Disclosure Rules, Manager Discretion and the Relative Informativeness of Earnings Components

Abstract

This is a study of earnings quality, examining whether components of earnings based on New Zealand (N.Z.) accounting classification systems have different information parameters. The N.Z. environment provides a unique opportunity to examine a period with no legislative backing of accounting standards and a flexible accounting standard. Combined, this gave mangers the ability to clearly identify earnings components they considered to be differentially informative. Informativeness is assessed by the ability of current period earnings to predict next period earnings and the contemporaneous relation between returns and earnings. The results indicate that disaggregated reported earnings are more informative than aggregated earnings in a non-trivial way. In one of the sample periods disaggregated earnings explained 29% of the variance in returns, more than twice the explanatory power of aggregated earnings. N.Z. accounting standard setters replaced SSAP7 with FRS7 in 1994 contending that the discretion available to mangers reduced the informativeness of earnings. Not only do the results not support that contention but earnings informativeness has fallen since FRS7 came into effect, suggesting that standard setters should revisit that decision. The results also have implications for the content and form of the N.Z. Stock Exchange (NZSE) preliminary announcement. "Unusual earnings" reported to the NZSE by companies are shown to be differentially informative to investors yet the NZSE does not always identify these components when the preliminary announcement is summarised and disseminated to market participants. To summarise, the effective codification of earnings brought about by FRS7 has reduced the informativeness of earnings – locking differences between components into total earnings. The N.Z. results beg the question as to whether similar economic events are locked into the COMPUSTAT summary earnings variables for U.S. data.