Further Evidence on the Risk-Return Relationship

Abstract

Recent tests of the capital asset pricing model by Fama and French (1992) showed that there is no significant relationship between the average return and systematic risk of cornmon stocks. We propose two econometric methods to improve the efficiency of the estimation and provide more powerful test statistics: joint pooled cross-section and time-series estimation and generalized least squares. Using these techniques, we find a highly significant relationship between average portfolio returns and systematic risk., In a recent article, Fama and French (1992) (FF) estimated the relationship between stock returns and beta, the measure of systematic risk, in order to test the capital asset pricing model (CAPM) of Sharpe (1964), Lintner (1965), Mossin (1966) and Black (1972). They found no significant relationship between return and beta even when beta was the only regressor in the return equation, hence casting doubt on the validity of the CAPM. The press coverage of this study concluded that “beta ... is dead” .~ In this paper we suggest that the press reports of beta’s death were greatly exaggerated. Our results show that the absence of a significant relationship between average stock returns and beta is due to the estimation methodology ofFarna and MacBeth (1973) (FM), which produces tests that are not sufficiently powerful. We develop the methodology of Amihud and Mendelson (1986, 1989) (AM) to test the CAPM, and find that the power of the resulting tests is greater than that of FM’s tests. This methodology is applied to examine the return-beta relationship using data for the period 1953-1990. When we employ the FM methodology, we find an insignificant relationship, consistent with FF. However, when we apply the improved AM methodology, we find a positive and highly significant return-beta relationship. We conclude that beta is still alive and well. In what follows, we present the test methodologies in section I. Section II describes our empirical study and presents the results. Concludingremarks are offered in section III. 1 New York Time3, February 18, 1992. See also “Beta beaten” (The Economist, March 7, 1992) and “Bye-bye to Beta” (Forbes, March 30, 1992). Recently, Chan and Lakonishok (1992), employing a similar methodology as FF, found that for data spanning over 19321992, the return-/3 relationship was weakly significant.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 47 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.