Forecasting, Box Jenkins method, Earnings forecasts, EPS

In recent years, there has been an increased emphasis on the forecasting of accounting earnings using the Box-Jenkins method of forecasting via autoregressive integrated moving average (ARIMA) models.' Generally, however, these models are univariate by definition and do not provide for the statistical modeling of events which occur outside of the earnings series. The purpose of this study is to explore the impact of this limitation by employing a more general approach which incorporates market and industry index data into the forecast model. One reason for exploring this more general approach is that Financial analysts have long recognized that economy-wide and industry-wide factors affect the financial numbers of individual firms. Index models enable quantification of the effects of these factors. Such quantification can be important when assessing financial trends in a firm and forecasting financial variables (Foster [1978, p. 155]; see also Brown and Ball [1967] for further motivation for index models). This objective can be achieved through the use of the single-input transfer-function method developed by Box and Jenkins [1970]. The transfer function provides a more generalized form of the ARIMA model by incorporating an additional predictor variable, in addition to past earnings, in the form of a market or industry price index. Section 1 contains a brief discussion of the transfer function and

Using secure, return cross-sectional association tests, empirical studies to date have yet to show significant Incremental Information In total current cost Income (Including all holding gains) beyond that provided by historical cost Income. The theoretical work of Revsine [1973] and others, however, Indicates that positive unexpected holding gain Information may be good news for some firms but not for others depending on the firms' ability to respond to cost Increases. Therefore, if firms differ in their responses to cost changes, cross-sectional studies may be unable to detect an Incremental Informational effect for total current cost Income. The present study provides evidence which supports the existence of such differential responses across firms. A sample of firms reporting SFAS No. 33 data is partitioned based upon past associations of operating Income with input cost changes. This association provides an empirical measure of firms' relative abilities to adjust operating flows in response to input cost changes. Statistical tests Indicate that such a partition yields significant differences in how equity returns and current cost Income are associated.

[Using security return cross-sectional association tests, empirical studies to date have yet to show significant incremental information in total current cost income (including all holding gains) beyond that provided by historical cost income. The theoretical work of Revsine [1973] and others, however, indicates that positive unexpected holding gain information may be good news for some firms but not for others depending on the firms' ability to respond to cost increases. Therefore, if firms differ in their responses to cost changes, cross-sectional studies may be unable to detect an incremental informational effect for total current cost income. The present study provides evidence which supports the existence of such differential responses across firms. A sample of firms reporting SFAS No. 33 data is partitioned based upon past associations of operating income with input cost changes. This association provides an empirical measure of firms' relative abilities to adjust operating flows in response to input cost changes. Statistical tests indicate that such a partition yields significant differences in how equity returns and current cost income are associated.]

Given the quantity of nonearnings data disclosed in firms' annual reports, and the many dimensions of performance measured, it is likely that such information is used in establishing equilibrium prices in the market for firms' shares. This study empirically tests the hypothesis that equity price‐relevant information conveyed by annual reports includes several measures other than earnings. The marginal impact of both earnings‐ and nonearnings‐based financial ratios is analyzed and reported. The ratio information is first partitioned into distinct sets using an a priori linear components (LISREL) model. Association tests then show incremental information effects for the earnings‐based ratio set as well as for several nonearnings‐based ratio sets. Résumé. Étant donné la quantité de données étrangères aux bénéfices présentées dans les rapports annuels des entreprises et les nombreuses dimensions sous lesquelles le rendement est mesuré, il est probable que cette information soit utilisée dans l'établissement de prix d'équilibre dans le marché des actions des entreprises. Dans la présente étude, les auteurs procèdent à des vérifications empiriques de l'hypothèse selon laquelle l'information pertinente aux prix relative aux participations que livrent les rapports annuels comporte plusieurs mesures étrangères aux bénéfices. Les auteurs analysent l'incidence marginale tant des ratios financiers basés sur les bénéfices que de ceux qui ne le sont pas, et ils en exposent les résultats. L'information indiciaire est d'abord partagée en jeux distincts à l'aide d'un modèle de composants linéaires a priori. Les tests d'association montrent ensuite les conséquences de l'information marginale pour le jeu des ratios basés sur les bénéfices ainsi que pour plusieurs jeux de ratios qui ne le sont pas.

Earnings forecasts, Analyst forecast, Time-series model, Multivariate techniques

The article focuses on comments on alternative interim reporting techniques within a dynamic framework. In an interesting recent paper, Dov Fried Joshua Livnat, conducted a formal analysis of alternative methods for interim earnings reporting. Their purpose was to investigate the conditions and assumptions under which these methods individually produce the most desirable results, given assumed sets of objectives. It is, however, essential to go further than this and ask on whose information these distributions are conditioned, that is, management or users. The point is not trivial, since management will have more information than users. In fact, the subsequent development in paper by Fried and Livnat makes clear that conditioning is on users' information, and moreover that users must be able to calculate in order to modify their forecasts when new information becomes available. Naturally, any conclusions will depend on the ability of management to forecast. For any given objective-reporting combination, Fried and Livnat derive the variance of the predictive distribution. One conclusion that emerges when the problem is viewed in this light is that estimates of the individual cash flows, under the integral approach, can be very poor.

Jane F. Mutchler, William Hopwood, James M. McKeown, The Influence of Contrary Information and Mitigating Factors on Audit Opinion Decisions on Bankrupt Companies, Journal of Accounting Research, Vol. 35, No. 2 (Autumn, 1997), pp. 295-310

William S. Hopwood, Paul Newbold, Peter A. Silhan, The Potential for Gains in Predictive Ability Through Disaggregation: Segmented Annual Earnings, Journal of Accounting Research, Vol. 20, No. 2, Part II (Autumn, 1982), pp. 724-732