Economic subjects | Finance » Abid-Bahloul-Mroua - Financial Development and Economic Growth in MENA countries

Source: http://www.doksinet Financial Development & Economic Growth in MENA countries Fathi Abid a, Slah Bahloul b, Mourad Mroua c a b Professor of Finance, University of Sfax, Faculty of Economics and Business, Road of the Airport 4, 3018, Sfax, Tunisia E-mail: Fathi.Abid@fsegsrnutn Associate Professor of Finance, University of Sfax, Higher School of Business Administration, Road of the Airport 4, 3018, Sfax, Tunisia E-mail: SlahBahloul@gmail.com c Associate professor of Finance, University of Sfax, Institute of High Business Studies, Road Sidi Mansour Km 10,B.P 43 – 3061 Sfax, Tunisia E-mail: mroua mourad@yahoo.fr 1 Source: http://www.doksinet Abstract The aim of this paper is twofold. It is first to evaluate the comparative performance of ten MENA (Middle East and North Africa) countries according to GDP growth and stock market return indicators using the non-parametric stochastic dominance approach. We will then use a multivariate vector autoregressive (VAR) model

to investigate the shock transmission from the most performing country to other dominated countries. Results, over the period from June 2005 to December 2013, show that the dominance of MENA countries differs across indicators. Qatar and Morocco are respectively the most performing countries according to GDP growth and stock market return. Furthermore, the GDP growth response to Qatar GDP growth shock is statistically significant for all countries while, the stock market response to Morocco stock market shock is insignificant in Qatar, Saudi Arabia, and UAE. The response time of stock market return is also short when it is significant Finally, this finding shows that the GDP growth dominance has the greater effect than the stock market dominance. Keywords: GDP growth, stock market return, stochastic dominance, VAR model, shock, generalized impulse response. JEL classification: G10, O40. 2 Source: http://www.doksinet 1. Introduction The economic performance measured by the GDP

growth rate of one country is generally perceived as the cause but sometimes the consequence of the financial performance measured by the stock market return of the same country. With the wave of financial liberalization, the economies of the MENA region have become increasingly integrated with each other and exhibit different performance levels and respond differently to shocks in times series especially in crisis times. That is why it is important for each country in the MENA region to know its strengths and weaknesses in terms of dominance and response to shocks compared to other countries in the same region, for better arrangements and negotiations. Contrarily to other regions, there are very few papers that evaluate the comparative economic and stock market performance of the MENA countries. For instance, Ramanathan (2006) uses data envelopment analysis (DEA) to evaluate the comparative performance of eighteen MENA countries according to several economic, educational and health

attributes and identifies four of them as the most efficient: Bahrain, Jordan, Kuwait, and the United Arab Emirates (UAE). Cheng, Jahan-Parvar and Rothman (2010) analyze excess market returns in nine MENA countries within the context of three variants of the Capital Asset Pricing Model (CAPM): static international CAPM; constant-parameter intertemporal CAPM; and Markov-switching intertemporal CAPM. They find that investment in most of the Arabic MENA markets provides returns uncorrelated with global markets, and thus would serve as financial instruments with which portfolio diversification could be improved. Andreano et al (2013) analyze the economic growth in the MENA countries over the period 1950-2007, with particular emphasis on the convergence process in terms of long-term trend of per-capita GDPs and confirm the hypothesis of conditional convergence. The objective of this paper is twofold. First, we evaluate the comparative GDP growth and stock market return of ten MENA countries

(Bahrain, Jordan, Kuwait, Morocco, Oman, Saudi Arabia, Tunisia, and UAE) over the period from June 2005 to December 2013. We use a nonparametric stochastic dominance (SD) approach based on Davidson and Duclos (2000 -DD) model. The latter is one of the most powerful tests (Lean, Wong, and Zhang, 2008) This approach provides a general framework for assessing alternative choice without any assumption on the distribution, and it satisfies the general utility function and takes into consideration all the distributional moments in the comparison (Hadar and Russel, 1969). Second, based on the stochastic dominance relationship results, we investigate the impact of a shock in the volatility of GDP growth or stock market return of the most performing country on the others dominated MENA countries by using a multivariate vector auto-regression (VAR) methodology. Several studies investigate the causality direction between the financial and economic indicators and the response of the GDP growth to

a probably stock market volatility shocks and vice versa. For example, Kar, Nazlıoğlu and Ağır (2011) investigate empirically the direction of causality between financial development and economic growth in the MENA countries using the panel Granger causality testing procedure for fifteen MENA countries over the period 1980–2007. They show that there is no clear consensus on the direction of causality between financial development and economic growth. Using the multivariate VAR methodology, Beetsma and Giuliodori (2012) show that the macroeconomic response pattern to stock market volatility shocks has changed substantially and the negative response of GDP growth to such shocks has become smaller over time. Ben Jedidia et al. (2014) study the impact of financial development on economic growth in Tunisia over the period 1973-2008 and find that neither stock market development nor the interventions of banks in the stock market have positive effects on the economic growth. Results

show that the dominance between MENA countries differs across indicators. Qatar and Morocco are respectively the most performing countries according to GDP growth and stock market return. Furthermore, the GDP growth response to Qatar GDP growth shock is 3 Source: http://www.doksinet statistically significant for all countries while, the stock market response to Morocco stock market shock is insignificant in Qatar, Saudi Arabia, and UAE. The response time of stock market return is also short when it is significant. The remainder of this paper is organized as follows: Section 2 describes the data and the methodology. Section 3 presents and analyses the results Finally, section 4 concludes the paper and gives recommendations to policy makers. 2. Data and methodology 2.1 Data description Our data set concerns monthly stock market price index in US dollar and annual GDP in current US dollar for 10 MENA countries; namely, Bahrain, Jordan, Kuwait, Lebanon, Morocco, Oman, Qatar, Saudi

Arabia, Tunisia and UAE. The other Arabian countries of the region are not considered due to insufficient observations during the whole period. Annual GDP are converted to monthly data using cubic spline interpolation. The sample period is from June 2005 to December 2013. Data on stock market indexes are from the Morgan Stanley Capital International. GDP data are from the World Bank databases Monthly index return and monthly GDP growth are computed on a continuous basis as the difference in logarithm between two consecutive observations. Descriptive statistics for each series of monthly returns and GDP growth include mean, standard deviation, skewness, Kurtosis, Jarque–Bera test, and Augmented Dickey–Fuller (ADF) test are reported in table 1. Table1 Descriptive statistics for return and GDP growth series (June 2005-December 2013) Bahrain Jordan Panel A: Stock market indices returns Mean -0.0199 -00098 Std. Dev 0.0751 00602 Skew -1.1921 -09484 Kurt 6.9747 60875 J-B 92.1986 563540

