The Economic Journal, Vol. 107, pp. 771-82.

Further reading Baltagi, B.H. (1995), Econometric Analysis of Panel Data, John Wiley & Sons, New York, NY. Baltagi, B.H. and Chang, Y.J. (1994), ‘‘Incomplete panels: a comparative study of alternative estimators for the unbalanced one-way error component regression model’’, Journal of Econometrics, Vol. 62, pp. 67-89. Demirgu¨c¸-Kunt, A. and Levine, R. (1996b), ‘‘Stock market development and financial intermediaries: stylized facts’’, World Bank Economic Review, Vol. 10, pp. 291-21. Sevestre, P. (2002), Econome´trie des donne´es de panel, Dunod, Paris. Wachtel, P. (2003), ‘‘How much do we really know about growth and finance?’’, Economic Review, Vol. 88, pp. 33-48.

Appendix 1

Table AI. Sample description

Bahrain Egypt Iran Jordan Kuwait Lebanon Morocco Oman Qatar Saudi Arabia Tunisia Turkey

[1989-1999] [1981-1999] [1993-1999] [1979-1999] [1993-1999] [1995-1999] [1983-1999] [1989-1999] [1997-1999] [1991-1999] [1987-1999] [1988-1999]

Stock market development

Appendix 2

Regressions

(1)

(2)

(3)

(4)

Jordan Egypt Tunisia Morocco Saudi Arabia Kuwait Qatar Oman Bahrain Lebanon Turkey Iran


3.581
26.765
47.706
21.504
9.727
26.917
45.578
20.606 33.669
47.374
24.994
45.968


1.58
18.832
35.61
12.638 4.373
16.477
26.0718
7.722 50.933
50.384
14.779
32.884

19.225
19.73
12.765
11.776
2.539
32.396
39.545
4.754 49.104
32.836
10.39
10.251


8.419
31.205
54.998
25.687
15.229
28.53
57.793
26.196 25.507
54.268
29.0707
52.836

(1)

(2)

(3)

Regressions Income

(5)

(6)

1.723 31.322
21.753
7.256
43.53 2.337
18.817 0.518
5.894 14.771
18.508
12.334
34.514
9.66
17.0393 10.00778 39.0471 69.416
40.333
21.32
4.555 3.468
36.825 4.188

(4)

(5)

(6)

(7)

(8)


9.753
24.532
44.316
17.751
1.964
20.355
41.739
14.0221 41.599
62.197
19.55
40.702

1.976
14.898
32.463
10.696 7.359
9.945
17.346
4.885 55.246
45.653 30.637
25.115

(7)

489

Table AII. Constants from the fixed effects regressions

(8)

0.00165 (1.74) 0.43 (2.679)

0.00189 0.00162 0.00187 (1.918) (1.66) (1.883) 0.261 0.497 (1.305) (2.988)
0.0859 (
0.491) 0.836 0.811 0.932 (9.248) (8.442) (10.255) 0.243 (2.674) 0.287 0.316 0.546 (3.562) (3.822) (5.739) 0.0431 (1.2)

0.00125 0.00161 0.00227 0.00156 (1.284) (1.496) (2.239) (1.521) Saving rate 0.451 (2.833)
0.361
0.0367
0.0539 Investment rate (
1.715) (
0.2) (
0.31) Credit to private 0.787 0.924 0.774 sector (8.522) (9.492) (7.91) M3 0.237 (2.56) Value traded 0.316 0.558 0.341 (3.914) (5.952) (4.104) Turnover ratio 0.0433 (1.162) Inflation change
0.248
0.22 (
2.0383) (
1.751) Constant
23.952
13.748
3.878
29.363
17.113 10.169
20.437
8.445 (
2.229) (
1.115) (
0.346) (
2.663) (
1.5) (0.709) (
1.632) (
0.645) Nb. observations 134 R2 0.568 2 ^v 70.761 ^u2 772.978 Statistic H 0.416

134 0.545 74.061 806.503 0.865

134 0.322 114.223 561.481 2.487

134 0.531 77.06 823.437 0.347

134 0.582 69.206 821.932 0.416

134 0.328 112.486 811.618 0.223

134 0.497 81.77 801.1 1.354

134 0.557 72.958 923.479 0.704

Corresponding author Samy Ben Naceur can be contacted at: [email protected], [email protected]

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

Table AIII. Random effects specifications of determinants of market capitalization

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0307-4358.htm

MF 33,7

Market efficiency, time-varying volatility and the asymmetric effect in Amman stock exchange Haitham Al-Zoubi

490

The Hashemite University, Department of Banking and Finance, Jordan/Zarga, Jordan, and

Bashir Kh.Al-Zu’bi Department of Economics, University of Jordan, Jordan/Amman, Jordan Abstract Purpose – The purpose of this paper is to empirically examine the market efficiency, asymmetric effect and time varying risk–return relationship for daily stock return of Amman Stock Exchange (ASE). Design/methodology/approach – The Box–Jenkins selection model is used to determine the stochastic process of equity returns; the exponential generalized autogressive conditional heteroscedesticity (EGARCH) and threshhold autoregressive conditional heteroscedasticity in mean are utilized to measure the persistent of volatility, risk–return relationship and volatility magnitude to bad and good news. Findings – The univariate statistics show negative skewness, excess kurtosis and deviation from normality for the ASE index. The results show that stock return follows an ARMA (1, 1) stochastic process with significant serial correlation, implying stock market inefficiency. The results also show significant positive relationship between equity return and risk in the ASE, which is consistent with the portfolio theory. The EGARCH model suggests the existence of the asymmetric effect. Originality/value – The paper offers insights into market efficiency, time-varying volatility and asymmetric effect in the ASE. Keywords Stock exchanges, Financial markets, Stock returns, Jordan Paper type Research paper

Managerial Finance Vol. 33 No. 7, 2007 pp. 490-499 # Emerald Group Publishing Limited 0307-4358 DOI 10.1108/03074350710753762

1. Introduction While empirical studies of market efficiency and return–volatility behavior are plentiful for developed stock market, the attention of researchers on developing and emerging stock markets has only begun in recent years. The last decade has been one of rapid growth in international capital flows as a Foreign Direct Investment, with this growth has come attendant for all developing countries to improve their capital markets. Not surprisingly, financial markets grown in size and changed in nature to meet international investors wants of diversifying their portfolios across the globe. As a result, huge empirical studies examining the efficiency and other properties of these markets were introduced as an information window to rationalize the investor choice and to help policy makers in conducting economic policy. A number of papers (Harvey and Berart, 1995; Bekart, 1995; Kim and Singal, 1999; Choudhury, 1996) have examined market efficiency and the risk–return behavior in a number of emerging markets economizes. Fama (1965) has found that stock prices exhibit fatter tails than a normal distribution. While market efficiency and return– volatility behavior have been examined for many emerging markets, it has not been The authors are very grateful to Neal Maroney, Jennifer O’ Sullivan, Atsuyuki Naka, and Elton Daal for their comments and suggestions.

