Description
The purpose of this paper is to try to understand the reasons for the differences in
amplitude of monetary policy (MP) rate cycles in the USA and the euro area. Among the different
candidates, the paper aims to test the role of economic structures, macroeconomic shocks, and MP
behaviour
Journal of Financial Economic Policy
The Fed and the ECB: why such an apparent difference in reactivity?
Alexis Penot Grégory Levieuge
Article information:
To cite this document:
Alexis Penot Grégory Levieuge, (2009),"The Fed and the ECB: why such an apparent difference in
reactivity?", J ournal of Financial Economic Policy, Vol. 1 Iss 4 pp. 319 - 337
Permanent link to this document:http://dx.doi.org/10.1108/17576380911050052
Downloaded on: 24 January 2016, At: 21:36 (PT)
References: this document contains references to 35 other documents.
To copy this document: [email protected]
The fulltext of this document has been downloaded 273 times since 2009*
Users who downloaded this article also downloaded:
Abdelkader Boudriga, Neila Boulila Taktak, Sana J ellouli, (2009),"Banking supervision and nonperforming
loans: a cross-country analysis", J ournal of Financial Economic Policy, Vol. 1 Iss 4 pp. 286-318 http://
dx.doi.org/10.1108/17576380911050043
Mikael Bask, (2009),"Optimal monetary policy under heterogeneity in currency trade", J ournal of Financial
Economic Policy, Vol. 1 Iss 4 pp. 338-354http://dx.doi.org/10.1108/17576380911050061
David G. Tarr, (2010),"The political, regulatory, and market failures that caused the US financial
crisis: What are the lessons?", J ournal of Financial Economic Policy, Vol. 2 Iss 2 pp. 163-186 http://
dx.doi.org/10.1108/17576381011070210
Access to this document was granted through an Emerald subscription provided by emerald-srm:115632 []
For Authors
If you would like to write for this, or any other Emerald publication, then please use our Emerald for
Authors service information about how to choose which publication to write for and submission guidelines
are available for all. Please visit www.emeraldinsight.com/authors for more information.
About Emerald www.emeraldinsight.com
Emerald is a global publisher linking research and practice to the benefit of society. The company
manages a portfolio of more than 290 journals and over 2,350 books and book series volumes, as well as
providing an extensive range of online products and additional customer resources and services.
Emerald is both COUNTER 4 and TRANSFER compliant. The organization is a partner of the Committee
on Publication Ethics (COPE) and also works with Portico and the LOCKSS initiative for digital archive
preservation.
*Related content and download information correct at time of download.
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
The Fed and the ECB: why such
an apparent difference
in reactivity?
Alexis Penot
GATE, University of Lyon, CNRS, ENS-LSH, Centre Le´on Be´rard,
Lyon, France, and
Gre´gory Levieuge
Laboratoire d’Economie d’Orle´ans, Faculte´ de Droit, d’Economie et de Gestion,
Orleans, France
Abstract
Purpose – The purpose of this paper is to try to understand the reasons for the differences in
amplitude of monetary policy (MP) rate cycles in the USA and the euro area. Among the different
candidates, the paper aims to test the role of economic structures, macroeconomic shocks, and MP
behaviour.
Design/methodology/approach – The paper starts by estimating vector autoregressive models
both for the USA and the euro area to identify the economic structures, the MP rules, and the
macroeconomics shocks of both areas. Then, it runs counterfactual simulations (by injecting European
Central Bank’s (ECB) monetary rule in the US model for example) to examine which factors had the
most signi?cant impact on differences in MP activism (measured by the variance of interest rates).
Findings – The paper ?nds that differences implied by MP rules alone cannot explain the
dissimilarity of interest rates paths. In the same way, while cyclical shocks are different in each area,
they do not suf?ce to explain the factual divergences. Finally, it is the structural dissimilarities which
essentially explain the difference in interest rate variances.
Originality/value – The paper brings new informations on a controversial issue and it tends to
reject the of?cial explanations given by ECB’s governor who points out differences in shocks.
Keywords Monetary policy, United States of America, Europe
Paper type Research paper
Introduction
“Too little, too late,” is the frequent comment about the stance of the European
monetary policy (MP), in particular when the evolution of its policy rate is compared to
the fed funds rate (Figure 1). Indeed, the European Central Bank (ECB) seems
systematically late especially in slowdown periods and it appears to be less dynamic.
Between January 1999 and September 2007, the ECB moved its key interest rate
23 times, against 35 for the Fed. Even more, between 1999 and 2005, the US policy
rate changes were twice as frequent as the euro rates moves (respectively 30 against
15)[1]. And since interest rate variations have the same magnitude, the US rate
is logically more volatile. The fed funds rate variance is four times higher than the
European repo rate (see the ?rst two lines of Table AI). Can we deduce from this
observation that the ECB is apathetic per se? More basically, where does this
difference come from?
The current issue and full text archive of this journal is available at
www.emeraldinsight.com/1757-6385.htm
The Fed
and the ECB
319
Journal of Financial Economic Policy
Vol. 1 No. 4, 2009
pp. 319-337
qEmerald Group Publishing Limited
1757-6385
DOI 10.1108/17576380911050052
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Following Sahuc and Smets (2007) and Christiano et al. (2007), three main reasons may
explain why actual interest rates volatility differ between these two regions. Indeed, at
the macroeconomic level, economies can be modeled following a classical framework
including three primary elements:
(1) a structure (which de?nes how shocks and monetary impulses are passed on);
(2) temporary shocks; and
(3) a MP represented by an interest rate rule, which react to the variables that are
driven by the structure and the temporary shocks.
Within such a context, differences in one (or more) of these elements are likely to
explain differences in the interest rates dynamics.
On these grounds, we investigate in this paper why the ECB is apparently less
reactive than the Federal reserve (FED). To this end, we follow the methodology
applied by Sahuc and Smets (2007) and Christiano et al. (2007), which is based on
counterfactual analysis. This method consists in simulating a pseudo Euro Area (EA)
interest rate in the cases in which the ECB monetary rule, the European shocks, and the
European structure would successively have been equivalent to their US counterparts.
Thereby, comparing the simulated EA rate with the actual EA and US rates, it is
possible to identify which one of these three components of the EA and US economies
mainly provoke the observed differences.
But contrary to these two papers, the classical framework MP rule-structure-shocks
in this paper relies on vector autoregressive (VAR) models, instead of dynamic
stochastic general equilibrium (DSGE) ones. We assert that DSGE models can be
unsuitable in such a comparative context, especially when the structural models are
supposed to be exactly similar (excepted the calibration). In that case, the conclusion of
structural similarity between the USA and the EA, as obtained by the pre-cited papers,
is almost trivial. On the contrary, aside from ?tting the data well, VAR models allow to
let the data speak (in terms of coef?cients but also in terms of dynamics through the
Figure 1.
Fed fund and ECB
repo rates
7
6
5
4
3
2
1
0
Jan. - 99 Jan. - 00 Jan. - 01 Jan. - 02 Jan. - 03 Jan. - 04 Jan. - 05
ECB repo rate
US federal funds rate
JFEP
1,4
320
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
choice of the p-order of the model), without structural preconceptions, which is an
advantage in this methodological grounds. We intend in this way to revisit the existing
results, and we show in particular that the questionable conclusion of structural
similarity between the USA and the EA does not hold with such an alternative model.
The reminder of the paper is organized as follows. Section 1 deals with the literature
related to this paper. Section 2 aims at formally delivering the three sources of
divergence between both areas. Section 3 explains the characteristics of the VARmodels
which we estimate for both areas. Then, we proceed to the counterfactual simulations,
comparing successively the monetary policies (Section 3), the shocks (Section 4), and the
structures (Section 5) of the EA and the USA. We conclude that economic structures do
explain the differences observed ex post between the ECB and the Fed interest rates,
while potential differences in shocks and MPrules do not. However, it does not mean that
the ECBand the FEDshare the same policy preferences. To this respect, in Section 6, we
evaluate what would have been the (optimal) EA interest rates with the US policy
preferences.
1. Related literature
Some papers have already studied the differences between the EA and the US interest
rates paths. Most of them consists in applying the method of counterfactual
simulations to DSGE models. Following this lines, on one hand, Christiano et al. (2007)
?nd that structural and cyclical differences could explain the diverging evolution of
output and in?ation. But since the ECB monetary rule implies more inertia than the US
one, these divergences do not turn up on interest rate paths.
On the other hand, Sahuc and Smets (2007) startlingly echoes Trichet’s (2006)
arguments by claiming that the interest rate divergences are explained by differences
in shocks, more than differences in structure or in MP. Using a similar structural model
estimated with US and EA data. Smets and Wouters (2005) also conclude that both
areas are structurally similar.
These contributions suffer from factual and methodological inconsistencies. First,
given the institutional and historical features which differentiate them, the ECB and
the Fed MPs are unlikely to be conducted similarly[2]. So, it is worth examining the
robustness of this result.
Moreover, regarding institutional disparities, it is hard to concede that the US and the
EA share the same economic structure. Relying on structural DSGE models and despite
some ad hoc accommodations (like backward-looking price setting, capital adjustment
costs, autoregressive shocks, etc.) these last references are likely to underestimate the
inertial behaviours of agents and then the sluggishness of data (Fuhrer, 1997; Estrella
and Fuhrer, 2002). Then, only “shocks” account for inertia, with the risk to overestimate
them if the economy is rigid. More basically, these models have not demonstrated their
superiority in terms of shock identi?cations. More embarrassing, the conclusion of
structural similarity is almost trivial when the structure of both economies is
theoretically identical and established a priori, while only the calibration is supposed to
be different. In this respect, Smets and Wouters (2005, p. 162) acknowledge: “the more
structure is imposed on the estimated model, the more the results will be colored by the
selected theoretical speci?cation.”
An alternative approach is then suitable to reassess these conclusions. Since imposing
an a priori structure is likely to pervert any comparison, it is interesting to use VAR
The Fed
and the ECB
321
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
models which precisely intend to let the data speak with few structural preconceptions.
Moreover, these models duly consider the sluggishness of macroeconomic variables. In
this respect, their ?tting and explanatory power are so largely recognized that they have
constituted benchmark models for revealing stylized facts in macroeconomics for several
decades. In particular, these models are used to assess the quality of structural
macroeconomic models (like DSGE); impulse responses functions (IRF) of VAR models
then constitute a reference which any structural model ought to reproduce[3].
We will then apply the method of counterfactual simulations to VAR models
estimated for the euro area and the USA. Moreover, the results of our simulations will
be submitted to econometrical tests that we expect to be more rigorous than the mere
graphical analyses used in the pre-cited papers. Beforehand, the next section provides a
formal representation of the method of counterfactual simulations.
2. What are the potential sources of divergence?
This section describes an economy reduced to four elements: a structure ðFð · ÞÞ,
temporary shocks (1
X
), a MP governed by a rule ð f ð · ÞÞ, and some MP shocks (1
i
). In
order to formalize Sahuc and Smets (2007) and Christiano’s et al. (2007) method, we can
write the linear system:
X
c
¼
c
F
c
F
c
ðX
c
; i
c
Þ þ1
c
X
i
c
¼
b
f
c
f
c
ðX
c
Þ þ1
c
i
8
<
:
ð1Þ
with c ¼ {US; EA} and X
c
a vector of n state variables (;n). Concretely,
c
F
c
F
c
ð · Þ is an
estimated multidimensional function which represents the economic structures, and 1
c
X
the vector of the associated error terms (the difference between the actual values and
the values predicted by the estimated structure) which represent by de?nition the
shocks that hit this structure. Next, i
c
is the short-term interest rate, i.e. the MP
instrument which depends, through an estimated MP rule
^
fð · Þ, on the state variables of
the system. The difference between the actual and the implied by the rule interest rates
represents the MP shocks (including the discretionary part of the MP), noted 1
i
. It is
important to note that, a priori, structures
c
F
c
F
c
ð · Þ, shocks 1
c
X
, MP rules
b
f
c
f
c
and/or MP
shocks 1
c
i
are distinct in the two regions.
Thereby, we resort to counterfactual simulations to detect the main cause(s) of
divergence. Concretely, successively substituting the EA structure, shocks, MP rule,
and MP shocks with their US counterparts, we calculate pseudo European interest
rates and test if they are signi?cantly different from the actual rate (diagram (1) in
Figure 2) and if they come closer to the actual US rate. This implies four cases.
First, what would have happened if the ECB had followed the US MP rule (diagram
(2))? The response is given by the following simulated interest rate:
Figure 2.
Counterfactual scenarios
e
X
EA
e
X
EA
e
X
US
f
EA
f
EA
f
EA
f
EA
f
US
X
US
X
EA
X
EA
X
EA
X
EA
i
EA
i
EA
i
EA
i
EA
i
EA
e
X
EA
e
X
EA
e
i
EA
e
i
US
e
i
EA
e
i
EA
(1) (2) (3) (4) (5)
e
i
EA
JFEP
1,4
322
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
~
i
EA
¼
d
f
US
f
US
ðX
EA
Þ þ1
EA
i
ð2Þ
Second, what would be the EA interest rate if the European MP surprises were similar
to the US ones (diagram (3))? The solution is given by:
~
i
EA
¼
d
f
EA
f
EA
ðX
EA
Þ þ1
US
i
ð3Þ
Third, given that the state variables can be decomposed as X
c
¼
c
X
c
X
c
þ1
c
X
, we can
determine to what extent the cyclical differences are likely to explain the observed
divergences between actual US and EA interest rates. 1
EA
X
has to be substituted with
1
US
X
in the European system (diagram (4)) to deduce a pseudo interest rate as:
~
i
EA
¼
d
f
EA
f
EA
d
X
EA
X
EA
þ1
US
X
þ1
EA
i
ð4Þ
In these cases, MP rules, MP shocks, and structures remain proper to each area.
