Do investors overreact to patterns of past financial performance measures?

摘要

The objectives of this thesis are threefold. First, this dissertation examines whether patterns (growth and consistency in growth) of firms' past financial performance influence investors' perceptions about stock values and future performance of these firms. Second, multiple estimation horizons of past performance variables (ranging from one to five years) are used to assess whether the interaction between growth patterns and measurement interval lengths of these variables influence investor expectations. Third, this thesis examines whether an intermediate price drifts (e.g. Jegadeesh and Titman [1993]) and subsequent long-horizon price reversal (e.g. DeBondt and Thaler (1985)] are manifestations of a market over-reaction as suggested in recent studies (e.g. Lee and Swaminathan [2000]).; The evidence provided in this study shows that low-growth firms outperform their high-growth firm counterparts across different performance variables, estimation intervals, and investment horizons except in the first post-formation year for firms ranked by their prior one-year financial growth rate (except for sales growth).; Furthermore, when ranking firms based on the consistency of their prior financial growth rates over multiple estimation periods, this study finds that a firm consistently achieving low (high) growth rates that places it in the lowest (highest) growth 40 percent earns high (low) stock returns. The consistency in a firm's prior financial performance influences the behavior of its future stock returns, i.e. the longer the consistency of exceptionally strong (weak) performance of a firm, the greater (lower) its subsequent stock returns.; First year returns for firms except SG exhibit a strong financial drift. The price drift seems to persist even with longer estimation horizons. This result indicates that the evidence of a price drift reported in the first post-formation year might be due to a market over-reaction, a finding consistent with results reported by Lee and Swaminathan (2000).; For all variables (except for past sales growth and to some degree past stock returns), the financial drift in year one return for portfolios based on prior-one year of past performance data, is significantly stronger than that reported in Chan et al. (2004). Results reported in this thesis indicate that the average abnormal return differential between low and high growth firms for the five-year estimation intervals (with exception of past sales growth) is greater than 10 percentage points. (Abstract shortened by UMI.)