Portfolio diversification: an empirical investigation of commercial banks

Abstract

This study had set out to address the problem of determining portfolio composition that ensures efficient diversification against risk by commercial banks in Kenya. Currently, most banks apply intuitive or subjective methods in deriving the number of distinct government securities to hold in a portfolio. Such methods not only lack objectivity but may not achieve the near portfolio diversification efficiency. It is with this view, that the study examined some factors that were deemed to be affecting portfolio diversification of Commercial banks in Kenya, and by using such factors to develop an empirical method of ascertaining the required portfolio composition. Data was obtained from 31 out of 37 commercial banks registered in Kenya by the end of 1994. Multiple Regression Analysis statistical technique, with the aid of Macromanager and Macrostat computer software packages, was used to conduct the data analysis. Based on the results of the Regression Analysis, there were indications of a relationship between the selected independent variables: Networth, Deposits not allocated to the customers (Deposits less loans) and the amount of loans, and the independent variable, number of distinct government securities. The relationship indicated a marked improvement over the years from 31% in 1993 to 67% in 1994.

In the absence of any laid down objective rule of determining the number of distinct government securities to hold, the 1994 model obtained, provides an estimate and/or a beginning ground for obtaining bank portfolio composition. It is also hoped that the results of this study will be read

and/or interpreted in light of the limitations