Modeling Asymmetric Effect in African Currency Market: The Case of Kenya

Abstract

At present, results on asymmetric volatility in the currencies of African economies are at best mixed. Although the bulk of the studies suggested the presence of asymmetric effect, a strand of studies reported ‘no asymmetric effect’. The present study seeks to examine empirically the possibility of asymmetry in Kenyan exchange rate volatility in the light of the two recent crises: the global financial crisis of 2008-09 and the election violence of January-February 2008. Two quantitative techniques were employed; the dichotomous-EGARCH specification and the standard method of subsample comparison. Empirical results showed that returns exhibit some skewness and excess kurtosis, giving justification for the generalized error distribution that was applied in the variance equations. EGARCH estimates showed the presence of asymmetry in the volatility of Kenyan shillings in which positive shocks tend to increase volatility more than the negative. In addition, the estimated effect of the crises on conditional variance is about 12.5%. Tests of specification indicated that EGARCH(1,1) fits the daily currency series well. Further, subsample comparisons revealed a significant increase in volatility during the crises. While ARCH effects increased by 39.8%, GARCH effects also went up by 2.5%. Finally, the article recommends some measures that would promote greater exchange rate stability in the economy.

Keywords: Exchange rate volatility, Financial crisis, EGARCH, Election violence, Kenya