The purpose of this paper is to further understanding of how new information impacts the market value of financial assets.
The paper uses a Bayesian approach to asset valuation, whereby investors use signals conveyed by new information to update their estimate of a structural valuation parameter. The underlying distributions – i.e. the distribution of the information signal and the prior distribution of the valuation parameter – are allowed to exhibit a degree of kurtosis greater than that of the normal distribution.
The revision in asset value as a function of the realization of the information signal is an S‐shaped function (in the local region centred on the zero‐surprise level of the signal), if the distribution of the information signal features excess kurtosis; conversely, if the prior of the valuation parameter features excess kurtosis, the revision in asset value is an inverted S‐shaped function.
The paper generates clear implications with respect to the shape of the function relating the revision in asset value to the realization of the signal only in the local region centred on the zero‐surprise level of the signal.
The paper helps to shed light on the well‐known empirical result that the stock price reaction to earnings' announcements is an S‐shaped function, centred on the zero‐surprise level of reported earnings.
In the financial accounting literature, the paper helps one to understand the role of the distributional assumptions underlying the stock price reaction to earnings' announcements, namely, the role of excess kurtosis both in reported earnings and in the prior of means earnings.