Prob. 0.0000 00000 ADF (t-Statistic) -6.3226 -83107 ADF (Prob) 0.0000 0.0000 Panel B: GDP growth Mean 0.0081 00102 Std. Dev 0.0079 00050 Skew -1.1886 11962 Kurt 4.0192 33746 J-B 28.7120 251649 Prob. 0.0000 00000 ADF (t-Statistic) -4.3311 -40484 ADF (Prob) 0.0007 00018 Kuwait Lebanon Morocco Oman Qatar S. Arabia Tunisia UAE -0.0020 0.0700 -0.4575 3.9184 7.2131 0.0271 -7.0606 0.0000 0.0026 0.0917 0.7377 7.0845 80.9397 0.0000 -8.4870 0.0000 0.0040 -0.0031 00017 0.0630 0.0640 00873 0.0442 -1.7741 -06089 3.7808 107147 51133 2.6501 3094558 255312 0.2658 0.0000 00000 -9.4209 -4.3909 -96181 0.0000 0.0006 00000 -0.0028 00034 -00048 0.0882 00528 01104 -0.8223 00875 -03981 4.2102 64145 47960 17.8934 501682 165639 0.0001 00000 00003 -8.9249 -94955 -80169 0.0000 00000 00000 0.0093 0.0171 -1.2031 3.8548 27.9837 0.0000 -4.1690 0.0012 0.0072 0.0052 0.5686 2.1577 8.5945 0.0136 -3.0747 0.0317 0.0057 0.0056 -0.1752 2.0597 4.3218 0.1152 -4.1397 0.0013 0.0093 0.0122 -0.8687 3.3540 13.4913

0.0012 -4.8869 0.0001 0.0105 00167 0.0149 00152 -0.5792 -07774 3.7198 30221 7.9820 103764 0.0185 00056 -5.5840 -43451 0.0000 00007 0.0039 00090 0.0050 00119 0.4608 -14092 2.7230 44200 3.9740 427459 0.1371 00000 -3.5724 -44273 0.0080 00005 Panel A of table 1 shows that the mean stock market return is negative for most countries. This mean of return is positive only for Lebanon, Morocco, Qatar and Tunisia. Morocco has the highest monthly mean returns (0.4 percent) Standard deviation varies from 528 percent for Tunisia to 11.04 percent for UAE Referring to the Skewness and Kurtosis coefficients and J-B statistics, the hypothesis of normality is rejected for all countries, except Morocco. The t-statistics of the ADF test are significant at 1% level implying that all stock market return series are stationary. Panel B of table 1 shows that Qatar has the highest mean GDP growth (1.67 percent) followed by Oman, Jordan, Kuwait, Saudi Arabia, UAE, Bahrain, Lebanon, Morocco, and Tunisia.

Tunisia has the lowest GDP growth risk level followed by Jordan, Lebanon, Morocco, Bahrain, UAE, Saudi Arabia, Oman, Qatar, and Kuwait. Based on Skewness and Kurtosis coefficients and J-B statistics, the hypothesis of normality is rejected for all markets 4 Source: http://www.doksinet except for Morocco and Tunisia. The ADF t-statistics show that all the GDP growth series are stationary. The correlation matrices between GDP growth and stock market return of the ten MENA countries are summarized in table 2. Table 2 Correlation matrices (June 2005-December 2013) Bahrain Jordan Kuwait Lebanon Morocco Panel A: Stock market return Bahrain 1 0.3288 05762 0.3135 0.3326 Jordan 1.0000 03343 0.3361 0.2818 Kuwait 1.0000 0.2472 0.3467 Lebanon 1.0000 0.3296 Morocco 1.0000 Oman Qatar S. Arabia Tunisia UAE Panel B: GDP growth Bahrain 1.0000 0.5708 09214 -0.4949 0.5421 Jordan 1.0000 04944 0.1552 0.7061 Kuwait 1.0000 -0.5535 0.4326 Lebanon 1.0000 0.0820 Morocco 1.0000 Oman Qatar S. Arabia Tunisia

UAE Panel C: GDP growth and stock market return Bahrain 0.1044 -0.1038 0.0551 -0.1608 -0.0220 Jordan -0.1197 0.0304 -0.0609 -0.0457 Kuwait -0.0302 -0.1474 -0.1259 Lebanon -0.0163 -0.0273 Morocco 0.0888 Oman Qatar S. Arabia Tunisia UAE Oman Qatar S. Arabia Tunisia UAE 0.4805 0.4265 0.5375 0.3683 0.2567 1.0000 0.4095 0.5159 0.4341 0.3098 0.1699 0.6006 1.0000 0.3783 0.4220 0.3053 0.3293 0.1733 0.4811 0.5392 1.0000 0.1914 0.2457 0.1554 0.1236 0.3188 0.3437 0.1591 0.1548 1.0000 0.4861 0.5484 0.4783 0.2935 0.2096 0.6221 0.6899 0.5915 0.2430 1.0000 0.9150 0.6970 0.9086 -0.3267 0.4567 1.0000 0.9677 0.6373 0.9466 -0.4144 0.5386 0.9480 1.0000 0.9047 0.4666 0.9433 -0.4745 0.3332 0.9395 0.9492 1.0000 0.7626 0.7845 0.5819 0.0386 0.8893 0.6981 0.7272 0.5447 1.0000 0.9737 0.5353 0.9707 -0.586 0.5091 0.9158 0.9659 0.9333 0.6807 1.0000 0.0408 0.0138 -0.0602 -0.0034 0.1153 -0.0353 -0.0331 -0.1388 -0.0213 0.0879 0.0292 -0.0016 -0.0091 0.1162 -0.1070 -0.0580 -0.1236 -0.0462 -0.1197

