Analysis of volatility spillover from market to market in the East Africa capital markets

Abstract

The existence of spillover effects is created by co-movement in the international stock prices and for international portfolio diversification to be considered effective, the level of spillovers among stock markets need to be so low or close to nonexistent so that one national market that is performing poorly can be hedged by the international market. The study purposed to model volatility effects between stock markets in the East African securities markets and analyze the behavior of volatility spillover and volatility persistence. The study was guided by three theories, the Efficient Markey Hypothesis, Random Walk Hypothesis, and the Arbitrage Pricing Theory. Four markets in East Africa were studied namely NSE, USE, DSE and RSE. Data comprising of the closing daily stock indices was obtained from secondary sources for the sample period of 2009 to 2019. The Exponential Generalized Autoregressive Conditional Heteroscedastic (E-GARCH 1,1) model and the Glosten, Jagannathan and Runkle Generalized Autoregressive Conditional Heteroscedastic (GJR-GARCH 1,1) model were used to model the asymmetric volatility and volatility persistence between the markets and was estimated using R. These models were selected based on empirical results which showed the GJR-GARCH model as the best fit for RSE and the E-GARCH model as the best fit for NSE, DSE, and USE. The findings show existence of volatility asymmetry in NSE, USE, and DSE, RSE. NSE, USE, and DSE showed positive volatility asymmetry with fat right tails. For RSE, bad news has larger effects than good news. Further, the results also show the presence of volatility persistence in the four markets, with some experiencing larger persistence than others. This indicates that turbulence takes a long duration to settle down in these markets. The study recommends regime specific volatility spillover modelling or Granger causality models for further analysis.