An empirical test of the relative value theory as approach to asset selection (a case of Nairobi stock exchange)

Abstract

The study set out to investigate whether Investment strategy in choosing assets to include in a portfolio using geometric mean is significantly different from the one used in choosing assets to include in a portfolio using arithmetic mean if the target return is the same. The need for the study emanated from the apparent competition between maximization of arithmetic mean of returns and maximization of geometric mean of returns. Such a competition would leave the problem of correctly choosing assets by investors unresolved and therefore very necessary to determine which of the two techniques above is better as far as maximizing returns in the long run is concerned. Shares' historical returns were computed and both their yearly arithmetic mean and geometric mean determined using secondary data obtained from the companies' financial statements available at the Nairobi Stock Exchange. Portfolios were constructed and two efficient frontiers designed; one on the basis of the arithmetic mean and the other on the basis of the geometric mean. The findings were that, firstly, the arithmetic mean return and geometric mean return are not statistically significantly different. This finding is strongly supported by the argument that there is an insignificant difference in the respective proportion of weights in the shares constituting the portfolios in both the geometric mean and arithmetic mean cases. Secondly, there was no shift in shares that constitute portfolios constructed based on arithmetic mean and those based on the geometric mean. During the six years under "study, the geometric mean efficient frontier was found to lie below the arithmetic mean efficient fronti-;· following higher arithmetic mean returns throughout the years than the geometric mean returns. The combination of the above findings led to the conclusion that the investment strategy in choosing assets to include in a portfolio using the geometric mean is not significantly different from the one used in choosing assets to include in a portfolio using arithmetic mean if the target return is the same. The implication of this study is that the best asset selection approach is yet to be determined and therefore investors ought to be very careful on the measure of average return they would wish to use in future depending on their risk preferences.