0.0113 0.0170 -0.0833 0.0151 0.0479 -0.1568 0.0658 0.0348 -0.0176 0.0281 0.1300 -0.1023 -0.0546 0.0275 0.0341 -0.0299 0.0003 0.0465 -0.1366 -0.0369 -0.0906 Panels A, B, and C of table 2 report respectively, the correlation matrices between stock market returns, between GDP growth rates, and between stock market returns and GDP growth rates for the ten MENA countries. The correlation coefficients between all stock market returns are positive and vary from 15.48 percent between Tunisia and Saudi Arabia to 68.99 percent between Qatar and UAE The correlation coefficients between GDP growth rates are in average more important than those between stock market returns. Furthermore, negative correlation coefficients are reported between Lebanon and Gulf Cooperation Council GCC countries 1. As shown in Panel C, stock market returns and GDP growth rates are negatively correlated in 58.18 percent of the cases The correlation coefficient between stock market return and GDP growth in the same

country is positive only in Bahrain and Morocco. 2.2 Methodology The methodology adopted in this paper is based on the non-parametric stochastic dominance approach to evaluate the comparative stock market return and GDP growth for 1 Bahrain, Kuwait, Oman, Qatar, Saudi Arabia and UAE constitute the Gulf Cooperation Council (GCC). 5 Source: http://www.doksinet MENA countries. We use the VAR model to study the impulse response of dominated countries to a shock in the dominant country. SD criterion was applied to both financial and economic indicators for each pair of countries in the MENA region. Countries that have more SD relationships more than others are ranked first in terms of the first order SD level. The dominant countries with respect to each performance criterion leave the competition by keeping the rank they got in the first round (first order SD level). The second order SD criterion is then applied to countries that are dominated in the first round. The

performance-dominant countries in the second round of the competition leave the competition keeping their ranking but coming just after the countries already ranked in the first round. The application of the third-order SD criterion concerns only the dominated countries during the first and second order SD. Finally, a country specific classification is obtained with respect to both stock market returns and economic growth rates. After obtaining this general SD-based MENA countries’ ranking, we tried to check by applying the VAR model if countries that are stochastically dominant, influence positively or negatively the financial and economic performance in other countries. In other words, it is to study the response function of stochastically dominated countries to financial and/or economic volatility-additive shocks in the economies of stochastically dominant countries. 2.21 The non-parametric stochastic dominance approach To overcome the limits of Mean Variance (M-V) approach, we

use SD approach which requires less restrictive assumptions and incorporates information on the entire distribution. The non-parametric SD approach does not explicit specifications of the agents utility function or restrictions on the functional form of the probability distribution. Testing for SD among distributions is an important issue in the study of asset management, income inequality, and market efficiency (Fong, Wong and Lean 2005; Gasbarro, Wong, and Zumwalt, 2007; Wong, Phoon and Lean, 2008; Zhuo, Wong and Fung, 2013; Al-Khazali, Lean and Samet, 2014). In this paper, SD approach based on Davidson and Duclos (2000) test is applied. Firstly, we compute the corresponding probability density functions (f and g) and the empirical distribution functions (CDFs) of two stock market returns or GDP growth rates series’, noted Y and Z respectively. The distributions functions have a common support of [a,b] where a < b. Referring to Wong, Phoon, and Lean (2008), we compute the CDFs

as follow: H 0 = h, H j (x ) = ∫ H j −1 (t )dt x (1) a th for h = f, g; H = F, G; and j = 1, 2, 3. We call the integral H j the j order CDF Three SD orders will be tested, First (FSD), Second (SSD) and Third (TSD)-order respectively. The dominance decision rules are defined as follow: X dominates Y at FSD (X≽ 1 Y), SSD (X≽ 2 Y), and TSD (X≽ 3 Y) order respectively if and only if F 1 (x)≤ G 1 (x), F 2 (x)≤ G 2 (x), and F 3 (x)≤ G 3 (x) for all x, and the strict inequality holds for at least one value of x in [a, b]. Secondly, we compute the order-j DD statistic, T j (x) (j = 1, 2, and 3) for a grid of preselected points x 1 ,x 2 ,,x k . For any two stock market indices returns or GDP growth rates of two MENA countries, Y and Z with CDFs F and G, respectively we define: Fˆ ( x) − Gˆ j ( x) (2) T j ( x) = j Vˆ ( x) j N where 1 ( x − hi ) +j −1 , H = F , G; h = y, z; ∑ N ( j − 1)! i =1 Vˆj ( x) = VˆYj ( x) + VˆZj ( x) − 2VˆYj, Z ( x) ; Hˆ j ( x)

= 6 Source: http://www.doksinet N  1 1 VˆHj ( x) =  ( x − hi ) 2+( j −1) − Hˆ j ( x) 2 ; 2 ∑ N  N (( j − 1)!) i =1  N  1 1 ( x − yi ) +j −1 ( x − zi ) +j −1 − Fˆ j ( x)Gˆ j ( x) . VˆYj, Z ( x) =  2 ∑ N  N (( j − 1)!) i =1  in which F j and G j are defined in (1), and (x) + = max{x,0}. It is empirically impossible to test the null hypothesis for the full support of the distributions. We test the null hypothesis for a pre-designed finite numbers of values x. The following hypotheses are tested: H 0 : F j ( xi ) = G j ( xi ), for all x i , i=1,.,k; H A : F j ( xi ) ≠ G j ( xi ), for some x i ; H A1 : F j ( xi ) ≤ G j ( xi ), for all x i , F j ( xi ) < G j ( xi ) for some x i ; H A1 : F j ( xi ) ≥ G j ( xi ), for all x i , F j ( xi ) > G j ( xi ) for some x i . Thirdly, following Bishop, Formby, and Thistle (1992), we test the null hypothesis for a predesigned finite numbers of values x which will be

rejected if the DD statistic is significant at any grid points. Under the null hypothesis H 0 , DD show that T j (x) is asymptotically distributed as the Studentized Maximum Modulus (SMM) distribution. The SMM distribution with k and infinite degrees of freedom M ∞j ,α is used to control for the probability of rejecting the overall null hypothesis. We define the following decision rules: T j ( xk ) < M ∞j ,α , accept H 0 : Y= j Z; T j ( xk ) < M ∞j ,α for all k and − T j ( xk ) > M ∞j ,α for some k, accept H A : Y≠ j Z; − T j ( xk ) < M ∞j ,α for all k and T j ( xk ) > M ∞j ,α for some k, accept H A1 : Y> j Z; and T j ( xk ) > M ∞j ,α for all k and T j ( xk ) > − M ∞j ,α for some k, accept H A2 : Z> j Y where M ∞j ,α is the 1-α percentile of M ∞j tabulated by Soline and Ury (1979). We specify k equal-interval grid points {x i , i=1,2,k}, to cover the common support of random samples {Yi} and {Zi}, and we propose k = 100

