ILO Research Paper No. 12
Labour market reforms in the
Euro area:
A DSGE approach
Giorgio Presidente
December 2015
ILO Research Paper No. 12
Labour market reforms in the Euro area:
A DSGE approach
Giorgio Presidente
December 2015
International Labour Office
Copyright © International Labour Office 2015
First published 2015
Publications of the International Labour Office enjoy copyright under Protocol 2 of the Universal Copyright
Convention. Nevertheless, short excerpts from them may be reproduced without authorization, on condition that
the source is indicated. For rights of reproduction or translation, application should be made to ILO Publications
(Rights and Permissions), International Labour Office, CH - 1211 Geneva 22 (Switzerland) or by email:
pubdroit@ilo.org. The International Labour Office welcomes such applications.
Libraries, institutions and other users registered with reproduction rights organisations may make copies in
accordance with the licences issued to them for this purpose. Visit http:/ifrro.org to find the reproduction rights
organisation in your country.
ISSN 2306-0875 (web pdf)
The designations employed in ILO publications, which are in conformity with United Nations practice, and the
presentation of material therein do not imply the expression of any opinion whatsoever on the part of the
International Labour Office concerning the legal status of any country, area or territory or of its authorities, or
concerning the delimitation of its frontiers.
The responsibility for opinions expressed in signed articles, studies and other contributions rests solely with
their authors, and publication does not constitute an endorsement by the International Labour Office of the
opinions expressed in them. Reference to names of firms and commercial products and processes does not imply
their endorsement by the International Labour Office, and any failure to mention a particular firm, commercial
product or process is not a sign of disapproval.
ILO publication and electronic products can be obtained through major booksellers or ILO local offices in many
countries, or direct from ILO Publications, International Labour Office, P.O. Box 6, CH-1211 Geneva 22
(Switzerland). Catalogues or lists of new publications are available free of charge from the above address, or by
email: pubvente@ilo.org. Visit our site: http://www.ilo.org/pubins
Table of contents
Acknowledgements
Abstract
1 Introduction ................................................................................................................................... 1
2 The model ...................................................................................................................................... 2
2.1 Matching................................................................................................................................ 3
2.2 Intermediate goods firm......................................................................................................... 3
2.3 Final goods firm..................................................................................................................... 4
2.4 Households ............................................................................................................................. 5
2.5 Government and resource constraint...................................................................................... 6
2.6 Determination of Wages......................................................................................................... 6
3 Long run impact of labour market reforms .................................................................................... 7
4 Short run impact of labour market reforms ................................................................................... 9
5 Sources of fluctuations in the Euro area ........................................................................................ 11
5.1 Estimation: Data and priors.................................................................................................. 11
5.2 Estimation results .................................................................................................................. 13
5.3 Decomposing the volatility of unemployment, output and inflation in the Euro area .......... 18
6 Conclusions .................................................................................................................................... 21
References............................................................................................................................................ 22
Acknowledgements
This paper benefited from the comments of three anonymous referees, and of the participants at the
Macroeconomic Workshop at Toulouse School of Economics, the seminars at the Graduate Institute of
International Studies, and the Research Department, International Labour Organization.
Abstract
Empirical work on the impacts of labour market institutions has produced mixed results. Much of this
literature is based on reduced form regressions that are subject to severe econometric and measurement
issues. This paper develops a framework to study the impact of labour market institutions in the
context of a DSGE model. The advantage of using a DSGE model is that one can observe the general
equilibrium outcomes of truly exogenous shifts in labour market policy. In addition, this class of models
are flexible, can be easily estimated and one can undertake policy simulations and counterfactual exercises.
After inspecting the short and long run response of key variables to several labour market reforms, an
application to the Euro area reveals that between 1970 and 2003, changes in labour market institutions
had only a limited impact on the volatility of output, inflation and unemployment. These results
stand in contrast to theories attributing to excessive regulation for the rise and persistence of European
unemployment. In addition, they suggest that labour market frictions are at most of marginal interest
for central banks.
Keywords: Unemployment, labour market institutions, DSGE, Bayesian estimation
JEL classification: E240; J20
Labour market reforms in the Euro area: A DSGE approach 1
1 Introduction
There are several issues about labour market institutions on which theory and evidence are inconclusive.
