Accounting information and the pricing of initial public offerings

Two issues pertaining to the pricing of initial public offerings (IPOs) are investigated in this dissertation: (1) The first issue investigated is whether accounting measures of total risk are associated with the initial returns on IPOs. Some researchers posit that the initial returns on IPOs are related to the ex ante uncertainty about the aftermarket prices of the securities. A set of four accounting measures of total firm risk--comprising financial leverage, operating leverage, growth and size--is used to proxy for ex ante uncertainty and is regressed on the initial returns. Tests are conducted on a sample of 240 firms from twelve industries that made IPOs between 1981 and 1984 using firm commitment contracts. The results with least-squares regression using a heteroskedasticity-consistent covariance matrix (White (1980)) and approximate randomization indicate that the set of accounting variables has significant explanatory power. Multicollinearity makes it difficult to make inferences about the explanatory power of the individual variables. When industry indicator variables are included in the regression, the set of accounting variables is not significant at traditional levels, although some of the variables in the set show evidence of explanatory power. (2) The second issue investigated is whether issuers of IPOs make income increasing discretionary accruals in the periods before going public. Economic incentives exist for issuers to obtain the highest possible proceeds from their issues and anecdotal evidence indicates that accounting numbers are used in the pricing of IPOs. Thus, it is predicted that issuers use accounting discretion to increase earnings in the periods before going public. The results allow rejection of the null hypothesis that issuers do not make income increasing discretionary accruals before going public. The results show that issuers make income increasing discretionary accruals only in the last accounting period reported in prospectuses. Firms that provide interim accounting data in their prospectuses make income increasing discretionary accruals in the interim period, but not in the last full year reported. In contrast, firms that do not provide interim data do make income increasing discretionary accruals in the last full year before going public, which is the last accounting period reported in the prospectus. Discretionary accruals are estimated using a modification of the method introduced by DeAngelo (1986). Total accruals in the benchmark period are adjusted for growth by multiplying by the ratio sales in the test period to sales in the benchmark period. The difference between total accruals in the test period and growth-adjusted accruals in the benchmark period is the estimate of discretionary accruals.