3:30 pm - 4:30 pm
550 General Services Building, 550 General Services Building University of Alberta , Edmonton Alberta
Title: Estimating Structural Models with Loose Constraints
Speaker: Dr. Dan Yu, Assistant Professor, Department of REES, University of Alberta
Date: Thursday, January 18, 2023
Time: 3:30 pm – 4:30 pm
Location: GSB 550
Abstract:
This paper proposes a method for estimating nonlinear economic models by simultaneously estimating endogenous objects —such as dynamic values and decision rules —in addition to the structural parameters of interest. This is achieved by simultaneously maximizing an estimation criterion and minimizing deviations from the constraints that define the solution of the model. In finite settings, we show consistency and asymptotic normality under standard conditions and derive a weighting scheme that allows the estimator to achieve the same efficiency as the standard approach (in which the estimation criterion is maximized while imposing a strict solution of the model). While other computationally convenient estimators require first-stage estimates of choice probabilities, this “Loose Constraints” method does not. It is, therefore, particularly useful in settings where the size of the state space is large relative to sample size or when the data do not reliably identify agents’ behavior without additional model structure.
Categories: