I. Introduction
This article deals with an adverse selection problem of the general class considered by
Baron and Myerson (1982). Agents have private information regarding a “type” parameter.
The uninformed principal acts as a Stackelberg leader, designing the best contract possible,
given the information asymmetry. There is a rich theoretical literature extending this basic
model. Typically the precise terms of an optimal contract schedule depend on the principal’s
prior beliefs regarding the probability distribution of types and the manner in which type
affects agents’ reservation utility. What is lacking in the literature is an easily implemented
methodology by which the principal can develop robust beliefs regarding these two items
with data that do not include agents’ private information. This article develops such a
methodology and applies it to the problem of reducing the cost of one of the United States’
largest environmental policies, the Conservation Reserve Program (CRP). Administered by
the U.S. Department of Agriculture (USDA), this program pays agricultural producers to
refrain from cultivating environmentally-sensitive cropland. The CRP is one of the largest
environmental programs in the U.S., both in terms of budget ($1.7 billion per year) and
scope (34.7 million acres) (Farm Service Agency, 2005b).
The key econometric problem is estimating the parameters of the type-dependent
production technology and the probability distribution of agent types when type is inherently
unobservable both to the principal and the econometrician. One branch of the empirical con-
tract theory literature, such as Wolak (1994), Thomas (1995), and Lavergne and Thomas
(2005) conducts estimation under the hypothesis that existing contracts are optimally de-
signed by a sophisticated principal. The authors first derive an optimal mechanism. This
analysis provides equations that can be econometrically estimated with observable data to
infer the principal’s beliefs regarding the distribution of types and the impact of type on the