Risk Aversion,Uncertain Information,and Market Efficiency

Abstract. We reexamine and extend tests of the uncertain information hypothesis (UIH) proposed by Brown, Harlow, and Tinic (1988, 1993). We find that their empirical results are sensitive to the sampling procedure employed and that their particular methodology does not sufficiently distinguish between event and nonevent periods. When the sampling procedure is modified to identify only relatively large, isolated events, the test results generally do not support the UIH. Instead, significant price shocks are consistently followed by short-lived price reversals. We observe this behavior following positive and negative events regardless of whether the event is classified as risk increasing or risk decreasing.