examined for many capital markets like that of Jordan. The examining of market efficiency, persistence of volatility, risk–return behavior, and the magnitude of volatility are very important to Jordan as a country looks for attracting foreign investment as a main channel of boosting economic growth. Stock market plays a vital role in modern society. It increases the investment opportunities by reallocating capital to the most productive users. By facilitating diversification across large number of assets, a stock market can be concerned as the most important tool in reducing the risk that investors bear. This reduces the cost of capital that enhances investment and economic growth. However, volatility and market efficiency are two important features that ultimately determine the effectiveness of stock market in the developing economize. Inefficiency in the stock market creates barriers to investors, which could lead to market failure. For example, market inefficiency may slow down the flow of information on corporate performance for the participants, which creates difficulties for investors to allocate their funds optimally among different types of investments. The resulting uncertainty may induce the investors to withdraw from the market until this uncertainty is resolved or discourage them from investing for the long-term. Moreover, if investors are not rewarded for taking on the higher risk of the stock market, or if the excess volatility weakens investors’ confidence, they will not invest their savings in the stock market, and hence deter economic growth. The aim of this study is to examine stock return distribution, sationarity and stochastic process of Amman financial market (AFM). In particular, it examines the issue of market efficiency, time-varying risk–return and news effect on the magnitude of volatility for this emerging equity market. The daily data set dating back to 1990 has not been utilizing before and part of it is collected handily from different monthly publication of Amman stock exchange (ASE) (Tables I–III). We employ exponential and threshold generalized conditional heteroskedasticity in the mean introduced by Nelson (1991) and Zakoian (1991), respectively, to examine time persistent of volatility, time-varying risk–return relationship and volatility magnitude to bad and good news in ASE. The autoregressive conditional heteroscedesticity (ARCH) family modes is efficiently capable to examine the relationships when the hetroscedasticity appears in time-series models as a result of the reflection of the way in which the variability of the dependent variable changes systematically over-time (Engle, 1982). This may result in Year

Industrial

Banking

Services

Insurance

Total

1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998

26 32 39 44 42 42 43 40 45 42 47

11 13 16 21 23 23 22 18 18 19 16

7 12 13 16 18 22 20 19 19 19 19

13 14 18 22 20 20 17 17 17 17 16

57 71 86 103 103 107 102 94 99 98 98

Source: Various Amman bourse annual reports

Market efficiency in ASE

491

Table I. Number of listed companies in ASE

MF 33,7

492

leptokurtosis, skewness, and volatility clustering which is observed on general financial data. One advantage of the exponential ARCH (EARCH) and threshold ARCH (TARCH) models, used in this study, over the generalized ARCH (GARCH) model is that the former models can capture what is called the asymmetric effect of returns introduced by Black (1976) where stock returns are negatively correlated with changes in return volatility, which implies that bad news has much stronger effect on stock prices than that of good news. This is well know in the literature as the leverage effect which if it exist implies that market participants face a utility function; that is increasing at decreasing rate with respect to the level of consumption. That is the real value of one unit of the medium of exchange is a decreasing function of investors’ wealth (i.e. investors have a decreasing marginal rate of substitution in their utility function). The paper is divided into six sections. Following the introduction in section 1, section 2 provides a brief overview of ASE. Section 3 discusses the methodology and data. Section 4 discusses the statistical properties of the stock prices and returns in ASE. Section 5 analyzes the empirical results of the models. Section 6 concludes the paper. 2. ASE: a brief description The AFM was established through a temporary Law No. 31 of the year 1976, and started its business on the first day of year 1978. It’s establishment was an important step toward regulating and developing the financial sector in Jordan, in order to achieve a better utilization of domestic as well as foreign financial resources through the development of an efficient capital market. It will ease, facilitate and speed the trading to safeguard the financial interest of Jordan. The visibility of the AFM is superior to other markets in the region; it has undergone accelerated growth especially during the last decade. Also Jordan is

Table II. Trading volume of the regular market by sector (1995-1999) (JD million)

Table III. Price index weighted by market capitalization (1995-1999) (1991 ¼ 100)

Year

Banks

Insurance

Services

Industry

Total

% change

1995 1996 1997 1998 1999

149.6 83.1 165.4 191.0 125.1

7.4 3.1 4.5 5.1 6.8

82.0 42.5 31.4 24.1 29.7

123.1 82.0 102.8 193.4 181.0

362.1 210.7 304.1 413.6 342.6

(15.8) (41.8) 44.3 36.0 (17.2)

Source: Various Amman bourse annual reports

Year

Banks

Insurance

Services

Industry

Total

% change

1995 1996 1997 1998 1999

1,883 194.4 232.3 281.7 251.9

130.8 120.8 122.9 124.7 123.7

129.9 115.3 116.6 109.0 110.6

130.0 116.2 115.5 76.9 96.2

159.2 153.5 169.2 170.1 167.4

10.8 (3.6) 10.3 0.5 (1.6)

Source: Various Amman bourse annual reports

represented on the board of International accounting standards. The objectives of the AFM were laid in the Law as follows:

Market efficiency in ASE

(1) To mobilize and to channel savings in order to serve the interest of the national economy through encouraging investment in securities. (2) To regulate issuance of and trading in securities. (3) To provide the necessary data and statistics to achieve the AFM objectives. The Jordanian Government adopted a sound capital market reforming policy in 1997. The objective was to restructure and to regulate the Jordanian market in accordance with the international standards and to ensure transparency and safe trading in securities. This restructuring of capital market implies a separation of the function of Jordan Securities Commission, as a supervisor and regulator from the executive role of the capital market. Therefore two new institutions were established; the ASE and the Securities Depository Center, to play the executive role and to be run by the private sector. The ASE shifted into electronic trading system on March 2000, in order to increase the efficiency and transparency in the security market. In fact this system creates a suitable environment for trading by using the high technology, which will enhance the ASE development. In addition to that the internal information network was established in ASE in order to provide a database for achieving the goals of the security market. A closer look at the ASE activities, indicates that in year 2001, the number of traded shares reached 340.6 million with a 49.1 per cent increase compared to that in year 2000. The volume of trading in year 2001 doubled compared to that in year 2000, in addition to that the share price index in year 2001 reached 172.7 compared to 133.1 in year 2000 (closing December 1991 ¼ 100). 3. Methodology and data 3.1 Methodology The study examines the distribution of equity return by comparing the descriptive statistics of ASE index. Market efficiency is tested with reference to the structure of ARCH in mean. While most researchers agree that volatility is predictable in many asset markets, they differ on how this volatility should be modeled. In recent years, the evidence for predictability has led to a variety of approaches. The most interesting of these approaches is the ‘‘asymmetric’’ or leverage volatility models, in which good news and bad news have different predictability for future volatility. These models are examined by empirical work of Black (1976), Frinch et al. (1987), Nelson (1991), Zakoian (1991) and Glosten et al. (1993). Black (1976) finds evidence that stock returns are negatively correlated with changes in return volatility, i.e. volatility tends to rise in response to ‘‘bad news’’ (negative excess returns) and to fall in response to good news (positive excess returns). However, if the magnitude of volatility response to bad and good news has the same absolute value of correlation we say that bad and good news have a symmetric effect on stock volatility. One implication of the symmetric effect of the arrival of new information on stock returns is that investors have constant marginal rate of substitution. Since the economic theory suggests that investors with convex utility function face a decreasing marginal rate of substitution while there consumption of goods is rising, one could predict that the response of stock returns will be much bigger to the bad news than that to good news (the leverage effect). On the other hand, if good

493

MF 33,7

494

news has a larger effect on stock return than the good news we say that stock prices display an asymmetric effect[1]. The ARCH model introduced by Engle (1982) allows the variance of the error term to vary over-time, in contrast the classical regression model, which assumes a constant variance. Bollerslev (1986) generalized the ARCH process by allowing for a lag structure for the variance, since stock returns are highly fluctuated the generalized ARCH models, i.e. the GARCH models, have been found to be valuable in modeling the time series behavior of stock returns (Akgiray, 1989; Frinch et al., 1987). Bollerslev (1986) allows the conditional variance to be a function of the lag’s squared errors as well as of its past conditional variances. The higher order ARCH and GARCH models introduced by Engle (1982) and Bollerslev (1986), respectively, are
2t ¼ ! þ

