The relation between the square of the quoted bid-ask spread and two serial covariances—the serial covariance of transaction returns and the serial covariance of quoted returns—is modeled as a function of the probability of a price reversal, π, and the magnitude of a price change, ∂, where ∂ is stated as a fraction of the quoted spread. Different models of the spread are contrasted in terms of the parameters, π and ∂. Using data on the transaction prices and price quotations for NASDAQ/NMS stocks, π and ∂ are estimated and the relative importance of the components of the quoted spread—adverse information costs, order processing costs, and inventory holding costs—is determined.

The bid‐ask spread can be decomposed into two parts: one part due to asymmetric information and the other part due to other factors such as monopoly power. The part due to asymmetric information attenuates statistical biases in mean return, variance, and serial covariance. Thus, using spread data to adjust for biases in return moments requires knowing not only the spread but the composition of the spread. Furthermore, any spread‐estimation procedure using transaction prices must estimate two spread components. On the other hand, the appropriateness of some previously suggested statistical corrections is independent of the spread composition.

We consider an Arrow‐Debreu model with agents who have different subjective probabilities. In general, asset prices will depend only on aggregate consumption and the distribution of subjective probabilities in each state of nature. If all agents have identical preferences then an asset with “more dispersed” subjective probabilities will have a lower price than an asset with less dispersed subjective probabilities if risk aversion does not decline too rapidly. It seems that this condition is likely to be met in practice, so that increased dispersion of beliefs will generally be associated with reduced asset prices in a given Arrow‐Debreu equilibrium.