Finally, a pseudo interest rate can be calculated, under the assumption that the euro
area works under the US structure (diagram (5)), everything else equal:
~
i
EA
¼
d
f
EA
f
EA
d
X
US
X
US
þ1
EA
X
þ1
EA
i
ð5Þ
This method allows to identify which sources of divergence among the four explain the
lower volatility of the European interest rate. Beforehand, it is necessary to explain the
models used to implement the generic method exposed so far.
3. Modeling the euro area and the USA
In the light of the arguments presented in introduction, VAR models are particularly
relevant in this problematic. They usually are reference models for stylized facts,
whereas DSGE models are inadequately discriminant and sometimes empirically
unsatisfying[4]. Finally, in the perspective of counterfactual simulations, VAR models
bene?t from few theoretical preconceptions.
3.1 Models, data, and estimation
The two VAR models can be written as follows:
X
t
¼ A
1
X
t21
þ · · · þ A
p
X
t2p
þ BZ
t2k
þ1
t
ð6Þ
with X a ?ve-element vector composed of the long-term real interest rate r, the output
gap y, the wage annual growth rate w, the in?ation rate p and the short-term nominal
interest rate i. For degrees of freedom purposes, we only consider the most important
variables at stake in the MP transmission mechanism and add the wage growth to
capture the institutional and cyclical divergences often observed between the US and
EA labor markets (Sahuc and Smets, 2007; Christiano et al. (2007)). 1 is the vector of
innovations associated to the endogenous variables, i.e. ð1
r
; 1
y
; 1
w
; 1
p
; 1
i
Þ, which,
respectively, represent a shock on the long-term interest rate (or ?nancial shock), a
demand shock, a labor shock, a supply shock and a MP shock. As the equation de?ning
the short-term interest rate is comparable to a MP rule[5], the shock 1
i
stands for a
temporary deviation to the rule, that is central bank discretionary decisions. Z stands
for exogenous commodity prices, inserted to successfully smooth the price puzzle effect
(with k depending on its informational contribution).
The Fed
and the ECB
323
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
US data stem from Datastream and EA from the area-wide model database (Fagan
et al., 2005), updated in autumn 2006, up to the fourth quarter of 2005. All variables are
seasonally adjusted via the standard X-11 method. The exogenous variable is the
Goldman Sachs Commodity Prices Index. For both areas, because of a possible unit
root in the real gross domestic product (GDP) series, we de?ne the output gap as the
residuals of the regression of the GDP on a constant and a quadratic trend, instead of
using a usual Hodrick-Prescott ?lter. Short- and long-term interest rates are the
three-month monetary rate and the ten-year benchmark rate, respectively[6].
Maximum-likelihood estimation covers the 1985-2005 period. Certainly, a
non-parametric estimation method could be applied, like a bayesian VAR (BVAR)
estimation, in particular, as suggested for instance by Litterman (1986). Besides, it is
shown that a BVAR estimation is more accurate than a maximum likelihood (ML)
estimation, provided that the a priori information (i.e. the priors) is adequate. But
imposing conditions, as required by this method, means setting restrictions
(i.e. theoretical preconceptions) on the estimated model. Instead, we want to let the
data speak. We want the differences to be entirely determined as a result and not as a
consequence of starting hypothesis, contrary to bayesian estimations of DSGE models,
which use to impose the same prior conditions for the EA and the USA, and which
inevitably lead to conclude that both regions are structurally similar.
Moreover, prior probability distributions raise numerous questions. Unknowing
about model parameters is often important, what is source of numerous pitfalls, as
shown by the discussion between Sims and Uhlig (1991), Phillips (1991a, b) and
Sims (1991). The choice of a given prior density is rarely duly justi?ed nor discussed
in the DGSE literature. While a lot of robustness checks would be required,
economists content themselves with replicating the beliefs applied in the existing
literature.
On the contrary, estimating a VAR through ML method imposes only one
restriction: it is implicitly assumed that the relations between the variables are linear.
A potential misspeci?cation of this order is however unlikely with regard to the vast
literature asserting the good ?tting of VAR models.
The p-order of the VAR models is determined with the parsimonious Bayesian
information criterion, with respect to the white noise hypothesis of disturbances.
Finally, we ?nd p ¼ 3 for the EA and p ¼ 5 for the USA.
Once estimated, the VAR models are used to generate IRFs[7], in order to asses their
dynamics (in the light of some well-known stylized facts) and to have an insight on the
differences between both areas.
3.2 The main characteristics of the US and EA models
Figure A1 represents the essential IRFs in the perspective of studying the transmission
of shocks and MP impulses. As a whole, it actually describes the textbook MP
transmission mechanism. In the euro area, the short-term interest rate has a positive
effect on the long-term interest rate, even if it is not contemporaneous because of the
Cholesky decomposition. Next, the long-term interest rate impacts the output gap
negatively. In accordance with the expected transmission delay of MP, this in?uence is
only signi?cant after the sixth quarter following the initial shock on the long-term
interest rate. Then a positive output gap has a huge and positive incidence on in?ation,
still visible after 15 quarters.
JFEP
1,4
324
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
For institutional reasons, the standard MP transmission mechanism mainly works
through the short-term interest rate in the USA. Strictly speaking, the real short-term
interest rate should have been considered. But the interpretation of an interest rate rule
in real terms is less direct, and passing from a real to a nominal short-term interest rate
would have rendered dif?cult the implementation of the counterfactual simulations.
Certainly, IRF con?rm the negative impact of the nominal short-term interest rate on
the output gap, with a signi?cant effect starting one year after the shock and still
effective three years latter. But the effect of the real short-term interest rate is also
veri?ed: the response of the output gap is then signi?cantly negative from the
seventh quarter and remains during the six following quarters (Appendix, Figure A1).
Finally, the output gap has a positive incidence on in?ation for three years after the
initial shock.
Before examining the results of the counterfactual simulations, it is worth noting
that such a method applied to VAR models is subject to the Lucas criticism.
Nevertheless, according to Hansen global stability and Chow predictive tests (with a
break point in 1999Q1), the VAR relations estimated for the EA are strikingly stable,
even concerning the MP rule equation (and this despite the transition from national
monetary policies to a common one).
This result is consistent with the reserves of numerous studies which, without
reviewing its theoretical logic, attenuate the practical signi?cance of the Lucas
criticism (Altissimo et al., 2002; Estrella and Fuhrer, 1999). Moreover, the Lucas
criticism is here dampened by the uncertainty surrounding the ECB implementation
and policy. Around 1999, economic agents were not able to formulate well-founded
expectations about the European MP. As a result, their attitude had probably not be
different if the ECB had chosen to follow the Fed’s rule for instance. As a whole, in the
?rst time of the ECB exercise, uncertainty (about MP, structures, and shocks),
observation and learning would have been the same whatever the MP rule.
4. Comparing monetary policy rules
The ?rst lines of the Table AI in the Appendix show that explaining the differences in
interest rate paths between the EA and the USA does not require explaining
differences in levels, but in variations. This being said, can the behaviour of central
banks explain the difference in interest rate paths?
In accordance with equation (2), we determine a pseudo EA interest rate, assuming
the ECB had followed the Fed’s rule (estimated over the 1999-2005). As indicated
in Table I, we are lead to accept the equivalence of the EA and the USA. MP rules,
1999:1-2005:4 Variance Variance equality
Euro area with the Fed MP rule
Actual EA interest rate 0.92 Yes
Simulated EA interest rate 0.90 (SL ¼ 0.47)
Euro area with the Fed MP shocks
Actual EA interest rate 0.92 Yes
Simulated EA interest rate 0.69 (SL ¼ 0.22)
Note: Monetary policy scenarios
Table I.
Tests on counterfactual
interest rate series
The Fed
and the ECB
325
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
in terms of interest rate volatility. So, with the Fed’s rule, the ECB would not have
appeared as more “reactive.” Therefore, MP rule differences do not explain the
divergences of interest rate paths[8].
We also examine the differences due to MP shocks, according to equation (3).
Logically, given the mean and variance equalities of these disturbances (see the last
lines of Table AI), the discretionary stance of monetary policies neither explain the
observed divergence of interest rate paths, as indicated in the second part of the Table I.
Consequently, the US and EA interest rates differences might be of structural and/or
cyclical order(s).
5. Comparing shocks
Is the larger inertia of the European MP due to smaller and less frequent shocks?
Preliminary statistical investigations reveal that long-term interest rates are correlated,
mainly because of international ?nancial spillovers, and ?nancial globalization makes
this correlation even higher for the last ten years. Supply shocks are also strongly
correlated, especially for the 1999-2005 period. Therefore, demand and labor shocks are
the only potential explanation for different MP stances. Next, according to the
historical decomposition of shocks, it appears that shocks are not of the same nature in
the two regions. This can be a sign of cyclical divergences, as asserted by Sahuc and
Smets (2007) and Trichet (2006).
But this preliminary analysis is not likely to explain the smaller volatility of the
European interest rate. Indeed, European shocks should be smaller or less volatile to
this end. And yet, variance equality tests reported in Table II indicate this is not the
case: shocks on long-term interest rates, in?ation, and labor market[9] are signi?cantly
larger in the euro area. Only demand shocks are slightly higher in the USA over
1999-2005, as already found in Sahuc and Smets (2007)[10]. These differences cannot
justify a stronger interest rate inertia (that is even the opposite).
The counterfactual simulations based on equation (4), summarized in Table III,
con?rm these inferences. If European shocks would have been equal to US ones, the
European interest rate would have not be statistically different from its actual value.
Besides, the same exercise for the USA is very instructive. Indeed, the second part of
the Table III indicates that the US interest rate would have been even more volatile if
driven by the European shocks (instead of the US ones). So, the USA seem to be more
sensitive to cyclical shocks than the EA. Therefore, if both regions do not
similarly respond to shocks, the interest rate differential must stem from structural
differences.
Varð1
EA
r;t
Þ= Varð1
US
y;t
Þ= Varð1
EA
w;t
Þ= Varð1
EA
p;t
Þ= Varð1
EA
i;t
Þ=
Varð1
US
r;t
Þ Varð1
EA
y;t
Þ Varð1
US
w;t
Þ Varð1
US
p;t
Þ Varð1
US
i;t
Þ
1985-2005 3.53
* *
1.15 7.26
* *
4.74
* *
1.59
*
1999-2005 6.61
* *
1.61
*
4.24
* *
6.21
* *
1.23
a
Notes: Rejection of the null hypothesis of variance equality at:
*
10 and
* *
1 percent level, respectively;
a
corresponds to Vð1
US
i;t
Þ=Vð1
EA
i;t
Þ
Table II.
Tests of variance equality
JFEP
1,4
326
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
6. Comparing economic structures
First, Figure A3 shows the inertia degree of the two regions, on the basis of the
response functions of the real variables of the model to idiosyncratic shocks. Table IV
offers a measure of this adjustment speed[11].
In the euro area, subsequently to a cost-push supply shock, in?ation returns to the
equilibrium level after seven quarters. The European in?ation dynamics is more
sluggish in the very short run. Moreover, as indicated in the last line of the Table IV, a
shock on the European labor market has a strongly persistent effect on wages. It takes
about two years for the wage impulse-response function to be cut by half in the euro
area. The wage growth is the most viscous variable of the European system.
Conversely, such a shock is quickly absorbed by the US labor market (where more than
half of the total effect is absorbed in the ?rst six months, compared to only a fourth in
the euro area in the same time). Finally, in the short run (until two years), the output
gap has a higher persistence in the euro area. So, differences of reactivity are obvious
between the EA and the USA.
Now, the counterfactual experience consists in simulating an European interest rate
between 1995 and 2005 under a US economic structure (relation (5)). Table V shows
that, in this case, the pseudo European interest rate is signi?cantly higher. The
variance equality test even indicates that the simulated European interest rate is
statistically equivalent to the US one. European in?ation and output gap would also be
more volatile if US and European structures would have matched[12].
Therefore, differences in structure do explain a large part of the differences in
interest rate volatilities between the two regions. This result is in contradiction with
Euro area (%) USA (%)
Reponse of Six months One year Two years Six months One year Two years
p to a shock on p 62 81 100 49 79 100
y to a shock on y 24 43 72 35 57 75
w to a shock on w 27 36 54 53 65 80
Notes: Numbers represent the relative adjustment made by each variable, in terms of cumulated
impulse response, second, fourth, and eight quarters after an initial idiosyncratic shock; full
adjustment is supposed to be achieved when the corresponding IRF is no longer signi?cantly different
from 0
Table IV.
Speed of adjustment of
in?ation (p), output gap
( y), and wages (w)
1999:1-2005:4 Variance Variance equality
Euro area with the US shocks
Actual EA interest rate 0.92 Yes
Simulated EA interest rate 1.21 (SL ¼ 0.24)
USA with the EA shocks
Actual US interest rate 3.72 No
Simulated US interest rate 18.8 (SL ¼ 0.00)
Note: Shocks scenario
Table III.
Tests on counterfactual
interest rate series
The Fed
and the ECB
327
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
previous studies. But yet, although the euro area is itself made of heterogeneous
countries, its global structure obviously differs from that of the USA.
7. Discussion on central bank preferences
Figure 3 shows the interest rate paths implied by the various scenarios considered. The
scenario in which European monetary authorities conduct their policy with the US
structures (everything else being equal) is the only case in which the variance of the
European interest rate is comparable to the variance of the US rate[13]: however, it is
worth noting that in this case the interest rate would have come up against the lower
zero bound for interest rate. On the contrary, the differences between the simulated and
the actual interest rates are not signi?cant in average when shocks are switched. In the
same way, MP rules also do not suf?ce to explain differences in interest rates.
But this last result does not mean that the ECB and the FED have the same objective
function. If central banks behave optimally, it can be theoretically demonstrated that
1999:1-2005:4 Variance Variance equality
Euro area with the US structure
Actual EA interest rate 0.92 No
Simulated EA interest rate 3.68 (SL ¼ 0.05)
Actual EA output gap 1.02 No
Simulated EA output gap 9.45 (SL ¼ 0.00)
Actual EA in?ation 0.76 No
Simulated EA in?ation 3.99 (SL ¼ 0.00)
Actual US interest rate 3.72 Yes
Simulated EA interest rate 3.68 (SL ¼ 0.00)
Note: Structural scenario
Table V.
Tests on counterfactual
interest rate series
Figure 3.
Simulated EA interest
rates
10
9
8
7
6
5
4
3
2
1
0
Q
2
-
1
9
9
9
Q
3
-
1
9
9
9
Q
4
-
1
9
9
9
Q
1
-
2
0
0
0
Q
2
-
2
0
0
0
Q
3
-
2
0
0
0
Q
4
-
2
0
0
0
Q
1
-
2
0
0
1
Q
2
-
2
0
0
1
Q
3
-
2
0
0
1
Q
4
-
2
0
0
1
Q
1
-
2
0
0
2
Q
2
-
2
0
0
2
Q
3
-
2
0
0
2
Q
4
-
2
0
0
2
Q
1
-
2
0
0
3
Q
2
-
2
0
0
3
Q
3
-
2
0
0
3
Q
4
-
2
0
0
3
Q
1
-
2
0
0
4
Q
2
-
2
0
0
4
Q
3
-
2
0
0
4
Q
4
-
2
0
0
4
Q
1
-
2
0
0
5
Q
2
-
2
0
0
5
Q
3
-
2
0
0
5
Q
4
-
2
0
0
5
Q
1
-
1
9
9
9
I
n
%
Actual EA interest rate
EA interest rate with US monetary policy Rul
EA interest rate with US shocks
EA interest rate with US structure
EA interest rate with US monetary policy shocks
Actual US interest rate
JFEP
1,4
328
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
their (estimated) MP rule responds to the optimization of an objective function, subject
to a given structural model and a given variance of shocks. Now, it is shown that in a
welfare view a` la (Woodford, 2003), objective functions precisely depend on the
structural parameters of the economies. And as Section 6 led to conclude that the
structures between the two regions is different, the objective functions may also do.
So, in order to strengthen the conclusion of structural dissimilarity, the last exercise
consists in determining and comparing the preference parameters of policymakers.
More precisely, we will simulate what the FED would have (optimally) done, had it
been in charge of the European MP, conditional on its own objective function[14].
Revealing ?rst the preference parameters of the FED is needed to this end[15].
7.1 The method for revealing preference parameters
We assume the Fed had to minimize the following usual quadratic loss function, where
l
y
and l
i
represent, respectively, the preference parameter for output and interest rate
variabilities:
Loss
t
¼ l
y
Varð y
t
Þ þ ð1 2l
y
ÞVarðp
t
Þ þl
i
Varði
t
2i
t21
Þ ð7Þ
subject to a given structural model Z
t
¼ FZ
t21
þU
t
, a variance-covariance matrix M
associated to the innovations’ vector U, and a MP rule i
t
¼ f ðXÞ. It can be
demonstrated, following Svensson (1998) and Ball (1997) that the unconditional
contemporaneous covariance matrix of Z, noted V is given in vector form by:
VecðVÞ ¼ ½I 2F^F?
21
VecðMÞ ð8Þ
So, the optimal MP rule is given by the parameters of the vector f that minimize the
weighted sum – based on the given preference parameters ðl
y
; l
i
Þ – of the
unconditional variances for the in?ation rate, the output gap and the interest rate.
These variances appear in VecðVÞ, following equation (8).
This robust method can be used to reveal the US preference parameters. To make
the algorithm optimization tractable and to focus on the main usual targets of MP
rules, we limit the US model to three main variables: interest rate, in?ation and output
gap. We note
^
i the interest rate deduced from the US estimated MP rule (MP shocks are
ruled out). The effective preference parameters of the Fed, ðl
*
y
; l
*
i
Þ, determine a MP
rule yielding an “optimal” interest rate
^
i
*
which is the nearest to the estimated rate
^
i.
The optimization program is then given by:
{l
y
;l
i
}
Min distance ¼
X
N
t¼0
^
i
*
t
ðl
y
; l
i
Þ 2
^
i
t
h i
2
with
^
i
*
t
¼ f
*
Z such as
{f
*
}
Min L ¼ l
y
Varðy
t
Þ þ ð1 2l
y
ÞVarðp
t
Þ þl
i
VarðDi
t
Þ
with Z
t
¼ F½ f ðl
y
; l
i
Þ? þ U
t
and Z ¼ ½ y
t21
. . .y
t25
; p
t21
. . .p
t25
; i
t21
. . .i
t25
?
8
>
>
>
>
>
>
>
<
>
>
>
>
>
>
>
:
7.2 What would have been the ECB rates with the US preferences?
Applying this method, we ?nd l
*
y
¼ 0:42 and l
*
i
¼ 5:6. These preference parameters
generate an optimal MP rule with an interest rate path that ?ts very well the estimated
interest rate, as it is shown by Figure 4 (distance ¼ 11.62). This result is consistent.
The Fed
and the ECB
329
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
On the one hand, the Fed has an explicit objective of output stabilization (while
in?ation stability is its main objective). On the other hand, a high value of l
i
is required
to reproduce interest rate inertia (Dennis, 2001). When these parameters are applied to
the euro area, they generate an optimal interest rate path which is more volatile than it
is in fact, as shown by Figure 5.
To such an extend that, as shown in Table VI, the variance equality between the
optimized EA interest rate and the US effective (estimated) interest rate is accepted.
Implicitly, this suggests that the parameter l
i
is higher in the ECB’s loss function. This
result can be interpreted twofold. Either it means that the European authorities are very
cautious, in the sense they display a high preference for interest rate smoothing. This
behaviour can be justi?ed in the light of Brainard’s arguments, because of the large
uncertainty that surrounds the total effects of a common European MP. Or, following
Woodford(2003, Chapter 6), it canbe demonstratedthat anobjective functionlike equation
(7) can be derived fromthe discounted sumof households’ utility. In this case, parameters
ðl
i
; l
y
Þ do not represent the monetary authorities’ preferences per se but the results of an
optimization implying the structural parameters of the underlying model. According to
this point of view, concluding that differences in objective functions can explain the
difference of volatilitybetweenthe USandthe EAinterest rates is equivalent toaf?rmthat
structural divergences explain it, what strengthens the conclusion of Section 6.
Conclusion
Given the dissimilar variances of the European and US interest rate, the ECB is often
blamed for not being reactive enough in comparison with the Fed. But such a hasty
Figure 4.
US optimal and actual
interest rates
–4
–3
–2
–1
0
1
2
3
4
5
Q
1
-
1
9
8
6
Q
3
-
1
9
8
6
Q
1
-
1
9
8
7
Q
3
-
1
9
8
7
Q
1
-
1
9
8
8
Q
3
-
1
9
8
8
Q
1
-
1
9
8
9
Q
3
-
1
9
8
9
Q
1
-
1
9
9
0
Q
3
-
1
9
9
0
Q
1
-
1
9
9
1
Q
3
-
1
9
9
1
Q
1
-
1
9
9
2
Q
3
-
1
9
9
2
Q
1
-
1
9
9
3
Q
3
-
1
9
9
3
Q
1
-
1
9
9
4
Q
3
-
1
9
9
4
Q
1
-
1
9
9
5
Q
3
-
1
9
9
5
Q
1
-
1
9
9
6
Q
3
-
1
9
9
6
Q
1
-
1
9
9
7
Q
3
-
1
9
9
7
Q
1
-
1
9
9
8
Q
3
-
1
9
9
8
Q
1
-
1
9
9
9
Q
3
-
1
9
9
9
Q
1
-
2
0
0
0
Q
3
-
2
0
0
0
Q
1
-
2
0
0
1
Q
3
-
2
0
0
1
Q
1
-
2
0
0
2
Q
3
-
2
0
0
2
Q
1
-
2
0
0
3
Q
3
-
2
0
0
3
Q
1
-
2
0
0
4
Q
3
-
2
0
0
4
Q
1
-
2
0
0
5
Q
3
-
2
0
0
5
Estimated US interest rate (demeaned, no shocks)
Optimal US interestrate (demeaned)
JFEP
1,4
330
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
appraisal forgets that MP must be evaluated in its context, i.e. considering both the
economic structures and the cyclical shocks. In this prospect, several studies searched
to explain the reasons for such an apparent difference in reactivity. As a whole, they
conclude that the cyclical shocks are mainly responsible. Beyond that, monetary
policies and, more surprisingly, economic structures would be the same in both areas.
But these conclusions are specious as they a priori assume that the US and EA models
are structurally identical.
So, in this paper, we reassess these results in the light of a VAR model, which is
precisely exempted from a priori theoretical preconceptions. Counterfactual
simulations con?rm that differences implied by MP rules alone cannot explain the
dissimilarity of interest rates paths. In the same way, while cyclical shocks are
different in each area, they do not suf?ce to explain the factual divergences. Finally, it
is the structural dissimilarities which essentially explain the differences in interest rate
1999:1-2005:4 Variance Variance equality
Estimated EA interest rate 1.08 No
Optimized EA interest rate with US preferences 4.25 (SL ¼ 0.00)
Estimated US interest rate 4.33 Yes
Optimized EA interest rate with US preferences 4.25 (SL ¼ 0.48)
Note: Preference scenario
Table VI.
Tests on counterfactual
interest rate series
Figure 5.
Simulated EA optimal
interest rate with US
preferences
–10
–6
–2
2
6
10
Q
1
-
1
9
8
6
Q
3
-
1
9
8
6
Q
1
-
1
9
8
7
Q
3
-
1
9
8
7
Q
1
-
1
9
8
8
Q
3
-
1
9
8
8
Q
1
-
1
9
8
9
Q
3
-
1
9
8
9
Q
1
-
1
9
9
0
Q
3
-
1
9
9
0
Q
1
-
1
9
9
1
Q
3
-
1
9
9
1
Q
1
-
1
9
9
2
Q
3
-
1
9
9
2
Q
1
-
1
9
9
3
Q
3
-
1
9
9
3
Q
1
-
1
9
9
4
Q
3
-
1
9
9
4
Q
1
-
1
9
9
5
Q
3
-
1
9
9
5
Q
1
-
1
9
9
6
Q
3
-
1
9
9
6
Q
1
-
1
9
9
7
Q
3
-
1
9
9
7
Q
1
-
1
9
9
8
Q
3
-
1
9
9
8
Q
1
-
1
9
9
9
Q
3
-
1
9
9
9
Q
1
-
2
0
0
0
Q
3
-
2
0
0
0
Q
1
-
2
0
0
1
Q
3
-
2
0
0
1
Q
1
-
2
0
0
2
Q
3
-
2
0
0
2
Q
1
-
2
0
0
3
Q
3
-
2
0
0
3
Q
1
-
2
0
0
4
Q
3
-
2
0
0
4
Estimated EA interestrate (demeaned, no shocks)
Optimal EA interestrate with US preferences
The Fed
and the ECB
331
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
variances. This conclusion is corroborated by a simulation indicating that the
European interest rate would have been more volatile if the ECB had the same deep
preference parameters as the FED. In a welfare perspective of a utility-based loss
function, this result tends to con?rm the structural divergences between both areas.
Besides, this conclusion is consistent with the largely spread wisdom that the USA and
the EA have dissimilar labour, ?nancial, and good markets.
So, a less reactive Central Bank is not necessarily an institution which is not prompt
to move its policy rates, nor per se an apathetic Central Bank. In this precise case, the
ECB conducts the MP of an inertial economy, which slowly and partially reacts to
shocks.
Notes
1. The timing and magnitude of MP responses to the subprime crises is no exception for the
most recent period.
2. Unlike the ECB, the Fed is explicitly mandated to take care of the output cycle (and not only
to ?ght in?ation). Moreover, the composition of MP committees and the way they come to a
decision can induce more or less inertia (Gerlach-Kristen, 2005). At least, the preferences of
the monetary authorities should be different.
3. For instance, Rotemberg and Woodford (1997) and Christiano et al. (2005) suggest to test the
reliability of DSGE models with respect to the IRFs of VAR models.
4. See for instance Blanchard (2008), Chari et al. (2008) and Iskrev (2008) for criticisms on DSGE
models which are particularly relevant in the perspective of this problematic.
5. The backward looking feature of this MP rule can be justi?ed by the unavailability of data in
real time (McCallum and Nelson, 1999). Moreover, as it is shown by a lot of reaction function
estimates, the past interest rate has to be added to the baseline Taylor rule for a good ?tting,
as i
t
¼ b
1
i
t21
þb
2
b pp
t
þb
3
y
t
, with b pp
t
¼ ðp
t
2 p
t
Þ. In this case, the MP rule can be written:
i
t
¼ b
2
P
n
n¼0
b
n
1
b pp
t2n
þb
3
P
n
n¼0
b
n
1
y
t2n
, where reference to the past is obvious.
6. As unit-root hypothesis is rejected by usual ADF tests, the cointegration study is useless.
7. IRFs are computed through bootstrapping and based on a Cholesky decomposition with the
variables ordered as speci?ed in equation (6).
8. In the margin of this exercise, we ?nd that, for the US simulated rate under the European
rule, the variance equality is accepted at the 5 percent but rejected at the 10 percent level.
However, even less volatile, the simulated US rate does not come closer to the EA one.
9. It can be argued that, since the early 1990s, all European countries have been promoting
structural reforms while the USA only experienced the earned income tax credit.
10. However, we do not share Smets and Wouters (2005) conclusion about the euro area more
affected by negative supply shocks. Our shocks have the same variance in the two regions.
11. Denoting f
ij,t
, the dynamic multiplier deriving from the VMA representation of a VAR
model, it represents the effect of an innovation 1
j
occurring at date t on the ith variable
observed in t þ m. The measure of adjustment delay of the ith variable submitted to an
idiosyncratic shock is then given by:
P
m
t¼0
jf
ij;t
j=
P
1
t¼0
jf
ij;t
j with i ¼ j and m ¼ {2, 4, 8}
quarters. For the denominator, the sum stops in fact when f
ij,t
is no longer signi?cant.
12. In the same way, we have found that the variance of the simulated US interest rate, subject to
the European structure, is signi?cantly lower than its actual variance. Moreover, the
variance equality between this pseudo US rate and the effective EA rate cannot be rejected.
13. In this case, the difference in level is largely explained by the higher US initial conditions.
JFEP
1,4
332
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
14. This section goes into detail the arguments mentioned by Christiano et al. (2007, footnote 3,
p. 2478): “it is possible that if the FED were literally in charge of the ECB, it might not have
applied the same monetary policy strategy that it uses in the US [. . .] [This] question would
require the identi?cation of the FED’s objective function, and then the computing of the
monetary policy rule that optimizes it, conditional on the economy corresponding to the
estimated EA economy.” It is exactly what is done in this section.
15. Certainly, this task is dif?cult to conceive for the euro area., where MP was national before
1999 and raises the problem of national preference parameters aggregation.
References
Altissimo, F., Sieviero, S. and Terlizzese, D. (2002), “How deep are the deep parameters?”,
Annales d’Economie et de Statistiques, Vol. 67-68, pp. 207-26.
Ball, L. (1997), “Ef?cient rules for monetary policy”, NBER Working Paper Series No. 5952.
Blanchard, O. (2008), “The state of Macro”, NBER Working Paper Series No. 14259.
Chari, V., Kehoe, P. and McGrattan, E. (2008), “New-Keynesian models are not yet useful for
policy analysis”, Federal Reserve Bank of Minneapolis Working Papers.
Christiano, L., Eichenbaum, M. and Evans, C. (2005), “Nominal rigidities and the dynamic effects
of a shock to monetary policy”, Journal of Political Economy, Vol. 113 No. 1, pp. 1-45.
Christiano, L., Motto, R. and Rostagno, M. (2007), “Shocks, structures or policies? A comparison
of the EA and the US”, Journal of Economic Dynamics and Control, Vol. 32, pp. 2476-506.
Dennis, R. (2001), “The policy preferences of the US federal reserve”, Federal Reserve Bank of
San Francisco, Working Paper in Applied Economic Theory 2001-08 (accessed July 2004).
Estrella, A. and Fuhrer, J. (1999), “Are ‘Deep’ parameters stable ? The Lucas critique as an
empirical hypothesis”, Federal Reserve Bank of Boston Working Papers No. 99-4.
Estrella, A. and Fuhrer, J. (2002), “Dynamic inconsistencies: counterfactual implications of a class
of rational expectations models”, American Economic Review, Vol. 92 No. 4, pp. 1013-28.
Fagan, G., Henry, J. and Mestre, R. (2005), “An area-wide model (AWM) for the euro area”,
Economic Modelling, Vol. 22, pp. 39-59.
Fuhrer, J. (1997), “The (un)importance of forward-looking behavior in price speci?cations”,
Journal of Money, Credit, and Banking, Vol. 29 No. 3, pp. 338-50.
Gerlach-Kristen, P. (2005), “Too little, too late: interest rate setting and the cost of consensus”,
Economics Letters, Vol. 88, pp. 376-81.
Iskrev, N. (2008), “How much do we learn from the estimation of DSGE models? A case study of
identi?cation issues in a new Keynesian business cycle model”, mimeo.
Litterman, R. (1986), “Forecasting with Bayesian vector autoregressions: ?ve years of
experience”, Journal of Business & Economic Statistics, Vol. 4 No. 1, pp. 25-38.
McCallum, B. and Nelson, E. (1999), “Performance of operational policy rules in an estimated
semi-classical structural model”, in Taylor, J.-B. (Ed.), Monetary Policy Rules, University of
Chicago Press, Chicago, IL, pp. 15-45.
Phillips, P. (1991a), “Bayesian routes and unit roots: De Rebus Prioribus Est Disputandum”,
Journal of Applied Econometrics, Vol. 6 No. 4, pp. 453-73.
Phillips, P. (1991b), “To criticize the critics: an objective Bayesian analysis of stochastic trends”,
Journal of Applied Econometrics, Vol. 6 No. 4, pp. 333-64.
Rotemberg, J. and Woodford, M. (1997), “An optimization-based econometric framework for the
evaluation of monetary policy”, in Bernanke, B. and Rotemberg, J. (Eds), Macroeconomics
Annual, MIT Press, Cambridge, MA, pp. 297-346.
The Fed
and the ECB
333
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Sahuc, J. and Smets, F. (2007), “Differences in interest rate policy at the ECB and the Fed:
an investigation with a medium-scale DSGE model”, Journal of Money, Credit and
Banking, Vol. 40 No. 2-3, pp. 505-21.
Sims, C. (1991), “Comment on ‘To Criticize the Critics’ by Peter c.b. Phillips”, Journal of Applied
Econometrics, Vol. 6 No. 4, pp. 423-34.
Sims, C. and Uhlig, H. (1991), “Understanding unit rooters: a helicopter tour”, Econometrica,
Vol. 59 No. 6, pp. 1591-9.
Smets, F. and Wouters, R. (2005), “Comparing shocks and frictions in US and EURO business
cycles: a Bayesian DSGE approach”, Journal of Applied Econometrics, Vol. 20, pp. 161-83.
Svensson, L. (1998), “Open-economy in?ation targeting”, CEPR Discussion Paper No. 1989.
Trichet, J. (2006), “Activism and alertness in monetary policy”, paper presented at the Conference
on Central Banks in the 21st Century, Bank of Spain, June 8.
Woodford, M. (2003), Interest and Prices, Foundations of a theory of Monetary Policy, Princeton
University Press, Princeton, NJ.
Further reading
A
´
lvarez, L., Dhyne, E., Hoeberichts, M., Kwapil, C., Le Bihan, H., Lu¨nnemann, P., Martins, F.,
Sabbatini, R., Stahl, H., Vermeulen, P. and Vilmunen, J. (2005), “Sticky prices in the euro
area: a summary of new micro evidence”, ECB Working Papers No. 563.
Bils, M. and Klenow, P. (2004), “Some evidence on the importance of sticky prices”, Journal of
Political Economy, Vol. 112, pp. 947-85.
Calmfors, L. and Drif?ll, J. (1988), “Bargaining structure, corporatism and macroeconomic
performance”, Economic Policy, Vol. 6 No. 3, pp. 13-62.
Ehrmann, M., Gambacorta, L., Martinez-Pages, J., Sevestre, P. and Worms, A. (2002), “Financial
systems and the role of banks in monetary policy transmission in the euro area”, Notes
d’Etudes et de Recherche de la Banque de France No. 93.
Hall, P. and Soskice, D. (2001), Varieties of Capitalism: The Institutional Foundations of
Comparative Advantage, Oxford University Press, New York, NY.
Kashyap, A. and Stein, J. (1997), “The role of banks in monetary policy: a survey with
implications for the European monetary union”, Federal Reserve Bank of Chicago
Economic Perspectives, Vol. 21 No. 5, pp. 2-18.
Nicoletti, G., Scarpetta, S. and Boylaud, O. (2000), “Summary indicators of product market
regulation with an extension to employment protection legislation”, OECD Working Paper
No. 226.
OECD (1997), “Economic performance and the structure of collective bargaining”, Employment
Outlook, Chapter 3, OECD, Paris, pp. 70-100.
Sellon, G. (2002), “The changing US ?nancial system: some implications for the monetary
transmission mechanism. Federal Reserve Bank of Kansas City”, Economic Review, Vol. 97
No. 1, pp. 5-35.
Siebert, H. (1997), “Labor market rigidities: at the root of unemployment in Europe”, The Journal
of Economic Perspectives, Vol. 11 No. 3, pp. 37-54.
JFEP
1,4
334
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Appendix
Figure A1.
Monetary policy
transmission channels in
euro area and the USA
Response of the long term interest rate to
a shock on the short term interest rate (EA model)
Response of the output gap to a shock on
the long term interest rate (EA model)
Response of the output gap to a shock on
the short term nom inal interest rate (US model)
Response of the output gap to a shock on
the short term real interest rate (US model)
Response of inflation to a shock on the
output gap (US model)
Response of the inflation to a shock on the
output gap (EA model)
0.35
0.3
0.25
0.2
0.15
0.05
–0.05
–0.15
–0.1
0
0.1
0.15
0.05
–0.05
–0.15
–0.1
0
0.1
0.35
0.3
0.25
0.2
0.15
0.05
0
0.1
0.35
0.3
0.25
0.2
0.15
0.05
–0.05
0
0.1
0.25
0.2
0.15
0.05
–0.05
–0.15
–0.25
–0.3
–0.2
–0.1
0
0.1
0 1 2 3 4 5 6 7 8 9 10 11 12 13
0 1 2 3 4 5 6 7 8 9 10 11 12 13
14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
0 1 2 3 4 5 6 7 8 9 10 11 12 13
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
14
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
–0.2
–0.25
–0.3
–0.35
0.15
0.05
–0.05
–0.15
–0.1
0
0.1
–0.2
–0.25
–0.3
–0.35
–0.4
0.4
The Fed
and the ECB
335
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Figure A2.
Monetary policy reactions
Response of the short term interest rate
to a one-unit shock on the output gap (EA model)
Response of the short term interest rate
to a one-unit shock on the output gap (US model)
Response of the short term interest rate
to a one-unit shock on inflation (EA model)
Response of the short term interest rate
to a one-unit shock on inflation (US model)
Response of the short term interest rate
to a one-unit idiosyncratic shock (EA model)
Response of the short term interest rate
to a one-unit idiosyncratic shock (US model)
2.5
1.5
0.5
0
0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10
0 1 2 3 4 5 6 7 8 9 10
0 1 2 3 4 5 6 7 8 9 10
11 12 13 14 15
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
0 1 2 3 4 5 6 7 8 9 10 11
2
1
1.8
1.6
1.4
1.2
0.8
0.6
0.4
0.2
0
1
1.6
1.4
1.2
0.8
0.6
0.4
0.2
0
1
1.6
1.4
1.2
0.8
0.6
0.4
0.2
0
1
1.6
1.4
1.2
0.8
0.6
0.4
0.2
0
1
1
0.8
0.6
0.4
0.2
–0.2
0
11
JFEP
1,4
336
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Corresponding author
Alexis Penot can be contacted at: [email protected]
1999:1-2005:4 Mean Variance Mean equality Variance equality
Actual EA interest rate 3.08 0.93 Yes No
Actual US interest rate 3.20 3.72 (SL ¼ 0.76) (SL ¼ 0.00)
Estimated EA interest rate 3.11 1.01 Yes No
Estimated US interest rate 3.20 3.71 (SL ¼ 0.83) (SL ¼ 0.00)
EA MP shocks 20.032 0.055 Yes Yes
US MP shocks 0.006 0.068 (SL ¼ 0.57) (SL ¼ 0.29)
Table AI.
Tests on effective interest
rate series
Figure A3.
Responses of variables
to idiosyncratic shocks
Response of the output gap to an
idiosyncratic shock (EA model)
Response of the output gap to an
idiosyncratic shock (US model)
Response of inflation to an
idiosyncratic shock (EA model)
Response of inflation to an
idiosyncratic shock (US model)
Response of wages to an
idiosyncratic shock (EA model)
Response of wages to an
idiosyncratic shock (US model)
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
–0.1
0
0.6
0.5
0.4
0.3
0.2
0.1
–0.1
–0.2
0
0.5
0.4
0.3
0.2
0.1
–0.1
0
0 1 2 3 4 5 6 7 8 9
0 1 2 3 4 5 6 7 8 9
10 11 12 13 0 1 2 3 4 5 6 7 8
0 1 2 3 4 5 6 7 8
0 1 2 3 4 5 6 7 8
9 10 11 12 13 14 15
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
0.35
0.25
0.05
–0.05
–0.1
0.15
0.1
0
0.2
0.05
–0.05
–0.1
0.15
0.1
0
0.2
0.3
1.4
1.2
1
0.8
0.6
0.4
0.2
0
The Fed
and the ECB
337
To purchase reprints of this article please e-mail: [email protected]
Or visit our web site for further details: www.emeraldinsight.com/reprints
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
doc_534462139.pdf
The purpose of this paper is to try to understand the reasons for the differences in
amplitude of monetary policy (MP) rate cycles in the USA and the euro area. Among the different
candidates, the paper aims to test the role of economic structures, macroeconomic shocks, and MP
behaviour
Journal of Financial Economic Policy
The Fed and the ECB: why such an apparent difference in reactivity?
Alexis Penot Grégory Levieuge
Article information:
To cite this document:
Alexis Penot Grégory Levieuge, (2009),"The Fed and the ECB: why such an apparent difference in
reactivity?", J ournal of Financial Economic Policy, Vol. 1 Iss 4 pp. 319 - 337
Permanent link to this document:http://dx.doi.org/10.1108/17576380911050052
Downloaded on: 24 January 2016, At: 21:36 (PT)
References: this document contains references to 35 other documents.
To copy this document: [email protected]
The fulltext of this document has been downloaded 273 times since 2009*
Users who downloaded this article also downloaded:
Abdelkader Boudriga, Neila Boulila Taktak, Sana J ellouli, (2009),"Banking supervision and nonperforming
loans: a cross-country analysis", J ournal of Financial Economic Policy, Vol. 1 Iss 4 pp. 286-318 http://
dx.doi.org/10.1108/17576380911050043
Mikael Bask, (2009),"Optimal monetary policy under heterogeneity in currency trade", J ournal of Financial
Economic Policy, Vol. 1 Iss 4 pp. 338-354http://dx.doi.org/10.1108/17576380911050061
David G. Tarr, (2010),"The political, regulatory, and market failures that caused the US financial
crisis: What are the lessons?", J ournal of Financial Economic Policy, Vol. 2 Iss 2 pp. 163-186 http://
dx.doi.org/10.1108/17576381011070210
Access to this document was granted through an Emerald subscription provided by emerald-srm:115632 []
For Authors
If you would like to write for this, or any other Emerald publication, then please use our Emerald for
Authors service information about how to choose which publication to write for and submission guidelines
are available for all. Please visit www.emeraldinsight.com/authors for more information.
About Emerald www.emeraldinsight.com
Emerald is a global publisher linking research and practice to the benefit of society. The company
manages a portfolio of more than 290 journals and over 2,350 books and book series volumes, as well as
providing an extensive range of online products and additional customer resources and services.
Emerald is both COUNTER 4 and TRANSFER compliant. The organization is a partner of the Committee
on Publication Ethics (COPE) and also works with Portico and the LOCKSS initiative for digital archive
preservation.
*Related content and download information correct at time of download.
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
The Fed and the ECB: why such
an apparent difference
in reactivity?
Alexis Penot
GATE, University of Lyon, CNRS, ENS-LSH, Centre Le´on Be´rard,
Lyon, France, and
Gre´gory Levieuge
Laboratoire d’Economie d’Orle´ans, Faculte´ de Droit, d’Economie et de Gestion,
Orleans, France
Abstract
Purpose – The purpose of this paper is to try to understand the reasons for the differences in
amplitude of monetary policy (MP) rate cycles in the USA and the euro area. Among the different
candidates, the paper aims to test the role of economic structures, macroeconomic shocks, and MP
behaviour.
Design/methodology/approach – The paper starts by estimating vector autoregressive models
both for the USA and the euro area to identify the economic structures, the MP rules, and the
macroeconomics shocks of both areas. Then, it runs counterfactual simulations (by injecting European
Central Bank’s (ECB) monetary rule in the US model for example) to examine which factors had the
most signi?cant impact on differences in MP activism (measured by the variance of interest rates).
Findings – The paper ?nds that differences implied by MP rules alone cannot explain the
dissimilarity of interest rates paths. In the same way, while cyclical shocks are different in each area,
they do not suf?ce to explain the factual divergences. Finally, it is the structural dissimilarities which
essentially explain the difference in interest rate variances.
Originality/value – The paper brings new informations on a controversial issue and it tends to
reject the of?cial explanations given by ECB’s governor who points out differences in shocks.
Keywords Monetary policy, United States of America, Europe
Paper type Research paper
Introduction
“Too little, too late,” is the frequent comment about the stance of the European
monetary policy (MP), in particular when the evolution of its policy rate is compared to
the fed funds rate (Figure 1). Indeed, the European Central Bank (ECB) seems
systematically late especially in slowdown periods and it appears to be less dynamic.
Between January 1999 and September 2007, the ECB moved its key interest rate
23 times, against 35 for the Fed. Even more, between 1999 and 2005, the US policy
rate changes were twice as frequent as the euro rates moves (respectively 30 against
15)[1]. And since interest rate variations have the same magnitude, the US rate
is logically more volatile. The fed funds rate variance is four times higher than the
European repo rate (see the ?rst two lines of Table AI). Can we deduce from this
observation that the ECB is apathetic per se? More basically, where does this
difference come from?
The current issue and full text archive of this journal is available at
www.emeraldinsight.com/1757-6385.htm
The Fed
and the ECB
319
Journal of Financial Economic Policy
Vol. 1 No. 4, 2009
pp. 319-337
qEmerald Group Publishing Limited
1757-6385
DOI 10.1108/17576380911050052
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Following Sahuc and Smets (2007) and Christiano et al. (2007), three main reasons may
explain why actual interest rates volatility differ between these two regions. Indeed, at
the macroeconomic level, economies can be modeled following a classical framework
including three primary elements:
(1) a structure (which de?nes how shocks and monetary impulses are passed on);
(2) temporary shocks; and
(3) a MP represented by an interest rate rule, which react to the variables that are
driven by the structure and the temporary shocks.
Within such a context, differences in one (or more) of these elements are likely to
explain differences in the interest rates dynamics.
On these grounds, we investigate in this paper why the ECB is apparently less
reactive than the Federal reserve (FED). To this end, we follow the methodology
applied by Sahuc and Smets (2007) and Christiano et al. (2007), which is based on
counterfactual analysis. This method consists in simulating a pseudo Euro Area (EA)
interest rate in the cases in which the ECB monetary rule, the European shocks, and the
European structure would successively have been equivalent to their US counterparts.
Thereby, comparing the simulated EA rate with the actual EA and US rates, it is
possible to identify which one of these three components of the EA and US economies
mainly provoke the observed differences.
But contrary to these two papers, the classical framework MP rule-structure-shocks
in this paper relies on vector autoregressive (VAR) models, instead of dynamic
stochastic general equilibrium (DSGE) ones. We assert that DSGE models can be
unsuitable in such a comparative context, especially when the structural models are
supposed to be exactly similar (excepted the calibration). In that case, the conclusion of
structural similarity between the USA and the EA, as obtained by the pre-cited papers,
is almost trivial. On the contrary, aside from ?tting the data well, VAR models allow to
let the data speak (in terms of coef?cients but also in terms of dynamics through the
Figure 1.
Fed fund and ECB
repo rates
7
6
5
4
3
2
1
0
Jan. - 99 Jan. - 00 Jan. - 01 Jan. - 02 Jan. - 03 Jan. - 04 Jan. - 05
ECB repo rate
US federal funds rate
JFEP
1,4
320
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
choice of the p-order of the model), without structural preconceptions, which is an
advantage in this methodological grounds. We intend in this way to revisit the existing
results, and we show in particular that the questionable conclusion of structural
similarity between the USA and the EA does not hold with such an alternative model.
The reminder of the paper is organized as follows. Section 1 deals with the literature
related to this paper. Section 2 aims at formally delivering the three sources of
divergence between both areas. Section 3 explains the characteristics of the VARmodels
which we estimate for both areas. Then, we proceed to the counterfactual simulations,
comparing successively the monetary policies (Section 3), the shocks (Section 4), and the
structures (Section 5) of the EA and the USA. We conclude that economic structures do
explain the differences observed ex post between the ECB and the Fed interest rates,
while potential differences in shocks and MPrules do not. However, it does not mean that
the ECBand the FEDshare the same policy preferences. To this respect, in Section 6, we
evaluate what would have been the (optimal) EA interest rates with the US policy
preferences.
1. Related literature
Some papers have already studied the differences between the EA and the US interest
rates paths. Most of them consists in applying the method of counterfactual
simulations to DSGE models. Following this lines, on one hand, Christiano et al. (2007)
?nd that structural and cyclical differences could explain the diverging evolution of
output and in?ation. But since the ECB monetary rule implies more inertia than the US
one, these divergences do not turn up on interest rate paths.
On the other hand, Sahuc and Smets (2007) startlingly echoes Trichet’s (2006)
arguments by claiming that the interest rate divergences are explained by differences
in shocks, more than differences in structure or in MP. Using a similar structural model
estimated with US and EA data. Smets and Wouters (2005) also conclude that both
areas are structurally similar.
These contributions suffer from factual and methodological inconsistencies. First,
given the institutional and historical features which differentiate them, the ECB and
the Fed MPs are unlikely to be conducted similarly[2]. So, it is worth examining the
robustness of this result.
Moreover, regarding institutional disparities, it is hard to concede that the US and the
EA share the same economic structure. Relying on structural DSGE models and despite
some ad hoc accommodations (like backward-looking price setting, capital adjustment
costs, autoregressive shocks, etc.) these last references are likely to underestimate the
inertial behaviours of agents and then the sluggishness of data (Fuhrer, 1997; Estrella
and Fuhrer, 2002). Then, only “shocks” account for inertia, with the risk to overestimate
them if the economy is rigid. More basically, these models have not demonstrated their
superiority in terms of shock identi?cations. More embarrassing, the conclusion of
structural similarity is almost trivial when the structure of both economies is
theoretically identical and established a priori, while only the calibration is supposed to
be different. In this respect, Smets and Wouters (2005, p. 162) acknowledge: “the more
structure is imposed on the estimated model, the more the results will be colored by the
selected theoretical speci?cation.”
An alternative approach is then suitable to reassess these conclusions. Since imposing
an a priori structure is likely to pervert any comparison, it is interesting to use VAR
The Fed
and the ECB
321
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
models which precisely intend to let the data speak with few structural preconceptions.
Moreover, these models duly consider the sluggishness of macroeconomic variables. In
this respect, their ?tting and explanatory power are so largely recognized that they have
constituted benchmark models for revealing stylized facts in macroeconomics for several
decades. In particular, these models are used to assess the quality of structural
macroeconomic models (like DSGE); impulse responses functions (IRF) of VAR models
then constitute a reference which any structural model ought to reproduce[3].
We will then apply the method of counterfactual simulations to VAR models
estimated for the euro area and the USA. Moreover, the results of our simulations will
be submitted to econometrical tests that we expect to be more rigorous than the mere
graphical analyses used in the pre-cited papers. Beforehand, the next section provides a
formal representation of the method of counterfactual simulations.
2. What are the potential sources of divergence?
This section describes an economy reduced to four elements: a structure ðFð · ÞÞ,
temporary shocks (1
X
), a MP governed by a rule ð f ð · ÞÞ, and some MP shocks (1
i
). In
order to formalize Sahuc and Smets (2007) and Christiano’s et al. (2007) method, we can
write the linear system:
X
c
¼
c
F
c
F
c
ðX
c
; i
c
Þ þ1
c
X
i
c
¼
b
f
c
f
c
ðX
c
Þ þ1
c
i
8
<
:
ð1Þ
with c ¼ {US; EA} and X
c
a vector of n state variables (;n). Concretely,
c
F
c
F
c
ð · Þ is an
estimated multidimensional function which represents the economic structures, and 1
c
X
the vector of the associated error terms (the difference between the actual values and
the values predicted by the estimated structure) which represent by de?nition the
shocks that hit this structure. Next, i
c
is the short-term interest rate, i.e. the MP
instrument which depends, through an estimated MP rule
^
fð · Þ, on the state variables of
the system. The difference between the actual and the implied by the rule interest rates
represents the MP shocks (including the discretionary part of the MP), noted 1
i
. It is
important to note that, a priori, structures
c
F
c
F
c
ð · Þ, shocks 1
c
X
, MP rules
b
f
c
f
c
and/or MP
shocks 1
c
i
are distinct in the two regions.
Thereby, we resort to counterfactual simulations to detect the main cause(s) of
divergence. Concretely, successively substituting the EA structure, shocks, MP rule,
and MP shocks with their US counterparts, we calculate pseudo European interest
rates and test if they are signi?cantly different from the actual rate (diagram (1) in
Figure 2) and if they come closer to the actual US rate. This implies four cases.
First, what would have happened if the ECB had followed the US MP rule (diagram
(2))? The response is given by the following simulated interest rate:
Figure 2.
Counterfactual scenarios
e
X
EA
e
X
EA
e
X
US
f
EA
f
EA
f
EA
f
EA
f
US
X
US
X
EA
X
EA
X
EA
X
EA
i
EA
i
EA
i
EA
i
EA
i
EA
e
X
EA
e
X
EA
e
i
EA
e
i
US
e
i
EA
e
i
EA
(1) (2) (3) (4) (5)
e
i
EA
JFEP
1,4
322
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
~
i
EA
¼
d
f
US
f
US
ðX
EA
Þ þ1
EA
i
ð2Þ
Second, what would be the EA interest rate if the European MP surprises were similar
to the US ones (diagram (3))? The solution is given by:
~
i
EA
¼
d
f
EA
f
EA
ðX
EA
Þ þ1
US
i
ð3Þ
Third, given that the state variables can be decomposed as X
c
¼
c
X
c
X
c
þ1
c
X
, we can
determine to what extent the cyclical differences are likely to explain the observed
divergences between actual US and EA interest rates. 1
EA
X
has to be substituted with
1
US
X
in the European system (diagram (4)) to deduce a pseudo interest rate as:
~
i
EA
¼
d
f
EA
f
EA
d
X
EA
X
EA
þ1
US
X
þ1
EA
i
ð4Þ
In these cases, MP rules, MP shocks, and structures remain proper to each area.
