Middle Eastern Finance and Economics ISSN: 1450-2889 Issue 2 (2008) EuroJournals Publishing, Inc. 2008 http://www.eurojournals.com/finance.

htm

Modelling Volatility using GARCH Models: Evidence from Egypt and Israel
Christos Floros Lecturer in Banking and Finance in the Department of Economics University of Portsmouth, Portsmouth Business School, Portsmouth, PO1 3DE, UK E-mail: christos.floros@port.ac.uk Tel: +44 (0) 2392 844244. Abstract This paper examines the use of GARCH-type models for modelling volatility and explaining financial market risk. We use daily data from Egypt (CMA General index) and Israel (TASE-100 index). Various time series methods are employed, including the simple GARCH model, as well as exponential GARCH, threshold GARCH, asymmetric component GARCH, the component GARCH and the power GARCH model. We find strong evidence that daily returns can be characterised by the above models. For both markets, we conclude that increased risk will not necessarily lead to a rise in the returns. The most volatile series is CMA index from Egypt, because of the uncertainty in prices (and economy) over the examined period. These findings are strongly recommended to financial managers and modellers dealing with international markets. Keywords: Middle East stock markets, Israel, Egypt, GARCH, volatility. JEL Classification Codes: C13, C32, C52, G15.

I. Introduction
There has been considerable volatility (and uncertainty) in the past few years in mature and emerging financial markets worldwide. Most investors and financial analysts are concerned about the uncertainty of the returns on their investment assets, caused by the variability in speculative market prices (and market risk) and the instability of business performance (Alexander, 1999). Recent developments in financial econometrics require the use of quantitative models that are able to explain the attitude of investors not only towards expected returns and risks, but towards volatility as well. Hence, market participants should be aware of the need to manage risks associated with volatility. This requires models that are capable of dealing with the volatility of the market (and the series). Due to unexpected events, uncertainties in prices (and returns) and the non-constant variance in the financial markets, financial analysts started to model and explain the behaviour of stock market returns and volatility using time series econometric models. One of the most prominent tools for capturing such changing variance was the Autorgressive Conditional Heteroskedasticity (ARCH) and Generalized ARCH (GARCH) models developed by Engle (1982), and extended by Bollerslev (1986) and Nelson (1991). Two important characteristics within financial time series, the fat tails1 and volatility clustering (or volatility pooling), can be
1

(G)ARCH processes appear to fit the empirical data of stock returns, since they can have sharp modes and fat tails, i.e. they can exhibit different degrees of leptokurtosis for the same variances.

Middle Eastern Finance and Economics - Issue 2 (2008)

