Maximum Likelihood Estimation and Inference in DSGE Model with Possible Weak Identification

• Activity Number: 689
• Type: Contributed
• Date/Time: Thursday, August 8, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Business and Economic Statistics Section
• Author(s): Linchun Chen*+
• Companies: University of California, San Diego
• Keywords: Maximum Likelihood ; Weak Identification ; DSGE model ; Asymptotic Size ; Confidence Set

Abstract

This paper investigates likelihood-based estimation and inference on weakly identified Dynamic Stochastic General Equilibrium (DSGE) models. We suggest weak identification (usually regarded as less curvature of likelihood function) be divided into three types, model-based, data-based and computation-based. Focusing on model-based weak identification, we propose a measure of strength of identification and analyze the impact of degree of identification on inconsistency of structural estimators and size distortion of several common used tests in asymptotic framework. To remedy such issues, under certain drifting sequences, we suggest two identification-robust tests and their confidence sets, whose asymptotic sizes are equal to their nominal sizes. To conclude, we use both Monte Carlo simulation and empirical example of a medium-scale DSGE model (Smets and Wouters, 2007) to illustrate our results.