For example, what is the impact of an increase in firing costs on unemployment? If increasing the cost of
dismissal provides an incentive for firms to retain workers even during difficult times, it might also make
them reluctant in hiring because of the scarce flexibility of adjusting employment.
Lack of agreement on the causal links between regulation and market performance fuels the policy debate.
Several European countries have recently undertaken labour market reforms in order to reduce the high
unemployment rates triggered by the financial crisis, but which particular institution should be reformed,
whether it should aim at making the labour market more flexible, or increase social protection are issues
still to be debated.
The point is that much of the existing literature on labour market institutions is based on reduced form
equations. These studies often regress the unemployment rate on some index of labour market institutions
and use proxies for shocks that are likely to impact the labour market. For example, Nickell and Layard
(1999) discuss extensively the impact of several labour market institutions on output and unemployment.
Based on a panel of OECD countries, they conclude that unions and the social security system are the
only institutions having a significant impact on the economy. Unfortunately, their evidence is far from
conclusive, since one can easily find papers contradicting their results.1
Another influential study on labour market institutions is by Blanchard and Wolfers (2000). They propose
a simple theory of unemployment based on the interaction between structural shocks and labour market
institutions: following an adverse shock, countries with “rigid” labour markets will experience higher
and more persistent unemployment. As simple and appealing as it stands, their work has an important
role in the literature, because it conciliates theories attributing shocks to the heterogeneous evolution
of unemployment in the OECD, with those pointing to excessive labour market regulation. But how
robust can their evidence be considered? Blanchard (2007) puts it succinctly: “... asking these panel
data regressions to tell us conclusively about causal effects of institutions, shocks, and interactions of
shocks and institutions, on unemployment is beyond what they can deliver. Causality is next to impossible
to establish, as many institutional changes are triggered by labour market developments.” Furthermore,
data on labour market institutions consist of indexes that are supposed to summarize in a single number
complex regulations and thus they are unlikely to deliver all the relevant information. On a more general
ground, reduced form regressions cannot take into account general equilibrium effects put in place by
changes of institutions.
For these reasons, this paper follows a different approach. It develops an otherwise standard DSGE model
with a rich characterization of the labour market, modeling explicitly key labour market institutions that
are believed to have a significant impact on the economy. These are: i) firing costs; ii) active labour
market policies; iii) labour taxes; iv) union strength, and v) unemployment benefits. This paper departs
from most of the existing literature in modeling labour market institutions as stochastic process, rather
than simply as parameters. It does so for two main reasons. First, by treating institutions as shocks, it is
possible to compute impulse response functions to truly exogenous changes in labour market regulation.
As emphasized by Blanchard (2007), this is not possible by adopting a reduced form approach. Second,
using a DSGE model allows to shed some light on the general equilibrium effects triggered by changes in
institutions, effects which cannot be disentangled in standard regression analysis.
1 A survey of the literature on labour market institutions is beyond the scope of this paper. However, see for example
Lazear (1990) and Bentolila and Bertola (1990) for contradicting results.
2 ILO Research Paper No. 12
The analysis conducted in this paper is complementary to the work of Zanetti (2011), which studies
the propagation of shocks in labour markets characterized by firing costs and unemployment benefits,
and the long run behavior of the model that follows their change. However, he focuses on firing costs
and unemployment benefits, without considering the impact of several other institutions, such as labour
taxes and unions. Moreover, by treating firing costs and unemployment benefits as parameters, he is
not able to assess the direct impact of institutions on unemployment and other variables. This can be
relevant for the political economics of the model. Even if not explicitly considered here, such an exercise
might provide insights on the short and long term cost and benefits associated to a reform, and so on
its political viability. Another issue that is possible to explore within this setting is how the qualitative
and quantitative response of unemployment depends on the persistence of the shocks to labour market
institutions. In some sense, the persistence of such shocks can be interpreted as the credibility of a reform.
For example, a very transitory shock to firing costs it is roughly equivalent to a not-credible reform of
the employment protection system. Since agents hold rational expectations, they know that the firing
costs will return to their pre-shock (pre-reform) value in a few periods only.