32

captured by the GARCH family models. A series with some periods of low volatility and some periods of high volatility is said to exhibit volatility clustering. Volatility clustering can be thought of as clustering of the variance of the error term over time: if the regression error has a small variance in one period, its variance tends to be small in the next period, too. In other words, volatility clustering implies that the error exhibits time-varying heteroskedasticity (unconditional standard deviations are not constant). In this paper, we capture financial time series characteristics by employing GARCH(p,q) model, and its EGARCH, Threshold GARCH (TGARCH), Asymmetric component (AGARCH), Component GARCH (CGARCH) and Power GARCH (PGARCH) extensions. These models have the advantage of permitting investigation of the potentially asymmetric nature of the response to past shocks. Several studies investigate the performance of GARCH models on explaining volatility of mature stock markets (e.g. Sentana and Wadhwani, 1992; Kim and Kon, 1994; Kearney and Daly, 1998; Floros, 2007; Floros et al., 2007), but few have tested GARCH models using daily data from Middle East stock markets. Mecagni and Sourial (1999) examine the behaviour of stock returns as well as the market efficiency and volatility effects in the Egyptian stock exchange using GARCH models. The results show significant departures from the EMH, tendency for returns to exhibit volatility clustering and a significant positive link between risk and returns. Furthermore, Tooma (2003) investigate the impact of price limits on volatility dynamics in the Egyptian Stock Exchange using several GARCH models under four different error distributions (Normal, Student-t, GED and Skewedt). The data covers the period 1993 2001. The empirical results suggest significant changes in the time varying volatility process. Alberg et al. (2006) estimate stock market volatility of Tel Aviv Stock Exchange indices, for the period 1992-2005, using asymmetric GARCH models. They report that the EGARCH model is the most successful in forecasting the TASE indices. Finally, Meric et al. (2007) study the co-movements of the US, UK and Middle East stock markets (Egyptian, Israeli and Turkish) during the period 1996 to 2006, and report a very low correlation. They also present the average weekly returns and volatility of returns. According to Meric et al. (2007), Israelis average weekly returns is 0.24%, while the Egyptian stock market has a high average weekly return (0.45%). Also, the Egyptian stock market has the highest return per unit of volatility risk (0.1%), while the Israeli stock market has the lowest return per unit of volatility risk (0.06%) over the examined period. The purpose of this paper is twofold: (i) to explain volatility modelling using recent daily data from Middle East emerging markets, and (ii) to evaluate the performance of the GARCH-family models in explaining financial market risk. The analysis focuses on two Middle East stock indices: the Egyptian CMA index and the Israeli TASE-100 index. The main reason we consider data from emerging markets of Middle East is because they are continue to be of empirical interest from practitioners and investors. According to Slater (2007), the Middle East region is a new frontier for global investors. The motivation for our paper is to add new evidence from two Middle East stock markets to the modelling of financial time series by explaining volatility clustering in these markets. It is not only important to understand the changes of prices of emerging financial markets over time, but also the process by which financing decisions through volatility modelling are reached. The paper is organised as follows: Section II provides data information. Section III presents the methodology, while Section IV presents the main empirical results. Finally, Section V concludes the paper and summarises our findings.

II. Data
The data employed in this study comprise 1987 daily observations on the Egyptian stock market (CMA General index) covering the period 2/7/1997 21/8/2007, and 2063 daily observations on the Israeli stock market (TASE-100 index) covering the period 1/7/1997 21/8/2007. Closing prices for stock

33

Middle Eastern Finance and Economics - Issue 2 (2008)

indices were obtained from Datastream International. Table 1 gives the descriptive statistics for daily stock market prices and returns. Daily returns are computed as logarithmic price relatives: Rt = ln(Pt / Pt 1 ) , where Pt is the daily price at time t. Figure 1 (TASE-100 index) and Figure 2 (CMA index) present the plots of price indices and returns over time. Furthermore, we present four statistics which are calculated using the observations in the full sample: Skewness, Kurtosis, Jarque-Bera and Augmented Dickey Fuller (ADF). Both price series have positive skewness implying that the distribution has a long right tail. On the other hand, the return series have negative skewness implying that the distribution has a long left tail. The values for kurtosis are high (close to three) in all cases. So, the distributions are peaked relative to normal. The JarqueBera test rejects normality at the 5% level for all distributions. So, the samples have all financial characteristics: volatility clustering and leptokurtosis. The daily returns for both indices (presented in Figure 1 and Figure 2) show that volatility occurs in bursts. Furthermore, in terms of stationarity, the results from the ADF tests indicate that both series are I(1), and therefore, time-series models can be used to examine the behaviour of volatility over time.
Figure 1: Plot of daily prices and returns for TASE-100 stock index (1997 2007)
Price
1,200 TASE-100 stock index

1,000

800

600

400

200 250 500 750 1000 1250 1500 1750 2000

Returns
.08 TASE-100 stock index

.04

.00

-.04

-.08

-.12 250 500 750 1000 1250 1500 1750 2000

Middle Eastern Finance and Economics - Issue 2 (2008)

Figure 2: Plot of daily prices and returns for CMA stock index (1997 2007)

