Empirical Distribution Of Returns Of Nairobi Stock Exchange 20 Share Index: 1998-2011

Abstract

The assumption that daily share index prices are normally distributed has long been disputed by the data. In this project, the normality assumption has been tested using time series data of daily NSE 20-Share Index for the period 1998-2011. It has been confirmed that the share price index does not follow the normal distribution. Other symmetrical distributions have been fit to the data i.e. logistic distribution and t- location scale distribution. With the aid of a programming language; Matlab we have computed the various Maximum Likelihood (ML) estimates from this distributions and tested how well they fit to the data. It has been established that the NSE 20 Share Index returns follows a t-Iocation scale distribution. We recommend that since we have found that the normal . inverse gamma mixture best fits the NSE 20 Share Index return, other normal mixtures can be investigated how well they fit this data.