(Fong, Lean, and Wong, 2008; Gasbarro, Wong, and Zumwalt, 2007). We note that if we accept either H 0 or H A , this implies that there are no SD relationships between the two specified countries performance measurements and neither one is preferred to the other. However, if H A1 or H A2 is accepted, a particular MENA country stochastically dominates another one (referring to its economic or financial performance measurement) 2. 2.22 The multivariate vector autoregressive model Based on the stochastic dominance results, we use the VAR methodology proposed by Sims (1980) to study the impact of a shock in the volatility of GDP growth or stock market returns of the most dominant country on the GDP growth and stock market returns of all MENA countries. The VAR models of order p, where the order p represents the number of lags are expressed as follows: p Gt = A0 + ∑ Ai Gt −1 + ut (3 i =1 2 The existence of SD implies that the expected utility of investors is always higher when

investing in a dominant country than in a dominated one. Consequently, the dominated country should not be chosen 7 Source: http://www.doksinet p Yt = A0 + ∑ AiYt −1 + ut (4) i =1 p Rt = A0 + ∑ Ai Rt −1 + ut (5) i =1 p Z t = A0 + ∑ Ai Z t −1 + ut (6) i =1 where Gt = [G1t .Gkt ] is a column vector of monthly GDP growth for the ten MENA markets; Yt = [G1t R1t .Rkt ] contains the GDP growth of the most dominant country and the stock market returns for all considered MENA countries; Rt = [R1t .Rkt ] is a vector of stock market return series; Z t = [R1t G1t .Gkt ] contains the stock return of the most dominant country and the GDP growth series; K is the number of countries; Ai is the matrix of unknown coefficients; A0 is a column vector of deterministic constant terms; ut is a vector of innovations that may be contemporaneously correlated but are uncorrelated with their own lagged values and uncorrelated with all of the right-hand side variables. The lag

length is determined using the Akaike information criteria (AIC) and the Schwarz criteria (SC). After estimating the earlier models, generalized impulse response functions are derived from the estimates. An impulse response function measures the time profile of the effect of a shock on the behavior of a series. When the upper and lower bands carry the same sign, the response is interpreted as being statistically significant at the 95% confidence level (Abugri, 2008). 3. Results analysis 3.1 Comparative stock market return results The results of SD relationships between the stock market returns of the ten MENA countries are reported in table 3. Panels A, and B respectively report the results of the First and Second-order SD relationships. Panel C presents the ranking of the ten MENA countries following the generated SD relationships. Table 3 Stochastic dominance relationships between stock market returns series (June 2005-December 2013) Panel A: First-order Stochastic Dominance (FSD)

relationships MENA countries Bahrain Bahrain Jordan Kuwait Lebanon Morocco Oman Qatar S. Arabia Tunisia UAE Dominated by Total ND FSD ND FSD FSD FSD FSD ND FSD 6 Dominates Jordan Kuwait Lebanon Morocco Oman Qatar S. Arabia Tunisia UAE Total Diff Rank ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND FSD ND ND ND ND ND ND ND ND ND ND ND ND ND FSD FSD ND FSD ND ND 0 0 1 1 4 1 2 3 1 1 -6 -2 1 1 4 0 2 1 1 -2 8 7 4 4 1 5 2 3 4 6 ND ND FSD ND ND FSD ND ND 2 ND ND ND ND ND ND ND 0 ND ND ND ND ND ND 0 ND ND ND ND ND 0 ND FSD ND ND 1 ND ND ND 0 FSD ND 2 ND 0 3 Panel B: Second-order Stochastic Dominance (SSD) relationships MENA countries Bahrain Bahrain Jordan Oman UAE Dominated by Total ND ND ND 0 Dominates Jordan Oman UAE Total Diff Rank ND ND ND ND SSD SSD 0 1 1 0 0 1 1 -2 2 1 1 3 ND ND 0 ND 0 2 Panel C: Country ranking according to stock market return Country ranking Morocco 1 Qatar 2 S. Arabia 3 Kuwait 4

8 Lebanon 4 Tunisia 4 Jordan 7 Oman 7 Bahrain 9 UAE 10 Source: http://www.doksinet Note: This table reports the First and Second-order stochastic dominance results to test the performance between all the stock market returns of the ten MENA countries: Bahrain, Jordan, Kuwait, Lebanon, Morocco, Oman, Saudi Arabia, Tunisia, and United Arab Emirates (UAE), over the period from June 2005 to December 2013 and their ranking through the SD relationships generated. The test is based on DD test statistics (refer to Equation 2) significance for the First and Second SD orders (FSD-SSD). The results in this table are read on a row-versuscolumn basis We report the number of the stochastic dominant (dominated) relationships, the difference number between the dominant and dominated relationships and the performance ranking for each country. From FSD test to SSD, we eliminate the countries which reveal positive difference FSD number relationships. For example, we do not consider Morocco,

Qatar, Saudi Arabia, Tunisia, Lebanon, and Kuwait in the SSD test since they display positive difference FSD number relationships. Panel A of table 3 shows that Morocco stock market return stochastically dominates at the First-order (FSD) its counterpart of Bahrain, Jordan, Saudi Arabia, and UAE. More precisely, the investment in Morocco stock market is better than in Bahrain, Jordan, Saudi Arabia, and UAE. Relying on the difference between the FSD dominant and dominated relationships, we can define an order of preferences in which Morocco ranks first followed by Qatar, Saudi Arabia, Tunisia, Kuwait and finally Lebanon. Differently, results show that there are no SD relationships between Bahrain stock market index return and the entire stock market returns of the other nine MENA countries. Besides, results reveal that Oman, UAE, Jordan, and Bahrain stock market returns seem to be more dominated by the other MENA stock market returns showing negative difference number of FSD