34

Price
3,000 CMA stock index

2,500

2,000

1,500

1,000

500

0 250 500 750 1000 1250 1500 1750

Returns
.3 CMA stock index

.2

.1

.0

-.1

-.2

-.3 250 500 750 1000 1250 1500 1750

35
Table 1: Descriptive Statistics & ADF Tests

Middle Eastern Finance and Economics - Issue 2 (2008)

A. Prices Mean Median Maximum Minimum Std. Dev. Skewness Kurtosis Jarque-Bera Probability Observations ADF (Level) ADF (1st diff.) B. Returns Mean Median Maximum Minimum Std. Dev. Skewness Kurtosis Jarque-Bera Probability Observations
Notes:

TASE-100 index 541.9466 463.2900 1150.780 249.1900 227.3553 0.884522 2.758381 274.0265 0.000000 2063 -0.576881 -11.64264 TASE-100 index 0.000606 0.000347 0.076922 -0.103816 0.014365 -0.381173 6.916253 1367.640 0.000000 2062

CMA General index 996.1721 635.5300 2902.370 326.7200 705.8703 1.194585 3.004005 472.5868 0.000000 1987 0.765703 -19.98205 CMA General index 0.001015 0.000661 0.271107 -0.290292 0.013269 -1.109135 207.6767 3467014. 0.000000 1986

Skewness is a measure of asymmetry of the distribution of the series around its mean. Kurtosis measures the peakedness or flatness of the distribution of the series. Jarque-Bera is a test statistic for testing whether the series is normally distributed. ADF regressions include intercept but not trend. We employ ADF test on the logarithms of stock indices. ADF critical values: (1%) 3.4334, (5%) 2.8627, (10%) 2.5674.

III. Methodology
In financial markets, fluctuation of prices (or returns) goes under the name of volatility - how much prices (or returns) are changing over a given period. Linear models are unable to explain a number of important features common to much financial data, including leptokurtosis, volatility clustering, long memory, volatility smile and leverage effects. That is, because the assumption of homoscedasticity (or constant variance) is not appropriate when using financial data, and in such instances it is preferable to examine patterns that allow the variance to depend upon its history. Therefore, to model the nonconstant volatility parameter, we consider GARCH-type models. Bollerslev (1986) proposed a GARCH(p,q) random process, which can represent a greater degree of inertia in its conditional volatility or risk. Following the literature (Akgiray, 1989; Connolly, 1989; Baillie and DeGennaro, 1990; Bera and Higgins, 1993; Bollerslev et al., 1992; Floros, 2007, among others), a simple GARCH model is parsimonious and gives significant results. GARCH allows the conditional variance of a stock index to be dependent upon previous own lags. The GARCH (p,q) model is given by: Rt = + t

t2 = + i t2i + j t2 j
i =1 j =1

(1)

where p is the order of GARCH while q is the order of ARCH process. Error, t , is assumed to be normally distributed with zero mean and conditional variance, t2 . Rt are returns, so we expect their mean value () to be positive and small. We also expect the value of to be small. All parameters in variance equation must be positive, and + is expected to be less than, but close to, unity, with >.

Middle Eastern Finance and Economics - Issue 2 (2008)

36

News about volatility from the previous period can be measured as the lag of the squared residual from the mean equation (ARCH term). Also, the estimate of shows the persistence of volatility to a shock or, alternatively, the impact of old news on volatility. Financial theory suggests that an increase in variance results in a higher expected return. To account for this, GARCH-in-Mean models are also considered, see Kim and Kon (1994). Standard GARCH-M model is given by: Rt = + 2 t2 + t

t ~ N (0, t2 )

