Toward enhancing beta estimates

The Application of Errors-In-Variables Methodology to Capital Market Research: Evidence on the Small-Firm Effect

“…Much recent capital markets research has concerned problems of measurement. In particular, attention has focused on the issues of nonsynchronous trading and benchmark error (see, e.g., [2], [3], [8], [10], [12], and [13]). The importance of measurement error as an issue is reflected in the fact that numerous capital market studies of security returns have been criticized (often trivialized) by the conjecture that reported findings are due to some type of measurement error.…”

“…Techniques for dealing with benchmark error are still being developed. One approach, suggested by Roll [15] and Lavely, Wakefield, and Barrett [8], is the use of a James-Stein estimator that formally recognizes that the expected bias in an estimated beta depends on how the estimate compares with the average beta of the population. Although the technique helps insure against extremely large errors in estimated coefficients, ambiguity in the interpretation of results still remains.…”

Benchmark Error and the Small Firm Effect: A Revisit

“…Roll (1980, 1981a) and Lavely, Wakefield, and Barrett (1980) suggest using a James‐Stein estimator to deal with the benchmark error problem. Unfortunately, ambiguity in interpreting the results is a problem.…”

“…Reinganum (1982) tests Roll's proposition and concludes that the small firm effect is significant in spite of this problem. Roll (1980Roll ( , 1981a and Lavely, Wakefield, and Barrett (1980) suggest using a James-Stein estimator to deal with the benchmark error problem. Unfortunately, ambiguity in interpreting the results is a problem.…”

Investment Horizon, Diversification, and the Efficiency of Alternative Beta Forecasts

“…The standard construction involves adjusting the observed beta through the Bayes‐Vasicek (1973), Blume (1971 and 1975), or Merrill Lynch (see Klemkosky and Martin, 1975) methods. Recently, Lavely, Wakefield, and Barrett (1980) have used the James‐Stein method (see Efron and Morris, 1973) to adjust betas.…”