relationships. Panel B of table 3 reports SSD relationships between Bahrain, Jordan, Oman, and UAE stock market returns. The results exhibit that only Jordan and Oman stock market return distributions stochastically dominate at the Second-order their UAE counterpart. Empirical findings reveal that there are no SSD relationships between Bahrain stock market return distribution and its counterpart of Jordan, Oman, and UAE. Panel B results prove that Jordan and Oman are the highest financially performing countries comparing to Bahrain and UAE. Referring to the results of panels A and B, panel C of table 3, reports MENA countries’ ranking according to stock market return. Results show that Morocco is the most performing country followed respectively by Qatar, Saudi Arabia, Kuwait-Lebanon-Tunisia, JordanOman, Bahrain, and UAE. This result can be explained by the degree of development of the financial stock market in Morocco comparing to the other MENA countries. For a more detailed

explanation the SD relationships summarized in Panels A and B of the table 3, we plot the DD statistics and the corresponding CDFs for the stock market monthly returns of UAE and Oman. 9 Source: http://www.doksinet Figure 1 Plot of the CDF of the Monthly stock market returns and DD statistics of UAE and Oman (June 2005-December 2013) 10 1.1 9 1 8 7 0.9 6 5 0.8 4 0.7 3 0.6 1 0 -1 0 0.5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 CDF DD Statistics 2 0.4 -2 -3 0.3 -4 0.2 -5 -6 0.1 -7 -8 0 -9 -10 -0.1 % return T1 T2 T3 stock Mark Ind return of UAE country Stock Mark Ind Retun of OMAN country As shown in the Figure 1, result of the T 1 statistic is statistically insignificant which shows the absence of FSD relationships. This result is confirmed since we find that the CDF of the two indices returns distribution’s across. Nevertheless, T 2 and T 3 DD statistics are significantly positive. Oman stock market return distribution dominates stochastically at

the Second and Third-order (SSD-TSD) UAE stock market return 3. This implies that Oman is more performed than UAE. 3.1 Comparative GDP growth results First, Second and Third-order SD relationships between the GDP growth rates of the ten MENA countries and their classification over the sample period are reported in table 4. As shown in Panel A of table 4, Qatar GDP growth rate reveals the highest number of dominant FSD relationships (4 SD relationships) and Tunisia exhibits the highest number of dominated FSD relationships (3 SD relationships). This implies that, among FSD relationships, Qatar is the most performing country and Tunisia is the lowest. Jordan stochastically dominates Lebanon, Morocco, and Tunisia at the FSD-order. Panel A of table 4 shows also that there are no SD relationships between Kuwait GDP growth distribution and its counterpart of the other nine MENA countries. Empirical findings reveal that Qatar is followed respectively by Jordan, and Lebanon. However, Bahrain,

Kuwait, Morocco, Oman, Saudi Arabia, Tunisia, and UAE cannot be ranked since they have dominated FSD relationships higher than their dominant counterparts. 3 Since the hierarchical relationship exists in SD that SSD implies TSD, we report only SSD if both SSD and TSD relationships are found. 10 Source: http://www.doksinet Table 4 Stochastic dominance tests between GDP growth series (June 2005-December 2013) Panel A: First-order Stochastic Dominance (FSD) relationships MENA countries Bahrain Bahrain Jordan Kuwait Lebanon Morocco Oman Qatar S. Arabia Tunisia UAE Dominated by Total ND ND FSD ND ND ND ND ND ND 1 Dominates Jordan Kuwait Lebanon Morocco Oman Qatar S. Arabia Tunisia UAE Total Diff Rank ND ND ND ND FSD ND ND FSD ND FSD ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND FSD ND FSD ND FSD FSD ND ND ND ND ND ND ND ND ND FSD ND ND 0 3 0 3 1 0 4 0 0 0 -1 3 -1 2 -1 -1 4 -1 -3 -1 5 2 5 3 4 5 1 5 6 5 ND ND ND ND ND ND ND ND 0 ND ND ND FSD ND ND ND 1

ND ND ND ND ND ND 1 ND ND ND ND ND 2 FSD ND ND ND 1 ND ND ND 0 ND ND 1 ND 3 1 Panel B: Second-order Stochastic Dominance (SSD) relationships MENA countries Bahrain Bahrain ND ND ND ND ND ND 0 Kuwait Morocco Oman S. Arabia Tunisia UAE Dominated by Total Dominates Kuwait Morocco Oman S. Arabia Tunisia UAE Total Diff Rank SSD ND ND SSD ND ND SSD ND ND ND ND ND ND ND ND SSD ND ND ND SSD ND 4 0 1 1 2 0 1 4 -5 1 0 1 0 -1 1 6 2 4 3 4 5 SSD SSD SSD ND SSD 5 ND ND ND ND 0 ND ND ND 1 ND ND 1 ND 0 2 Panel C: Third-order Stochastic Dominance (SSD) relationships TSD GDP Kuwait Dominated by Kuwait Oman Tunisia UAE Total ND TSD ND 1 MENA countries Oman Tunisia ND ND ND TSD ND ND 1 0 Dominates Total 0 0 4 0 UAE ND ND TSD Diff -1 -1 4 -1 Rank 2 2 1 2 1 Panel D: Country ranking according to GDP growth Country ranking Qatar 1 Jordan 2 Lebanon 3 Bahrain 4 Morocco 5 S. Arabia 6 Tunisia 7 Kuwait 8 Oman 8 UAE 10 Note: This table reports the First,

Second and Third-order stochastic dominance results to test the performance between all the GDP growth rates of the ten MENA countries: Bahrain, Jordan, Kuwait, Lebanon, Morocco, Oman, Saudi Arabia, Tunisia, and United Arab Emirates (UAE), over the period from June 2005 to December 2013 and their ranking through the SD relationships generated. The test is based on DD test statistics (refer to Equation 2) significance for the First, Second and Third SD orders (FSD-SSD-TSD). The results in this table are read on a row-versus-column basis. We report the number of the stochastic dominant (dominated) relationships, the difference number between the dominant and dominated relationships and the performance ranking for each country. From FSD test to SSD, we eliminate the countries which reveal positive difference FSD number relationships. From SSD test to TSD, we eliminate the countries which reveal positive difference SSD number relationships. For example, we do not consider the Qatar,

