Additional Evidence On the Abatement of Errors In Predicting Beta Through Increases In Portfolio Size and On the Regression Tendency

Foster, Taylor W.; Hansen, Don R.; Vickrey, Don

doi:10.1111/j.1468-5957.1988.tb00129.x

Cited by 9 publications

(5 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For example, Krueger and Johnson (1991), using 60 months' data to estimate historic betas and a simple adjustment for their regression tendency, state that their 'results indicate that anomaly studies are sensitive to the characterisation of beta' (p. 578). The error in estimating betas due to this phenomenon can be minimised by using a split estimation period before and after the event period and is also smaller the more shares in the portfolio and the longer the return interval (Foster, Hansen and Vickrey, 1988). Brown, Lockwood and Lummer (1985) present a statistic for testing for a significant change in market model error variances and coefficients between estimation period and event period.…”

Section: Beta Stabilitymentioning

confidence: 99%

Event Study Methods and Evidence on Their Performance

Armitage

1995

Journal of Economic Surveys

288

115

View full text Add to dashboard Cite

The paper outlines widely used methods of estimating abnormal returns and testing their significance, highlights respects in which they differ conceptually, and reviews research comparing results they produce in various empirical contexts. Direct evidence on the performance of different methods is available from simulation experiments in which known levels of abnormal return are added. The market model is most commonly used to generate expected returns and no better alternative has yet been found despite the weak relationship between beta and actual returns. Choice of procedure for significance testing depends on the characteristics of the data. The evidence indicates that in many cases the best procedure is to standardise market model abnormal returns by their time series standard errors of regression and use the ?-test. Alternatively a rank test appears to be at least as powerful. If errors are crosscorrelated or increase in variance during the test period, other methods discussed should be used.

show abstract

Section: Beta Stabilitymentioning

confidence: 99%

Event Study Methods and Evidence on Their Performance

Armitage

1995

Journal of Economic Surveys

288

115

View full text Add to dashboard Cite

show abstract

“…The exact distribution of beta depends on the values of p and fl. When p=O, betas are generated by the random coefficient process and the beta normality assumption invoked by Vasicek (1973), Blume (1971 and1975), Dimson and Marsh (1983), Foster, Hansen and Vickrey (1988) and Kolb and Rodriguez (1989) is justified. When p = I, betas are generated by the random walk process with a stationary mean value, but unstable variance.…”

Section: The Modelmentioning

confidence: 99%

“…Reliability of beta prediction models has been linked with stability of beta estimates. Blume (1971 and1975) was the first to identify the regression tendency of OLS beta estimates and later investigated by Elgers, Haltiner andHawthorne (1979), D' lmson andMarsh (1983), Foster, Hansen and Vickrey (1988) and Kolb and Rodriguez (1989). Blume (1971) examined the predictive ability ofOLS beta estimates for NYSE betas to change gradually over time.…”

Section: Introductionmentioning

confidence: 99%

“…Studies on stability of cross-sectional betas over time have used different parametric and nonparametric techniques such as the correlation coefficient test by Blume (1971); portfolio regression tests by Blume (1975) and Dimson and Marsh (1983); the mean squared prediction error (hlSE) criterion by Dimson and Marsh (1983) and Foster, Hansen and Vickrey (1988); and the use oftransition matrices by Foster, Hansen and Vickrey (1 988) and Kolb and Rodriguez (1989). Their conclusions were mixed.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

An Adjustment Procedure for Predicting Betas When Thin Trading is Present: Canadian Evidence

Boabang¹

1996

Business Fin & Account

View full text Add to dashboard Cite

We have incorporated effects of the process that generates true betas for TSE stocks, as well as thin trading effects, into the beta adjustment model. We note the Blume and Dimson and Marsh beta adjustment techniques aim at eliminating beta forecast error through regression tendency bias. Effects of other sources of forecast error have been ignored. We show the process generating security betas affects both cross‐sectional correlation coefficient and order bias, while thin trading affects only cross‐sectional correlation coefficient. We demonstrate that when OLS beta estimates are used to forecast their future risk levels, order bias accounts for 86% of forecast error, while thin trading effects account for 14% of forecast error. A beta regression tendency model which properly accounts for effects of cross‐sectional correlation (which is a function of thin trading) and order bias completely abates forecast error. Our results have implications for the use of correlation coefficient to measure stability of betas across time, for beta adjustment models proposed in the literature, and for event study methodologies that rely on prediction errors.

show abstract

“…Furthermore, the intervalling period employed here is a month, whereas Value Line uses weekly observations. Although a rather naive adjustment is being used here, more complex beta adjustments are often not efficacious (e.g., Foster et al, 1988). Application of this equation to the highest beta anomaly portfolio, consisting of small firms with high price/earnings ratios that carried a Timeliness ranking of 1 , would reduce this portfolio's average beta from 1.35 to 1.25.…”

Section: Methods For Analyjis Of Systematic Risk Mkasurcmentioning

confidence: 99%