The Best Econometrics Model for Forecasting Equity Market Returns in Developing Countries

The emerging market economies are fast improving in terms of the real sector and financial sector growth. This is due to the role played by equity market that facilitates re-allocation of funds. This paper aims to find the best GARCH model for forecasting stock returns of emerging markets, and besides to use maximum likelihood estimation method based on the Marquardt algorithm to estimate how returns respond to market news. It was observed the best model for predicting return in equity markets of Tunisia, Kenya, and Sudan is exponential GARCH with general error distribution (GED). For Egypt, Mauritius, South Africa, Namibia, and Nigeria, the gjrGARCH (1,1) with Student’s-t distributions performs best. These market returns react differently to market news relating to them. Whereas, sGARCH with Gaussian normal distribution is mostly suitable for analysing symmetric responses of return to market news, implying returns in these markets does not react differently to market news. These findings have policy implications for investors in these respective economies. Amongst others, the study advises investors, particularly those in the equity market where volatility decays slowly and the market where volatility responds asymmetrically to be watchful as these could pose significant threat to their market portfolios. Investors in these markets, particularly those in the equity market where volatility decays slowly and the market where volatility responds asymmetrically, be watchful, as these could pose a significant threat to their market portfolio.