Landes & Posner on the Economics of Presidential Pardons

William M. Landes and Richard A. Posner (University of Chicago Law School and University of Chicago Law School) have posted The Economics of Presidential Pardons and Commutations on SSRN. Here is the abstract:

This paper develops a simple economic model of the demand for and supply of presidential pardons (including commutations and other clemency grants). The model assumes that the number of pardon applications depends on the expected benefits and costs of applying and the number of pardons depends on the president's calculation of his net political benefits from approving an application. This yields an equilibrium number of applications and pardons that can be estimated from time series data over the 1900 to 2005 period. Overall, the regression results support the model. For example, we find that the likelihood of receiving a pardon has a positive effect on applications; and that increases in the number of persons paroled (a substitute for a pardon) and in the time from conviction to pardon (which lowers the benefits of a pardon) reduces the number of applications. We also observe a positive time trend in applications that corresponds to the increase in the number of federal prisoners (and persons released from prison) that make up the pool of potential pardon applicants. With respect to the number of pardons, we find that democratic presidents (who we expect to be less tough on crime) are more likely to grant pardons; that the crime rate has a negative impact on pardons; and that the number of pardons increased during Prohibition and during wartime and postwar periods.

I wonder if anyone who has worked in the "Office of Pardon Attorney"--which must be the de facto decision maker for most pardon decisions--would have a comment on the assumption that the supply of pardons is a function of political benefits?