Measuring stability and growth in annual EPS

T he measurement of stability and growth in earnings per share, or EPS, has been an important concern of researchers and practitioners of financial analysis for years. The generally accepted measures of EPS stability and growth are derived from the loglinear regression model. Logarithms cannot be used, however, if the EPS series has a negative or zero value. In this paper, we introduce measures of EPS stability and growth derived from the simple OLS linear regression model for raw-score data. The presence of negative or zero EPS values leaves these measures unaffected. Our new measures will be compared to the log-linear regression measures using 10 years of annual EPS data, from the period 1970-1980, collected for each of the Dow Jones 30 industrial companies.' Our choice of 10 years of data is consistent with the usual time frame employed in this type of stability and growth analysis. The EPS data have been adjusted for stock dividends and splits.