Calculating earnings per share: An empirical analysis estimating the dilutive impact of convertible securities on a current basis

This study investigated the accuracy of distinguishing on a current basis convertible debentures between dilutive and non-dilutive groups. The test was based upon actual dilution from subsequent conversions, not just some level of conversion taking place. The study\u27s premise was to determine whether the accuracy of the common stock equivalency requirements could be enhanced within the current levels of objectivity and cost effectiveness or whether the concept should be discontinued as some commentators have suggested. A random sample of 191 convertible debenture issues outstanding on December 31, 1984, was taken with a review for actual dilution over the subsequent three years made. Issues showing potential dilution of three percent or more (the materilaity level included in Accounting Principles Board Opinion No. 15) and resulting in one-half or more conversion during the test period were classified as dilutive (since Opinion No. 15 requires the issue be classified as one, the dilutive impact should be included only when at least one-half of the issue is converted). All other issues selected with a potential for material dilution were classified as non-dilutive. Two ratios were tested on a current basis regarding their ability to predict the actual classifications. The results show that the more theoretically supported conversion value/call price ratio with a cutoff value of one (other cutoff amounts examined showed no additional success) and the yield to maturity/Aa bond rate measure with a cutoff value of 2/3 correctly predicted 85.6 and 80.8 percent of the classifications respectively. The difference in results was found insignificant. However, both success rates were determined as significantly greater than the 60 percent properly predicted as non-dilutive assuming the equivalency concept is discontinued. The results thus show that equally significant accuracy in primary earnings per share can be achieved by either continuing all earnings per share requirements other than frequency of the equivalency test or by utilizing annually a more theoretically supported conversion value/call price measure