P X

j "2t
1 ¼

j¼1


2t ¼ ! þ

q X

p X

2 j Zt
j
2t
j

ð1Þ

2 j Zt
j
2t
j

ð2Þ

j¼1

Bi
2t
1 þ

i¼1

p X j¼1

where: "t ¼
t Zt Zt  i:i:d: with EðZt Þ ¼ 0;
2t

VarðZt Þ ¼ 1

and

2

¼
ð"t
1 ; "t
2 ; "t
3 ; . . . . . .Þ

Nelson (1991) shows that Equation (2) can be written as: 1 X  2
2t ¼ ! þ k Zt
1
2t
k k
1  where ! and k are non-negative parameters. Nelson contradicts generalized ARCH models in which GARCH models assume that only magnitude and not the positivity or negativity of unanticipated excess returns determine the conditional variance
2t . If the distribution of Zt is asymmetric, the future change in variance is conditionally uncorrelated with excess returns today. In Equations (1) and (2),
2t is a function of lagged
2t and lagged Zt2, and so invariant to changes in the algebraic sign of Zt (i.e. only changes in the algebraic sign of lagged residuals determines conditional variance). This suggests a model in which
2t respond asymmetrically to positive and negative disturbance is preferable for empirical studies of assets pricing. Another limitation that Nelson (1991) suggests, results from the non-negativity  constraints on ! and k in Equation (3) that is imposed in GARCH and GARCH-M models to insure non-negative conditional variance. The Exponential GARCH (EGARCH) introduced by Nelson (1991) and Threshold ARCH (TARCH), which is introduced independently by Zakoian (1990) and Glosten et al. (1993) is free of all these shortcomings, and they can robustly capture for asymmetric effect that GARCH-M cannot. Lewis et al. (1992) mention that the existence of asymmetric effect may give wrong estimate of the risk–return relationship if

GARCH-M is used. The higher order specification of the conditional variance in TARCH model is:
2t ¼ ! þ

q X i¼1

i "2t
i þ "2t
1 dt
1 þ

p X

Market efficiency in ASE

j
2t
j

j¼1

where dt ¼ 1 if "t  0, and dt ¼ 0 otherwise. In this model, good news ("t  0) and bad news ("t  0), have differential effects on the conditional variance, good news has an impact of , while bad news has an impact of  þ . If   0, we say that the leverage effect exists. If  6¼ 0 the news effect is asymmetric. The EGARCH or exponential GARCH model was proposed by Nelson (1991). The specification for the higher order conditional variance is: !

p q X X
"t
i
"t
i 2 2

þ i Bj logð
t
j Þ þ i
Logð
t Þ ¼ ! þ
t
i

t
i j¼1 i Note that the left-hand side of the equation is the log of the conditional variance. This implies that the asymmetric effect is exponential, rather than quadratic, and that forecasts of the conditional variance are generated to be non-negative. The presence of leverage effects can be tested by the hypothesis that   0. The impact is asymmetric if  6¼ 0. As the original GARCH model, the size and significance of j indicates the magnitude effect imposed by the lagged error term ("t
1) on conditional variance
t2. In other words, the size and significance of j implies the existence of the ARCH process in the error term (volatility clustering). Engle and Bollerslev (1986) show that the persistence of shocks to volatility can be examined depending on the sum of the parameter of the Autoregressive and Moving average component of
2t , i þ Bj, values of the sum lower than unity implies a tendency for the volatility response to decay over-time. In contrast, values of the sum equal or greater than unity imply indefinite (or increasing) volatility persistence to shocks over-time. The order of all GARCH models in this paper are estimated by generating a random error series from ARMA (1, 1) process, that stock returns in ASE follow, and then regress the squired residuals on lagged residuals and implying collologram analysis and Breuch–Godfrey LM test to check for the order of the process. 3.2 Data We gathered data for AFM (the lonely stock market in Jordan) from two sources. First, we collected electronic data on ASE index from first of January 1994 to the 31st of July 2001 from (AFM) CD report. For the period spanning from January 1992 to July 2001, the data was collected by hand from AFM monthly bulletin. The ASE index is composed of all traded stocks (value waited) in the exchange. The index excludes dividends. 4. Statistical properties of (ASE) equity price and returns The study examines the stability of returns by employing CUSM-test to determine the structural brake in the series. The Chow test is then estimated to insure the existence of

495

MF 33,7

496

structural brake. Both tests suggest no structural change during the period of the study. Tables IV and V provide the statistical properties of ASE daily equity prices and returns. The Jarque–Bera test is examined to test for normality, as it appears from the table, ASE index shows negative skewness, excess kurtosis, and deviation from normality. Fama (1965) shows that daily and monthly return distribution of Dow Jones and NYSE indices are leptokurtic, negatively skewed and depart from normality. Bekaert et al. (1998) provides evidence that the majority of emerging markets have positive skewness, excess kurtosis, and non-parametric distribution. 5. Time-varying volatility and equity return of ASE 5.1 ARMA (1,1) process and capital market efficiency Table VI shows the empirical findings of return process and market efficiency. The equity returns are calculated as the first-order log difference of ASE index. The Box– Jenkins selection method shows that stock returns follow significant ARMA (1,1) stationary trend process, none of the autocorrelation functions or the partial autocorrelation functions for the first 20 lags are significant. The Philips–Peron test is utilized to check for the stationarity of the stock returns, the P-value of the test is below 0.05, which implies that stock returns are stationary. Since ARMA (1, 1) process is significant the returns are predictable on the basis of past returns. As a result, we reject the weak form of market efficiency (i.e. the past information is not reflected quickly on stock price). This could arise from frictions in trading process, limited provision of information of firms’ performance to market participants, and the lack of professional financial analyst and brokerage houses that can analyze stock market returns for investors. These findings suggest the modernization of trading system by setting more requirements on listed companies, like an increasing in the disclosure requirements,

Table IV. Univariate statistics for ASE index

Mean

Table V. Univariate statistics for ASE daily return

Mean

Table VI. ARMA (1, 1) process for ASE daily return

150.3480

0.000154

Standard deviation

Minimum

Maximum

Skewness

Kurtosis

Jarque–Bera

18.76767

100.6482

187.6626


0.781089

3.364288

251.5210

Standard deviation

Minimum

Maximum

Skewness

Kurtosis

Jarque–Bera

0.006831


0.043102

0.047449

0.422161

8.781970

3336158

AR(1) parameter MA(1) parameter Mean of ACF Lijung-Box Q Statistic for 48 lags RMSE SBC Philips–Peron unit root test Note: *Significant at 1 per cent level

0.266463 (0.01992) 0.999783 (0.00034) 0.153894 Insignificant 0.0065872
16.7500
36.29910*

regulating general meeting between the shareholders and managers and forcing listed companies to provide timely audited financial statements. 5.2 Time-varying, persistent and clustering of volatility We present the empirical results of volatility and risk in Table VII. The hypothesis that volatility is a significant determinant of stock returns is confirmed as the parameter , which is positive and statistically significant in both EGARCH and TARCH in mean models, the result is consistent with the portfolio theory. The squared error parameter () is positive and significant implies the existence of ARCH process in the error term, this suggests the tendency of the shocks to persist (volatility clustering). Since the sum of () and (B) is greater than or approximately equal to unity, the tendency for volatility response to shocks display a long trend, implying time-varying volatility in the equity returns in ASE.

