SEPARATION HYPOTHESIS TESTS IN THE AGRICULTURAL HOUSEHOLD MODEL

Abstract

In an agricultural household model, farmers’ production decisions can be either separated or nonseparated from preferences. Since previous studies on agricultural household behavior and policy effects have shown that the model with separation yields different results than the model with nonseparation, it is important to have a test to identify the correct model. This paper provides new tests that extend current tests in two directions. First, the new tests avoid issues that current tests have to address, such as simultaneity bias and the estimation of the production function. Second, the new tests make use of more information implied from the separation hypothesis than current tests