Equilibrium Unemployment Dynamics in a Panel of OECD Countries
| DOI | http://doi.org/10.1111/obes.12061 |
| Published date | 01 April 2015 |
| Date | 01 April 2015 |
| Author | Ragnar Nymoen,Victoria Sparrman |
164
©2014 The Department of Economics, University of Oxford and JohnWiley & Sons Ltd.
OXFORD BULLETIN OF ECONOMICSAND STATISTICS, 77, 2 (2015) 0305–9049
doi: 10.1111/obes.12061
Equilibrium Unemployment Dynamics in a Panel of
OECD Countries*
Ragnar Nymoen† and Victoria Sparrman‡
†Department of Economics, University of Oslo, Norway (e-mail: ragnar.nymoen@econ.uio.no)
‡Research Department, Statistics Norway (e-mail: victoria.sparrman@ssb.no)
Abstract
We focus on the equilibrium unemployment rate as a parameter implied by a dynamic
aggregate model of wage and price setting. The equilibrium unemployment rate depends
on institutional labour market institutions through mark-up coefficients. Compared with
existing studies, the resulting final equation for unemployment has a richer dynamic struc-
ture. The empirical investigation is conducted in a panel data framework and uses OECD
data up to 2012. We propose to extend the standard estimation method with time dum-
mies to control and capture the effects of common and national shocks by using impulse
indicator saturation (WG-IIS), which has not been previously used on panel data. WG-IIS
robustifies the estimators of the regression coefficients in the dynamic model, and it affects
the estimated equilibrium unemployment rates. We find that wage co-ordination stands
out as the most important institutional variable in our data set, but there is also evidence
pointing to the tax wedge and the degree of compensation in the unemployment insurance
system as drivers of equilibrium unemployment.
I. Introduction
The concept of equilibrium unemployment in the OECD area has been the subject of both
analytical and empirical research. One influential analytical approach, whichalso underlies
our research, combines a model of monopolistic price setting among firms with collec-
tive bargaining over the nominal wage level; see Layard, Nickell and Jackman (2005)
and Layard and Nickell (2011). Intuitively, when the system is not in a stationary situ-
ation, nominal wage and price adjustments constitute a wage–price spiral that leads to
increasing or falling inflation. According to Layard et al. (2005), the equilibrium for real
wages requires that unemployment becomes equal to the non-accelerating inflation rate
of unemployment (NAIRU). The equilibrium unemployment rate is not interpretable as a
*We wouldlike to thank two anonymous referees and the Editor, Christopher Bowdler. Previous versions of this
article have been presented at the Annual Meeting for NorwegianEconomists, 2010, Oslo, Norway, at the research
seminar at Statistics Norway and at the Cross-sectional Dependence Conference, 2013, Cambridge, UK.Thank you
also to Camilla KartoAbrahamsen, Ellen Marie Rossvoll, Erik Biørn, Neil Ericsson, Øyvind Eitrheim, Steinar Holden
and Bill Russell, for comments and discussions during the process of writing this article. The numerical results in
this article were obtained by use of OxMetrics 6/PcGive 13 and Stata 12.
JEL Classification numbers: C22, C23, C26, C51, E02, E11, E24.
Equilibrium unemployment dynamics in OECD countries 165
completely invariantparameter that is pinned down once and for all by autonomous market
structures. Specifically, it can vary with changing labour market institutions.
We attempt to bridge the gap between the formal, but static, theoretical framework,
which is common to several empirical studies, Belot and van Ours (2001), Bassanini
and Duval (2006), Nickell, Nunziata and Ochel (2005) and Blanchard and Wolfers (2000)
among others, and the dynamic specification of the estimated models. Our theoretical model
leads to a cointegrated vector autoregression (VAR) with three endogenous variables; the
rate of unemployment, the real exchange rate and the wage share.The VAR-based equilib-
rium rate of unemployment is also a NAIRU because it is determined jointly with a constant
rate of inflation. Our main parameter of interest is the equilibrium rate of unemployment,
rather than all the other parameters of the full system. We therefore base our empirical model
on the main aspects of a final equation for the unemploymentrate that is implied by the VAR.
The theoretical model gives some guidelines for the empirical specification that differ
from those in the conventional empirical literature. First, the autoregressive order of the
final equation is shown to be three, whereas the custom has been to impose first-order
dynamics. The higher-order dynamics can be verified or refuted by econometric testing,
which we do in the empirical parts of the article. Second, the expected sign and magnitudes
of the autoregressive coefficients stem from the theoretical model. Third, the theoretical
model also has implications for the error term, which is affected by foreign price and
technology shocks, which,when not controlled for, will be disturbances. We therefore allow
for a heteroscedastic and autocorrelated error term in the empirical part of the analysis. Our
dynamic theory model also predicts that changes in institutional features will affect un-
employment rates gradually, not by sudden shifts, which is realistic and in line with earlier
studies.
One of the contributions of this article is to estimate the equilibrium level of unemploy-
ment for each of twenty OECD countries from a panel dataset, dependent on the current
level of the labour market institutions. The estimation results, based on data for the period
1960–2012, show that different levels of wage-bargaining co-ordination, the generosity of
the unemployment insurance system and tax wedges may have contributed towards differ-
ent levels of equilibrium unemployment.We find that the equilibrium unemployment rates
range from just above three percentage points to 13 percentage points.
The result is based on an estimation method that controls for all major shocks that have
hit the individual unemploymentrates in the sample. It includes the onset of the credit crisis
and the transformation of the financial crisis into an international job crisis. To account for
both national and common shocks, we have augmented the model using country-specific
indicator variables of years that represent temporary location shifts in the mean of the
individual unemployment rates (see Doornik (2009) and Johansen and Nielsen (2009)).
This is a first application of so-called impulse indicator saturation to an econometric panel
data model. We suggest this as a way of obtaining robust estimators of the regressions
coefficients (and the standard errors) of the theoretical explanatory variables in the model
and as a means to investigate the effect of shocks on the equilibrium unemployment rates.
We consider that our findings are in line with earlier panel data analysis in which equi-
librium unemployment is determined by labour market institutions. Specifically, Nickell
et al. (2005) find a strong role for institutional variables in explaining the increase in
unemployment rates in Europe over the their sample period, 1960–95. This corresponds to
©2014 The Department of Economics, University of Oxford and JohnWiley & Sons Ltd
To continue reading
Request your trial