(2)

t2 = + a t21 + 1 t21 if 2 is positive (and significant), then increased risk leads to a rise in the mean return ( 2 t2 can be interpreted as a risk premium). Exponential-GARCH models were designed to capture the leverage effect noted in Black (1976) and French et al. (1987). A simple variance specification of EGARCH is given by: (3) log t2 = + log t21 + a t 1 + t 1 t 1 t 1 The logarithmic form of the conditional variance implies that the leverage effect is exponential (so the variance is non-negative). The presence of leverage effects can be tested by the hypothesis that < 0 . If 0 , then the impact is asymmetric. Furthermore, the Threshold-GARCH model was introduced by Zakoian (1994) and Glosten, Jaganathan and Runkle (1993). The TGARCH specification for the conditional variance is given by:

t2 = + ai t2i + t21 d t 1 + j t2 j
i =1 j =1

(4)

where d t = 1 if t < 0 and d t = 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 a , while bad news has an impact of a + . If > 0 then the leverage effect exists and bad news increases volatility, while if 0 the news impact is asymmetric. An alternative specification for the conditional volatility process is Component-GARCH. The conditional variance in the CGARCH(1,1) model is given by (5.1):

t2 = + a( t21 ) + ( t21 )

(5.1) (5.2)

t2 qt = a( t21 qt 1 ) + ( t21 qt 1 )
qt = + (qt 1 ) + 1 ( t21 t21 )

(5.3) The component model shows mean reversion to (constant over time), while it allows mean reversion to a varying level qt , see (5.2) and (5.3). In equations (5.2) and (5.3), t is volatility and qt is the time varying long run volatility. Equation (5.2) describes the transitory component, t2 q t , while equation (5.3) describes the long run component qt . An extension of CGARCH model is Asymmetric component GARCH. The AGARCH model combines the component model with the asymmetric TGARCH model. This specification introduces asymmetric effects in the transitory equation. The AGARCH model is given by: Rt = xt + t
qt = + (qt 1 ) + ( t21 t21 ) + 1 z1t

(6)

t2 qt = ( t21 qt 1 ) + 1 ( t21 qt 1 )d t 1 + 2 ( t21 qt 1 ) + 2 z 2t

37

Middle Eastern Finance and Economics - Issue 2 (2008)

where z1t and z 2t are the exogenous variables and d is the dummy variable indicating negative shocks. 1 > 0 implies transitory leverage effects in the conditional variance. Finally, Taylor (1986) and Schwert (1989) introduced the standard deviation GARCH model, where the standard deviation is modelled rather than the variance. This model is generalized in Ding et al. (1993) with the Power ARCH specification. In the Power-GARCH model, the power parameter 1 of the standard deviation can be estimated rather than imposed, and the optional parameters are added to capture asymmetry of up to order r :
t = + j t j + ai ( t i i t i )
q p
1 1

(7)

j =1

i =1

where 1 > 0, i 1 for i = 1...r , i = 0 for all i > r, and r p. The symmetric model sets i = 0 for all i . Note that if 1 = 2 and i = 0 for all i , the PGARCH model is simply a standard GARCH specification. As in the previous models, the asymmetric effects are present if 0 .