Market efficiency in ASE

497

5.3 Asymmetric and leverage effect in ASE For equities, it is often observed that downward movements in the market are followed by higher volatilities than upward movements of the same magnitude. Table VII presents the asymmetric effect parameter (). In the EGARCH model, it is positive and statistically different from zero, indicating the existence of asymmetric effect. Since () is positive, there is no leverage effect i.e. good news has higher impact on volatility than that of bad news. In comparison with the TARCH model () is less than zero, indicates the same result of () > 0 in the EGARCH model, but it is statistically insignificant at 5 per cent level of significant (P-value ¼ 0.12). The shape of news impact curve for ASE is differ than the normal phenomena, that bad news followed by higher volatilities than that of good news, this might be explained by market inefficiency in that companies tries to spread good news and hide that of bad news. Another reason that we suggest is due to economic condition that Jordan economy faces during the study period, since the economic sanction on Iraq, the Jordan economy fall in stagnation, no intensive growth has been achieved since 1993; some temporary expectations about removing the economic sanctions on Iraq may have a strong effect on ASE. GARCH-M (1, 1)  0 1 B1 1 þ B 1  SBIC Ljung-Box Q test for 20 lags Breusch–Godfry LM test Number of observations

EGARCH-M (1, 1)

0.197915 (3.117431*) 0.00000414 (4.611661*) 0.253329 (7.302486*) 0.670827 (15.84923*) 0.924156 –
7.393548

0.196468
1.494074 0.428976 0.885029 1.314005 0.047821
7.395502

Insignificant 0.027209 P-value ¼ 0.868981

Insignificant 2.591410 P-value ¼ 0.273705

2346

(3.106497*) (
14.61822*) (18.91895*) (95.24230*) (3.925557*)

2346

Notes: *Significant at 5 per cent level; **significant at 1 per cent level

TARCH-M (1, 1) 0.263572 (4.057484*) 0.00000413 (4.644982*) 0.296945 (6.010262*) 0.673620 (16.62133*) 0.970565
0.098207
(1.554708**)
7.393890 Insignificant 0.018589 P-value ¼ 0.891551 2346

Table VII. Estimates of GARCH (1, 1), EGARCH (1, 1) and TARCH (1, 1) in mean for ASE

MF 33,7

498

6. Conclusion The paper has empirically examined the stock return behavior in ASE, market efficiency, the time-varying risk–return relationship, the persistence of the stock volatility and the leverage effect for the holding period 1990-2000. The ASE shows negative skeweness, excess kurtosis, and deviation from normality. The ASE volatility tends to change over-time, and is serially correlated. In addition, the ASE returns show a significant serial correlation, implying stock market inefficiency. The risk–return parameter is positive and significant, which is consistent with the portfolio theory. The EGARCH model shows asymmetric effect in that good news has a higher impact on stock price than that of bad news. The market inefficiency may exist as a result of a lake of information resource that investors are trying to find. To improve capital market efficiency, the government should emphasis a policy of timely disclosure and make incentives for business to invest in the financial services sector. Since Iraqi market absorbed more than 40 per cent of Jordan exports before 1991 (The Gulf War), a temporally expectations about finishing economic sanctions on Iraq may create the asymmetric effect on ASE. In most cases, bad news has a stronger effect on volatility than that of good news. The opposite appears in the case of Jordan. Note 1. The existence of the asymmetric effect does not necessarily implies the leverage effect since the leverage effect only implies that bad news has a higher effect than good news while the opposite does not hold. References Akgiray, V. (1989), ‘‘Conditional heteroscedasticity in time series stock returns: evidence and forecast’’, Journal of Business, Vol. 62, pp. 55-80. Bekaert, G. (1995), ‘‘Market integration and investment barriers in emerging equity markets’’, The World Bank Economic Review, Vol. 9, pp. 75-107. Bekaert, G. and Harvey, C.R. (1997), ‘‘Emerging equity market volatility’’, Journal of Financial Economics, Vol. 43, pp. 29-77. Bekaert, G., Erb, C.B., Harvey, C.R. and Viskanta, T.E. (1998), ‘‘Distributional characteristics of emerging market returns and asset allocations’’, Journal of Portfolio Management, Risk Loving, Vol. 24, pp. 102-16. Black, F. (1976), ‘‘Studies of stock market volatility changes’’, Proceedings of the American Statistical Association, Business and Economic Statistics Section, pp. 177-81. Bollerslev, T. (1986), ‘‘Generalized autoregressive conditional heteroskedasticity’’, Journal of Econometrics, Vol. 31, April, pp. 307-27. Choudhry, T. (1996), ‘‘Stock markets volatility and the crash of 1987: evidence from six emerging markets’’, Journal of International Money and Finance, Vol. 15 No. 6, pp. 969-81. Engle, R.F. (1982), ‘‘Autoregressive conditional heteroscedasticity with estimates of the variance of United Kingdom inflation’’, Econometrica, Vol. 50, July, pp. 987-1007. Engle, R.F. and Bollerselv, T. (1986), ‘‘Modeling the persistence of conditional variances’’, Econometric Reviews, Vol. 5, pp. 81-7. Fama, E.F. (1965), ‘‘The behavior of stock market Prices’’, Journal of Business, January, Vol. 38, pp. 34-105. Frinch K., Schwert, W. and Stambauhg, R. (1987), ‘‘Expected returns and volatility’’, Journal of Financial Economics, Vol. 19, pp. 3-30.

Glosten, L.R., Jagannathan, R. and Runkle, D.E. (1993), ‘‘On the relation between the expected value and the volatility of the nominal excess return on stocks’’, Journal of Finance, Vol. 48, pp. 1779-1801. Kim, E.H. and Singal, V. (1999), ‘‘Opening up of stock market by emerging economics: effect on portfolio flows and volatility of stock prices’’, working paper, The World Bank. Nelson, D.B. (1991), ‘‘Conditional heteroskedasticity in asset returns: a new approach’’, Econometrica, Vol. 59, pp. 347-70. Zakoian, J.-M. (1990), ‘‘Threshold heteroskedastic models’’, manuscript, CREST, INSEE, Paris. Zakoian, J.-M. (1991), ‘‘Threshold heteroskedastic model’’, unpublished paper, institue National de la statistique el des Etudes Economiques, Paris. Further reading Harvey, C.R. (1995), ‘‘Predictable risk and returns in emerging markets’’, Review of Financial Studies, Vol. 8, pp. 773-816. Corresponding author Haitham Al-Zoubi can be contacted at: [email protected]

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

Market efficiency in ASE

499

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0307-4358.htm

MF 33,7

Empirical testing of the loss provisions of banks in the GCC region Taisier A. Zoubi

500

School of Business and Management, American University of Sharjah, Sharjah, UAE, and

Osamah Al-Khazali School of Business and Management, American University of Sharjah, Sharjah, UAE Abstract Purpose – The purpose of this study is to examine the factors which affect loss provision for loans and investment in Murabaha, Musharka, and Mudarabah for banks in the Gulf Cooperation Council (GCC) region. The effect of prior period earnings, legal and statutory reserves, size of the bank, level of debt, and loan and investment to deposit ratio on the loss provisions of banks are examined for the period 2000-2003. Design/methodology/approach – To test the factors that explain the loan loss provision and to test the income smoothing hypothesis, debt to equity hypothesis, and reserve hypothesis, a single stage regression model was developed and tested. Findings – The results indicate that when return on assets (ROA) before tax and loss provisions for the current year is higher than the prior year ROA and the actual capital reserve is below the legal required reserve, then management is expected to increase loss provisions for the current year. This result is robust for all the years of this study. Originality/value – While prior research has examined the issue of the loan loss provision in USA, Japan, and Europe, no research has examined the issue of the loss provisions in the GCC region. This study demonstrates that the income smoothing hypothesis is relevant across different regulatory requirements, economic conditions, and different accounting standards. Managers of banks in the GCC region use the loss provision, among other things, to smooth earnings to achieve certain objectives. Keywords Banks, Islam, Loans, Investments, Persian Gulf States Paper type Research paper

Managerial Finance Vol. 33 No. 7, 2007 pp. 500-511 # Emerald Group Publishing Limited 0307-4358 DOI 10.1108/03074350710753771