Finally, a pseudo interest rate can be calculated, under the assumption that the euro
area works under the US structure (diagram (5)), everything else equal:
~
i
EA
¼
d
f
EA
f
EA
d
X
US
X
US
þ1
EA
X
þ1
EA
i
ð5Þ
This method allows to identify which sources of divergence among the four explain the
lower volatility of the European interest rate. Beforehand, it is necessary to explain the
models used to implement the generic method exposed so far.
3. Modeling the euro area and the USA
In the light of the arguments presented in introduction, VAR models are particularly
relevant in this problematic. They usually are reference models for stylized facts,
whereas DSGE models are inadequately discriminant and sometimes empirically
unsatisfying[4]. Finally, in the perspective of counterfactual simulations, VAR models
bene?t from few theoretical preconceptions.
3.1 Models, data, and estimation
The two VAR models can be written as follows:
X
t
¼ A
1
X
t21
þ · · · þ A
p
X
t2p
þ BZ
t2k
þ1
t
ð6Þ
with X a ?ve-element vector composed of the long-term real interest rate r, the output
gap y, the wage annual growth rate w, the in?ation rate p and the short-term nominal
interest rate i. For degrees of freedom purposes, we only consider the most important
variables at stake in the MP transmission mechanism and add the wage growth to
capture the institutional and cyclical divergences often observed between the US and
EA labor markets (Sahuc and Smets, 2007; Christiano et al. (2007)). 1 is the vector of
innovations associated to the endogenous variables, i.e. ð1
r
; 1
y
; 1
w
; 1
p
; 1
i
Þ, which,
respectively, represent a shock on the long-term interest rate (or ?nancial shock), a
demand shock, a labor shock, a supply shock and a MP shock. As the equation de?ning
the short-term interest rate is comparable to a MP rule[5], the shock 1
i
stands for a
temporary deviation to the rule, that is central bank discretionary decisions. Z stands
for exogenous commodity prices, inserted to successfully smooth the price puzzle effect
(with k depending on its informational contribution).
The Fed
and the ECB
323
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
US data stem from Datastream and EA from the area-wide model database (Fagan
et al., 2005), updated in autumn 2006, up to the fourth quarter of 2005. All variables are
seasonally adjusted via the standard X-11 method. The exogenous variable is the
Goldman Sachs Commodity Prices Index. For both areas, because of a possible unit
root in the real gross domestic product (GDP) series, we de?ne the output gap as the
residuals of the regression of the GDP on a constant and a quadratic trend, instead of
using a usual Hodrick-Prescott ?lter. Short- and long-term interest rates are the
three-month monetary rate and the ten-year benchmark rate, respectively[6].
Maximum-likelihood estimation covers the 1985-2005 period. Certainly, a
non-parametric estimation method could be applied, like a bayesian VAR (BVAR)
estimation, in particular, as suggested for instance by Litterman (1986). Besides, it is
shown that a BVAR estimation is more accurate than a maximum likelihood (ML)
estimation, provided that the a priori information (i.e. the priors) is adequate. But
imposing conditions, as required by this method, means setting restrictions
(i.e. theoretical preconceptions) on the estimated model. Instead, we want to let the
data speak. We want the differences to be entirely determined as a result and not as a
consequence of starting hypothesis, contrary to bayesian estimations of DSGE models,
which use to impose the same prior conditions for the EA and the USA, and which
inevitably lead to conclude that both regions are structurally similar.
Moreover, prior probability distributions raise numerous questions. Unknowing
about model parameters is often important, what is source of numerous pitfalls, as
shown by the discussion between Sims and Uhlig (1991), Phillips (1991a, b) and
Sims (1991). The choice of a given prior density is rarely duly justi?ed nor discussed
in the DGSE literature. While a lot of robustness checks would be required,
economists content themselves with replicating the beliefs applied in the existing
literature.
On the contrary, estimating a VAR through ML method imposes only one
restriction: it is implicitly assumed that the relations between the variables are linear.
A potential misspeci?cation of this order is however unlikely with regard to the vast
literature asserting the good ?tting of VAR models.
The p-order of the VAR models is determined with the parsimonious Bayesian
information criterion, with respect to the white noise hypothesis of disturbances.
Finally, we ?nd p ¼ 3 for the EA and p ¼ 5 for the USA.
Once estimated, the VAR models are used to generate IRFs[7], in order to asses their
dynamics (in the light of some well-known stylized facts) and to have an insight on the
differences between both areas.
3.2 The main characteristics of the US and EA models
Figure A1 represents the essential IRFs in the perspective of studying the transmission
of shocks and MP impulses. As a whole, it actually describes the textbook MP
transmission mechanism. In the euro area, the short-term interest rate has a positive
effect on the long-term interest rate, even if it is not contemporaneous because of the
Cholesky decomposition. Next, the long-term interest rate impacts the output gap
negatively. In accordance with the expected transmission delay of MP, this in?uence is
only signi?cant after the sixth quarter following the initial shock on the long-term
interest rate. Then a positive output gap has a huge and positive incidence on in?ation,
still visible after 15 quarters.
JFEP
1,4
324
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
For institutional reasons, the standard MP transmission mechanism mainly works
through the short-term interest rate in the USA. Strictly speaking, the real short-term
interest rate should have been considered. But the interpretation of an interest rate rule
in real terms is less direct, and passing from a real to a nominal short-term interest rate
would have rendered dif?cult the implementation of the counterfactual simulations.
Certainly, IRF con?rm the negative impact of the nominal short-term interest rate on
the output gap, with a signi?cant effect starting one year after the shock and still
effective three years latter. But the effect of the real short-term interest rate is also
veri?ed: the response of the output gap is then signi?cantly negative from the
seventh quarter and remains during the six following quarters (Appendix, Figure A1).
Finally, the output gap has a positive incidence on in?ation for three years after the
initial shock.
Before examining the results of the counterfactual simulations, it is worth noting
that such a method applied to VAR models is subject to the Lucas criticism.
Nevertheless, according to Hansen global stability and Chow predictive tests (with a
break point in 1999Q1), the VAR relations estimated for the EA are strikingly stable,
even concerning the MP rule equation (and this despite the transition from national
monetary policies to a common one).
This result is consistent with the reserves of numerous studies which, without
reviewing its theoretical logic, attenuate the practical signi?cance of the Lucas
criticism (Altissimo et al., 2002; Estrella and Fuhrer, 1999). Moreover, the Lucas
criticism is here dampened by the uncertainty surrounding the ECB implementation
and policy. Around 1999, economic agents were not able to formulate well-founded
expectations about the European MP. As a result, their attitude had probably not be
different if the ECB had chosen to follow the Fed’s rule for instance. As a whole, in the
?rst time of the ECB exercise, uncertainty (about MP, structures, and shocks),
observation and learning would have been the same whatever the MP rule.
4. Comparing monetary policy rules
The ?rst lines of the Table AI in the Appendix show that explaining the differences in
interest rate paths between the EA and the USA does not require explaining
differences in levels, but in variations. This being said, can the behaviour of central
banks explain the difference in interest rate paths?
In accordance with equation (2), we determine a pseudo EA interest rate, assuming
the ECB had followed the Fed’s rule (estimated over the 1999-2005). As indicated
in Table I, we are lead to accept the equivalence of the EA and the USA. MP rules,
1999:1-2005:4 Variance Variance equality
Euro area with the Fed MP rule
Actual EA interest rate 0.92 Yes
Simulated EA interest rate 0.90 (SL ¼ 0.47)
Euro area with the Fed MP shocks
Actual EA interest rate 0.92 Yes
Simulated EA interest rate 0.69 (SL ¼ 0.22)
Note: Monetary policy scenarios
Table I.
Tests on counterfactual
interest rate series
The Fed
and the ECB
325
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
in terms of interest rate volatility. So, with the Fed’s rule, the ECB would not have
appeared as more “reactive.” Therefore, MP rule differences do not explain the
divergences of interest rate paths[8].
We also examine the differences due to MP shocks, according to equation (3).
Logically, given the mean and variance equalities of these disturbances (see the last
lines of Table AI), the discretionary stance of monetary policies neither explain the
observed divergence of interest rate paths, as indicated in the second part of the Table I.
Consequently, the US and EA interest rates differences might be of structural and/or
cyclical order(s).
5. Comparing shocks
Is the larger inertia of the European MP due to smaller and less frequent shocks?
Preliminary statistical investigations reveal that long-term interest rates are correlated,
mainly because of international ?nancial spillovers, and ?nancial globalization makes
this correlation even higher for the last ten years. Supply shocks are also strongly
correlated, especially for the 1999-2005 period. Therefore, demand and labor shocks are
the only potential explanation for different MP stances. Next, according to the
historical decomposition of shocks, it appears that shocks are not of the same nature in
the two regions. This can be a sign of cyclical divergences, as asserted by Sahuc and
Smets (2007) and Trichet (2006).
But this preliminary analysis is not likely to explain the smaller volatility of the
European interest rate. Indeed, European shocks should be smaller or less volatile to
this end. And yet, variance equality tests reported in Table II indicate this is not the
case: shocks on long-term interest rates, in?ation, and labor market[9] are signi?cantly
larger in the euro area. Only demand shocks are slightly higher in the USA over
1999-2005, as already found in Sahuc and Smets (2007)[10]. These differences cannot
justify a stronger interest rate inertia (that is even the opposite).
The counterfactual simulations based on equation (4), summarized in Table III,
con?rm these inferences. If European shocks would have been equal to US ones, the
European interest rate would have not be statistically different from its actual value.
Besides, the same exercise for the USA is very instructive. Indeed, the second part of
the Table III indicates that the US interest rate would have been even more volatile if
driven by the European shocks (instead of the US ones). So, the USA seem to be more
sensitive to cyclical shocks than the EA. Therefore, if both regions do not
similarly respond to shocks, the interest rate differential must stem from structural
differences.
Varð1
EA
r;t
Þ= Varð1
US
y;t
Þ= Varð1
EA
w;t
Þ= Varð1
EA
p;t
Þ= Varð1
EA
i;t
Þ=
Varð1
US
r;t
Þ Varð1
EA
y;t
Þ Varð1
US
w;t
Þ Varð1
US
p;t
Þ Varð1
US
i;t
Þ
1985-2005 3.53
* *
1.15 7.26
* *
4.74
* *
1.59
*
1999-2005 6.61
* *
1.61
*
4.24
* *
6.21
* *
1.23
a
Notes: Rejection of the null hypothesis of variance equality at:
*
10 and
* *
1 percent level, respectively;
a
corresponds to Vð1
US
i;t
Þ=Vð1
EA
i;t
Þ
Table II.
Tests of variance equality
JFEP
1,4
326
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
6. Comparing economic structures
First, Figure A3 shows the inertia degree of the two regions, on the basis of the
response functions of the real variables of the model to idiosyncratic shocks. Table IV
offers a measure of this adjustment speed[11].
In the euro area, subsequently to a cost-push supply shock, in?ation returns to the
equilibrium level after seven quarters. The European in?ation dynamics is more
sluggish in the very short run. Moreover, as indicated in the last line of the Table IV, a
shock on the European labor market has a strongly persistent effect on wages. It takes
about two years for the wage impulse-response function to be cut by half in the euro
area. The wage growth is the most viscous variable of the European system.
Conversely, such a shock is quickly absorbed by the US labor market (where more than
half of the total effect is absorbed in the ?rst six months, compared to only a fourth in
the euro area in the same time). Finally, in the short run (until two years), the output
gap has a higher persistence in the euro area. So, differences of reactivity are obvious
between the EA and the USA.
Now, the counterfactual experience consists in simulating an European interest rate
between 1995 and 2005 under a US economic structure (relation (5)). Table V shows
that, in this case, the pseudo European interest rate is signi?cantly higher. The
variance equality test even indicates that the simulated European interest rate is
statistically equivalent to the US one. European in?ation and output gap would also be
more volatile if US and European structures would have matched[12].
Therefore, differences in structure do explain a large part of the differences in
interest rate volatilities between the two regions. This result is in contradiction with
Euro area (%) USA (%)
Reponse of Six months One year Two years Six months One year Two years
p to a shock on p 62 81 100 49 79 100
y to a shock on y 24 43 72 35 57 75
w to a shock on w 27 36 54 53 65 80
Notes: Numbers represent the relative adjustment made by each variable, in terms of cumulated
impulse response, second, fourth, and eight quarters after an initial idiosyncratic shock; full
adjustment is supposed to be achieved when the corresponding IRF is no longer signi?cantly different
from 0
Table IV.
Speed of adjustment of
in?ation (p), output gap
( y), and wages (w)
1999:1-2005:4 Variance Variance equality
Euro area with the US shocks
Actual EA interest rate 0.92 Yes
Simulated EA interest rate 1.21 (SL ¼ 0.24)
USA with the EA shocks
Actual US interest rate 3.72 No
Simulated US interest rate 18.8 (SL ¼ 0.00)
Note: Shocks scenario
Table III.
Tests on counterfactual
interest rate series
The Fed
and the ECB
327
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
previous studies. But yet, although the euro area is itself made of heterogeneous
countries, its global structure obviously differs from that of the USA.