IV. Empirical Results

First, we filter conditional mean structure in the data by estimating ARMA(p,q) models. The dependent variable is returns, Rt . AR(p) and MA(q) orders are determined by reference to the Akaike information criterion (AIC). For both indices, we select an ARMA(1,1) model. The results from mean equations (not presented here) show significant parameters, indicating that lagged values of returns and lagged errors depend on the current returns of stock indices. Furthermore, we estimate a number of different GARCH-family models to explain conditional variance and volatility clustering. An iterative procedure is used based upon the method of Marquardt algorithm. Heteroskedasticity Consistent Covariance (HCC) option is used to compute quasi-maximum likelihood (QML) covariances and standard errors using the methods described by Bollerslev and Wooldridge (1992). This is normally used if the residuals are not conditionally normally distributed. Table 2 reports the parameter estimates of all conditional volatility (GARCH-family) models defined in the previous section. For both indices, the sum of ARCH and GARCH coefficients is very close to one, indicating that volatility shocks are quite persistent. The coefficient of the lagged squared returns is positive and statistically significant for most specifications. We conclude that strong GARCH effects are apparent for both financial markets. Also, the coefficient of lagged conditional variance is significantly positive and less than one, indicating that the impact of old news on volatility is significant. The magnitude of the coefficient, , is especially high for TASE-100 index, indicating a long memory in the variance. In addition, the coefficients of the conditional variance in the mean equation of GARCH-M models, denoted as 2 , are positive but insignificant for both indices. This suggests that higher market-wide risk, proxied by the conditional variance, will not necessarily lead to higher returns. Furthermore, EGARCH models show a negative and significant parameter for both indices, indicating the existence of the leverage effect in returns during the sample periods. However, the TGARCH leverage effect term is not significant in the case of CMA, while the news impact is asymmetric (for both markets). For TASE, the leverage effect exists and bad news increases volatility. In addition, bad news has an impact of 0.2263 (TASE) and 2.1071 (CMA). Also, the estimate of a is smaller than the estimate of in both cases, which implies that negative shocks havent a larger effect on conditional volatility than positive shocks of the same magnitude. The results of the estimation of CGARCH(1,1) and AGARCH(1,1) models show mixed findings for both indices. AGARCH models show weak transitory leverage effects in the conditional variances. However, the estimates of the persistence in the long run component are significant and close to unity, indicating that the long run component converges very slowly to the steady state. Short run volatility is governed by 0.8587 (TASE-100) and 0.6492 (CMA).

Middle Eastern Finance and Economics - Issue 2 (2008)

38

The results of the estimation of PGARCH(1,1) models confirm that the asymmetric effects are present for TASE-100 index (the asymmetric parameter is positive and significant). For CMA, only the ARCH and GARCH parameters are significant. Finally, the mean values of the volatility (GARCH variance series) from the above GARCH models are 0.02% and 0.05% for TASE-100 and CMA, respectively. Hence, the Egyptian market shows higher volatility (variance) than the Israeli market over the examined period. This is in line with Meric et al. (2007).
Table 2: GARCH-family Models for Volatility (Variance Specifications)

INDEX/MODEL PART A. TASE-100 INDEX ARMA(1,1)1.63e-06 GARCH(1,1) (3.1852)* ARMA(1,1)-1.1871 EGARCH(1,1) (-3.9507)* ARMA(1,1)2.58e-05 TGARCH(1,1) (4.4862)* ARMA(1,1)0.0001 CGARCH(1,1) (1.6680) ARMA(1,1)0.0002 AGARCH(1,1) (7.4768)* ARMA(1,1)1.63e-05 GARCH(1,1)-M (3.1838)* ARMA(1,1)0.0007 PGARCH(1,1) (0.6098) PART B. CMA INDEX ARMA(1,1)7.51E-06 GARCH(1,1) (3.9245)* ARMA(1,1)-3.6945 EGARCH(1,1) (-16.0631)* ARMA(1,1)9.53E-06 TGARCH(1,1) (3.4169)* ARMA(1,1)0.0844 CGARCH(1,1) (0.1005) ARMA(1,1)0.0001 AGARCH(1,1) (1.2770) ARMA(1,1)0.0001 GARCH(1,1)-M (446.0587)* ARMA(1,1)0.0042 PGARCH(1,1) (0.3882)

a
0.1098 (3.2787)* 0.2450 (4.1583)* 0.0340 (1.3942) 0.1029 (2.4912)* 0.9189 (42.5220)* 0.1095 (3.2597)* 0.1320 (4.1319)* 0.5819 (4.4848)* 1.2571 (6.7167)* 0.8476 (3.1826)* 0.24117 (2.8375)* 0.5410 (1.7709) 0.6998 (1.3264) 0.2628 (2.2812)*

0.8130 (16.7740)* 0.8833 (28.2098)* 0.7456 (16.1609)* 0.9990 (515.7019)* -0.0740 (-2.6502)* 0.8131 (16.7301)* 0.7736 (14.0724)* 0.3493 (1.9686)* 0.6920 (30.4891)* 0.3613 (2.1179)* 0.9999 (99946)* 0.2357 (0.6621) -0.0333 (-1.0543) 0.7800 (20.6920)*