Christoffel, Kuester and Linzert (2009) in their study of optimal monetary policy in an estimated DSGE
model consider several labour market institutions in a stylized way. For example, they consider exogenous
shocks to separation rate as a proxy for firing costs. In contrast, this paper models explicitly the process
of endogenous separation and how it is affected by firing costs, leading to a richer and more realistic
analysis. The impact of search frictions on inflation has been extensively studied by Krause, Lopez-Salido
and Lubik (2008a, 2008b). However, they ignore labour market institutions which potentially have an
impact on inflation. An improvement in this direction is made by Thomas and Zanetti (2009), which
explicitly considers the effect of reforming employment protection legislation and the unemployment
insurance system on inflation. However, they ignore a number of other institutions and do not address
the direct impact of changes in regulation on the labour market.
The framework developed in this paper allows to conduct several experiments. The theory is developed
in Section 2. Section 3 uses a calibrated version of the model to study how changes in labour market
institutions affect the steady state. Firing costs and labour taxes increase output and decrease unemployment
in the long run, while unemployment benefits and higher bargaining power of workers have the
opposite effect. We then treat labour market institutions as stochastic processes. Section 4 looks at the
impulse response functions of the model after temporary and persistent shocks. The main finding is that
whether the change in policy is expected to long-last, or be in place for a few periods only, it makes a
large difference in terms of unemployment. Section 5 estimates the model on the Euro area with Bayesian
methods, assesses the behavior of the model and performs a variance decomposition. It finds that only
a fraction of unemployment, output and inflation’s fluctuation is due to labour market reforms, while
the variance of output and inflation are almost exclusively explained by “standard” structural shocks.
Section 6 concludes.
2 The model
The model used in this paper is essentially the same as in Thomas and Zanetti (2009), with only a minor
difference discussed below. Their model has the advantage of being very tractable, but at the same time
rich enough to study the impact of the selected labour market institutions.2
The economy is populated
by four types of agents: a large household, a competitive intermediate goods producer, a monopolistically
competitive final goods producer, and the government. The government is assumed to undertake both
2 In particular, exhibiting endogenous separation allows to study the impact of firing costs.
Labour market reforms in the Euro area: A DSGE approach 3
fiscal and monetary policy. The proceeds of taxation finance government expenditure and unemployment
benefits. The household is composed by individuals that consume the final good and can be either
employed or unemployed. Each worker has an idiosyncratic productivity level and earns a wage that is a
function of such productivity. An unemployed member of the household receives unemployment benefits.
In the model there are two types of taxes: one is paid by the intermediate goods firm (payroll tax) and
the other by the worker (income tax). The intermediate goods firm uses only labour as a production
factor and has a reservation productivity level below which it finds profitable to fire the worker and to
pay the firing cost. Wages are collectively bargained between unions and firms. To what extent unions
are able to appropriate the economic surplus deriving from each job match depends on its institutional
power. The intermediate goods are purchased by the final goods producer, which repackages them and
resells them to the household.3
The final goods market is characterized by monopolistic competition and
price stickiness. The labour market is subject to search frictions.
The timing of the model is as follows. At the beginning of each period, agents learn about the realization
of all shocks. Then, intermediate producers chooses i) how many vacancies need to be posted to achieve
the optimal number of workers, and ii) the minimum productivity threshold to be imposed on the
workers. Once hiring has taken place, idiosyncratic productivity realizes and layoff, determination
of wages and production take place. Before the beginning of the next period, there is an exogenous
separation happening with a constant probability. Given the structure of the model, some members
of the household might be initially hired, but then displaced within the same period because it is not
productive enough. As a consequence, they will not engage in production and they will not earn the
production wage. Rather, they will join the pool of searching workers and get unemployment insurance.
There are also other possibilities, for example Zanetti (2011) assumes that all new matches begins with
the highest idiosyncratic productivity level. Which one of the two is more appropriate in this context
is probably a matter of choice. However, given that precarious contracts, short-term employment and
unpaid internships are widely used in Europe, this paper adopts the former view.
