Earnings per share and cash flow per share as determinants of share value : tests of significance using the bootstrap with Demsetz's method

Demsetz (1995) develops a method to determine the relationship between share prices and accounting profitability measures. Unfortunately, the probability distributions of the underlying statistics are unknown. This paper uses bootstrap methodology for obtaining standard errors associated with these statistics as well as percentiles for the respective distributions. We apply the bootstrap analysis to test for the relative importance of earnings per share (EPS) and cash flow per share (CPS) in explaining share prices. Contemporaneous year-on-year changes in EPS is statistically significantly better at explaining year-on-year changes in share prices than year-on-year changes in CPS. This confirms that the analysis of EPS is economically meaningful.