Length of sample period and optimal portfolios

Abstract

Scope and Method of Study: This study is designed to examine the effect of length of sample period on the selection of optimal portfolios which are derived by using the historical records as the sources of data. The major data are the S&P(500) Index and the prices of twenty five securities which are selected from the 500 Largest Industries in 1980 ranked by FORTUNE. The sample periods in this study range from two years (1979-1980) to twenty years ( 1961-1980). The major models for the selection of optimal portfolios are the single index model and the constant correlation coefficient model. Also, the regression analysis methods are used for determining the significance of time effects on the selection of optimal portfolios.

Findings and Conclusions: In general, there is no relationship between the length of the sample period and estimated security characteristics under the single index model and the constant correlation coefficient model. Moreover, the length of the sample period does not affect the composition of optimal portfolio.