An Examination of the Statistical Significance and Economic Relevance of Profitability and Earnings Forecasts from Models and Analysts

Abstract

In this paper, we propose and empirically test a cross-sectional profitability forecasting model which incorporates two major improvements relative to extant models. First, in terms of model construction, we incorporate mean reversion through the use of a twostage partial adjustment model and inclusion of a number of additional relevant determinants of profitability. Second, in terms of model estimation, we employ least absolute deviation (LAD) analysis instead of ordinary least squares because the former approach is able to better accommodate outliers. Results reveal that forecasts from our model are more accurate than three extant models at every forecast horizon considered and more accurate than consensus analyst forecasts at forecast horizons of two through five years. Further analysis reveals that LAD estimation provides the greatest incremental accuracy improvement followed by the inclusion of income subcomponents as predictor variables, and implementation of the two-stage partial adjustment model. In terms of economic relevance, we find that forecasts from our model are informative about future returns, incremental to forecasts from other models, analysts' forecasts, and standard risk factors. Overall, our results are important because they document the increased accuracy and economic relevance of a cross-sectional profitability forecasting model which incorporates improvements to extant models in terms of model construction and estimation.Les auteurs proposent et testent de fac ßon empirique un mod ele pr evisionnel de rentabilit e transversal comportant deux am eliorations importantes par rapport aux mod eles existants. En premier lieu, au chapitre de la structure du mod ele, les auteurs incorporent la r egression a la moyenne en utilisant un mod ele d'ajustement partiel a deux paliers et plusieurs d eterminants pertinents suppl ementaires de la rentabilit e. En second lieu, au chapitre de l'estimation du mod ele, les auteurs, plutôt que d'appliquer la m ethode classique des moindres carr es, ont recours a l'analyse du moindre ecart absolu parce qu'elle permet de mieux traiter les valeurs extrêmes. Ils constatent que les pr evisions produites par leur mod ele sont plus exactes que celles de trois mod eles existants, et cela pour tous les horizons pr evisionnels du plan de recherche, et qu'elles sont egalement plus exactes que les pr evisions consensuelles des analystes, pour les horizons pr evisionnels de deux a cinq ans. Une analyse plus pouss ee indique que l'estimation du moindre ecart absolu procure l'am elioration marginale la plus importante de l'exactitude, suivie de l'inclusion des sous-el ements du revenu a titre de variables ind ependantes et de l'emploi du mod ele d'ajustement partiel a deux paliers. Pour ce qui est de la pertinence economique, les auteurs observent que les pr evisions produites par leur mod ele sont plus r ev elatrices des rendements futurs que celles que fournissent les autres mod eles, que les pr evisions des analystes et que les facteurs de risque standard. Dans l'ensemble...

The usefulness of the double entry constraint for predicting earnings

“…OLS is also unable to place the required parameter structure on the endogenous predictor variable, which is required to obtain estimates consistent with the underlying accounting relationship. Evans et al (2017) and Hou et al (2012) provide evidence suggesting that a cross-sectional profitability forecasting model which incorporates the reversion of profitability to expected levels using methods that alleviate the effect of influential observations can lead to forecast accuracy improvement. They extend the literature regarding reliable, accurate, and value-relevant forecasts by employing least absolute deviation (LAD) analysis instead of ordinary least squares (OLS) because the former approach is able to better accommodate influential observations.…”

Tail-Heaviness, Asymmetry, and Profitability Forecasting by Quantile Regression

“…In contrast, a recent paper by Evans et al (2017) focuses on comparing model-based forecasts of ROE (both LAD and OLS) to analysts' explicit forecasts of ROE. They do not at the same time examine the profitability distributional shape's effects on the accuracy of quantile regression forecasts nor consider both the MAFE and MSFE criteria.…”

Exploiting ergodicity in forecasts of corporate profitability

“…More closely related to our approach, several papers employ cross-sectional profitability forecasting models. Fairfield et al (2009) predict the return on equity and net operating assets by means of dynamic panel models, while Evans et al (2017), Fama and French (2000), and Allen and Salim (2005) conduct forecasting analysis on profitability using two stage partial adjustment models. Similar to our approach, these models rely on cross-sectional data to predict changes in ROA.…”