1. Introduction The main objective of banks, whether Islamic or conventional, is to attract deposits and to reinvest those deposits to make profit. Banks in the Gulf Cooperation Council (GCC) region[1] have several investment opportunities. Conventional banks invest their funds in loans to customers and in securities. Islamic banks invest their fund jointly with customer through three different methods: Murabah, Musharaka, and Mudarabah[2]. Although the investment approach used by Islamic banks is drastically different than the ones used by conventional banks, the risk that some or all of the loans made to customers in the case of the conventional banks and some or all of the investment in Murabah, Mudarabah, or Musharaka in the case of the Islamic banks, may not be collected due to the inability of the borrowers or investors to repay their obligations to the banks. Therefore, banks should establish an allowance for loss provision to absorb any future losses. To do that, management estimates the amount of losses for loans and investment in Murabah, Mudarabah, and Musharaka that may not be collected in the future using the accrual basis. An allowance for losses is established as a contra-asset

account. The allowance of losses represents the amount of expected losses of the outstanding balance of loans or investments in Murabahah, Murabah, and Musharaka. The addition to the balance of the allowance for losses at the end of each period must be reported in the income statement of that period as loan loss provision. If the expected losses, due to past and present events, exceed the amount in the allowance for losses, then a bank increases its loss allowance and that increase should be reported in the income statement of the current period as an expense. Allowance for loss provision decreases by the actual write-off of loans and investment in Murabahah, Mudarabah, and Musharka. Any amount of recovery, that had been previously written-off, will increase the balance of the allowance for loss provision. Loss provision represents a material amount in the banks’ financial statements and potentially an area for management manipulation. If a bank misstated the loss provision, this will result in misstatements of assets, earnings, and capital. Prior literature on this subject has shown that banks’ managers manage earnings, using the loss provisions, to smooth income and to send signals to the users of the financial statements (see for example, Hasan and Hunter, 1994; Bhat, 1996; Lobo and Yang, 2001; Hasan and Wall, 2004). This study examines the factors which affect loss provision for loans and investment in Murabah, Musharkaha, and Mudarabah for banks in the GCC region. More specifically, this study investigates the effect of prior period earnings, legal and statutory reserves, level of debt, and loan and investment to deposit ratio on the loss provision of banks. While prior research has examined the issue of the loan loss provision in USA, Japan, and Europe, no research has examined the issue of the loss provisions in the GCC region. The remainder of this paper proceeds as follows: section 2 discusses the relevant literature, section 3 provides a description of the accounting for loss provision and allowance, section 4 explains the incentives of earnings managements via loss provision, section 5 presents a description of the data selection and methodology, section 6 discusses the results and interpretation, and section 7 provides summary and conclusion. 2. Review of the relevant literature Several studies have examined the use of loan loss allowance by banks as a tool to manage earnings, to achieve their compensation contracts, to manage capital adequacy, and to examine behavior of stock prices. For example, Greenwalt and Sinkey (1988) examined the effect of banks’ earnings on the loan loss provision and found that banks smooth their earnings via loan loss provisions. Bhat (1996) examined the use of loan loss provisions by large US banks for the period 1981-1991 to smooth earnings. His result shows that banks smooth their net income when banks experience low growth, low book-to-asset ratio, high loans-to-deposit (LD) ratio, high debt-to-asset ratio, low market-to-book value ratio, low return on assets, high loan-loss provisions to gross loans ratio, and low assets. Moyer (1990), Kim and Kross (1998), and Kiridaran et al. (2003) found that loan write-offs, loan loss provision, and securities gains and losses are used by banks to manage capital requirements. Beatty et al. (1995) found that loan loss provisions, writeoff of loans, and issuing securities are used to manage earnings and capital reserve ratio by banks. Hasan and Hunter (1994) examined the impact of several factors on the loan loss provision for 344 banks for the period 1985-1989. They found that the factors believed to affect loan loss provision are significant and consistent with income smoothing

Empirical testing of loss provisions 501

MF 33,7

502

hypothesis. Collins et al. (1995), Beaver and Engel (1996), and Ahmed et al. (1999) concluded that banks use loan loss allowance to manage capital to meet regulatory requirements. Beaver and Engel (1996) examined the relationship between loan loss allowance and security returns of banks in USA and found an increase in the loan loss allowance is associated with a decrease in stock prices. Healy (1985) and Degeorge et al. (1999) provided evidence that managers manipulate earnings to maximize their own compensation. Banks also manage earnings to reach a ‘‘target’’ earning. For example, Burgstahler and Dichev (1997) examined firms’ current period earnings in relation to zero earnings and last year’s earnings. Their results showed that managers manage earnings when their latent earnings are below their thresholds. Docking et al. (2000) examined the announcements of the loan loss reserve on the stock prices of 82 US commercial banks for the period 1985-1990 using the event study methodology. Their results show that stock prices of the sample banks reacted negatively to the announcements of the loan loss provisions. Hasan and Wall (2004) examined the factors that may affect the allowance for loan loss for four different samples (US, Europeaon, Canada, and Japan) for the period 19932000. Their results indicate that the loan loss reserve ratio is a significant factor in explaining the allowance for loan losses for only the US banks while it is not for the other three samples (Canada, Europe, and Japan). Lobo and Yang (2001) found that managers of banks use the loan loss provisions to signal some information to the user of the financial statements, to smooth earnings overtime, and manage regulatory reserve requirements. Kanagertnam et al. (2004) examined the effect of external financing needs, gains and losses of available-for-sale securities, and the regulatory requirements on loan loss provisions. Their sample consists of 2545 US banks for the period 1992-2001. Their results show that external financing and gains and losses of available-for-sale securities are important factors in explaining earnings smoothing overtime by banks. They argued that income smoothing reduces banks’ perceived risk, and as a result, cost of borrowing will be reduced. It is clear from the evidence provided by the prior research that banks use the loan loss provision to manage earnings and capital. 3. Accounting for loss provision and allowance Banks in the GCC region adopted the financial accounting rules established by the International Accounting Standards Board, previously International Accounting Standards Committee (IASC) (Hussain et al. 2002). The (IASC) issued IAS No. 30 ‘‘Disclosures in the financial statements of banks and similar financial institutions’’, in 1991. This standard deals with, among other things, loan loss provisions measurement, and recognition and disclosures in the financial statements. According to IAS No. 30, banks should disclose the following: (1) ‘‘The movements in the provision for losses on loans and advances during the period. It should disclose separately the amount recognized as an expense in the period for losses on uncollectible loans and advances, the amount charged in the period for loans and advances written off and the amount credited in the period for loans and advances previously written off that have been recovered’’. (2) ‘‘The aggregate amount of the provision for losses on loans and advances on the balance sheet date’’.

(3) ‘‘The aggregate amount included in the balance sheet date for loans and advances on which interest is not being accrued and the basis used to determine the carrying amount of such loans and advances’’ (IASC, 1991, IAS 30, para. 43).

Empirical testing of loss provisions

Although, the accounting standards for loan loss provision provide general guidelines, management must use its judgment and discretion in computing the amount of loan loss provisions. Management must estimate percentage of the total loans and the amount of investments in Murabaha, Musharka, and Mudaraha that may not be collected in the future and adjust the allowance balance to reflect the new estimate. Hence, management has great latitude in the estimation process of loss provisions.

