On properties of return distributions in İstanbul Stock Exchange

Abstract

In this thesis study, the distributional behavior of return series of Istanbul StockExchange (ISE) stocks and some ISE indices are analyzed. Normality tests are conducted by employing Kolmogorov-Smirnov and Jarque-Bera statistics, it is shown that the normality of ISE indices and most stock returns are rejected by both statistics for daily data. The stability of these series is also investigated by using different characteristic exponent estimation techniques. Estimated parameters are significantly larger than that are implied by the stable Paretian laws. To explore portfolio normality, a simulation experiment is conducted by generating random portfolios. It is observed that normal return series are dominated by the non-normal ones, resulting in increased deviation from normality as the number of firms in the portfolio increases.