This paper examines estimation issues associated with multivariate tests of asset pricing. Two issues are considered: (1) the constraint that the sample size (N) must be less than the time series (T), and (2) the relative effect on power of using the multivariate statistic versus a univariate counterpart. We find that an alternative statistic that allows for large N does not dominate the usual portfolio tests. More notably, we find that the power of a simple diagonal statistic usually dominates the multivariate statistic for cases considered in this study.

The robustness of the multivariate test of Gibbons, Ross, and Shanken (1986) to nonnormalities in the residual covariance matrix is examined. After considering the relative performance of various tests of normality, simulation techniques are used to determine the effects of nonnormalities on the multivariate test. It is found that, where the sample nonnormalities are severe, the size and/or power of the test can be seriously misstated. However, it is also shown that these extreme sample values may overestimate the population parameters. Hence, we conclude that the multivariate test is reasonably robust with respect to typical levels of nonnormality.

We investigate the relation between the Value Line enigma and post-earnings-announcement drift. The ability of Value Line's ‘timeliness’ ranks to predict future abnormal returns is well-documented. However, we show that most rank changes occur within eight trading days of an earnings announcement. Once we control for post-earnings-announcement drift, differences in abnormal returns across Value Line timeliness ranks are no longer significant. Moreover, we find that timeliness ranks have no predictive power for firms with small earnings ‘surprises’. We conclude that the Value Line enigma is a manifestation of post-earnings-announcement drift.

Previous research has shown that analysts' forecasts of quarterly earnings per share (EPS) are more accurate than those of accepted time‐series models. In addition, some previous research suggests that, on average, analysts' forecasts tend to be optimistic (i.e., biased). Two explanations for analysts' superiority have been proposed: (1) analysts use more recent information than can time‐series models and (2) analysts use forecast‐relevant information not included in the time‐series of past earnings. This paper provides evidence on a third potential source of analyst superiority: the possibility that humans can use past earnings data to predict future earnings more accurately than can mechanical time‐series models. We find that human judges do no worse than accepted time‐series models when both use the same information set: namely, the series of past EPS figures. To date, little or no research has attempted to determine why analyst bias might exist. Still, some possible reasons have been forwarded. First, pessimistic forecasts or reports may hinder future efforts of the analyst or the analyst's employer to obtain information from the company being analyzed. Second, forecast data bases may suffer a selection bias if analysts tend to stop following those firms that they perceive as performing poorly. This study proposes, and provides evidence regarding, a third possible explanation for analyst bias: the use of judgmental heuristics by analysts. Many studies have shown that human predictions are often biased because of the use of such heuristics. We present evidence that suggests this may be the case for analysts' forecasts of earnings per share. Résumé. De précédents travaux de recherche ont démontré que les prévisions des analystes relatives au bénéfice par action (BPA) trimestriel sont plus exactes que celles que permettent d'obtenir les modèles reconnus basés sur les séries chronologiques. De plus, les résultats de certains travaux de recherche laissent croire qu'en moyenne, les prévisions des analystes tendent à être optimistes (c'est‐à‐dire biaisées). Deux explications à cette supériorité ont été proposées: 1) l'information que les analystes utilisent est plus récente que celles utilisées dans les modèles fondés sur les séries chronologiques et 2) les analystes utilisent de l'information pertinente aux prévisions qui ne figure pas dans les séries chronologiques relatives aux bénéfices passes. Les auteurs attribuent à un troisième facteur potentiel cette supériorité: la possibilité pour les humains d'utiliser les données relatives aux bénéfices passés pour prédire les bénéfices futurs de façon plus précise que ne le peuvent les modèles fondés sur les séries chronologiques. Ils en viennent à la conclusion que les humains obtiennent des résultats tout aussi efficaces que les modèles chronologiques reconnus lorsqu'ils utilisent un jeu de renseignements identique, soit les données historiques relatives au BPA. Jusqu'à maintenant, peu de chercheurs, sinon aucun, ont tenté de déterminer à quoi tiendrait l'existence d'un biais chez l'analyste. Malgré tout, certaines explications possibles ont été proposées. Premièrement, les prévisions ou les rapports pessimistes peuvent faire obstacle aux efforts futurs de l'analyste ou de son employeur pour obtenir de l'information de la société faisant l'objet de l'analyse. Deuxièmement, les bases de données servant à la prévision peuvent être entachées d'un biais de sélection si les analystes ont tendance à cesser de suivre les entreprises qui leur semblent afficher une piètre performance. Les auteurs proposent et attestent une troisième explication possible du biais de l'analyste: l'utilisation de méthodes heuristiques fondées sur le jugement. De nombreuses études ont démontré que les prédictions humaines sont souvent biaisées par suite de l'utilisation de ces méthodes heuristiques. Les auteurs apportent des arguments qui permettent de croire que ce pourrait être le cas des prévisions des analystes du bénéfice par action.

We examine the annual stock performance of firms that disclose the use of derivatives to hedge over the period 1995 to 1999. We find that only 21.6% of publicly traded U.S. corporations in our sample hedged with derivative instruments over this period and their use is concentrated in the larger companies. Similar to other studies we find that when derivatives are used, interest rate and currency securities are used much more frequently than commodity products. Our sample of 1308 companies that hedge outperforms other securities by 4.3% per year on average over our sample period. This result is robust to several alternative methods of estimating abnormal returns. When we segment performance by the type of hedge used, however, we find that the over-performance is due entirely to larger firms that hedge currency. We find no abnormal returns for firms hedging either interest rates or commodities. The abnormal returns in firms hedging currency is robust to alternative models that seek to control for exchange rate fluctuations and global equity returns; however, we find no significant abnormal returns to currency hedgers when using an augmented model that controls for the role of intangible assets.

We compare the relative magnitudes of the components of the bid-ask spread for New York Stock Exchange (NYSE)/American Stock Exchange (AMEX) stocks to those of National Association of Securities Dealers Automated Quotations (NASDAQ)/National Market System (NMS) stocks. We find that the order-processing cost component is smaller, and the adverse selection component is greater on the NYSE/AMEX trading systems than on the NASDAQ/NMS system. The inventory holding component is also greater for exchange-traded stocks than for NASDAQ/NMS stocks, but this may be attributable to differences in the characteristics of the firms whose stocks trade on the respective systems.

We compare the relative magnitudes of the components of the bid‐ask spread for New York Stock Exchange (NYSE)/American Stock Exchange (AMEX) stocks to those of National Association of Securities Dealers Automated Quotations (NASDAQ)/National Market System (NMS) stocks. We find that the order‐processing cost component is smaller, and the adverse selection component is greater on the NYSE/AMEX trading systems than on the NASDAQ/NMS system. The inventory holding component is also greater for exchange‐traded stocks than for NASDAQ/NMS stocks, but this may be attributable to differences in the characteristics of the firms whose stocks trade on the respective systems.

We investigate the relation between earnings predictability, information asymmetry and the behavior of the adverse selection cost component of the bid‐ask spread around quarterly earnings announcements for NASDAQ firms. While we find an increase in the adverse selection component of the bid‐ask spread on the day of and the day prior to quarterly earnings announcements for firms with less predictable earnings, we find no evidence of such changes for firms with more predictable earnings. During a non‐announcement period, we find that firms with relatively less predictable earnings have consistently higher total bid‐ask spreads than firms with more predictable earnings. This finding suggests that firms with relatively less predictable earnings have a higher cost of equity capital than comparable firms with more predictable earning streams, ceteris paribus. Hence, earnings predictability may be a legitimate concern of managers who wish to minimize their cost of equity capital at least as it pertains to bid‐ask spreads.