Evidence Against Imposing Restrictions on Hurdle Models as a Test for Simultaneous vs. Sequential Decision Making
William J. Burke
Print
EmailAbstract
Agricultural economists frequently employ hurdle models to estimate the determinants of truncated outcomes such as market participation and adoption. A pervasive belief is that restrictions can be placed on hurdle models to test whether the decisions made in the underlying data generating process occurred sequentially or simultaneously. This article argues against the ability to draw this conclusion and further submits there is a negative correlation between failing to reject these restrictions and sample size. Evidence to support both proposals comes from data collected in a natural setting as well as simulated data with a known data generating mechanism.
