that the supplemental disclosures with respect to specific accruals can permit capital markets to form unbiased estimates of the implications of the accrual for future earnings and hence for valuation. For example, Beaver and McNichols (2001) show that increased disclosure regarding the history of policy loss reserves in the property casualty insurance companies can make the accruals transparent to investors, and that revisions policy of loss accruals (development) are not associated with subsequent abnormal returns. Unresolved Issues The magnitude and length of the abnormal returns is surprising. For example, Frankel and Lee (1998) report that in the 36 months after portfolio formation, the abnormal returns associated with market-to-value strategies are 31 percent, whereas strategies that also exploit the predictability of analysts' forecasts are associated with abnormal returns of 45 percent. There are several unresolved issues (1) How can widely disseminated and examined data used with simple portfolio strat- egies that require no knowledge of accounting be associated with abnormal returns? From an economic perspective, widely disseminated data are not likely candidates. (2) How can studies of arcane disclosures (e.g., nonperforming loans and pensions, as in Beaver et al. [1989] and Barth et al. [1992]) find that such disclosures are apparently reflected in prices, yet more visible variables, such as earnings and book value, are not? (3) How can studies of security returns in the very short run (e.g., intraday returns, as in Patell and Wolfson [1984]) show evidence of relatively rapid response (within hours, if not minutes), and yet have evidence of abnormal returns that appear to persist for years after the portfolio formation date?