An empirical examination of the time series properties of earnings per share using transfer function analysis at the industry level

The purpose of this study was to examine the time-series properties of accounting earnings at the industry level. The study utilized transfer function analysis to identify and estimate industry specific transfer functions. These functions utilized industry indices and macroeconomic time-series variables as input series. Several industry specific earnings models were identified for a broad sample of industries using both firm specific and cross-sectional approaches. Firms were grouped into industries based upon their four-digit Standard Industrial Classification codes. Several of these industry groupings displayed a high degree of heterogeneity among their earning-per-share series. For these industry groupings, identification of a common earnings model was not successful. In other groupings, distinctive patterns were detected in cross-sectionally derived crosscorrelation functions indicating a relatively high degree of homogeneity among their earnings series. For these industry groupings, the identification of a common earnings model was more successful.