“How financial liberalization impacts stock market volatility in Africa: evidence from Nigeria”

Understanding the impact of financial liberalization on stock market is important for decision making by investors. The neo-classical economists believe that financial liberalization reduces stock market volatility while the post-Keynesian economists argue that financial liberalization increases volatility of the stock market. This study investigates the effect of financial liberalization on the volatility of an emerging stock market in Africa, with particular focus on the Nigerian stock market. The estimation results reveal that financial liberalization has a significant positive impact on return volatility, thus indicating that it increases stock market volatility. Also, the study finds no evidence of asymmetry in the stock market. Patrick Olufemi Adeyeye (South Africa), Olufemi Adewale Aluko (Nigeria), Oladapo Fapetu (Nigeria), Stephen Oseko Migiro (South Africa) BUSINESS PERSPECTIVES LLC “СPС “Business Perspectives” Hryhorii Skovoroda lane, 10, Sumy, 40022, Ukraine www.businessperspectives.org How financial liberalization impacts stock market volatility in Africa: evidence from Nigeria Received on: 2nd of August, 2017 Accepted on: 27th of September, 2017


INTRODUCTION
Stock market volatility is the rate at which prices of stocks/assets drift in the stock market. Bekaert and Harvey (1997) argue that understanding of volatility in emerging stock markets is crucial for deciding on the cost of capital, evaluating direct investment and making decisions related to allocation of assets. Financial liberalization policy paved way for the influx of foreign investors into the domestic financial market. Foreign investors have shown great interest in emerging stock markets due to the desire for higher returns and international diversification. The activities of foreign investors in the stock market increase the market liquidity. Spyrou and Kassimatis (1999) note that financial liberalization will cause volatility to increase through increased market liquidity. Kaminsky, Lyons, and Schmukler (2000) found that stock markets that were highly liquid suffered more than markets with less liquid during the Asian and Russian crises. Volatility in the stock market tends to rise due to financial liberalization (Aggarwal, Inclan, & Leal, 1999;Miles, 2002;Haung, 2008).
The liberalization of the stock market refers to the elimination or removal of repressive policies existing in the market. Stock market liberalization creates a paradigm shift from administratively controlled system to a market-based system. In a liberalised stock market, the interaction of market forces of demand and supply acts as the mechanism for determination of stock prices. Hence, there is a tendency for stock prices to be more volatile when determined by these market forces. Financial liberalization is likely to attract short-term investors into an economy, thereby leading to bubbles in asset price and instability in the financial system (Arestis & Demetriades, 1999;Singh, 2003). Haung (2008) perceives stock markets to be more efficient after liberalization and this leads to higher level of volatility, because stock prices rapidly respond to relevant information. Also, Bhattacharyya (2014) argues that the volatility of the stock market is not determined by financial liberalization but by quality of market information. Financial liberalization integrates the stock market of a country with other countries, thus establishing the market as a member of the global financial market. At the country aggregate level, the expected return on stocks after financial liberalization is determined by the price of risk and covariance of the domestic stock market with the global market (Haung, 2008).
The neo-classical economists (Mckinnon, 1973;Shaw, 1973) believe that financial repressive policies such as capital controls distort financial prices, thus giving room for volatility to occur. They opine that financial liberalization results to reduction in stock market volatility. However, the post-Keynesian economists argued that stock market liberalization introduces volatility into the economy (Spyrou, 1999). This shows that controversy exists among economists on the effect of financial liberalization effect on the volatility of the stock market. Also, there is ambiguity in empirical literature on the effect of financial liberalization on emerging stock market volatility. Studies such as Grabel (1995), Levine and Zervos (1998), Miles (2002), Haung (2008), and Zhang and Ding (2008) provide evidence that volatility of emerging stock markets tends to increase due to financial liberalization, whereas Spyrou and Kassimatis (1999), Kim and Singal (2000), Kassimatis (2002), Cunado, Biscarri, and Pérez de Gracia (2006), and Ndako (2012) showed that financial liberalization is not a determining cause of volatility of emerging stock markets.