7. Discussion on central bank preferences
Figure 3 shows the interest rate paths implied by the various scenarios considered. The
scenario in which European monetary authorities conduct their policy with the US
structures (everything else being equal) is the only case in which the variance of the
European interest rate is comparable to the variance of the US rate[13]: however, it is
worth noting that in this case the interest rate would have come up against the lower
zero bound for interest rate. On the contrary, the differences between the simulated and
the actual interest rates are not signi?cant in average when shocks are switched. In the
same way, MP rules also do not suf?ce to explain differences in interest rates.
But this last result does not mean that the ECB and the FED have the same objective
function. If central banks behave optimally, it can be theoretically demonstrated that
1999:1-2005:4 Variance Variance equality
Euro area with the US structure
Actual EA interest rate 0.92 No
Simulated EA interest rate 3.68 (SL ¼ 0.05)
Actual EA output gap 1.02 No
Simulated EA output gap 9.45 (SL ¼ 0.00)
Actual EA in?ation 0.76 No
Simulated EA in?ation 3.99 (SL ¼ 0.00)
Actual US interest rate 3.72 Yes
Simulated EA interest rate 3.68 (SL ¼ 0.00)
Note: Structural scenario
Table V.
Tests on counterfactual
interest rate series
Figure 3.
Simulated EA interest
rates
10
9
8
7
6
5
4
3
2
1
0
Q
2
-
1
9
9
9
Q
3
-
1
9
9
9
Q
4
-
1
9
9
9
Q
1
-
2
0
0
0
Q
2
-
2
0
0
0
Q
3
-
2
0
0
0
Q
4
-
2
0
0
0
Q
1
-
2
0
0
1
Q
2
-
2
0
0
1
Q
3
-
2
0
0
1
Q
4
-
2
0
0
1
Q
1
-
2
0
0
2
Q
2
-
2
0
0
2
Q
3
-
2
0
0
2
Q
4
-
2
0
0
2
Q
1
-
2
0
0
3
Q
2
-
2
0
0
3
Q
3
-
2
0
0
3
Q
4
-
2
0
0
3
Q
1
-
2
0
0
4
Q
2
-
2
0
0
4
Q
3
-
2
0
0
4
Q
4
-
2
0
0
4
Q
1
-
2
0
0
5
Q
2
-
2
0
0
5
Q
3
-
2
0
0
5
Q
4
-
2
0
0
5
Q
1
-
1
9
9
9
I
n
%
Actual EA interest rate
EA interest rate with US monetary policy Rul
EA interest rate with US shocks
EA interest rate with US structure
EA interest rate with US monetary policy shocks
Actual US interest rate
JFEP
1,4
328
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
their (estimated) MP rule responds to the optimization of an objective function, subject
to a given structural model and a given variance of shocks. Now, it is shown that in a
welfare view a` la (Woodford, 2003), objective functions precisely depend on the
structural parameters of the economies. And as Section 6 led to conclude that the
structures between the two regions is different, the objective functions may also do.
So, in order to strengthen the conclusion of structural dissimilarity, the last exercise
consists in determining and comparing the preference parameters of policymakers.
More precisely, we will simulate what the FED would have (optimally) done, had it
been in charge of the European MP, conditional on its own objective function[14].
Revealing ?rst the preference parameters of the FED is needed to this end[15].
7.1 The method for revealing preference parameters
We assume the Fed had to minimize the following usual quadratic loss function, where
l
y
and l
i
represent, respectively, the preference parameter for output and interest rate
variabilities:
Loss
t
¼ l
y
Varð y
t
Þ þ ð1 2l
y
ÞVarðp
t
Þ þl
i
Varði
t
2i
t21
Þ ð7Þ
subject to a given structural model Z
t
¼ FZ
t21
þU
t
, a variance-covariance matrix M
associated to the innovations’ vector U, and a MP rule i
t
¼ f ðXÞ. It can be
demonstrated, following Svensson (1998) and Ball (1997) that the unconditional
contemporaneous covariance matrix of Z, noted V is given in vector form by:
VecðVÞ ¼ ½I 2F^F?
21
VecðMÞ ð8Þ
So, the optimal MP rule is given by the parameters of the vector f that minimize the
weighted sum – based on the given preference parameters ðl
y
; l
i
Þ – of the
unconditional variances for the in?ation rate, the output gap and the interest rate.
These variances appear in VecðVÞ, following equation (8).
This robust method can be used to reveal the US preference parameters. To make
the algorithm optimization tractable and to focus on the main usual targets of MP
rules, we limit the US model to three main variables: interest rate, in?ation and output
gap. We note
^
i the interest rate deduced from the US estimated MP rule (MP shocks are
ruled out). The effective preference parameters of the Fed, ðl
*
y
; l
*
i
Þ, determine a MP
rule yielding an “optimal” interest rate
^
i
*
which is the nearest to the estimated rate
^
i.
The optimization program is then given by:
{l
y
;l
i
}
Min distance ¼
X
N
t¼0
^
i
*
t
ðl
y
; l
i
Þ 2
^
i
t
h i
2
with
^
i
*
t
¼ f
*
Z such as
{f
*
}
Min L ¼ l
y
Varðy
t
Þ þ ð1 2l
y
ÞVarðp
t
Þ þl
i
VarðDi
t
Þ
with Z
t
¼ F½ f ðl
y
; l
i
Þ? þ U
t
and Z ¼ ½ y
t21
. . .y
t25
; p
t21
. . .p
t25
; i
t21
. . .i
t25
?
8
>
>
>
>
>
>
>
<
>
>
>
>
>
>
>
:
7.2 What would have been the ECB rates with the US preferences?
Applying this method, we ?nd l
*
y
¼ 0:42 and l
*
i
¼ 5:6. These preference parameters
generate an optimal MP rule with an interest rate path that ?ts very well the estimated
interest rate, as it is shown by Figure 4 (distance ¼ 11.62). This result is consistent.
The Fed
and the ECB
329
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
On the one hand, the Fed has an explicit objective of output stabilization (while
in?ation stability is its main objective). On the other hand, a high value of l
i
is required
to reproduce interest rate inertia (Dennis, 2001). When these parameters are applied to
the euro area, they generate an optimal interest rate path which is more volatile than it
is in fact, as shown by Figure 5.
To such an extend that, as shown in Table VI, the variance equality between the
optimized EA interest rate and the US effective (estimated) interest rate is accepted.
Implicitly, this suggests that the parameter l
i
is higher in the ECB’s loss function. This
result can be interpreted twofold. Either it means that the European authorities are very
cautious, in the sense they display a high preference for interest rate smoothing. This
behaviour can be justi?ed in the light of Brainard’s arguments, because of the large
uncertainty that surrounds the total effects of a common European MP. Or, following
Woodford(2003, Chapter 6), it canbe demonstratedthat anobjective functionlike equation
(7) can be derived fromthe discounted sumof households’ utility. In this case, parameters
ðl
i
; l
y
Þ do not represent the monetary authorities’ preferences per se but the results of an
optimization implying the structural parameters of the underlying model. According to
this point of view, concluding that differences in objective functions can explain the
difference of volatilitybetweenthe USandthe EAinterest rates is equivalent toaf?rmthat
structural divergences explain it, what strengthens the conclusion of Section 6.
Conclusion
Given the dissimilar variances of the European and US interest rate, the ECB is often
blamed for not being reactive enough in comparison with the Fed. But such a hasty
Figure 4.
US optimal and actual
interest rates
–4
–3
–2
–1
0
1
2
3
4
5
Q
1
-
1
9
8
6
Q
3
-
1
9
8
6
Q
1
-
1
9
8
7
Q
3
-
1
9
8
7
Q
1
-
1
9
8
8
Q
3
-
1
9
8
8
Q
1
-
1
9
8
9
Q
3
-
1
9
8
9
Q
1
-
1
9
9
0
Q
3
-
1
9
9
0
Q
1
-
1
9
9
1
Q
3
-
1
9
9
1
Q
1
-
1
9
9
2
Q
3
-
1
9
9
2
Q
1
-
1
9
9
3
Q
3
-
1
9
9
3
Q
1
-
1
9
9
4
Q
3
-
1
9
9
4
Q
1
-
1
9
9
5
Q
3
-
1
9
9
5
Q
1
-
1
9
9
6
Q
3
-
1
9
9
6
Q
1
-
1
9
9
7
Q
3
-
1
9
9
7
Q
1
-
1
9
9
8
Q
3
-
1
9
9
8
Q
1
-
1
9
9
9
Q
3
-
1
9
9
9
Q
1
-
2
0
0
0
Q
3
-
2
0
0
0
Q
1
-
2
0
0
1
Q
3
-
2
0
0
1
Q
1
-
2
0
0
2
Q
3
-
2
0
0
2
Q
1
-
2
0
0
3
Q
3
-
2
0
0
3
Q
1
-
2
0
0
4
Q
3
-
2
0
0
4
Q
1
-
2
0
0
5
Q
3
-
2
0
0
5
Estimated US interest rate (demeaned, no shocks)
Optimal US interestrate (demeaned)
JFEP
1,4
330
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
appraisal forgets that MP must be evaluated in its context, i.e. considering both the
economic structures and the cyclical shocks. In this prospect, several studies searched
to explain the reasons for such an apparent difference in reactivity. As a whole, they
conclude that the cyclical shocks are mainly responsible. Beyond that, monetary
policies and, more surprisingly, economic structures would be the same in both areas.
But these conclusions are specious as they a priori assume that the US and EA models
are structurally identical.
So, in this paper, we reassess these results in the light of a VAR model, which is
precisely exempted from a priori theoretical preconceptions. Counterfactual
simulations con?rm that differences implied by MP rules alone cannot explain the
dissimilarity of interest rates paths. In the same way, while cyclical shocks are
different in each area, they do not suf?ce to explain the factual divergences. Finally, it
is the structural dissimilarities which essentially explain the differences in interest rate
1999:1-2005:4 Variance Variance equality
Estimated EA interest rate 1.08 No
Optimized EA interest rate with US preferences 4.25 (SL ¼ 0.00)
Estimated US interest rate 4.33 Yes
Optimized EA interest rate with US preferences 4.25 (SL ¼ 0.48)
Note: Preference scenario
Table VI.
Tests on counterfactual
interest rate series
Figure 5.
Simulated EA optimal
interest rate with US
preferences
–10
–6
–2
2
6
10
Q
1
-
1
9
8
6
Q
3
-
1
9
8
6
Q
1
-
1
9
8
7
Q
3
-
1
9
8
7
Q
1
-
1
9
8
8
Q
3
-
1
9
8
8
Q
1
-
1
9
8
9
Q
3
-
1
9
8
9
Q
1
-
1
9
9
0
Q
3
-
1
9
9
0
Q
1
-
1
9
9
1
Q
3
-
1
9
9
1
Q
1
-
1
9
9
2
Q
3
-
1
9
9
2
Q
1
-
1
9
9
3
Q
3
-
1
9
9
3
Q
1
-
1
9
9
4
Q
3
-
1
9
9
4
Q
1
-
1
9
9
5
Q
3
-
1
9
9
5
Q
1
-
1
9
9
6
Q
3
-
1
9
9
6
Q
1
-
1
9
9
7
Q
3
-
1
9
9
7
Q
1
-
1
9
9
8
Q
3
-
1
9
9
8
Q
1
-
1
9
9
9
Q
3
-
1
9
9
9
Q
1
-
2
0
0
0
Q
3
-
2
0
0
0
Q
1
-
2
0
0
1
Q
3
-
2
0
0
1
Q
1
-
2
0
0
2
Q
3
-
2
0
0
2
Q
1
-
2
0
0
3
Q
3
-
2
0
0
3
Q
1
-
2
0
0
4
Q
3
-
2
0
0
4
Estimated EA interestrate (demeaned, no shocks)
Optimal EA interestrate with US preferences
The Fed
and the ECB
331
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
variances. This conclusion is corroborated by a simulation indicating that the
European interest rate would have been more volatile if the ECB had the same deep
preference parameters as the FED. In a welfare perspective of a utility-based loss
function, this result tends to con?rm the structural divergences between both areas.
Besides, this conclusion is consistent with the largely spread wisdom that the USA and
the EA have dissimilar labour, ?nancial, and good markets.
So, a less reactive Central Bank is not necessarily an institution which is not prompt
to move its policy rates, nor per se an apathetic Central Bank. In this precise case, the
ECB conducts the MP of an inertial economy, which slowly and partially reacts to
shocks.
Notes
1. The timing and magnitude of MP responses to the subprime crises is no exception for the
most recent period.
2. Unlike the ECB, the Fed is explicitly mandated to take care of the output cycle (and not only
to ?ght in?ation). Moreover, the composition of MP committees and the way they come to a
decision can induce more or less inertia (Gerlach-Kristen, 2005). At least, the preferences of
the monetary authorities should be different.
3. For instance, Rotemberg and Woodford (1997) and Christiano et al. (2005) suggest to test the
reliability of DSGE models with respect to the IRFs of VAR models.
4. See for instance Blanchard (2008), Chari et al. (2008) and Iskrev (2008) for criticisms on DSGE
models which are particularly relevant in the perspective of this problematic.
5. The backward looking feature of this MP rule can be justi?ed by the unavailability of data in
real time (McCallum and Nelson, 1999). Moreover, as it is shown by a lot of reaction function
estimates, the past interest rate has to be added to the baseline Taylor rule for a good ?tting,
as i
t
¼ b
1
i
t21
þb
2
b pp
t
þb
3
y
t
, with b pp
t
¼ ðp
t
2 p
t
Þ. In this case, the MP rule can be written:
i
t
¼ b
2
P
n
n¼0
b
n
1
b pp
t2n
þb
3
P
n
n¼0
b
n
1
y
t2n
, where reference to the past is obvious.