-0.1291 (-3.6702)* 0.1923 (3.7282)* 0.7558 (8.4741)* 0.12650 (3.5827)*

0.0080 (1.7293) 0.0531 (1.4445)

-0.8346 (-10.6764)* 0.0522 (0.5008)

0.54490 (3.2194)*

1.1801 (2.9332)*

-0.3073 (-2.5295)* 1.2595 (0.6275) 0.4081 (1.4711) 0.2828 (2.8863)*

0.0390 (0.4614) -0.2671 (-1.9845)*

0.0292 (0.0824) 0.2089 (0.9138)

-0.1233 (-0.9236)

0.3865 (0.6449)

Notes: We report the results from GARCH-type models using the method of maximum likelihood, under the assumption that the errors are conditionally normally distributed. T-statistics in the parentheses * Significant at the 5% level

V. Conclusions
The lognormality of the asset price distribution is not a satisfactory assumption. In fact, it is well documented that equity prices (and returns) do not follow such a distribution. Stylized facts about the distribution of returns have been reported in Bollerslev et al. (1994) and Pagan (1996). They include: leptokurtosis, leverage effects, volatility clustering (or pooling), volatility smile and long memory. Research examining financial returns has raised the question of whether GARCH-family models are able to capture volatility clustering. The results reported in this paper address this issue by providing estimates from symmetric and asymmetric GARCH models for daily price returns. This is a

39

Middle Eastern Finance and Economics - Issue 2 (2008)

full empirical investigation of financial market volatility using daily data from two Middle East stock indices, namely the Egyptian CMA index and the Israeli TASE-100 index. In summary of our results, we find strong evidence that daily returns can be characterised by the GARCH models. The sum of the GARCH coefficients is close to one in almost all cases. That implies persistence of the conditional variance. A large sum of the coefficients in the conditional variance equations implies that a large positive or a large negative return will lead future forecasts of the variance to be high. For TASE-100, we report that the leverage effect exists and bad news increases volatility. Finally, the estimates of the persistence in the long run component are significant, indicating that the long run component converges very slowly to the steady state. Furthermore, volatility (as measured by the variance or standard deviation of returns) is examined as a measure of the total financial risk. For both markets, we conclude that increased risk will not necessarily lead to a rise in the returns. The most volatile series is CMA index from Egypt, because of the uncertainty in prices (and economy) over the examined period. This finding is in line with Meric et al. (2007). Hence, the fluctuation of volatility in emerging markets is as important as in mature markets. These findings are strongly recommended to financial managers and modellers dealing with Middle East stock markets. Future research should examine the performance of multivariate time series models when using daily returns of international mature and emerging markets.

Middle Eastern Finance and Economics - Issue 2 (2008)