503

4. Incentives of earnings management via loss provisions and hypotheses development The review of the prior literature indicates that there are several factors that explain the determinants of banks’ loss provision. Some of the factors are non-performing loans (NPL), loan write-offs, beginning balance for allowance for loan losses, past and present earnings, debt to equity (DE) ratio, capital adequacy ratio, LD ratio, and size of the bank. According to the prior literature, banks use loan loss provisions to achieve several objectives. Some of the objectives are to smooth earnings, affect stock prices, affect management bonus, provides some signals to the users about their ability to absorb future losses, signal of what the expected future earnings, and to comply with the legal reserve requirements. Managers’ compensation of banks may be tied to net income. Hence, managers may decrease the estimate amount of loan loss provision to report higher net income than what it should be and consequently increase their bonus. Healy (1985) found a relationship between management bonuses and earnings manipulation of reported income. Management believes that earnings affect stock prices. Managers may use the loan loss provision to increase or decrease earnings and hence affect stock prices. It has been documented, in the accounting literature for the last three decades, that earnings affect stock prices. Management may underestimate loss provision during low year earning before the loss expense and overestimate the loss provision during high year earning. Hence, Managers, through loss provision, are able to shift earnings among periods to smooth income over time. Loan loss provision is merely an accounting adjustment to earnings that do not involve any direct cash outflows. Loss provision decreases the reported net income, retained earnings, and stockholders’ equity. However, loss provisions offer a signal to the users of the financial statements about the collectability of the loans and investment in Murabah, Musharka, and Mudarabah. Therefore, the income smoothing hypothesis predicts that a positive relationship between prior years earnings and the loan loss provision. In other words, high earnings in prior year relative to the current year will lead to lower loss provision for the current year. Since loss provision affects retained earning, which is part of the stockholders equity, DE ratio decreases, and therefore the perceived risk of the bank will increase. The DE hypothesis predicts that the higher the DE (high risk), the lower the loss provision for the current year. In other words, lower loss provision will lead to higher earnings, higher equity, and lower DE ratio (low risk).

MF 33,7

504

Any increase (decrease) in the loan loss allowance results in a decrease (increase) in retained earnings. The decrease (increase) in earnings will affect the amount that the bank can transfer to the reserve account, since the amount transferred to the legal or statutory reserve is a percentage of the net income of the current year for all banks in the GCC region (see the Appendix for regulatory reserve requirements of the GCC countries). If a bank’s reserve account balance is below the required amount set by the regulatory agency, then a bank will have an incentive to increase its reported earnings, and as a result, increase the amount to be transferred to the legal and statutory reserve. Hence, the gap between the minimum required reserve of the bank and the actual balance of the bank’s reserve will shrink. However, if the bank’s reserve balance equals the minimum required balance, then a bank will not have an incentive to increase its reported earnings from the reserve regulation compliance point of view. Therefore, the reserve hypothesis predicts a positive relationship between the gap of the reserve ratio and the loss provisions. 5. Data and methodology In order to test the factors that explain the loan loss provision and to test the income smoothing hypothesis, DE hypothesis, and reserve hypothesis, we collected the data from the annual reports of banks in the GCC region and the Institute of Banking Studies (Kuwait). The income statement, statement of change in stockholders’ equity, balance sheet, statement of cash flows, and the notes to the financial statements were obtained from the annual report of each bank. In the selection of the period of analysis, a crucial issue was the availability of data. Since the annual reports were obtained from the website of each bank, we were restricted with data available in those sites. We were able to obtain data for 47 banks for the years 2002 and 2003 (total 94 year/ bank observations). However, only 20 banks had data available for the years 2001 and 2000 (total 40 year/bank observation). Annual reports prior to 2000 were available for a very limited number of banks. Therefore, the period of this study covers only 2000-2003. We examined the notes of the financial statements to determine whether each bank was in compliance with the IAS No. 30. We found that all banks in our sample, whether conventional or Islamic, complied with the International Accounting Standards with respect to the loss provision and allowance as presented in IAS No. 30, during the period of this study (2000-2003). Hence, loss provisions were calculated by banks using the same accounting standards and consequently, the loan losses data obtained from the annual reports are comparable.

Table I. List of the number of banks in the sample by country/year

Bahrain Kuwait Oman Qatar Saudi Arabia UAE Islamic Total

2003

2002

2001

2000

Total

Total banks operating in each country

7 5 4 3 9 10 9 47

7 5 4 3 9 10 9 47

4 5 1 0 6 3 1 20

4 5 2 0 6 3 0 20

22 20 11 6 30 26 19 134

9 8 7 4 9 18 10 65

The distribution of sample banks by countries for 2000-2003 is shown in Table I. As can be seen, the sampled banks in this study are fairly representative to the number of banks operating in each country. However, the UAE banking industry is modestly represented (10 banks out of 18) in our sample. Several methodologies have been used in prior research to examine how management manages earnings via loan loss provisions and the factors which influence loan loss provision. For example, Collins et al. (1995) used time series data to examine loan loss provisions. This methodology requires several years of data. On the other hand, Barth et al. (1990) and Hasan and Wall (2004) used cross-sectional models to examine loan loss provisions. Since the data available only covers four years, this study employs a single stage model to examine the factors that affect loss provision of banks. This study examines the income smoothing, DE, and the legal reserve hypothesis for the banking industry in the GCC region. To do this, the following model is used: LLP ¼ CROA þ LD þ DE þ RD þ TYPE þ LOGTA þ TYPE where LLP, is the loss provision to total loans and investment in Murabah, Musharka, and Mudarabah; CROA, the (earnings before tax[3] and loss provision divided by asset) minus return on assets (ROA) for the last year (t – 1); LD, the gross loans and investment in Murabah, Musharka, and Mudarabah to deposits; DE, the total debt to common equity; RD, the current bank’s reserve minus legal and statutory required reserve/equity; LogTA, the log of total assets; and TYPE, the a dummy variable, it equals 0 if a bank is Islamic, 1 if a bank is conventional. The relationship between the loss provision and the explanatory variables are: (1) CROA: this variable is included to test the income smoothing hypothesis. The relationship between CROA, and LLP is predicted to be positive. If ROA before tax and loss provision for this year is higher than last year ROA, then management is expected to increase loss provision to decrease net earnings for the current year. A positive relationship is expected. (2) LD: this variable measures the relationship between loans and investment to customers’ deposits. The higher the ratio, the more the need for external funds. To attract external funds, a bank must reduce the perceived risk via decrease in the loss provision. A negative relationship between LLP and LD is expected. (3) DE: this variable measures the risk of the bank. It is calculated by dividing total debt by total equity. A negative relationship between LLP and DE is expected. This variable is used to test the DE hypothesis. (4) RD: this variable measures the degree of the bank’s compliance with the legal and statutory reserve requirement. RD equals current bank’s reserve minus legal required reserve divded by equity. This variable is used to test the legal reserve hypothesis. A positive relationship between LLP and RD is expected. (5) TYPE: type of banks (Islamic or conventional). This variable is included to see if there is a difference between loss provision of Islamic banks and loss provision of conventional ones. No relationship between LLP and TYPE is expected.