2.1 Matching
Matching between vacancies, vt, and the stock of searching workers at the beginning of period t, ut, takes
place through a constant returns to scale function.
mt = σm
t uγ
t v1−γ
t
where γ ∈ (0, 1) and σm
t describes the efficiency of matching and we will interpret it as the effectiveness
of active labour market policies. Given the matching function we define, respectively, the probability of
filling a vacancy for the firm and the probability of finding a job for the worker,
qt =
mt
vt
st =
mt
ut
2.2 Intermediate goods firm
There is a representative intermediate goods firm employing nt workers. Workers have idiosyncratic
productivity levels, not known by the firm at the time of hiring. Idiosyncratic productivity is i.i.d. across
time and workers, and it is governed by the cumulative distribution function G(z) with p.d.f. G (z) = g(z).
Firms produce output according to,
yt = Atnt
zR
t
z
g(z)
1 − G(zR
t )
dz
3 Having two separated typologies of firms is a well known artifact which makes the model more tractable.
4 ILO Research Paper No. 12
where At is an aggregate productivity shock. The variable zR
t is a threshold productivity level below
which the firm finds profitable to destroy a job match. Employment from the perspective of the firm
evolves according to,
nt = 1 − G(zR
t ) (ρnt−1 + qtvt)
where, 1 − ρ is an exogenous probability of separation. It can be shown that the Bellman equation
describing the value of a job match with productivity z for the firm is given by
Jt(z) = pw
t Atz − (1 + τf
t )wt(z) + ρβEtΛt,t+1
zR
t+1
Jt+1(z )g(z ) dz − G(zR
t+1)Ft+1
This equation says that in the current period, the value of each worker is given by its contribution to
production, where pw
t represents the real price of the intermediate good (but also real marginal costs for
the final goods producer), minus the wage paid net of payroll taxes. In the next period, if the match is not
exogenously destroyed, (event occurring with probability ρ) two things can happen: if the idiosyncratic
productivity z is again above the threshold, the match will deliver a future value Jt+1(z ), otherwise the
firm will fire him and pay the firing cost, F (event occurring with probability G(zR
)). The term βΛt,t+1
represents the discount factor of the firm. Similarly, the Bellman equation describing the value of posting
a vacancy for the firm is
Vt = −χ − qt + qt
zR
t
Jt(z)g(z) dz − G(zR
t )Ft + (1 − qt)βEtΛt,t+1Vt+1
In this equation, χ represents the cost of posting a vacancy. The value of posting a vacancy is the
expected value of a job match times the probability of filling the vacancy, plus the value of such a vacancy
in the next period times the probability of not filling it today. We assume that in the intermediate goods
market there is free entry. Consequently, firms will post vacancies until the value of doing it becomes
zero. Holding in each period, this fact delivers the condition Vt = Vt+1 = 0. As a result, by combining (1)
and (2), we find the hiring condition
χ
qt
=
zR
t
pw
t Atz − (1 + τf
t )wt(z) g(z) dz − G zR
t Ft + ρβ 1 − G zR
t EtΛt,t+1
χ
qt+1
For the firm, the economic surplus of holding a productive worker in the workforce (and thus, not firing
him) is given by Jt(z) − (−Ft). The threshold productivity zR
t is found by noticing that a firm will hold
the worker and not fire him until when the surplus of the match yields zero rent, Jt(zR
t ) + Ft = 0, that
implies the following separation condition
pw
t AtzR
t − (1 + τf
t )wt(zR
t ) + ρβEtΛt,t+1
χ
qt+1
= −Ft
This equation states that the firm is indifferent between holding and letting produce a worker with
productivity zR
t and firing him paying the firing cost.
2.3 Final goods firm
There is a representative monopolistically competitive retailer that purchases intermediate goods and
repackages, and resells them as final goods to the household. In addition, they set prices on a staggered
basis as in Calvo (1983). In each period, the firm has an exogenous probability of resetting its price
equal to 1 − λ.4
In case of resetting, the firm chooses the price that maximizes expected discounted
4 Notice that 1
1−λ
is the average time between each reset price.
