JBS Vol 04. Num 1. 2002 - SERIAL DEPENDENCE IN THE DHAKA STOCK EXCHANGE RETURNS: AN EMPIRICAL STUDY

Shah Saeed Hassan Chowdhury
M. Shibley Sadique
Abstract

This study reviews the econometric specification of the null hypothesis that stock returns are serially independent
against the alternative that they are not Tests used in this study are the well-known runs test and the variance ratio
test. The fact that stock returns are serially dependent has important practical and theoretical consequences for
financial economists and also for the market participants Firstly, if returns are serially dependent, their variance
will depend on the time interval used to estimate returns. Secondly, when asset returns are serially dependent, they
become predictable, at least partially Together, these consequences imply that active fund management can "beat
the market". Test results obtained in this study suggest significant linear dependence between weekly Ingged price
changes (returns) of the Dhaka Stock Exchange