Working Paper Series no. 83: Evaluation de l'estimation par les GMM de la fonction de réaction de la Federal Reserve Bank

Abstract

Estimating a forward-looking monetary policy rule by the Generalized Method of Moments (GMM) has become a popular approach since the influential paper by Clarida, Gali, and Gertler (1998). However, an abundant econometric literature underlines the unappealing small-samples properties of GMM estimators. Focusing on the Federal Reserve reaction function, we assess GMM estimates in the context of monetary policy rules. First, we show that three usual alternative GMM estimators yield substantially different results. Then, we compare the GMM estimates with two Maximum-Likelihood (ML) estimates, obtained using a small model of the economy. We use Monte-Carlo simulations to investigate the empirical results. We find that the GMM are biased in small sample, inducing an overestimate of the inflation parameter. The two-step GMM estimates are found to be rather close to the ML\ estimates. By contrast, iterative and continuous-updating GMM procedures produce more biased and more dispersed estimators.

Clémentine Florens, Eric Jondeau and Hervé Le Bihan
March 2001

Classification JEL : E52, E58, F41

Keywords : Forward-looking model, monetary policy reaction function, GMM estimator, FIML estimator, small-sample properties of an estimator