Labour market reforms in the Euro area: A DSGE approach 5
profits, subject to future constraints on price setting. As it is standard in the literature, one can show
that by log-linearizing the expression for the price index and for the optimal reset price, we obtain the
conventional New Keynesian Philips curve relating inflation to movements in real marginal cost, pw
t , and
expected inflation
ˆπt =
(1 − λ)(1 − λβ)
λ
(ˆpw
t + ˆµt) + βEtˆπt+1
where hatted variables denote log-deviation from the steady state. ˆµt represents a mark-up shock. It is
commonly interpreted as representing the intensity of product market competition. It has been shown,
however, that it is interpretable as a wedge coming from the labour market (see Blanchard et al., 1997).
This interpretation is discussed in the following sub-section.
2.4 Households
There is a representative household composed by a continuum of members with total number normalized
to one. Members of the household can be either employed or unemployed and they pool all their income
together to perfectly insure each others.5
The head of the household maximizes utility over aggregate
consumption.
max E0
∞
t=0
βt
U(ct)
The budget constraint is given by
ct +
Bt
pt
= nt
zR
t
(1 − τw
t )wt(z)
g(z)
1 − G(zR
it )
dz + (1 − nt)(bt + r) + Πt − Tt +
Bt−1
pt
(1 + it)
The variable ct is aggregate consumption and Bt represents nominal bonds holding. Because of the
income-pooling assumption, each member consumes the same quantity and therefore the head of the
family can optimize over aggregate consumption. The first term on the right hand side of the budget
constraint is labour income net of taxation, the second represents the income of the unemployed. The
variables bt and r represent, respectively, unemployment benefits and “home production”. There are
profits Πt, lump sum transfers Tt and the last term represents financial income. The nominal interest
rate set by the monetary authority is represented by it. Idiosyncratic productivity is drawn only after
being hired and so ut = 1 − ρnt−1. In fact, 1 − nt includes workers being hired but successively laid-off
and they are not allowed to participate in the pool of workers twice. We make the following assumption
on the hiring process: workers are screened after being hired, but wage payments are made after the
screening. Therefore, it might be that some members of the household are hired and then laid off without
receiving any compensation. The earlier papers on this issue have often assumed that all new matches
start at the highest level of productivity. We chose the alternative way because we believe it to be more
realistic in light of the precarious work contracts given to most young workers. Finally, pt represents the
price of the final good.
Denoting with S(z) the Lagrange multiplier associated with the law of motion of employment from the
perspective of the household
nt = 1 − G(zR
t ) (ρnt−1 + st(utet))
it can be shown that for a member of the household with productivity z, the net value of being employed
is given by
St(z) = (1 − τw
t )wt(z) − bt − r + βEtΓt,t+1[1 − G(zR
t+1)](ρ − st+1)St+1(z )
5 This assumption has become standard after Merz (1995).
6 ILO Research Paper No. 12
In the equation, τw
t is the income tax. Unemployment benefits are computed as the replacement rate
times the steady state average wage, bt = ωb
t ¯w. Home production, r, is assumed to be instantaneously
produced and to be constant.6
The last term in brackets represents the “option value” of not being
currently employed.
2.5 Government and resource constraint
The government, or monetary authority, conducts monetary policy following a simple Taylor rule
it = iρi
t−1
πt
π∗
(1−ρi)φπ
exp(σi
t)
The coefficient ρi describe the extent of policy inertia while φπ the responsiveness of the nominal interest
rate to deviation of inflation to some target level π∗
. The shock σi
t represents variations in the nominal
interest rate that are not explained by changes in inflation. The government uses the tax revenue to
finance unemployment benefits and public expenditure, gt. The latter is assumed to follow
gt = 1 −
1
g
t
yt
where g
t is stochastic. Thus, government’s balance is given by
gt + bt(1 − nt) = (τw
t + τf
t )wtnt + Ft
G(zR
t )
1 − G(zR
t )
nt + Tt
The term
G(zR
t )
1−G(zR
t )
nt is the total number of workers fired in each period. We complete the description of
the model by imposing the resource constraint
yt = ct + gt + χqtvt
For estimation purposes, it is convenient to enter government expenditures in the resource constraint as a
residual, as it is often observed in the DSGE literature.