Jordan, and Lebanon countries in the SSD test since they display positive difference FSD number relationships. Besides, we do not consider the Bahrain, Morocco, and Saudi Arabia countries in the TSD test since they display positive difference SSD number relationships. Panel B of table 4 summarizes Second-order SD relationships between Bahrain, Kuwait, Morocco, Oman, Saudi Arabia, Tunisia, and UAE over the period from June 2005 to December 2013. The results reveal that Bahrain GDP growth distribution stochastically dominates at the Second-order that of Kuwait, Oman, Saudi Arabia, and UAE. Besides, is not dominated by the GDP growth distributions of any of the six countries. Referring to the difference between dominant and dominated SSD relationships, we find that Bahrain is followed respectively by Morocco and Saudi Arabia. Panel C of table 4 reports Third-order SD relationships between Kuwait, Oman, Tunisia, and UAE GDP growth series over the period from June 2005 to December 2013 and

shows that Tunisia GDP growth distribution stochastically dominates at the Third-order its counterparts of Kuwait, Oman, and UAE. The ranking of the ten MENA countries according to GDP growth through FSD, SSD, and TSD relationships is summarized in Panel D of table 4. Qatar performs economically better than the other MENA countries and it is followed respectively by Jordan, Lebanon, Bahrain, Morocco, Saudi Arabia, Tunisia, Kuwait-Oman, and UAE. From the GDP growth indicator, UAE seems to be the lowest economically performing country. 11 Source: http://www.doksinet Our SD GDP growth results can be proved from table 5 which illustrates how we draw the different dominance conclusions displayed in Panels A, B, and C of table 4. Results show that 29 percent (42 percent) of the First-order DD statistic T 1 is significantly positive (negative); thus, the results lead us to reject the hypothesis that Tunisia GDP growth distribution FSDdominates the UAE GDP growth distribution and vice

versa. From table 5, we find that 26 percent (31 percent) of the Second-order DD statistic T 2 is significantly positive (negative) implying the absence of SSD relationships between the two GDP growth distributions. However, table 5 reveals that 22 percent of third-order DD statistic T 3 is significantly positive and none it is significantly negative implying that Tunisia GDP growth distribution TSDdominates the UAE one. Tunisia MENA country performs better economically than UAE Table 5 DD statistics of the Stochastic dominance test between UAE and Tunisia GDP growth distributions (June 2005-December 2013) Total (%) Positive Domain (%) Negative Domain (%) Max (|Tj|) T1 > 0 29 29 0 4.190 Total (%) Positive Domain (%) Negative Domain (%) Max (|Tj|) T2 > 0 26 26 0 3.716 Total (%) Positive Domain (%) Negative Domain (%) Max (|Tj|) T3 > 0 22 22 0 3.310 FSD T1 < 0 42 0 42 8.939 SSD T2 < 0 31 0 31 5.108 TSD T3 < 0 0 0 0 1.706 Notes: Readers may refer to equation (2)

for the formula of T j for j=1, 2, 3 with F = UAE GDP growth distribution and G = Tunisia GDP growth distribution. The period is from June 2005 to December 2013 From the results of tables 3 and 4, we find that MENA countries classification according to GDP growth differs from the one based on stock market return. Economic performance can not imply stock market performance and vice versa. This result can be justified by negative correlation coefficients between GDP growth and stock market return in most countries (table 2). 3.3 Shock transmission between MENA countries The generalized impulse response functions based on the estimation of the VAR models are reported in figure 2 to 6 4. 4 Only response functions are reported here because the results of the VAR estimates are similar to those obtained by plotting the impulse response functions. VAR estimate results are available on request 12 Source: http://www.doksinet Figure 2 Generalized impulse response functions of GDP growth to

Qatar GDP growth shock Jordan Bahrain Lebanon .0006 .0003 .00012 .0004 .00008 .0002 .00004 .0002 .00000 .0001 .0000 -.00004 .0000 -.00008 -.0002 -.00012 -.0004 -.0001 2 4 6 8 10 12 2 4 Morocco 6 8 10 12 2 4 S. Arabia .0004 6 8 10 12 8 10 12 8 10 12 Tunisia .0004 .0012 .0003 .0003 .0008 .0002 .0002 .0001 .0004 .0001 .0000 .0000 .0000 -.0001 -.0001 -.0004 2 4 6 8 10 12 -.0002 2 4 Kuwait 6 8 10 12 2 4 Oman .0016 .0016 .0012 .0012 .0008 .0008 .0004 .0004 .0000 .0000 6 UAE .0012 .0008 .0004 -.0004 .0000 -.0004 2 4 6 8 10 12 -.0004 2 4 6 8 10 12 2 4 6 Notes. Solid line represents the generalized response function Dotted lines are 95% confidence bounds Figure 2 shows the generalized impulse response functions of GDP growth in each MENA country to Qatar GDP growth shock. The response is statistically significant and more important in the first month. After that, the effect decreases over time

and becomes null after four to ten month depending on the country. This response is positive for all countries except Lebanon where the response is negative in the first two months. This shows that an increase in Qatar GDP growth enhances GDP growth in the other MENA countries. Figure 3 plots the response of stock returns in all MENA countries to a GDP growth shock in Qatar. The response is short but positive and significant only for Tunisia in the first two months. For other MENA countries, the response is insignificant This result shows that GDP growth and stock market return in Qatar are not correlated. 13 Source: http://www.doksinet Figure 3 Generalized impulse response functions of stock market return to Qatar GDP growth shock Morocco S. Arabia Qatar .03 .02 .02 .01 .01 Kuwait .03 .03 .02 .02 .01 .01 .00 .00 .00 .00 -.01 -.01 -.02 -.02 -.01 -.01 -.02 -.02 -.03 1 2 3 4 5 6 7 8 9 -.03 1 10 11 12 2 3 4 5 Lebanon 6 7 8 9 -.03 1 10 11