40

References
[1] [2] Akgiray, V., (1989) Conditional heteroscedasticity in time series of stock returns, Journal of Business, Vol 62, pp. 55-80. Alberg, D., Shalit, H., and Yosef, R. (2006) Estimating stock market volatility using asymmetric GARCH models, Discussion Paper No. 06-10, Monaster Center for Economic Research, Ben-Gurion University of the Negev, Israel. Alexander, C. (1999) Risk Management and Analysis,Volume 1: Measuring and Modelling Financial Risk, John Wiley and Sons, New York, NY. Baillie, R. T. and DeGennaro, R. P. (1990) Stock return and volatility, Journal of Financial and Quantitative Analysis, Vol 25, pp. 203-214. Bera, A. K. and Higgins, M. L. (1993) ARCH models: properties, estimation and testing, Journal of Economic Surveys, Vol 7, pp. 305-362. Black, F. (1976), Studies of stock price volatility changes, in Proceedings of the 1976 Meetings of the Business and Economics Statistics Section, American Statistical Association, pp. 177-81. Bollerslev, T. (1986) Generalised autoregressive conditional heteroscedasticity, Journal of Econometrics, Vol 31, pp. 307-27. Bollerslev, T., Chou, R. Y. and Kroner, K. F. (1992) ARCH modelling in finance, Journal of Econometrics, Vol 52, pp. 5-59. Bollerslev, T., Engle, R. and Nelson, D. (1994), ARCH models, in Engle, R. and McFadden, D. (Ed.), Handbook of Econometrics, Vol. 4, North Holland, Amsterdam. Bollerslev, T. and Wooldridge, J. M. (1992) Quasi maximum likelihood estimation and inference in dynamic models with time varying covariances, Econometric Reviews, Vol 11, pp. 143-179. Connolly, R. A. (1989) An examination of the robustness of the weekend effect Journal of Financial and Quantitative Analysis, Vol 24, pp. 133-69. Ding, Z., Granger, C. W. J., and R. E., Engle (1993) A Long Memory Property of Stock Market Returns and a New Model, Journal of Empirical Finance, 1, pp. 83 - 106. Engle, R. F. (1982) Autoregressive conditional heteroscedasticity with estimates of the variance of UK inflation, Econometrica, Vol 50, pp. 987-1008. Floros, C. (2007) The use of GARCH models for the calculation of Minimum Capital Risk Requirements: International Evidence, International Journal of Managerial Finance, Vol. 3(4), pp. 360-371. Floros, C., Jaffry, S. and G.V. Lima (2007), Long memory in the Portuguese Stock Market, Studies in Economics and Finance, Vol. 24 (3), pp. 220-232. French, K. R., Schwert, G. W. and Staumbaugh, R. F. (1987) Expected stock returns and volatility, Journal of Financial Economics, Vol 19, pp. 3-29. Glosten, L., Jaganathan, R. and Runkle, D. (1993) Relationship between the expected value and volatility of the nominal excess returns on stocks, Journal of Finance, Vol 48, pp. 1779802. Kearney, C. and Daly, K. (1998) The causes of stock market volatility in Australia, Applied Financial Economics, Vol 8, pp. 597-605. Kim, D. and Kon, S. I. (1994) Alternative Models for Conditional Heteroscedasticity of Stock Returns, Journal of Business, Vol 67, pp. 563-598. Mecagni, M. and Sourial, M. S. (1999) The Egyptian stock market: Efficiency tests and volatility effects, IMF Working paper No. 99/48. Meric, G., Ratner, M., and Meric, I. (2007) Co-movements of the U.S., U.K., and Middle East Stock markets, Middle Eastern Finance and Economics, Issue 1, pp. 60-73. Nelson, D. (1991) Conditional heteroscedasticity in asset returns: a new approach, Econometrica, Vol 59, pp. 347-70.

[3] [4] [5] [6]

[7] [8] [9] [10]

[11] [12] [13] [14]

[15] [16] [17]

[18] [19] [20] [21] [22]

41 [23] [24] [25] [26] [27] [28] [29]

Middle Eastern Finance and Economics - Issue 2 (2008) Pagan, A. (1996) The econometrics of financial markets, Journal of Empirical Finance, Vol 3, pp. 15-102. Schwert, G. W. (1989) Why does stock market volatility change over time?, Journal of Finance, Vol 54, pp. 1115-51. Sentana, E. and Wadhwani, S. (1992) Feedback traders and stock return autocorrelations: evidence from a century of daily data, Economic Journal, Vol 102, pp. 415-425. Slater, J. (2007) The New Frontier: Mideast, Heard on the Street Column in Section C1 Money & Investing, in Wall Street Journal. Taylor, S. (1986) Modelling Financial Time Series, John Wiley & Sons, New York. Tooma, E. A. (2003) Modeling and forecasting Egyptian stock market volatility before and after price limits, Working Paper 0310, Economic Research Forum, Cairo, Egypt. Zakoian, J. M. (1994) Threshold heteroscedastic models, Journal of Economic Dynamics and Control, Vol 18, pp. 931-955.