| This dissertation tests the Hong and Stein (1999) model of gradual information diffusion leading to returns momentum. I hypothesize that investors who are confident with their own assessments of the implications about firms' prospects are slow to extract information from other investors' trading behaviors, leading to slow information diffusion in the market. Drawing on psychology literature, input consistency and input amount increase confidence. I refer to analysts' forecast as inputs to investors' decisions and measure input consistency as forecast conformity to a trend line and input amount as the number of forecasts issued throughout the period.; I find that returns momentum depends on forecast conformity but I fail to find any evidence that returns momentum depends on forecast frequency. In further analysis, the results suggest that when analysts adjust their forecasts over time in a manner consistent with a positive trend, investors adjust their beliefs about firms' prospects and their confidence in these new beliefs increases with the degree of conformity of the analysts' forecasts to the trend line. However, forecast conformity does not appear to explain returns momentum when the trend is extremely negative. When analysts' forecasts suggest a zero trend, returns momentum does not exist. |