Empirical testing of loss provisions 505

MF 33,7

506

(6) LOGTA: natural logarithm of total assets. A positive relationship between LLP and LOGTA is expected. Larger banks are expected to afford larger loss provision than smaller banks. 6. Results and interpretations Table II provides the descriptive statistics for all the variables used in this study. The average LLP for the sample banks is relatively low, 1.31 per cent for the period 20002003. This implies that banks in the GCC region make a very low estimate of loss provision. By examining the maximum LLP, we found that some banks estimated their loss provision to be 25.68 per cent of their total outstanding loans. The average ROA before tax and loss provision is 2.23 per cent. Total DE is very high, debt is 7.48 times the equity for the period 2000-2003. It appears that a large portion of the investment in assets by banks in the GCC region is financed by debt. Average NPL is 7.53 per cent of the total loans and investments. Table III presents the average simple correlation coefficients among LLP, CROA, LD, DE, and LOGTA. The correlation coefficients among the independent variables are very low, except LD and TYPE. The low correlation coefficients among the independent variables suggest that there was not a multicollinearity problem among the variables. The average correlation coefficient between LLP and CROA is 0.255 and significant at 0.05 level of significance for the entire sample. Thus, if this year earning is more than last year, then loss provision for the current period would be higher. The average correlation coefficients between LLP on the one hand and LD, RD, and TYPE on the other hand are
0.028, 0.110, and 0.086, respectively. Those correlation coefficients are not significant at any reasonable level. The correlation between LLP and LOGTA is
0.217 and significant at 0.01 level of significance. LLP and DE are positively correlated and significant. The results of the regression model, which examines the effect of banks’ specific variables on loss provision for the entire period (2000-2003), are presented in Table IV. The results suggest that ROA of the prior period relative to the ROA before tax and loss provision for the current period (CROA), DE, and the size of the bank (LOGTA) are significant in explaining the variation of the loss provision (LLP) banks in the GCC region. About 23 per cent of the variation in LLP can be explained by the variation in the three variables, as measured by the coefficient of determination (R2). The overall model is significant at the 1 per cent level of significance. The results indicate that the coefficients of the two variables have the predicted signs and are significant at the 1 per cent level. The sign of DE is not as predicted but the variable is significant at 1 per cent level. The results of the regression model are consistent with the income

LLP ROA BEFO TAX AND LOSS CROA LD DE RD Table II. Descriptive statistics for LOGTA all the variables included NPL TYPE in the regression model

N

Minimum

Maximum

Mean

Standard deviation

134 134 134 132 133 134 134 134 134

0.0000
0.0116
0.0898 0.0654 0.6268
15.0488 4.4611 0.0000 0

0.2568 0.13912 0.2181 67.8503 23.3239 4,975.1244 8.0698 0.5178 1

0.0131 0.0223 0.0042 1.3755 7.4819 64.773 6.7743 0.0753 0.86

0.0265 0.0143 0.0252 5.8732 3.9266 520.467 0.7738 0.0853 0.350

smoothing hypothesis but not with the legal reserve and the debt to the equity hypothesis. To test the effect of the type of financial institution (Islamic vs conventional) on loss provision (LLP), the type of institution (TYPE) was included in the regression model. The results, shown in Table IV, reveal that the type of institution has no effect on the amount of loss provision. In other words, Islamic and conventional banks behave in the same manner in dealing with loss provision. To investigate the effects of combining the data for the four years, the regression model was re-estimated for 2003 and 2002 separately[4]. The results of the regression models for 2003 and 2002 are presented in Tables V and VI, respectively. The overall models performance, measured by R2, has improved to 77.4 and 88.1 per cent for the years 2003 and 2002, respectively. However, it is worth noting that the only significant variable in each year (2003 and 2002) was CROA, which is consistent with the income smoothing hypothesis prediction. The major finding to note is that the single year regression model provides a better explanation to the income smoothing hypothesis than the model which includes data for the four years. The R2 for the regression model for the year 2003 (R2 ¼ 77.4 per cent) and the year 2002 (R2 ¼ 77.4 per cent) exceed the R2 for the model for all combined data of all the years (R2 ¼ 23.0 per cent). Thus, the use of the model for each year can explain the loss provision better than a model using the entire time period.

Empirical testing of loss provisions 507

7. Summary and conclusion The findings of this study indicate that managers of banks in the GCC region smooth income via loss provision. The results, as shown in the previous section, LLP

CROA

LD

DE

RD

LOGTA

TYPE

1 0.255*
0.028 0.278* 0.110
0.217** 0.086

0.255* 1 0.090
0.058
0.079 0.021 0.081


0.028 0.090 1 0.022 0.047
0.019
0.234**

0.278*
0.058 0.022 1 0.127 0.133 0.192*

0.110 0.079 0.047 0.127 1
0.030 0.048

0.217** 0.021 -0.019 0.133
0.030 1 0.085

0.086 0.081
0.234** 0.192* 0.048 0.085 1

Table III.

Notes: *Correlation is significant at the 0.01 level (2-tailed); **correlation is significant at the 0.05 level (2-tailed)

Correlations among all the variables that are included in the model

LLP CROA LD DE RD LOGTA TYPE

Coefficient T-statistics R2 Adjusted R2 F-value

Intercept

CROA

LD

DE

RD

LOGTA

TYPE

0.0570 3.021* 0.230 0.193 6.189*

0.3070 3.650*


0.0031
0.854

0.0021 3.907*

0.00044 1.099

0.0090 3.303*

0.00078 0.012

Notes: Model: LLP = CROA+LD+DE+RD+TYPE+LOGTA. * indicates significance at p = 0.01 or better

Table IV. Result of the regression model using data of all banks/years (134 observations)

MF 33,7

508

Table V. Result of the regression model using 2003 data (47 banks)

Table VI. Result of the regression model using 2002 data (47 banks)

indicate that when ROA before tax and loss provision for a current year is higher than a previous year’s ROA, then management is expected to increase loss provision for the current year. The results support the income smoothing hypothesis. However, the results of this study are not consistent with DE hypothesis and the legal reserve hypothesis. The results also indicate that the higher the LD, the lower the loss provision taken in the current year in order to attract more deposits by customers. The type of bank (Islamic or conventional) is not an important factor in the determination of the loss provision, consistent with our prediction. The results presented in the paper lead us to two conclusions. First, significant results for the income smoothing hypothesis are robust to the time sub-periods. The relative ROA of the prior period to the ROA before tax and loss for the current period is a significant factor in determining the amount of loss provision to be charged to the current period. Second, the legal reserve hypothesis was not supported either in the entire period or separate periods. This result suggests that banks in the GCC region do not use loss provision to manage their legal and statutory reserve to meet reserve requirements specified by the banks regulators in the GCC region. The results of this study demonstrate that the income smoothing hypothesis is relevant across different regulatory requirements, economic conditions, and different accounting standards. Managers of banks in the GCC region use the loss provision, among other things, to smooth earnings to achieve certain objectives.

Coefficient T-statistics R2 Adjusted R2 F-value

Intercept

CROA

LD

DE

RD

LOGTA

TYPE

0.0139 0.814 0.774 0.740 22.369*

0.570 10.894*


0.0015
0.395

0.0031 0.634

0.0008 1.121

0.0016 0.769

0.0035 0.814

Notes: Model: LLP = CROA þ LD þ DE þ RD þ TYPE þ LOGTA. * indicates significance at p = 0.01 or better

Coefficient T-statistics R2 Adjusted R2 F-value

Intercept

CROA

LD

DE

RD

LOGTA

TYPE

0.0139 0.814 0.881 0.776 22.369*

0.570 10.894*


0.0015
0.395

0.0031 0.634

0.0008 1.121

0.0016 0.769

0.0035 0.814

Notes: Model: LLP=CROA þ LD þ DE þ RD þ TYPE þ LOGTA. * indicates significance at p ¼ 0.01 or better

Notes 1. GCC region consists of Bahrain, Kuwait, Oman, Qatar, United Arab Emirates, and Saudi Arabia. 2. Haqiqi and Pomeranz (1987) explain the different investment arrangements that are used by Islamic banks: .

.

Mudarabah: The bank acts as a partner to the customer who borrowed the money from the bank. The bank shares in the profits and losses of the business. There is no term for the loan. Murabah: The bank provides the cash for purchasing goods by customer for a share of the profit once the goods are sold back to the customer.

Musharaka: The bank provides part of the capital needed for the business and shares in profits and/or losses. 3. The only country which imposes income tax on banks’ profit is Oman. Banks in other countries may have income tax expense imposed on their overseas operation by other countries. In either case, the net income was adjusted by the income tax expense. 4. The regression model for the years 2000 and 2001 was not estimated due to the small sample size in those two years. .

