The Dynamics of Financial Reporting: The Case of IPOs

We examine the relevance of earnings announcements in an IPO setting. Given the cross-sectional differences in earnings informativeness that can be identified in a setting of mature firms, an important but yet unexplored question is how these firms arrived at their respective equilibrium levels. We find considerable cross sectional variation in the post-IPO trend in earnings informativeness. For this purpose, we employ a latent class model to group firms into types on the basis of the starting level and the trend in earnings relevance. We then examine whether we can predict the firm’s “type” with variables know at or shortly after the IPO. We find that despite the significance of some IPO variables (e.g. underwriter quality, underpricing, IPO analyst following), the predictive power of the model is low, which implies that it is very difficult to separate firms at the outset. Additional tests reporting that over time the cross-sectional variance in earnings informativeness increases and that the predictive ability of lagged earnings relevance for current earnings relevance increases support this view. The next question we address is what explains the development of earnings informativeness over time. We find that both analysts and institutional owners drive increases in earnings informativeness rather than them being attracted to firms with informative earnings. Furthermore, initial earnings relevance levels are positively linked to post-IPO earnings informativeness trends, which is inconsistent with changes in inherent earnings informativeness being the main driver of differences in earnings relevance trends. It is however consistent with investor learning and reputation building by high type firms.