This study provides evidence that differential interpretations are an important stimulus for speculative trading. We measure differential interpretations using data on analysts' revisions of forecasts of annual earnings after the announcement of quarterly earnings that are components of those annual earnings numbers. We find two conditions under which differential interpretations play a significant role in explaining trading. First, we present empirical evidence supporting Kandel and Pearson's (1995) argument that trading coincident with small price changes reflects investors' differential interpretations of information. This evidence is important because it is inconsistent with conventional models of trade that assume homogeneous interpretations. Second, we also find that differential interpre? tations explain a significant amount of the trading occurring in a sample where trading volume is higher than the (firm-specific) non-announcement period average. This result is consistent with informed traders acting on their differential interpretations when there is enough liquidity trading to help camouflage their own information-based trades. In sum, the study's results confirm Bachelier's (1900) intuition that differential interpretations are an important stimulus for trading.

Beginning in the 1990s, firms often continue to trade on the major national exchanges after Chapter 11 bankruptcy filings. For bankruptcies filed from 1993–2003, we find that the more negative the filing period price reaction, the more favorable the immediate post-filing returns, on average. This reversal is not attributable to bid-ask bounce, it holds after controlling for other factors associated with post-filing returns, and it appears more attributable to the activities of large traders than to small traders. Supplementary tests reveal that the pattern of post-filing returns differs significantly for bankruptcies filed in bull versus bear markets. Bankruptcies filed during the 1993 to 1999 bull market enjoy substantial but short-lived reversals averaging one-third of the filing period price plunge. These reversals are inconsistent with efficient assimilation of the bankruptcy information. In contrast, we find no evidence of post-filing reversals for bankruptcies filed from 2000 to 2003.