2.6 Determination of Wages
The firm bargains over wages with a union representing the workers. Wages are chosen to maximize the
joint surplus of the match, according to the Nash Bargaining rule
Jt(z) + Ft
1−ηt
St(z)ηt
The variable ηt represents the bargaining power of the workers, or equivalently union strength. The
modeling choice of using a “right to manage” model of bargaining, as opposed to other specification is
dictated by the wide use made of it in the literature. In particular, it constitutes a convenient choice
because it makes it easier to compare results with Zanetti (2011) and Thomas and Zanetti (2009).
Substituting (1) and (3) in (5), and differentiating with respect to wages we obtain
ηt
St(z)
(1 − τw
t ) =
1 − ηt
Jt(z) + Ft
(1 + τf
t )
6 The inclusion of home production is needed to obtain more accurate estimates of the replacement rate, as discussed in
Hagedorn and Manovskii (2008).
Labour market reforms in the Euro area: A DSGE approach 7
The solution of this problem delivers the wage bill equation, expressed as the sum of the surpluses for the
firm and for the worker, weighted according to their relative bargaining power
wt =
ηt
(1 + τf
t )
pw
t Atzt + Ft + ρβEtΛt,t+1
χ
qt+1
+
(1 − ηt)
(1 − τw
t )
bt + r − βEt [1 − G(zR
t+1)](ρ − st+1)St+1
where
wt =
zR
t
wt(z)
g(z)
1 − G(zR
t )
dz
Given the assumptions of constant returns in production, that idiosyncratic productivity is i.i.d. across
workers, and that there is a continuum of workers at the firm, the law of large numbers allows us to think
equivalently of the firm bargaining with each individual with idiosyncratic productivity, or to a union
bargaining over average wages. The presence of taxes introduces a distortion in the bargaining power
between workers and firms.
Following den Haan, Ramey and Watson (2000) we define job creation rate as the proportion of new
hires on total employment, minus those vacancies reposted within the period and filled that correspond
to exogenous quits
jct =
mt
nt
− (1 − ρ)qt
Similarly, the job destruction rate equals total separation minus reposted and filled vacancies relative to
exogenous quits
jdt = (1 − ρ) + ρG(zR
t ) − (1 − ρ)qt
Variation in total employment is given by the difference between the two flows
nt+1 − nt
nt
= jct − jdt
3 Long run impact of labour market reforms
This section studies the long run impact of changes in institutions by looking at the steady state of the
model. The quarterly discount rate is set to β = 0.99, consistent with a steady state inflation rate of
approximately 1%. As in Trigari (2009), the average markup is assumed to be of 20% over marginal costs.
Normalizing the price of the final good producer to one, delivers a real marginal cost pw
= 0.83. The
steady state government to GDP ratio is set to 20%. As in Thomas and Zanetti (2009), quarterly total
separation for the Euro area is assumed to equal 3.2%. Endogenous and exogenous separation happens
with equal probability, ρ = 1 − 0.032
2 . As in Thomas and Zanetti (2009), the matching function elasticity
is specified to be γ = 0.4, and a log-normal distribution governing workers’ idiosyncratic productivity is
assumed. Lacking direct evidence for the parameters governing such a distribution, we impose unitary
mean and shape equal to 0.1: these numbers ensure that all the endogenous variables are economically
meaningful. Finally, average values of labour market institutions are computed as sample averages over
the period 1970-2003.7
We then construct artificial “Euro area labour market institutions” by computing
labour force-weighted averages of the four largest Euro members, Germany, France, Italy and Spain.8
Of the six labour market institutions considered, replacement rates and labour taxes have a direct map
7 Data on labour market institutions are taken from CEP-OECD Institution Data in Nickell (2006).
8 See section 5.1 for more details.
8 ILO Research Paper No. 12
to the data: b = 0.25, τf
= 0.21, and τw
= 0.14. Lacking a direct link between the available index of
employment protection legislation and firing costs, we proceed as follows. The OECD EPL indicators
range from 0 to 6 (corresponding to the most stringent regulation). We assume that the most strict
EPL would force the firm to pay a cost equal to the full wage in case of lay-off. The Euro-average is
1.04, so we compute the replacement rate as 1.04
6 = 0.17. Although there is some information on ALMP
expenditure and coverage, there is not an obvious way to map these to matching efficiency. Therefore,
matching efficiency is left as free parameter in the computation of the steady state. Finally, the parameter
representing workers’ bargaining power, η, is set to 0.8. Also in this case, even if information on union
density or coverage is available, it is not obvious how to map it to the value of η. Although in the
literature η would typically range from 0.4 to 0.6, there is no direct evidence on what a realistic value
should be. Often, it is assumed η = γ, i.e. the Hosios condition to hold. However, here we follow Gertler,
Sala and Trigari (2008) which estimate a full-fledged search and matching model with several frictions
and find η = 0.8. All other parameters and steady state variables are computed by solving a 21 equations
system.
Table 1: Long run impact of labour market regulation
Benchmark Marginal Integral
Variable FC UB PAYROLL INCOME BARGAIN All
Output 2.29 2.33 2.20 2.30 2.35 2.27 2.20
Unemployment 0.18 0.17 0.21 0.18 0.16 0.19 0.20
Vacancies 0.10 0.12 0.10 0.11 0.09 0.09 0.11
Tightness 0.56 0.72 0.46 0.62 0.57 0.47 0.53
Reserv. prod. 2.23 2.21 2.28 2.23 2.20 2.24 2.28
Real wage 1.89 1.89 1.87 1.81 1.89 1.90 1.81
Job destruction 0.15 0.14 0.17 0.15 0.13 0.15 0.17
Job creation 0.18 0.17 0.21 0.18 0.16 0.19 0.22
Note: Benchmark refers to the calibrated steady state described in Section 3.
Each institution is increased by 5 percentage points. Bargaining power is increased from benchmark value (η = 0.8) to η = 1.
Table 1 shows the impact of labour market regulation on the model steady state. Under the label
“marginal” the effect of increasing firing costs, unemployment benefits, payroll and income taxes by 5
percentage points is shown. In the simulation, the bargaining power of the workers is increased from 0.8
to 1, i.e. the case in which workers are paid their marginal products leaving no rent to be extracted by
the firm. The last column labeled as “integral” shows the impact of a comprehensive reform in which all
the institutions are increased together at the same time. Consistent with the results in Zanetti (2009),
increasing firing costs from 0.17 to 0.22 increases output and lowers unemployment. Job destruction
decreases reflecting the decline in reservation productivity induced by the more expensive lay-off. Job
creation decreases as well, but not enough to have a negative impact on employment. This is due to the
fact that the real wage remains relatively unaffected by the reform. As in Zanetti (2009), increasing
the replacement rate - in our case from 0.25 to 0.30 - has opposite results. Output decreases and
unemployment rises. Reservation probability increases producing an increase in job destruction. On the
one hand, increasing the replacement rate increases the outside option of the worker, so that the real wage
should increase. However, the fall of labour market tightness implies a higher vacancy filling probability
for the firm. As a consequence, hiring costs fall and so do wages, with the latter effect dominating the
former. The third column simulates an increase in payroll taxes, the contributions paid by the firm for
each employed worker, from 0.21 to 0.26. Naturally, this reform increases substantially the cost of labour
for the firm. As a consequence, labour demand falls and so does the real wage. However, output and
employment remain substantially unchanged with respect to the benchmark case. The model seems
Labour market reforms in the Euro area: A DSGE approach 9
much more sensitive to an increase of income taxes, from 0.14 to 0.19. The reform increases output and
decreases unemployment. The reason behind these trends is difficult to understand based on these results.
One important aspect is that in the model labour taxes are used to finance government expenditure.
Therefore, an increase in income taxation could have boosted public spending and so sustained output
and employment through demand side. Increasing the bargaining power of the workers from 0.8 to 1, i.e.
to the point in which wages equal the marginal product of labour, net of hiring and firing costs. After
the reform, the real wage and unemployment both increase, and output falls. This is due to a decline in
labour demand - as shown in a reduction of vacancies and due to the reduced surplus appropriated by
the firm from each job match. Finally, the last column simulates the outcome of an “integral” reforms
involving all the changes at a time. Output falls and unemployment increases, with both changes being
quite significant, respectively -5% and +2%. labour market tightness falls, despite the mild increase in
vacancies. Average workers productivity increases despite the increase in firing costs. Nevertheless, real
wages falls. Therefore, according to the model, such reform would surely worsen welfare.
