Professional Documents
Culture Documents
Resear
h Interests
Ma
roe
onomi
s and Monetary E
onomi
s, Labor E
onomi
s, Applied E
ono-
metri
s.
Current Position
sin e 09/2006 E onomist, Monetary Poli y Strategy Division, European Central Bank.
Edu
ation
2001 2006 Dr. rer. pol (summa
um laude), Goethe University Frankfurt, Germany.
Thesis title: The Role of Labor Markets in Monetary DSGE Models and Ro-
bust Monetary Poli
y.
2000 2001 M.S
. Finan
ial E
onomi
s and E
onometri
s (with distin
tion), University of
Essex, UK.
1999 2000 Study of E
onomi
s. University of Kiel and University of Bonn, Germany.
1997 1999 Vordiplom (B.S. E
onomi
s), University of Kiel, grade 1.0 (s
ale from 0.7-5.0).
Publi
ations
2006 Value-at-Risk Predi
tion: A Comparison of Alternative Strategies , Journal of
Finan
ial E
onometri
s 4(1), p. 5389, with S. Mittnik and M. Paolella.
2005 Optimal Monetary Poli
y Rules for the Euro Area , Journal of Common Mar-
ket Studies, 43(3), p. 507537, with A. Dieppe and P. M
Adam.
Working Papers
2007 Resus
itating the Wage Channel in Models with Unemployment Flu
tuations,
with K. Christoel. Prepared for Carnegie Ro
hester Conferen
e, Nov. 2007.
2007 The (Un)Importan
e of Unemployment Flu
tuations for Welfare, mimeo. Eu-
ropean Central Bank and University of Amsterdam, with P. Jung.
2007 Is the New Keynesian Phillips Curve Flat?, European Central Bank Working
Paper No. 809, with G. Mueller and S. Stoelting.
2006 Real Pri
e and Wage Rigidities in a Model with Mat
hing Fri
tions, European
Central Bank Working Paper No. 720.
2006 Identifying the Role of Labor Markets for Monetary Poli
y in an Estimated
DSGE Model, ECB Working Paper No. 635, with K. Christoel and T. Linzert.
2005 Insuran
e Poli
ies for Monetary Poli
y in the Euro Area , ECB Working Paper
No. 480, with V. Wieland.
2007 Carnegie Ro
hester Conferen
e Labor Markets, Ma
roe
onomi
Flu
tuations,
and Monetary Poli
y, Pittsburgh; Annual Meeting of the Verein für So
ialpoli-
tik, Muni
h; Conferen
e New Challenges in E
onomi
Resear
h, Louvain-la-
Neuve, Belgium; European E
onomi
Asso
iation, Budapest; Conferen
e Us-
ing E
onomi
Models to Make Poli
y Re
ommendations, San Sebastian, Spain;
North Ameri
an Summer Meetings of the E
onometri
So
iety, Duke Univer-
sity; Cze
h National Bank, Prague.
2006 International Resear
h Forum on Monetary Poli
y, Federal Reserve Board,
Washington D.C.; European E
onomi
Asso
iation, Vienna; So
iety for E
o-
nomi
Dynami
s, Van
ouver; So
iety for Computational E
onomi
s, Cyprus.
2005 Deuts
he Bundesbank, Frankfurt; E
onometri
So
iety World Congress, Lon-
don; So
iety for Computational E
onomi
s, Washington D.C.
2004 Spring Meeting of Young E
onomists, Warsaw.
2004 2006 Resear
h and tea
hing assistant, Goethe University, Frankfurt.
(tea
hing Ph.D. and third year ma
roe
onomi
s
lasses)
2005/Jan-Jun Resear
h visitor, Deuts
he Bundesbank, Frankfurt.
2002 2003 Intern (Graduate Resear
h Program), European Central Bank (4 months).
1998 2000 Tea
hing assistant (undergrad. linear algebra and statisti
s), University of Kiel.
Referee A tivity
Journal of E
onomi
Dynami
s and Control, Berkeley Press Ele
troni
Jour-
nals, S
ottish Journal of Politi
al E
onomy.
