Chaos and Nonlinear Dynamics in Emerging Capital Market

Abstract

Capital market efficiency of emerging markets has been investigated widely in recent years, but to-date the empirical results remain inconclusive because most empirical studies fail to consider the institutional features of emerging markets and employ efficiency tests which have been developed for highly liquid markets of developed countries. Furthermore, these studies use empirical tests that are designed to detect linear structure in financial time series. However, recent developments in econometrics of financial markets show evidence of nonlinear relationships in asset returns in developed markets. Given the defining characteristics of emerging capital markets, nonlinearity is most likely to be even more evident in these developing markets compared to developed ones. Using BDS test, the present paper rejects the random walk hypothesis (RWH) for the Tunisian Stock Market (TSE). Despite the multitude of economic and financial reforms, the rejection of the RWH seems to be the result of substantial non-linear dependence and not to non-stationarity in the returns series, which in turn implies a GARCH modeling. Results from Hsieh test show that the source of nonlinearity structure is multiplicative, not additive. Further investigations suggest the use of a FIEGARCH model to cope with the evidence of high volatility persistence and long memory in the conditional variance. Our empirical results also show that despite a high leverage in the TSE index the leverage parameter is insignificantly different from zero. Finally, we argue that the common assumptions of constant