Long memory in the Ukrainian stock market

Abstract

This paper examines the dynamics of stock prices in Ukraine by estimating the degree of persistence of the PFTS stock market index. Using long memory techniques we show that the log prices series is I(d) with d slightly above 1, implying that returns are characterised by a small degree of long memory and thus are predictable using historical data. Moreover, their volatility, measured as the absolute and squared returns, also displays long memory. Finally, we examine if the time dependence is affected by the day of the week; the results indicate that Mondays and Fridays are characterised by higher dependency, consistently with the literature on anomalies in stock market prices.