On the firm-specific component of conservatism and its determinants.

摘要

This dissertation addresses the fundamental question of whether there are systematic differences in the level of conservatism across firms. I develop a conceptual framework to identify three components of conservatism - a constant time-invariant component, a firm-specific and time-invariant component and a firm-specific and time-variant component. The conceptual framework explains that recent concerns raised in the literature on whether conservatism and earnings management are two sides of the same coin are mostly driven by this last component of conservatism. By removing discretionary accruals from earnings (effectively this last component of conservatism), the paper documents asymmetric response to good news and bad news. I then construct an over-time firm-specific conservatism index and show that this measure is influenced by the level of competition at an industry level, growth, leverage, firm-specific risk and the predictive ability of earnings. Finally, I also examine how the stock market reacts to good (bad) news coming from more (less) conservative firms. I find that the earnings response coefficient is higher for good news coming from more conservative firms. On the other hand, the market reaction to bad news is stronger for less conservative firms, consistent with the idea that the market considers these earnings to be more credible.