12 2 3 4 5 Tunisia .04 7 8 9 1 10 11 12 2 3 4 5 Jordan .03 6 7 8 9 10 11 12 8 9 10 11 12 Oman .02 .02 .02 6 .03 .02 .01 .01 .01 .00 .00 .00 -.02 .00 -.01 -.01 -.04 -.02 1 2 3 4 5 6 7 8 9 -.02 1 10 11 12 -.01 2 3 4 5 Bahrain 6 7 8 9 10 11 12 8 9 10 11 12 -.02 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 UAE .03 .04 .02 .02 .01 .00 .00 -.02 -.01 -.02 -.04 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 Notes. Solid line represents the generalized response function Dotted lines are 95% confidence bounds Figure 4 displays the generalized impulse response functions of stock return in MENA countries to a shock in the stock return of Morocco. There is a short but positive and significant response in most countries except, Qatar, Saudi Arabia, and UAE. The response shows that an increase in the return of Morocco stock market return leads to arise in the return of the other markets. This result

proves the presence of positive correlation coefficients between all considered MENA stock market return series (panel A of table 2). Figure 5 plots the response of GDP growth in all MENA countries to a stock market return shock in Morocco. The response is significant only for Morocco for the first two months and Jordan and Lebanon in the first month. A decline in stock market return in Morocco leads to a decline in the GDP growth. Figure 6 plots the generalized impulse response function of GDP growth of Tunisia, Jordan and Lebanon to a shock in the GDP growth of GGC countries. The response is significant for all countries. It is negative for Lebanon implying that a decline in GCC GDP growth enhances Lebanon GDP growth during the first eight months. 14 Source: http://www.doksinet Figure 4 Generalized impulse response functions of stock market return to Morocco stock market return shock Qatar S. Arabia Kuwait .03 .04 .04 .03 .03 .02 .02 .02 .01 .01 .01 .00 .00 .00 -.01

-.02 -.01 2 4 6 8 10 12 -.01 2 4 Lebanon 6 8 10 12 2 Tunisia .03 .03 .04 .02 .02 .02 .01 .01 .00 .00 .00 -.02 -.01 4 6 8 10 12 4 6 8 10 12 2 4 Bahrain 10 12 6 8 10 12 8 10 12 UAE .05 .04 .04 .03 .02 8 -.01 2 Oman .03 6 Jordan .06 2 4 .03 .02 .01 .02 .01 .01 .00 .00 -.01 .00 -.01 2 4 6 8 10 12 -.01 2 4 6 8 10 12 2 4 6 Notes. Solid line represents the generalized response function Dotted lines are 95% confidence bounds 15 Source: http://www.doksinet Figure 5 Generalized impulse response functions of GDP growth to Morocco stock market return shock Jordan Lebanon .0003 Bahrain .00015 .0003 .0002 .00010 .0002 .0001 .00005 .0001 .0000 .00000 -.0001 .0000 -.00005 -.0001 -.0002 -.00010 2 4 6 8 10 -.0003 2 12 4 6 8 10 2 12 4 S.Arabia Morocco .0004 .0006 .0003 .0004 .0002 .0002 6 8 10 12 8 10 12 8 10 12 Tunisia .0002 .0001 .0001 .0000 .0000 -.0002 -.0001

-.0004 .0000 -.0001 -.0002 -.0006 2 4 6 8 10 -.0002 2 12 4 6 Kuwait 8 10 2 12 4 6 UAE Oman .0008 .0006 .0006 .0004 .0004 .0004 .0002 .0002 .0000 .0000 .0000 -.0002 -.0002 -.0004 -.0004 -.0004 -.0006 -.0008 2 4 6 8 10 -.0006 2 12 4 6 8 10 12 2 4 6 Notes. Solid line represents the generalized response function Dotted lines are 95% confidence bounds Figure 6 Generalized impulse response functions of GDP growth to GCC GDP growth shock Jordan Lebanon .00100 .0008 .00075 .0006 .00050 .0004 .00025 .0002 .00000 .0000 -.00025 -.0002 -.00050 -.0004 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 Morocco 6 7 8 9 10 11 12 8 9 10 11 12 Tunisia .0008 .0012 .0006 .0008 .0004 .0002 .0004 .0000 .0000 -.0002 -.0004 -.0004 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 Notes. Solid line represents the generalized response function Dotted lines are 95% confidence bounds 16 Source:

http://www.doksinet 4. Conclusion and policy implications The dependence between financial development and economic growth has long been one of the most well-known and contentious in the academic circle (Yang and Yi 2008). As the majorities of MENA countries has implemented structural adjustment programs in order to develop their financial sector and enforce their economic growth, this paper contributes to the literature by evaluating the comparative economic and stock market performance of ten MENA countries using the non-parametric Davidson and Duclos (2000) stochastic dominance test. Shock transmission from more dominant to dominated countries has been also investigated based on the multivariate vector autoregression (VAR) methodology. The MENA area is characterized by a high level of volatility, due to many factors, as oil price fluctuations, political instability, and regional conflicts. It is daily at the center of the economic and political debate, and this stylized fact

represents a further source of interest (Andreano et al. 2013). The mains results, over the period from June 2005 to December 2013, are summarized in four points. First, Qatar and Morocco are respectively the most performing countries according to GDP growth and stock market return while UAE is the least performing country. The dominance of Qatar over the study period can be due to the fact that this country relies on its energy sector to support its economy. Petroleum and liquefied natural gas are the cornerstones of Qatar’s economy and account for more than 90% of total government revenue (IMF Country Report 2015). The super performance of Qatars economy is now tainted with a systemic risk linked to excessive volatility in oil and gas prices and an irreversible direction of global economies to renewable energy to reduce the negative effects of global warming. This performance as it is dependent on the energy policy is characterized as fragile as the fluctuation of international oil

price is considered as the most important determinant of the economic growth. Thus, policy makers in Qatar should develop their economic fundamentals for a sustainable performance. As noted by Andreano et al (2013), the degree of international openness, technological development, and human capital are relevant for growth. For Morocco, it is not surprising to be the most performing according to stock market return while this country has one of the oldest stock markets on the MENA region founded in 1929. Furthermore, Morocco is the first MENA country which signed European Union partnership since 1996. This agreement leads to large increases of the European aid to Morocco and develops the integration and exchange degree between international financial markets. For the UAE, although it has the most diversified economy in the GCC, except for Dubai, most of their revenues are from oil. The least rank of this country can be due to the fact that Dubai suffered from a significant economic

crisis in 2007-2010 and is currently in extreme debt in addition to the decline of the real estate market and oil prices. Second, the dominance of MENA countries differs across performance indicators. Equity returns and GDP growth in the studied MENA countries are generally not linked showing that financial and real spheres are not integrated. This result confirms these of several studies that failed to find a positive correlation between a country’s economic growth and its stock market’s return (Dimson et al. 2002 and Ritter 2005) Economic growth is determined by growth in the supply of labor and increases in productivity while, stock returns are determined by the cost of capital, which is the rate of return required by investors to bear the risk of owning stocks. These results yield two relevant policy implications: First, investors should not invest in stocks based only on economic growth. Second, knowing the rank of the country among the other countries of the region according

