Testing the Structural Interpretation of the Price Puzzle with a Cost‐Channel Model*
| Author | Efrem Castelnuovo |
| DOI | http://doi.org/10.1111/j.1468-0084.2011.00658.x |
| Published date | 01 June 2012 |
| Date | 01 June 2012 |
425
©Blackwell Publishing Ltd and the Department of Economics, University of Oxford, 2011. Published by Blackwell Publishing Ltd,
9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA.
OXFORD BULLETIN OF ECONOMICS AND STATISTICS, 74, 3 (2012) 0305-9049
doi: 10.1111/j.1468-0084.2011.00658.x
Testing the Structural Interpretation of the Price
Puzzle with a Cost-Channel Model*
Efrem Castelnuovo†
†Department of Economics, University of Padova and Bank of Finland, Via del Santo 33, 35123
Padova (PD), Italy (e-mail: efrem.castelnuovo@unipd.it)
Abstract
Weestimate a variety of small-scale new-Keynesian DSGE models with the cost channel to
assess their ability to replicate the ‘price puzzle’, i.e. the inflationary impact of a monetary
policy shock typically arising in vector autoregression (VAR) analysis. To correctly identify
the monetary policy shock, we distinguish between a standard policy rate shifter and a shock
to ‘trend inflation’,i.e. the time-varying inflation target set by the Fed. Our estimated models
predict a negative inflation reaction to a monetary policy tightening. We offer a discussion
of the possible sources of mismatch between the VAR evidence and our own.
I. Introduction
What is the short-run reaction of inflation to an unexpected and temporary monetary policy
tightening? Macroeconomic textbooks suggest that inflation should react negatively to
such a monetary policy move (Woodford, 2003a; Gal`ı, 2008). As a matter of fact, how-
ever, empirical investigations based on the vector autoregression (VAR) methodology cast
doubts on this prediction.
Figure 1 (top panel) depicts the impulse response functions produced with a VAR esti-
mated with post-WWII US data.1An unexpected one-shot increase in the policy rate leads
to (i) a significantly positive reaction of the policy rate, (ii) a significantly negative reac-
tion of the output gap, and (iii) a significantly positive reaction of inflation. This evidence
stands in stark contrast with conventional wisdom. Eichenbaum (1992) labels this evidence
ÅThe author thanks John Knight (Editor), two anonymous referees, Eric Mayer and Pau Rabanal, and Paolo Surico
for insightful feedbacks on earlier versions of this article. We also thank St´ephane Adjemian, Guido Ascari, Jacopo
Cimadomo, Dudley Cooke, Juha Kilponen, Ola Melander, Gaia Narciso, Peter Tillmann, Matti Vir´en and seminar
participants at the Bank of Finland, CEPII (Paris), University of Brescia, Trinity College Dublin, University of Pavia
as well as conference participants at ASSET2007 (Padua) for useful comments and suggestions. Part of this research
was conducted while visiting the Sveriges Riksbank, whose kind hospitality is gratefully acknowledged. All remain-
ing errors are ours. Financial support by the Italian Ministry of University and Research (PRIN 2005-N. 2005132539)
and the University of Padua is gratefully acknowledged. The view expressed in this article is not necessarily that of
the Bank of Finland.
JEL Classification numbers: E30, E52.
1Impulse responses related to a trivariate VARwith GDP deflator inflation, Congressional Budget Office (CBO)
output gap, and federal funds rate, sample: 1954:III–2008:II. Similar evidence, originally reported by Sims (1992),
is also put forward by Stock and Watson (2001) and Rabanal (2007). Hanson (2004) shows that this result is robust
to the introduction of commodity prices as well as a variety of other inflation predictors in the VAR.
426 Bulletin
51015
–0.1
–0.05
0
0.05
0.1
0.15
Inflation
Output gap
Real unit labour cost
Policy ratePolicy rate
51015
–0.4
–0.2
0
0.2
0.4
51015
–0.05
0
0.05
0.1
0.15
0.2
0.25
0.3
51015
–0.1
–0.05
0
0.05
0.1
0.15
Inflation
51015
–0.4
–0.2
0
0.2
0.4
51015
–0.05
0
0.05
0.1
0.15
0.2
0.25
0.3
Figure 1. SVARimpulse response functions to a monetary policy shock. Sample: 1954:III–2008:II. Variables
– top panel: quarterly GDP inflation, CBO output gap, quarterly federal funds rate; bottom panel: variables
– top panel: Quarterly GDP inflation, Real unit labour costs (ULCs)(non-farm business sector), quarterly
federal funds rate. Source: FREDII. Identification of the monetary policy shock via Cholesky decomposition
(lower triangular matrix, ordering: quarterly inflation, output gap, quarterly federal funds rate). Solid line:
mean response. Dotted lines: 90% confidence bands (analytically computed). VAR estimated with four lags.
Real (ULC) measure considered in percentualized log-deviations with respect to its Hodrick–Prescott trend
(weight: 1600)
the ‘price puzzle’ (the ‘VAR evidence’ henceforth). Importantly, Castelnuovo and Surico
(2010) show that this result is robust to the implementation of an alternative identification
strategy, based on sign restriction, which does not assume recursiveness, and it is then con-
sistent with the timing of models such as the popular standard new-Keynesian framework.
In fact, a possible interpretation of this VAR empirical regularity is offered by models
embedding the ‘supply channel’,otherwise known as the ‘cost channel’. The idea is simple.
Cash-constrained firms must borrow money from financial intermediaries to pay the wage
bills to workers before the goods market opens. Consequently, the interest rate paid on
borrowings enters firms’marginal costs and influences firms’ price setting, so giving a struc-
tural role to the presence of the policy rate in the new-Keynesian Phillips curve (NKPC).
This creates an extra link between monetary policy moves and aggregate inflation fluctu-
ations (Chowdhury, Hoffmann and Schabert, 2006; Ravenna and Walsh, 2006; Kilponen
and Milne, 2007; Surico, 2008; Tillmann, 2009; and Llosa andTuesta, 2009). Clearly, if the
inflationary impact induced by monetary policy moves via the supply channel is stronger
than the one operating via the standard ‘demand channel’, a positive reaction of inflation
to a monetary policy tightening may very well realize.
The plausibility of such a structural interpretation, however, is ultimately an empirical
issue. This article employs Bayesian techniques to estimate a new-Keynesian small-scale
DSGE model embedding the cost channel. The model is an extension of the baseline,
©Blackwell Publishing Ltd and the Department of Economics, University of Oxford 2011
To continue reading
Request your trial