Financial liberalization has been given as the root cause of financial crises such as the Mexico crisis of 1991, Turkey crisis of 2001 and the 2008 global financial crisis experienced by countries other than the country where the crisis originated from. The global financial crisis that emanated from the United States in 2008 caused a decline in the stock market performance of most economies, especially emerging stock markets in Africa. The Nigerian stock market was overwhelmingly affected by this crisis and this led to increased uncertainty about return on investment in the market. A large exodus of investors from the Nigerian stock market was witnessed during the crisis, because equity became less attractive, thus resulting to a near collapse of the stock market.
The emerging stock markets possess higher level of volatility compared to the developed ones (Bekaert & Harvey, 1997). This can be as a result of greater procyclicality of capital flows, higher exposure to exogenous shocks and high country risk associated with emerging economies. Miles (2002) points out that the instability of emerging stock markets due to volatility has been an area of concern for academics and policymakers. Literature is replete on the effect of financial liberalization on stock market volatility in emerging economies, however, there is no substantial evidence on emerging stock markets in Africa. Therefore, this study intends to investigate the effect of financial liberalization on the volatility of an emerging stock market in Africa, focusing particular attention on the Nigerian Stock Exchange (NSE).
The NSE is one of the largest stock markets in Africa, thus findings from this study can be related to other emerging markets in Africa. Also, the findings would either validate or justify the assertions of the neo-classical and post-Keynesian economists. The motivation for this study was borne out of the lack of empirical attention on this area of discourse in the Nigeria despite being one of the most attractive investment grounds for foreign equity investors in Africa. This study would inform policy makers on how the openness of the financial sector affects equity investment in the Nigerian stock market. Also, it would provide investors with the knowledge of how financial liberalization is associated with investment uncertainty.
The rest of this paper is structured as follows: section 1 reports empirical evidence on the how financial liberalization affects stock market volatility in emerging economies. Section 2 focuses on data issues and preliminary analyses, while section 3 presents the model and estimations. Final section concludes the study.

Empirical review
Prior studies have provided divergent evidence on the impact of financial liberalization on volatility in emerging stock markets. Grabel (1995) observed that financial liberalization is positively correlated with volatility of nine emerging markets. However, Bekaert and Harvey (1997) found that there is no positive correlation between financial liberalization and stock market volatility in twenty emerging stock markets. Levine and Zervos (1998) evaluated the impact of financial liberalization on the volatility of sixteen emerging markets and observed that volatility became more pronounced in these markets after financial liberalization.
Aggarwal, Inclan, and Leal (1999) examined the kinds of events between 1985 and 1995 that led to high volatility in emerging stock markets and found that the crash of October 1987 is the only global event that led to significant rise in the volatility of several emerging stock markets. Spyrou and Kassimatis (1999) showed that there was low intensity in volatility of eight emerging stock markets subsequent to financial liberalization. Kim and Singal (2000) discovered that volatility in emerging stock markets declined after financial liberalization. Kassimatis (2002) found that stock market volatility in six emerging economies declined after important events of financial liberalization. Contrarily, Miles (2002) found that financial reforms have a significant impact in almost all the emerging stock markets in the sample and it does not reduce stock market volatility. Edwards, Biscarri, and Pérez de Gracia (2003) analyzed stock market cycles in six countries comprising of four Latin American and two Asian countries. The study found that volatility in the Latin American countries is lower in both the bull and bear cycle phases and behaves in a similar way to that of some more developed countries. However, it was evidenced from that the 1997 financial crises adversely affected the Asian countries. In a study of nine emerging stock markets, Kedmey (2005) showed that financial liberalization caused increase in volatility of some emerging stock markets while it led to decline in volatility of others. The author attributed the mixed results to be due to the country-specific factors.