References Ahmed, A., Takeda, C. and Thomas, S. (1999), ‘‘Bank loan loss provision: a reexamination of capital management, earnings management, and signaling effects’’, Journal of Accounting and Economics, Vol. 28, pp. 1-25. Barth, M., Beaver, W. and Wolfson, M. (1990), ‘‘Components of earnings and the structure of bank share prices’’, Financial Analysts Journal, Vol. 46, pp. 53-60. Beatty, A., Chamberlain, S. and Magliolo, J. (1995), ‘‘Managing financial reports of commercial banks: the influence of taxes, regulatory capital and earnings’’, Journal of Accounting Research, Vol. 33, pp. 231-61. Beaver, W. and Engel, E. (1996), ‘‘Discretionary behavior with respect to allowance for loan losses and behavior of securities prices’’, Journal of Accounting and Economics, Vol. 22, pp. 177-206. Bhat, V. (1996), ‘‘Banks and income smoothing: an empirical analysis’’, Applied Financial Economics, Vol. 6, pp. 505-10. Burgstahler, D. and Dichev, I. (1997), ‘‘Earnings management to avoid earnings decrease and losses’’, Journal of Accounting and Economics, Vol. 24, pp. 99-126. Collins, J., Shackelford, D. and Wahlen, J. (1995), ‘‘Banks differences in the coordination of regulatory capital, earnings, and taxes’’, Journal of Accounting Research, Vol. 33, pp. 263-91. Degeorge, F., Patel, J. and R. and Zechhauser (1999), ‘‘Earnings management to exceed thresholds’’, Journal of Business, Vol. 772, pp. 1-33. Docking, D., Hirschey, M. and Jones, E. (2000), ‘‘Reaction of bank stock prices to loan–loss reserve announcements’’, Review of Quantitative Finance and Accounting, Vol. 15, pp. 277-97. Greenawalt, M. and Sinkey, J. (1988), ‘‘Bank loan-loss provisions and the income-smoothing hypothesis: an empirical analysis, 1976-1984’’, Journal of Financial Services Research, Vol. 1, pp. 301-18. Hasan, I. and Hunter, W. (1994), ‘‘The income smoothing hypothesis: an analysis of thrift industry’’, working paper 94-3, Federal Reserve Bank of Atlanta. Hasan, I. and Wall, L. (2004), ‘‘Determinants of the loan loss allowance: some cross-country comparison’’, The Financial Review, Vol. 39, pp. 129-152.

Empirical testing of loss provisions 509

MF 33,7

510

Healy, P. (1985), ‘‘The effect of bonus schemes on accounting decisions’’, Journal of Accounting and Economics, Vol. 7, pp. 85-107. Hussain, M., Islam, M. Gunasekaran, A. and Maskooki, K. (2002), ‘‘Accounting standards and practices of financial institutions in GCC countries’’, Managerial Auditing Journal, Vol. 17, pp. 350-62. International Accounting Standards Committee (JASC): IAS No. 30. (1991), ‘‘Disclosures in the financial statements of banks and similar financial institutions’’. Kanagertnam, K., Lobo, G. and Mathieu, R. (2004), ‘‘Earnings management to reduce earnings variability: evidence from banks loan loss provisions’’, Review of Accounting and Finance, Vol. 3, pp. 128-48. Kim, M. and Kross, W. (1998), ‘‘The impact of the 1989 change in Bank Capital Standards on Loan Loss Provisions and loan write-off’’, Journal of Accounting and Economics, Vol. 25, pp. 69-99. Kiridaran, K., Lobo, G. and Yang, D. (2003), ‘‘Determinants of signaling by banks through loan loss provisions’’, working paper series, SSRN, pp. 1-37. Lobo, G. and Yang, D. (2001), ‘‘Bank managers’ heterogeneous decisions on discretionary loan loss provisions’’, Review of Quantitative Finance and Accounting, Vol. 16, pp. 223-50. Moyer, S. (1990), ‘‘Capital adequacy ratio regulations and accounting choices in commercial banks’’, Journal of Accounting and Economics, Vol. 13, pp. 123-54. Further reading Elliott, J., Hanna, J. and Shaw, W. (1991), ‘‘The evaluation by the financial markets of changes in bank loan loss reserve levels’’, The Accounting Review, Vol. 66, pp. 847-61. Haqiqi, A.W. and Pomeranz, F. (1987), ‘‘Accounting needs of Islamic banking’’, Advances in International Accounting, Vol. 1, pp. 153-68. Scholes, M., Wilson, G. and Wolfson, M. (1990), ‘‘Tax planning, regulatory capital planning and financial reporting strategy for commercial banks’’, The Review of Financial Studies, Vol. 3, pp. 625-50. Appendix Regulatory reserve requirements of the GCC countries. (1) Bahrain: statutory reserve – 10 per cent of the net profit for each year must be transferred to a statutory reserve until this reserve equals 50 per cent of the share capital. This reserve is not available for distribution. (2) Kuwait: statutory reserve – 10 per cent of the net profit for each year must be transferred to a statutory reserve until this reserve equals 50 per cent of the share capital. Distribution of this reserve is limited to the amount required to enable the payment of dividend of 5 per cent shares capital when accumulated profits are not sufficient for the payment of a dividend of that amount. (3) Oman: legal reserve – According to Oman Commercial Companies Law of 1974, banks must transfer 10 per cent of net profit of the year to legal reserve until this reserve equals one-third of the share capital. This reserve is not available for distribution. (4) Qatar: legal reserve – banks must transfer 20 per cent of net profit of the year to legal reserve until this reserve equals 100 per cent of the capital. This reserve is not available for distribution. (5) Saudi Arabia: statutory reserve – 25 per cent of the net profit for each year must be transferred to a statutory reserve until this reserve equals 100 per cent of the share capital. This reserve is not available for distribution.

(6) UAE: Statutory reserve – 10 per cent of the net profit for each year must be transferred to a statutory reserve until this reserve equals 50 per cent of the share capital. This reserve is not available for distribution. Special reserve – In accordance with the union Law No. 10 of 1980 concerning the central bank of UAE, banks must transfer 10 per cent of the net profit for each year to a special reserve until this reserve equals 50 per cent of the share capital. This reserve is not available for distribution. (7) Islamic banks: follow the reserve rules of the country of incorporation. Corresponding author Taisier A. Zoubi can be contacted at: [email protected]

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

Empirical testing of loss provisions 511

Call for papers Managerial Law: the International Journal of Law and Management Frequency: 6 issues per year Editors: Professor James Kirkbride and Professor Geraint Howells Managerial law: the International Journal of Law and Management is a leading journal addressing all aspects of regulation and law as they impact on organisational development, operations and leadership. Organisations and their leaders operate in an increasingly complex world of emerging regulation across national and international boundaries. Managerial Law seeks to acknowledge the dynamics of that environment and provide a platform for articles and contributions to stimulate scholarly debate in the development of law and practice. Led by a newly recruited expert Editorial Board, Managerial Law presents the latest research on policy, practice and theoretical perspectives and their impact on the development and leadership of organisations. Contributions of a multi-disciplinary nature are welcome – in particular those that appeal to an international readership. Of particular interest are contributions in: .

Employment and industrial law

.

Corporate governance and social responsibility

.

Intellectual property

. .

Corporate law and finance Insolvency

.

Commercial law and consumer protection

.

Environmental law

.

Taxation

.

Competition law

.

Regulatory theory

Papers can be between 2,500 and 15,000 words in length and will undergo a double blind peer review. For more information please refer to the journal’s Author Guidelines at www.emeraldinsight.com/info/journals/ml/notes.htm All articles should be submitted to: James Kirkbride, Dean, Faculty of Business and Law, Liverpool John Moores University, John Foster Building, Liverpool, UK. Email: [email protected]