Premium
Learning and Estimation of the New Keynesian Phillips Curve Models
Author(s) -
Liu Dandan
Publication year - 2011
Publication title -
southern economic journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.762
H-Index - 58
eISSN - 2325-8012
pISSN - 0038-4038
DOI - 10.4284/0038-4038-78.2.382
Subject(s) - output gap , phillips curve , economics , proxy (statistics) , new keynesian economics , econometrics , inflation (cosmology) , measure (data warehouse) , extant taxon , estimation , monetary policy , keynesian economics , mathematics , statistics , computer science , physics , management , database , evolutionary biology , theoretical physics , biology
The extant literature is split on the best measure of marginal cost in the New Keynesian Philips Curve, with the output gap and the labor share being the most commonly advocated proxy measures. Which one is the best measure? In this article, I assume that agents update their understanding and expectation period by period, a learning process. In terms of econometrics, I use a recursive Vector Autoregression approach and conduct a forecasting exercise that considers updating of information sets used for formation of expectation. I find that the traditional output gap measure is a more significant variable explaining the dynamics of the U.S. inflation rate, as compared with a measure of the labor income share. Furthermore, the role of the output gap cannot be replaced using lagged values of inflation. Instead, both the output gap and lags of the actual inflation rate are important determinants of inflation.