Earnings management and earnings predictability: A quantile regression approach

This study argues that the managerial choice of earnings management strategy could be contingent upon a firm's information asymmetry and such a strategy may affect the firm's earnings predictability. Measuring information asymmetry by earnings predictability based on the subsequent industry-adjusted dispersion in analysts' forecasts and employing a quantile regression to analyze 28,383 US firm-year observations from 1988 to 2014, this study reports that the effect of earnings management strategies on earnings predictability is nonuniform. Specifically, the amount of absolute discretionary accruals is negatively (positively) related to the subsequent industry-adjusted dispersion in the low (high) quantiles of analysts' forecasts. These results support the hypothesis that a firm could implement earnings management strategies according to the degree of information asymmetry between the firm's management and corporate outsiders.