OVERCONFIDENCE 1203 x104 a b 0.04 E() 0.02 0 2 1 2 。=/。 l=a/c 0 0 0.4 0.2 2 2 3 。 =/c, FIc.2.-Effect of information in signals:a,trading barrier;&duration between trades; G bubble;d,extra volatility component.Here,r=5 percent,A =0.0 0.1,=0.7, i=0,and c=10".The values of the bubble and the extra volatility component are computed at the trading point.The trading barrier,the bubble,and the extra volatility component are measured as multiples ofo(r+A),the fundamental volatility of the asset. In both cases,there is a monotonically increasing relationship be- tween the size of the bubble at the trading point and duration between trades.In addition,the extra price volatility either increases or does not decrease.We have also verified that this qualitative relationship holds for many other parameter values.In our risk-neutral world,we may consider several assets and analyze the equilibrium in each market in- dependently.In this way our comparative statics properties can be trans- lated into results about correlations among equilibrium variables in the different markets.Thus our model is potentially capable of explaining the observed cross-sectional correlation between log market/book and log turnover for U.S.stocks in the period 1996-2000 as documented by Cochrane (2002)and a similar cross-sectional correlation between the price ratio of China's A shares to B shares and turnover (see Mei, Scheinkman,and Xiong 2003). Reproduced with permission of the copyright owner.Further reproduction prohibited without permission.Reproduced with permission of the copyright owner. Further reproduction prohibited without permission