abstract: This dissertation analyzes the reliability of reported employee stock option (ESO) expense, the determination of expected life of ESOs, motivations to manipulate ESO expense, and the impact of noise in ESO expense on subsequent stock price returns. Based on unique data, this is the first paper to measure average historical ESO life for all employees of a broad set of firms. I find average life has a mean of 4.12 years. Average life is reduced by 0.38 years per 10 percentage point increase in volatility, and industry effects explain an additional 7% of the variation. Reported expected life increases 0.37 years per year of historical life and an additional 0.16 years per year of age of the outstanding options. Deviations of reported volatility and life from benchmarks have positive correlations with deviations from own reporting history. Using stated assumptions rather than benchmark assumptions drops (increases) ESO expense by 8.3% (17.6%) for the 25th (75th) percentile firm. The change in earnings per share decreases (increases) by $0.019 ($0.007) for the 25th (75th) percentile firm. Tests for motivations to manipulate stock option expense downward have mixed results. Absolute values of deviations from benchmarks have a positive relationship with subsequent stock price volatility suggesting noise in reported stock option expense results in stock price noise. Deviations from benchmarks and subsequent cumulative abnormal returns have statistically significant results but are difficult to interpret. / Dissertation/Thesis / Ph.D. Business Administration 2011

The objective of this research is to test the expensing of stock options as part of CEO compensation to earnings quality. Agency theory posits a conflict between the CEO's own self-interest and that of the owners who seek to maximize the long term value of their investment. To avoid this conflict compensation should align and bond these parties. Data was retrieved from Compustat, ExecuComp and Corporate Governance databases spanning the years of 2000 through 2009. The Dechow and Dichev (2002) earnings quality model using the change in working capital and error terms taken as the residuals was utilized. All hypotheses used earnings quality as a proxy for management choices and as the predictive power of accruals. The first hypothesis indicated granting of CEO stock options has a positive association to earnings quality. The second hypothesis tests the implementation of SFAS 123 (R) by expensing stock options and the association to earnings quality. The third and final hypothesis utilized the number of BOD members as to compare the association between expensing stock options as part of CEO compensation and earnings quality. Empirical support for all three hypotheses was found and consistent with expectations established by other research using earnings quality methodologies. Both the granting and expensing of stock options as part of CEO compensation has an association to earnings quality. There exists a stronger association between expensing stock options and earnings quality when firms have a larger number of BOD members. Support for agency theory was discovered because all three hypotheses were supported. This study was limited to U.S. firms that were publicly traded on major U.S. exchanges and only CEO compensation. Other executive compensation was not included. These limitations provide opportunities for future research. Knowledge was gained by exploring the earnings quality measures for evidence of bonding and alignment theory. This study extends the research in earnings quality by examining the relationship of granting and expensing of stock options as per SFAS 123 (R). It also contributes to the work in SFAS 123 (R) by testing four years before and after 2005, when implementation occurred.