SOAS Research Online

Abstract

In this paper, stock market volatility in the East European emerging markets of Hungary and Poland is investigated using daily indexes. The results suggest the presence of non-linearity in the indexes through the BDSL statistic, while the presence of conditional heteroscedasticity is detected through LM tests. Conditional volatility is then modelled as a GARCH process; however, as measured by a GARCH-M model, this does not seems to be priced in the Hungarian and Polish stock markets. Moreover, the evidence rejects the Martingale hypothesis that future changes of stock prices in the two markets are orthogonal to past information. The well-known day-of-the-week effect, reflected in significantly positive Friday and negative Monday returns, does not seem to be present in these markets. While a marked decline in conditional volatility in the Polish market after June 1995 may be explained by appreciating Zloty exchange rates against the German Mark and increasing integration with developed markets, a similar (but less consistent) pattern between exchange rates (Hungarian against German and UK currencies) and conditional volatility is found for the Hungarian market.