to GDP growth and stock market return, permit policy makers to increase their ability to attract foreign investors. Third, contagion analysis shows that the GDP growth response to Qatar GDP growth shock is statistically significant for all countries, but the stock market return response is generally insignificant. The shock in the economy of Qatar is propagated to other countries via the GDP 17 Source: http://www.doksinet growth and not the stock market returns. The reason behind this may well be the fact that Qatar, as one of the richest country in the world, has been extremely influential economically the MENA region. The oil-poor countries are oil price sensitive since a large part of their economies is dependent on worker remittances as well as on development aid and tourism revenues from the oil-rich labor poor countries. Referred to United Nations Comtrade Database, this country has a bilateral trade relationship with all studied countries. Fourth, stock market response to

Morocco stock market shock is insignificant in Qatar, Saudi Arabia, and UAE. The response time of stock market return is also short when it is significant. In the most of MENA countries, the financial system is largely directed and protected from foreign competition by keeping relatively high levels of government controls. This can explain the weakness of the financial system’s competitiveness and efficiency even if these financial and regulatory policies provide some stability to the financial system. Ben Naceur et al. (2014) show, that MENA region has fared well in growth primarily in connection with energy resources. Nonetheless, financial development has lagged, which is at the core of failure to diversify many economies in the region. Important policy implications from this result are that potential benefits from international portfolio diversification are significantly higher than one might expect. If one country falls, others are not This result confirms earlier study of

Lagoarde-Segot and Lucey (2007) who highlight the presence of outstanding diversification benefits in the MENA region. At the government level, policymakers can be better equipped to predict the impact of a shock in the dominant country on their economy and stock market and then to determine the way a country can be better prepared and mitigate the negative consequences of crises. Finally, the implication that arises from this study is that in order to promote growth, these countries should introduce financial policy changes in old government policies in order to make their economy progressively market-oriented and integrated with global economic structure in a more meaningful way. As noted by Cooray (2010), policy measures taken to increase the size, liquidity and activity of the stock market will further enhance growth. A more diversified economy is inherently more stable, more capable of creating jobs and opportunities for the next generation and less vulnerable to the boom and bust

cycles of oil and natural gas prices. References Abugri, B., 2008 Empirical relationship between macroeconomic volatility and stock returns: Evidence from Latin American markets. International Review of Financial Analysis17, 396–410 Al-Khazali, O., Lean, H, Samet, A, 2014 Do Islamic stock indexes outperform conventional stock indexes? A stochastic dominance approach. Pacific-Basin Finance Journal 28, 29-46 Andreano, M.S, Laureti, L, Postiglione, P, 2013 Economic growth in MENA countries: Is there convergence of per-capita GDPs?. Journal of Policy Modeling 35, 669-683 Beetsma, R., Giuliodori, M, 2012 The changing macroeconomic response to stock market volatility shocks. Journal of Macroeconomics 34, 281-293 Ben Jedidia, K. Boujelbène T and Helali, K, 2014 Financial development and economic growth: new evidence from Tunisia. Journal of Policy Modeling 36, 883-898 Ben Naceur, S., Cherif, M, Kandil, M, 2014 What drives the development of the MENA financial sector?. Borsa Istanbul Review

14, 212-223 Bishop, J., Formby, J, Thisle, J, 1992 Convergence of the South and non-South income distributions, 1969-1979. American Economic Review, 82, 262-272 Cheng, R., Jahan-Parvar, M, Rothman, P, 2010 An empirical investigation of stock market behavior in the Middle East and North Africa. Journal of Empirical Finance 14, 413-427 Cooray, A., 2010 Do stock markets lead to economic growth? Journal of Policy Modeling, 32, 448460 Davidson, R., Duclos, J-Y 2000 Statistical Inference for Stochastic Dominance and for the Measurement of poverty and Inequality. Econometrica 68, 1435-1464 18 Source: http://www.doksinet Dimson, E., Marsh, P, Staunton, M, 2002 Triumph of the Optimists: 101 Years of Global Investment Returns. Princeton University Press, Princeton Fong, W.M, Lean, HH, Wong, WK, 2008 Stochastic dominance and behavior towards risk: the market for internet stocks. Journal of Economic Behavior and Organization 68, 1, 194-208 Fong, W.M, Wong, WK, Lean, HH, 2005 International

Momentum Strategies: A Stochastic Dominance Approach. Journal of Financial Markets 8, 89–109 Gasbarro, D., Wong, WK, Zumwalt, JK, 2007 Stochastic dominance analysis of iShares The European Journal of Finance 13, 89–101. Hadar, J., Russel, R, 1969 Rules for ordering uncertain prospects American Economic Review 59, 25-34. Kar, M., Nazlıoğlu, Ş, Ağır, H, 2011 Financial development and economic growth nexus in the MENA countries: Bootstrap panel granger causality analysis. Economic Modelling 28, 685-693 Lagoarde-Segot, T., Lucey, B M, 2007 International portfolio diversification: Is there a role for the Middle East and North Africa?. Journal of Multinational Financial Management 17, 401-416 Ramanathan, R., 2006 Evaluating the comparative performance of countries of the Middle East and North Africa: A DEA application. Socio-Economic Planning Sciences 40, 156-167 Ritter, J.R, 2005 Economic growth and equity returns Pasific-Basin Finance Journal 13, 489-503 Sims, C.A, 1980

Macroeconomics and reality Econometrica 48, 1–48 Stoline, M.R, Ury, HK, 1979 Tables of the studentizes maximum modulus distribution and an application to multiple comparaisons among means. Technometrics, 21, 87-93 Wong, W.K, Phoon, KF, Lean, HH, 2008 Stochastic dominance analysis of Asian hedge funds Pacific-Basin Finance Journal 16, 204-223. Yang, Y., Yi, M, 2008 Does Financial Development Cause Economic Growth? Implication for Policy in Korea. Journal of Policy Modeling 30, 827-840 Zhuo, Q., Wong, WK, Fung, JKW, 2013 Stochastic dominance relationships between stock and stock index futures markets: International evidence. Economic Modeling 33, 552-559 19