DATA ISSUES AND PRELIMINARY ANALYSES
This study aims to determine the effect of financial liberalization on the volatility of the Nigerian stock market. Monthly data on Nigerian Stock Exchange (NSE) All-Share Index (ASI) was  Therefore, the pre-liberalization period is from January 1985 to July 1995, while the liberalization period stretches from August 1995 to September 2016. Returns (Rt) is used to capture the stock market instead of price, because investors tend to react to yields and it is derived from the stock market price index (i.e., ASI) and it is calculated as: where t R is return at time t, ∆ is the first difference operator, ln is natural logarithm. Table 1 shows that the mean return for the market in the pre-liberalization and liberalization period is 2.902747% and 0.741009%, respectively. This indicates that the market was more profitable before the implementation of the liberalization reform. In the full sample period, the mean return is 1.457796% which indicates that the market is relatively profitable. The maximum and minimum return of the market in the full sample period are 32.35158% and -36.58828% and both values were recorded in the liberalization period. Despite having a lower mean return, the liberalization period has a higher standard deviation statistic compared to the liberalization period. This indicates that the market is more volatile in the liberalization period. Returns in the pre-liberalization period are positively skewed, but otherwise in the liberalization period. This implies that there is a high tendency to obtain more positive values of returns in the pre-liberalization period than the liberalization period. The Kurtosis statistic in the pre-liberalization, liberalization and full period greatly exceeds 3 and this indicates that return has a leptokurtic (high-peaked) distribution in all periods. The pvalue of the Jarque-Bera statistic in the pre-liberalization, liberalization and full period is less than 0.01 in each case, hence, the rejection of the null hypothesis of normal distribution at 1% significance level. Note: * and *** denote rejection of null hypothesis at 1% significance level. Note: * and *** imply rejection of null hypothesis at 1% and 10% significance level, respectively, and p-values are in parentheses.
The presence of conditional heteroscedasticity is determined by performing the ARCH-LM test on the return series. The null hypothesis for the test is that there is homoscedasticity, i.e., there is no autoregressive conditional heteroscedasticity (ARCH) effects. Table 3 presents the results of the ARCH-LM test at lag 1, 4 and 8.
From Table 2, it can be seen that the ARCH-LM test rejects the null hypothesis of no ARCH effects in returns at lag order 1, 4 and 8. This implies that return volatility clustering exists in the Nigerian stock market, hence a volatility model can be modelled for the market.

MODEL AND ESTIMATION
This study built a first order autoregressive model [AR (1)] which specifies returns (R) as a function of past returns and a dummy variable (DUM). DUM is the stock market liberalization index and it is either 0 or 1 where 0 is assigned to each month in the pre-liberalization period and 1 is given to each month in the liberalization period. The model is transformed into generalized autoregressive conditional heteroscedasticity (GARCH) model. This study adopted four variants of GARCH model namely symmetric GARCH model, asymmetric GARCH or threshold GARCH (TGARCH) model, power GARCH (PGARCH) model and exponential GARCH (EGARCH) model. The models consist of a conditional mean equation and a conditional variance equation. The conditional mean equation in AR (1) form is presented as where 0 α is the constant parameter, 1 α is the coefficient of the one-month period lagged value of returns, 2 α is the coefficient of the dummy variable, and t ε is the error term.
The dummy variable was incorporated as a variance regress or in the conditional variance equation. The symmetric GARCH model was developed by Bollerslev (1986) and the conditional variance equation in GARCH (1,1) is stated as: where 2 t σ is the conditional variance, ϕ is the intercept (constant term); ρ is the coefficient of ARCH term, γ is the coefficient of GARCH term, λ is the coefficient of the stock market liberalization index (dummy variable).
The TGARCH model allows asymmetric impact of news (innovations) on volatility to be determined. The a symmetry impact of news can be said to exist if the asymmetric term is statistically significant. The conditional variance of the TGARCH (1,1) model is expressed as: where τ is the asymmetric term (if τ ≠ 0, news impact is asymmetric). If τ < 0, this implies bad news and if τ > 0, it implies good news.