Languages
Given the growing need for managing nan
ial risk, risk predi
tion plays an in
reasing role
in banking and nan
e. In this study we
ompare the out-of-sample performan
e of existing
methods and some new models for predi
ting value-at-risk (VaR) in a univariate
ontext. Us-
ingmore than 30 years of the daily return data on the NASDAQ Composite Index, we nd that
most approa
hes perform inadequately, although several models are a
eptable under
urrent
regulatory assessment rules for model adequa
y. A hybrid method,
ombining a heavy-tailed
generalized autoregressive
onditionally heteroskedasti
(GARCH) lter with an extreme value
theory-based approa
h, performs best overall,
losely followed by a variant on a ltered histor-
i
al simulation, and a new model based on heteroskedasti
mixture distributions. Conditional
autoregressive VaR (CAViaR) models perform inadequately, though an extension to a parti
ular
CAViaR model is shown to outperform the others.
Optimal Monetary Poli
y Rules for the Euro Area (2005), Journal of Common Market
Studies, 2005, 43(3), p. 507537, with A. Dieppe and P. M
Adam.
In this arti
le, we analyse the
ondu
t of optimal monetary poli
y for the new euro area. The
aggregate euro area e
onomy is modelled to have relatively sluggish adjustment properties and a
private se
tor with mainly ba
kward-looking expe
tations. In this e
onomy, we assume that the
entral bank sear
hes for its best-performing monetary poli
y rule, e.g. the optimal weight to give
to ination stabilization
ompared to that of output, the optimal degree of forward-looking in
the planning horizon, and so on. We rst nd that the optimal degree of gradualism in interest
rate-setting needs only be relatively mild and that the
entral bank should in
orporate new
information qui
kly into poli
y-making. Se
ond, there is substantial gain from implementing and
ommuni
ating quite forward-looking poli
ies. The optimal fore
ast horizon for ination ranges
around six quarters. In
ontrast to deliberately simple rule-based poli
y re
ommendations, fully
optimal poli
y is a
ompli
ated response to many dierent e
onomi
indi
ators. With regard
to this we nd, third, that optimal poli
y should be based on a broad information set, even if
the resulting poli
y framework is hard to
ommuni
ate to the outside world. Thus, the arti
le
ontributes to the debate on optimal monetary poli
y for the euro area, as well as to the
ondu
t
of monetary poli
y in fa
e of substantial persisten
e in the transmission me
hanism.
Keith KUESTER 4
Higher wages all else equal translate into higher ination. More rigid wage adjustment implies a
slower adjustment of pri
es. This view of the wage
hannel is deeply entren
hed in many
entral
banks' view of the labor market. In this paper, we present a model with equilibrium unem-
ployment whi
h (a) features su
h a proper wage
hannel, whi
h (b) reprodu
es the u
tuations
of unemployment over the business
y
le and whi
h (
) implies a reasonable (low) elasti
ity of
steady state unemployment with respe
t to
hanges in benets. Sear
h and mat
hing models
featuring right-to-manage a
ount for (a). We show that a small modi
ation of these, namely
the introdu
tion of a period-by-period xed
ost asso
iated with jobs, greatly magnies prot
u
tuations for any given degree of wage u
tuations, whi
h gives (b). While the
alibration still
relies on low prots, it does not ne
essarily demand a small gap between the value of working
and the value of unemployment for the worker, whi
h yields (
).
The (Un)Importan e of Unemployment Flu tuations for Welfare (2007), with P. Jung.
