Some personal observations on the debate on the link between financial reporting quality and the cost of equity capital. (December 2013)

• Record Type: Journal Article
• Authors: Shevlin, Terry
• Is Part Of: Australian journal of management. Volume 38:Number 3(2013)
• Page Start: 447
• Page End: 473
• Publication Date: 2013-12
• Subjects: Financial reporting quality -- information risk -- price-risk factor; Management -- Periodicals; Management -- Australia -- Periodicals; Business enterprises -- Finance -- Periodicals; Industrial management -- Australia -- Periodicals; 330
• DOI: 10.1177/0312896213510669
• Languages: English
• ISSNs: 0312-8962

Abstract

Within the last decade there has been much written about the possible link between the quality of a firm's external financial reporting and its cost of equity capital. In this paper I provide my personal observations about the literature. I make the following points. I find the Francis et al. (2005) paper interesting and innovative. This paper documents a positive association between portfolio returns formed on the basis of accrual quality and firms' realized returns after controlling for the Fama-French three factors. Francis et al. conclude that accrual quality is a priced risk factor. Second, while this conclusion might be too strong based on their results, I think the Core et al. (2008) critique of the Francis et al. claim is also too strong and does not lead me to reject accrual quality as a priced risk factor. Third, the evidence of an association between financial reporting quality and implied cost of capital estimates is robust. However several papers question the construct validity of these measures which are based on analyst earnings forecasts. I discuss these papers and conclude that the implied cost of capital does have empirical validity. Finally, various papers model the relation between financial reporting quality and cost of capital and, while interesting, mostly restrict themselves to a single factor asset pricing model. However, restricting the model to a single factor model by assumption rules out the possibility of information quality being a separately priced risk factor and further restricts the role information quality to informing about a firm's covariances (or estimation error in the CAPM beta) because it is assumed that information quality is diversifiable. I think whether something is diversifiable is an empirical question that cannot be resolved by assumption within a theoretical model.