The PGARCH model was developed by Ding, Granger, and Engle (1993) to also account for asymmetry impact of news on volatility. When the power is fixed at 1, it indicates that the conditional standard deviation is modelled and the PGARCH model is a GARCH model with asymmetric term when the power is fixed at 2. Using this model, the conditional standard deviation is modelled instead of the conditional variance modelled in other variants of the GARCH model. The PGARCH (1,δ ,1) model is expressed as: where δ is the power parameter fixed at 1 and t δ σ is the conditional standard deviation.
The EGARCH model was introduced by Nelson (1991) to also capture asymmetry impact of innovations or shocks on volatility and it is covariance stationary when the coefficient of the GARCH term is less than 1. The model is usually expressed in natural logarithm (ln) of the conditional variance so that the conditional variance is always positive even if the parameters are negative. The EGARCH (1,1) model is specified as: The GARCH models were estimated with maximum likelihood estimation (MLE) method. The Gaussian error distribution cannot be assumed for the model because of the non-normal distribution of the return series; hence, an alternative for series with non-normal distribution was applied. The Student's t distribution with fixed degree of freedom was employed. Table 4 presents the estimation results of the GARCH models.
The EGARCH model fits best, because its AIC, SC, HQ and Log Likelihood produced the lowest information compared to other models, hence, only the result of the EGARCH model is discussed. The coefficient of the dummy variable in the mean equation indicates that stock market liberalization has a significant negative impact on returns. This implies that increase in stock market liberalization reduces market returns. The statistical significance of the ARCH term provides evidence of returns volatility clustering, while the statistical significance of the GARCH term indicates that returns volatility is persistent. The asymmetry term is not significant and this suggests that the volatility of returns is nonasymmetric (i.e., good and bad news have the same effect on volatility). The coefficient of the stock market liberalization index in the variance equation is positive and significant, thus indicating that financial liberalization increases returns volatility in the stock market. The ARCH-LM test accepts the null hypothesis of no ARCH effects at lags 1 and 2. Note: * and ** indicate p-value < 1% and 5%, respectively, and p-values are reported in parentheses. Also, F-statistic is reported for ARCH-LM test.
It can be observed from Figure 3 and Figure 4 that there are notable spikes which provide evidence of structural breaks. Subsequent to the full liberalization of the Nigerian stock market in January 1995, the most prominent structural break occurred during the period the market experienced turmoil as a result of the global financial crisis. It can also be deduced from the plots of conditional variance and standard deviation of return that the risk associated with investment was relatively higher in the financial liberalization episode compared to the period with no financial liberalization episode.

CONCLUSION
There are two contrasting views on the effect of financial liberalization on stock market volatility. The neo-classical economists believe that financial liberalization reduces stock market volatility, while the post-Keynesian economists argue that financial liberalization increases volatility in the stock market. This study assessed the effect of financial liberalization on emerging stock market volatility in Africa, with special focus on the Nigerian stock market. It was found that financial liberalization in creases volatility of the Nigerian stock market. This implies that there is a positive association between financial liberalization and market risk. This study is in line with previous studies such as Grabel (1995), Levine   and Zervos (1998), Miles (2002), Kedmey (2005), Haung (2008), and Afef (2014). It also supports the argument of the post-Keynesian economists against financial liberalization. The financial liberalizationinduced volatility of the Nigeria stock market may be attributed to the large inflows of foreign portfolio investment which created excess liquidity, weak risk-based regulatory framework, macroeconomic instability, relatively low level of stock market development and high political risk. It was also revealed that there is no evidence of asymmetry in the Nigerian stock market, thus implying that investors react to good and bad news in the same way. In spite of financial liberalization increasing stock market volatility, it is important to note that financial liberalization still provides benefits to the Nigerian economy such as allowing the inflow of foreign capital, which, in turn, accelerates the rate of capital formation in the domestic economy, thus promoting economic growth. However, this study recommends that the government may increase capital controls as a way to suppress the negative effect of financial liberalization on the volatility of the Nigerian stock market.