This paper develops a New Keynesian model with mat
hing fri
tions and liquidity-
onstrained
onsumers. We estimate the model on six ma
roe
onomi
time-series using Bayesian te
hniques
and provide an estimate for the welfare
ost of endogenous unemployment risk. First, we nd
that Shimer's (2005) initial pessimism with respe
t to the working of ma
roe
onomi
labor
market sear
h and mat
hing models is unwarranted. The model repli
ates the entire time-path
of unemployment rates almost perfe
tly without having used this information in the estimation
pro
ess. In terms of welfare we nd, se
ond, that so
iety as a whole would pay at least 0.2of
steady state
onsumption to avoid business
y
le u
tuations
ompletely. These
osts rise for
workers who have no means to self-insure themselves against u
tuations and who
an only
repla
e part of their wage in
ome when unemployed. At the extreme, with a repla
ement rate
of only 10to avoid business
y
le risk. Of the full welfare
ost of business
y
les for liquidity-
onstrained workers about a quarter to a half is due to the
o-movement of unemployment risk
with the business
y
le.
Is the New Keynesian Phillips Curve at? (2007), with G. Müller and S. Stölting.
Ma
roe
onomi
data suggest that the New Keynesian Phillips
urve is quite at - despite mi
roe-
onomi
eviden
e implying frequent pri
e adjustments. While real rigidities may help to a
ount
for the
oni
ting eviden
e, we propose an alternative explanation: if pri
e markup/
ost-push
sho
ks are persistent and negatively
orrelated with the labor share, the latter being a widely
used measure for marginal
osts, the estimated pass-through of measured marginal
osts into
ination is limited, even if pri
es are fairly exible. Using a standard New Keynesian model, we
show that the GMM approa
h to the New Keynesian Phillips
urve leads to in
onsistent and
upward biased estimates if
ost-push sho
ks indeed are persistent. Monte Carlo experiments
suggest that the bias is quite sizeable: we nd average pri
e durations estimated as high as 12
quarters, when the true value is about 2 quarters. Moreover, alternative estimators appear to be
biased as well, while standard diagnosti
tests fail to signal a misspe
i
ation of the model.
Keith KUESTER 5
Real Pri e and Wage Rigidities in a Model with Mat hing Fri tions (2007).
This paper in
orporates a sear
h and mat
hing framework into the (Calvo-staggered) pri
e setting
se
tor. Mat
hing fri
tions lead pri
e setting rms to negotiate wage rates with their employees.
The negotiation of wages substantially in
reases strategi
omplementarity in pri
e setting among
suppliers of dierentiated goods. This leads to an in
rease in real rigidities as in Woodford (2003)
whi
h redu
es implied pri
e durations for a given estimate of the slope of the Phillips
urve thus
re
on
iling ma
ro- and mi
ro-eviden
e. A novel result of the paper is that the same fa
tors whi
h
indu
e smooth ination also dampen the adjustment of wages in response to sho
ks. In the sear
h
and mat
hing framework this is key for explaining the highly responsive nature of va
an
ies in
the data. Another interesting nding for the Phillips
urve is that ination is not only driven by
an output gap but also by an employment gap - a feature usually negle
ted in empiri
al resear
h.
The modied model mat
hes impulse responses of an SVAR for post Vol
ker-disination U.S.
data very well.
Identifying the Role of Labor Markets for Monetary Poli
y in an Estimated DSGE
Model (2006), with K. Christoel and T. Linzert.
Insuran e Poli ies for Monetary Poli y in the Euro Area (2005), with V. Wieland.
In this paper, we examine the
ost of insuran
e against model un
ertainty for the Euro area
onsidering four alternative referen
e models, all of whi
h are used for poli
y-analysis at the
ECB. We nd that maximal insuran
e a
ross this model range in terms of a Minimax poli
y
omes at moderate
osts in terms of lower expe
ted performan
e. We extra
t priors that would
rationalize the Minimax poli
y from a Bayesian perspe
tive. These priors indi
ate that full
insuran
e is strongly oriented towards the model with highest baseline losses. Furthermore, this
poli
y is not as tolerant towards small perturbations of poli
y parameters as the Bayesian poli
y
rule. We propose to strike a
ompromise and use preferen
es for poli
y design that allow for
intermediate degrees of ambiguity-aversion. These preferen
es allow the spe
i
ation of priors
but also give extra weight to the worst un
ertain out
omes in a given
ontext.