Uniformly Consistent Empirical Likelihood Estimation Subject to a Continuum of Unconditional Moment Inequalities

This paper extends moment-based estimation procedures to statistical models defined by a continuum of unconditional moment inequalities. The underlying probability distribution in the model is the (infinite dimensional) parameter of interest. For a general class of estimating functions that indexes the continuum of moments, we develop the estimation theory of this parameter using the method of empirical likelihood and a feasible value-function-approach. This estimator is uniformly consistent over the set of underlying distributions in the model, and for large sample sizes, it has smaller mean integrated squared error than the estimator that ignores the information in the moment inequality conditions for DGPs that have at least one moment inequality that is slack.