6. As unit-root hypothesis is rejected by usual ADF tests, the cointegration study is useless.
7. IRFs are computed through bootstrapping and based on a Cholesky decomposition with the
variables ordered as speci?ed in equation (6).
8. In the margin of this exercise, we ?nd that, for the US simulated rate under the European
rule, the variance equality is accepted at the 5 percent but rejected at the 10 percent level.
However, even less volatile, the simulated US rate does not come closer to the EA one.
9. It can be argued that, since the early 1990s, all European countries have been promoting
structural reforms while the USA only experienced the earned income tax credit.
10. However, we do not share Smets and Wouters (2005) conclusion about the euro area more
affected by negative supply shocks. Our shocks have the same variance in the two regions.
11. Denoting f
ij,t
, the dynamic multiplier deriving from the VMA representation of a VAR
model, it represents the effect of an innovation 1
j
occurring at date t on the ith variable
observed in t þ m. The measure of adjustment delay of the ith variable submitted to an
idiosyncratic shock is then given by:
P
m
t¼0
jf
ij;t
j=
P
1
t¼0
jf
ij;t
j with i ¼ j and m ¼ {2, 4, 8}
quarters. For the denominator, the sum stops in fact when f
ij,t
is no longer signi?cant.
12. In the same way, we have found that the variance of the simulated US interest rate, subject to
the European structure, is signi?cantly lower than its actual variance. Moreover, the
variance equality between this pseudo US rate and the effective EA rate cannot be rejected.
13. In this case, the difference in level is largely explained by the higher US initial conditions.
JFEP
1,4
332
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
14. This section goes into detail the arguments mentioned by Christiano et al. (2007, footnote 3,
p. 2478): “it is possible that if the FED were literally in charge of the ECB, it might not have
applied the same monetary policy strategy that it uses in the US [. . .] [This] question would
require the identi?cation of the FED’s objective function, and then the computing of the
monetary policy rule that optimizes it, conditional on the economy corresponding to the
estimated EA economy.” It is exactly what is done in this section.
15. Certainly, this task is dif?cult to conceive for the euro area., where MP was national before
1999 and raises the problem of national preference parameters aggregation.
References
Altissimo, F., Sieviero, S. and Terlizzese, D. (2002), “How deep are the deep parameters?”,
Annales d’Economie et de Statistiques, Vol. 67-68, pp. 207-26.
Ball, L. (1997), “Ef?cient rules for monetary policy”, NBER Working Paper Series No. 5952.
Blanchard, O. (2008), “The state of Macro”, NBER Working Paper Series No. 14259.
Chari, V., Kehoe, P. and McGrattan, E. (2008), “New-Keynesian models are not yet useful for
policy analysis”, Federal Reserve Bank of Minneapolis Working Papers.
Christiano, L., Eichenbaum, M. and Evans, C. (2005), “Nominal rigidities and the dynamic effects
of a shock to monetary policy”, Journal of Political Economy, Vol. 113 No. 1, pp. 1-45.
Christiano, L., Motto, R. and Rostagno, M. (2007), “Shocks, structures or policies? A comparison
of the EA and the US”, Journal of Economic Dynamics and Control, Vol. 32, pp. 2476-506.
Dennis, R. (2001), “The policy preferences of the US federal reserve”, Federal Reserve Bank of
San Francisco, Working Paper in Applied Economic Theory 2001-08 (accessed July 2004).
Estrella, A. and Fuhrer, J. (1999), “Are ‘Deep’ parameters stable ? The Lucas critique as an
empirical hypothesis”, Federal Reserve Bank of Boston Working Papers No. 99-4.
Estrella, A. and Fuhrer, J. (2002), “Dynamic inconsistencies: counterfactual implications of a class
of rational expectations models”, American Economic Review, Vol. 92 No. 4, pp. 1013-28.
Fagan, G., Henry, J. and Mestre, R. (2005), “An area-wide model (AWM) for the euro area”,
Economic Modelling, Vol. 22, pp. 39-59.
Fuhrer, J. (1997), “The (un)importance of forward-looking behavior in price speci?cations”,
Journal of Money, Credit, and Banking, Vol. 29 No. 3, pp. 338-50.
Gerlach-Kristen, P. (2005), “Too little, too late: interest rate setting and the cost of consensus”,
Economics Letters, Vol. 88, pp. 376-81.
Iskrev, N. (2008), “How much do we learn from the estimation of DSGE models? A case study of
identi?cation issues in a new Keynesian business cycle model”, mimeo.
Litterman, R. (1986), “Forecasting with Bayesian vector autoregressions: ?ve years of
experience”, Journal of Business & Economic Statistics, Vol. 4 No. 1, pp. 25-38.
McCallum, B. and Nelson, E. (1999), “Performance of operational policy rules in an estimated
semi-classical structural model”, in Taylor, J.-B. (Ed.), Monetary Policy Rules, University of
Chicago Press, Chicago, IL, pp. 15-45.
Phillips, P. (1991a), “Bayesian routes and unit roots: De Rebus Prioribus Est Disputandum”,
Journal of Applied Econometrics, Vol. 6 No. 4, pp. 453-73.
Phillips, P. (1991b), “To criticize the critics: an objective Bayesian analysis of stochastic trends”,
Journal of Applied Econometrics, Vol. 6 No. 4, pp. 333-64.
Rotemberg, J. and Woodford, M. (1997), “An optimization-based econometric framework for the
evaluation of monetary policy”, in Bernanke, B. and Rotemberg, J. (Eds), Macroeconomics
Annual, MIT Press, Cambridge, MA, pp. 297-346.
The Fed
and the ECB
333
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Sahuc, J. and Smets, F. (2007), “Differences in interest rate policy at the ECB and the Fed:
an investigation with a medium-scale DSGE model”, Journal of Money, Credit and
Banking, Vol. 40 No. 2-3, pp. 505-21.
Sims, C. (1991), “Comment on ‘To Criticize the Critics’ by Peter c.b. Phillips”, Journal of Applied
Econometrics, Vol. 6 No. 4, pp. 423-34.
Sims, C. and Uhlig, H. (1991), “Understanding unit rooters: a helicopter tour”, Econometrica,
Vol. 59 No. 6, pp. 1591-9.
Smets, F. and Wouters, R. (2005), “Comparing shocks and frictions in US and EURO business
cycles: a Bayesian DSGE approach”, Journal of Applied Econometrics, Vol. 20, pp. 161-83.
Svensson, L. (1998), “Open-economy in?ation targeting”, CEPR Discussion Paper No. 1989.
Trichet, J. (2006), “Activism and alertness in monetary policy”, paper presented at the Conference
on Central Banks in the 21st Century, Bank of Spain, June 8.
Woodford, M. (2003), Interest and Prices, Foundations of a theory of Monetary Policy, Princeton
University Press, Princeton, NJ.
Further reading
A
´
lvarez, L., Dhyne, E., Hoeberichts, M., Kwapil, C., Le Bihan, H., Lu¨nnemann, P., Martins, F.,
Sabbatini, R., Stahl, H., Vermeulen, P. and Vilmunen, J. (2005), “Sticky prices in the euro
area: a summary of new micro evidence”, ECB Working Papers No. 563.
Bils, M. and Klenow, P. (2004), “Some evidence on the importance of sticky prices”, Journal of
Political Economy, Vol. 112, pp. 947-85.
Calmfors, L. and Drif?ll, J. (1988), “Bargaining structure, corporatism and macroeconomic
performance”, Economic Policy, Vol. 6 No. 3, pp. 13-62.
Ehrmann, M., Gambacorta, L., Martinez-Pages, J., Sevestre, P. and Worms, A. (2002), “Financial
systems and the role of banks in monetary policy transmission in the euro area”, Notes
d’Etudes et de Recherche de la Banque de France No. 93.
Hall, P. and Soskice, D. (2001), Varieties of Capitalism: The Institutional Foundations of
Comparative Advantage, Oxford University Press, New York, NY.
Kashyap, A. and Stein, J. (1997), “The role of banks in monetary policy: a survey with
implications for the European monetary union”, Federal Reserve Bank of Chicago
Economic Perspectives, Vol. 21 No. 5, pp. 2-18.
Nicoletti, G., Scarpetta, S. and Boylaud, O. (2000), “Summary indicators of product market
regulation with an extension to employment protection legislation”, OECD Working Paper
No. 226.
OECD (1997), “Economic performance and the structure of collective bargaining”, Employment
Outlook, Chapter 3, OECD, Paris, pp. 70-100.
Sellon, G. (2002), “The changing US ?nancial system: some implications for the monetary
transmission mechanism. Federal Reserve Bank of Kansas City”, Economic Review, Vol. 97
No. 1, pp. 5-35.
Siebert, H. (1997), “Labor market rigidities: at the root of unemployment in Europe”, The Journal
of Economic Perspectives, Vol. 11 No. 3, pp. 37-54.
JFEP
1,4
334
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Appendix
Figure A1.
Monetary policy
transmission channels in
euro area and the USA
Response of the long term interest rate to
a shock on the short term interest rate (EA model)
Response of the output gap to a shock on
the long term interest rate (EA model)
Response of the output gap to a shock on
the short term nom inal interest rate (US model)
Response of the output gap to a shock on
the short term real interest rate (US model)
Response of inflation to a shock on the
output gap (US model)
Response of the inflation to a shock on the
output gap (EA model)
0.35
0.3
0.25
0.2
0.15
0.05
–0.05
–0.15
–0.1
0
0.1
0.15
0.05
–0.05
–0.15
–0.1
0
0.1
0.35
0.3
0.25
0.2
0.15
0.05
0
0.1
0.35
0.3
0.25
0.2
0.15
0.05
–0.05
0
0.1
0.25
0.2
0.15
0.05
–0.05
–0.15
–0.25
–0.3
–0.2
–0.1
0
0.1
0 1 2 3 4 5 6 7 8 9 10 11 12 13
0 1 2 3 4 5 6 7 8 9 10 11 12 13
14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
0 1 2 3 4 5 6 7 8 9 10 11 12 13
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
14
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
–0.2
–0.25
–0.3
–0.35
0.15
0.05
–0.05
–0.15
–0.1
0
0.1
–0.2
–0.25
–0.3
–0.35
–0.4
0.4
The Fed
and the ECB
335
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Figure A2.
Monetary policy reactions
Response of the short term interest rate
to a one-unit shock on the output gap (EA model)
Response of the short term interest rate
to a one-unit shock on the output gap (US model)
Response of the short term interest rate
to a one-unit shock on inflation (EA model)
Response of the short term interest rate
to a one-unit shock on inflation (US model)
Response of the short term interest rate
to a one-unit idiosyncratic shock (EA model)
Response of the short term interest rate
to a one-unit idiosyncratic shock (US model)
2.5
1.5
0.5
0
0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10
0 1 2 3 4 5 6 7 8 9 10
0 1 2 3 4 5 6 7 8 9 10
11 12 13 14 15
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
0 1 2 3 4 5 6 7 8 9 10 11
2
1
1.8
1.6
1.4
1.2
0.8
0.6
0.4
0.2
0
1
1.6
1.4
1.2
0.8
0.6
0.4
0.2
0
1
1.6
1.4
1.2
0.8
0.6
0.4
0.2
0
1
1.6
1.4
1.2
0.8
0.6
0.4
0.2
0
1
1
0.8
0.6
0.4
0.2
–0.2
0
11
JFEP
1,4
336
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Corresponding author
Alexis Penot can be contacted at: [email protected]
1999:1-2005:4 Mean Variance Mean equality Variance equality
Actual EA interest rate 3.08 0.93 Yes No
Actual US interest rate 3.20 3.72 (SL ¼ 0.76) (SL ¼ 0.00)
Estimated EA interest rate 3.11 1.01 Yes No
Estimated US interest rate 3.20 3.71 (SL ¼ 0.83) (SL ¼ 0.00)
EA MP shocks 20.032 0.055 Yes Yes
US MP shocks 0.006 0.068 (SL ¼ 0.57) (SL ¼ 0.29)
Table AI.
Tests on effective interest
rate series
Figure A3.
Responses of variables
to idiosyncratic shocks
Response of the output gap to an
idiosyncratic shock (EA model)
Response of the output gap to an
idiosyncratic shock (US model)
Response of inflation to an
idiosyncratic shock (EA model)
Response of inflation to an
idiosyncratic shock (US model)
Response of wages to an
idiosyncratic shock (EA model)
Response of wages to an
idiosyncratic shock (US model)
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
–0.1
0
0.6
0.5
0.4
0.3
0.2
0.1
–0.1
–0.2
0
0.5
0.4
0.3
0.2
0.1
–0.1
0
0 1 2 3 4 5 6 7 8 9
0 1 2 3 4 5 6 7 8 9
10 11 12 13 0 1 2 3 4 5 6 7 8
0 1 2 3 4 5 6 7 8
0 1 2 3 4 5 6 7 8
9 10 11 12 13 14 15
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
0.35
0.25
0.05
–0.05
–0.1
0.15
0.1
0
0.2
0.05
–0.05
–0.1
0.15
0.1
0
0.2
0.3
1.4
1.2
1
0.8
0.6
0.4
0.2
0
The Fed
and the ECB
337
To purchase reprints of this article please e-mail: [email protected]
Or visit our web site for further details: www.emeraldinsight.com/reprints
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
doc_534462139.pdf