Decomposing Beveridge Curve Dynamics By Correlated Unobserved Components
| Date | 01 December 2016 |
| Published date | 01 December 2016 |
| Author | Sabine Klinger,Enzo Weber |
| DOI | http://doi.org/10.1111/obes.12135 |
877
©2016 The Department of Economics, University of Oxford and JohnWiley & Sons Ltd.
OXFORD BULLETIN OF ECONOMICSAND STATISTICS, 78, 6 (2016) 0305–9049
doi: 10.1111/obes.12135
Decomposing Beveridge Curve Dynamics By
Correlated Unobserved Components*
Sabine Klinger†,‡ and Enzo Weber†,‡,§
†Institute for Employment Research, Regensburger Straße 104 90478, Nuremberg, Germany
(e-mail: sabine.klinger@iab.de, enzo.weber@iab.de)
‡University of Regensburg, Universit¨atsstraße 31 93053, Regensburg, Germany
§IOS, Landshuter Str. 493047, Regensburg, Germany
Abstract
Our paper provides a consistent framework to study the structural or cyclical nature of
Beveridge curve (BC) dynamics: We connect equilibrium unemployment theory to a flexi-
ble multivariate unobserved components model. We disentangle permanent and transitory
components of all series determining the BC and its position. Cointegration and identifica-
tion are addressed. The German curve is an ideal illustration as reforms of the institutional
setting and the Great Recession were accompanied by a remarkable labour market devel-
opment. We find an extraordinary increase in trend matching efficiency after the reforms,
which testifies to a permanent improvement. Matching efficiency accounts for about half
of the BC´s inward shift. As tightness also increased,a persistent upward movement masked
the inward shift.
I. Introduction
The Beveridge curve (BC) is well known as the downward sloping relationship between
vacancies and unemployment (Elsby, Michaels and Ratner, 2015a for an overview). Fol-
lowing the influential work by Abraham and Katz (1986) and Blanchard and Diamond
(1989), the traditional interpretation says that shifts of the curve indicate structural change
which worsens or improves the functionality of the labour market. Movements along the
curve, in contrast, reflect cyclical variation. The labour market development during and
after the Great Recession revitalized an interest in BC analyses (e.g. Bouvet, 2012, Daly
et al., 2012, Pissarides, 2013). Nonetheless, the dynamics of structural change or cyclicality
were hardly investigated with respect to the BC.
JEL Classification numbers: C32, E24, E32, J2, J69
*We are grateful to two anonymousreferees, Daniel Heuer mann, KatjaWolf as well as participants of the 2013
EALE conference, the 2014 conference of the German Economic Association, the 7th ifo Workshop on Macro-
economics and the Business Cycle, the 2013 Econometric Seminar of IAB and the University of Regensburg, the
Economic Seminar at the University of Leipzig and the IAB Scientific Advisory Council for insightful comments.
We are indebted toThomas Rothe and J ¨urgen Wiemers for data provisionand the programming of an identification
check.
878 Bulletin
However, notable arguments in the literature trigger deviations from the traditional BC
pattern. Davis, Fabermanand Haltiwanger (2013), Fujita and Ramey (2009), B ¨orsch-Supan
(1991) and Kosfeld, Dreger and Eckey (2008) theoretically and empirically motivated that
shifts of the BC could occur for cyclical reasons. Their focus lies on matching efficiency –
the parameter that indicates functionality of the labour market on the one hand but varies
with search intensity over the cycle on the other. As matching efficiency is unobservable,
investigating its dynamics in greater detail reasonably relies on unobserved components
(UCs) models (as in Dixon, Lim and Freebairn, 2014, Sedl´aˇcek, 2014).
Our paper augments this research and develops an analytical framework. We build on
the coherent BC theory with two labour market states (employment, unemployment) and
construct a new multivariate UCs model for the labour market on the fundamentals of
Morley, Nelson and Zivot (2003). This connection of economic theory and econometric
flexibility is essential: First, the theory guides us to a comprehensive modelling of eco-
nomic equilibrium and cointegration – we integrate the long-run equilibrium exactly for
the permanent components of the series forming the steady state BC. This also avoids
the usual problem that observed unemployment might be a bad approximation for equi-
librium unemployment (Elsby, Hobijn and Sahin, 2013). Second, the flexible UC model
lets the data decide which BC outcomes rest upon structural determinants and which on
cyclical – we disentangle each of the constituents (unemployment, vacancies) and each
of the shifting parameters (matching efficiency, separation rate, employment) into their
unobserved permanent and transitory components (stochastic trends and stationary cy-
cles). Third, the model represents complex interactions in the labour market and provides
stringent economic interpretations for the correlations between the trend and cycle shocks.
On this basis, we can consolidate our understanding of the labour market and evaluate
the design of institutions and policies from a macroeconomic perspective.With that aim, we
investigate the empirical relevanceof the theoretical deter minants of BC dynamics. So far,
we are aware of variance decomposition analyses and DSGE calibrations on the relevance
of certain labour market parameters for unemployment (e.g. Hertweck and Sigrist, 2015,
Jung and Kuhn, 2014). We add to the literature as we build on consistent theory-based
estimates to divide BC dynamics into their structural and cyclical nature, control for their
interaction and see how permanent and cyclical processes interfere in the development of
the BC.
We apply our framework to German data. Following severe labour market reforms, the
German BC has shifted inwards since the end of 2006, with only moderate movements
along the curve during the Great Recession and the adjacent recovery. This development
contrasts with pre-reform descriptions of the German labour market as being sclerotic
with high persistent unemployment. It also contrasts with the situation in the U.S. (Benati
and Lubik, 2014, Lubik, 2013, Pissarides, 2013, Daly et al., 2012, Sala, S¨oderstr¨om and
Trigari, 2013). Therefore, the German BC is especially worth to focus on in a period where
changing labour market institutions and atypical economic conditions may haveinterfered.
Our analysis underlines the requirement to extend the traditional BC interpretation.
The observation of a shift or movement along the curve does not necessarily reveal the
true nature of the change on the labour market. With respect to the German setting, the
model detects an outstanding upturn of matching efficiency in the aftermath of the labour
market reforms. While this has been proposed before using dummy variables (Fahr and
©2016 The Department of Economics, University of Oxford and JohnWiley & Sons Ltd
To continue reading
Request your trial