Do Investors Perceive Marking-to-Model as Marking-to-Myth? Early Evidence from FAS 157 Disclosure

Abstract

Using disclosure mandated by Statement of Financial Accounting Standards (FAS) 157 Fair Value Measurements, I evaluate the concern that fair value estimates for assets and liabilities not traded in active markets, aka mark-to-model, are too unreliable to be used in financial reporting. I document a significant positive association between stock prices and fair values of net assets measured using unadjusted market prices (Level 1), other observable inputs (Level 2), and significant unobservable inputs (Level 3). While the estimated coefficients on the mark-to-model estimates (Levels 2 and 3) are consistently lower than those on the mark-to-market fair values (Level 1), the difference is significant only for Level 3 net assets. Furthermore, even at its maximum, the difference does not exceed 35% of the coefficient on Level 1 net assets. Additional analysis suggests that the valuation gap is more pronounced for firms with lower equity capital and fewer financial experts on the Audit Committee, as well as for companies that develop their mark-to-model estimates internally. I also document a significant positive association between Level 3 net gains and both quarterly returns and returns for the three-day period surrounding the filing of Form 10-Q. Collectively, the results suggest that equity investors find mark-to-model fair value estimates sufficiently reliable to be reflected in firm value.