Bubbles and stock market behavior: Rational and irrational.

摘要

The purpose of this dissertation is to conduct an empirical analysis on the returns of the U.S. stock market during the 1990's. The important questions are: the existence of empirical evidence of rational speculative bubbles, examining evidence of irrational behavior, and considering evidence of shifts in investor's psychological expectations. The relationship between stock market value based on market fundamentals and the stock market price are analyzed in the context of rational speculative bubble theory. The theory of speculative bubbles predicts that stock market prices fluctuate around a fundamental value path and price bubbles develop as a series of small persistent steps away from their path. Any sudden movement back to the fundamental path is the popping or bursting of the bubble. I test this theory by attempting to capture these characteristics when present.;A direct test of the theory of rational bubbles is developed following the methodology of Charles Bischoff, et al. utilizing two different estimation models: the irrational model of Modigliani-Cohn (1979) and a Factor Model following the methodology of Chen, Roll and Ross (1986). I define a bubble as a movement away from the fundamental estimated valuation path for two or more consecutive periods. I test for a probability growth rate and the existences of a bubble using Monte-Carlo experiments. Rational theory requires that the bubble grows at a rate sufficient to compensate investors for the possibility of a crash.;My calculations suggest that I cannot formally reject at least some version of this rational bubble theory. However, over the time period studied (1968:1--2002:5), the rate of inflation of the U.S. equity bubble implied by a number of Monte Carlo studies is not large enough to regard complete adherence to the rational bubble theory as quantitatively plausible.;Empirical tests of the stability of the fundamental path estimation model suggest that a structural change occurred in the stock market during the 1990's. Explanation for this model structural change and subsequent stock market behavior is related to changing levels of average risk aversion and shifts in investors' expectations. I test the implications of changes in the degree of average risk aversion on post market returns. I find a significant 12-month lag between changes in average risk aversion and subsequent stock market returns.