The Pitfalls of Using Short-interval Betas for Long-run Investment Decisions

Abstract

We investigate empirical relationships between beta, the Treynor ratio, and the investment horizon for portfolios of small stocks, large stocks, and bonds. Betas and Treynor ratios are computed for holding periods of to 30 years. For both the stock and bond portfolios, beta and the Treynor ratio change substantially with the holding period. Furthermore, the relative Treynor rankings of the portfolios change. Therefore, betas and Treynor ratios cannot be calculated independently of the intended investment horizon. Our results suggest that the impact of one's assumed investment horizon has not received sufficient attention in computing systematic risk (beta) or interpreting reward-to-risk performance measures (such as the Treynor ratio, Sharpe ratio, or Jensen's alpha). Thus, investors with long-run investment horizons must interpret performance parameters obtained from investment advisory services with due consideration for horizon effects.