Of course, the results obtained in this section depend heavily on the parameterization and modeling
choices. Thus, they should not be taken at face value. However, this exercise has shown that studying
labour market reforms in a DSGE model can uncover general equilibrium effects producing outcomes that
are different from what conventional wisdom would suggest. This is because the effect of labour market
institutions is often thought in terms of partial equilibrium: shedding light on the general equilibrium
effects of labour market reforms should be central to the research agenda of those interested in the labour
market.
4 Short run impact of labour market reforms
In the previous section, we studied the impact of “marginal” and “integral” labour market reforms on the
long run equilibrium of the model. This section, investigates the dynamic response of the unemployment
rate after shocks to labour market institutions. To do so, each labour market institution is modeled
as an autoregressive stochastic process and the response of unemployment is analyzed through impulse
response analysis. This exercise is interesting for two main reasons. The first reason is related to the
political economy of the model. Although not explicitly considered here, the political viability of a reform
can be assessed by observing the behavior of key variables after a change of institutions. For example,
suppose that after an increase of firing costs, unemployment would initially rise but eventually decline.
This kind of reform would obviously be less popular as compared to one in which unemployment would
immediately decrease and then stay permanently low. In other words, by examining the transition path
between two steady states, we want to infer something about the viability of the reform. The second way
in which this exercise can be useful is to assess how unemployment reacts to permanent, as opposed to
transitory policy shocks. It is tempting to interpret the two kinds of shock as two reforms characterized
by different degrees of credibility. For example, a transitory shock is a change in policy that is expected to
be reversed in a relatively short time period. This might be due to political reasons (e.g. the credibility
of the government in place), or reasons linked to the economic outlook (such as in case of an increase in
unemployment insurance following an economic downturn). In practice, this amounts to running impulse
response functions with two different autoregressive (AR) coefficients for the policy shocks. The transitory
case corresponds to AR coefficients of magnitude 0.5, while the “quasi-permanent“ case to coefficients
of magnitude 0.99. Actually, changes in labour market institutions are usually thought as permanent.
However, following Krause and Lemke (2006), we approximate them as having an AR coefficient very
close to unity. This strategy has the advantage of approximating the permanent case as an AR coefficient
close to one implies
Ext+s ≈ xt
10 ILO Research Paper No. 12
for s < ∞, but at the same time it does not require to rewrite the system to accommodate a stochastic
trend.
Figure 1: Impulse response functions of unemployment to changes in labour market institutions
0 5 10 15
−2
−1
0
1
2
FIRING
0 5 10 15
−4
−2
0
2
ALMP
0 5 10 15
−1
0
1
2
BARGAINING
0 5 10 15
0
2
4
6
8
INCOME
0 5 10 15
0
1
2
3
BENEFITS
0 5 10 15
−1
0
1
2
PAYROLL
Note: Dashed line: temporary change (AR coeff. 0.5). Full line: persistent change (AR coeff. 0.99)
To simulate the model, we need to specify the missing three parameters, plus all the shock coefficients
representing labour market institutions. Following Thomas and Zanetti (2009), we impose a Calvo price
parameter λ = 0.88. For the Taylor rule, the response-coefficient to inflation is set φπ = 3.15, and a
monetary policy inertia equal to ρi = 0.70. In treating labour market institutions as stochastic processes,
we specify them as AR(1) processes. While the choice is admittedly arbitrary, this formulation has the
advantage of being simple but at the same time able to take into account the potential persistence of
labour market institutions. We use an AR coefficient of 0.5 for what is labeled “temporary change”,
and a coefficient equal to 0.99 for a “persistent change”. Figure 1 shows the response of unemployment
to 1% policy shocks for each labour market institution. The dashed line represents the transitory
shock, while the full one the persistent one. The results show that whether the reform is expected to
last only for a few periods only, or for a long time matters for the response of unemployment. This
depends on both the consumption smoothing decisions of the households, and the employment decisions
of firms. Not only the size of adjustment of unemployment changes radically in the two cases, but
Labour market reforms in the Euro area: A DSGE approach 11
even its qualitative response. An increase in replacement rate, payroll and income taxes unambiguously
increase unemployment, although by much more when the change is persistent. Conversely, the sign
of the response of unemployment for the other institutions depends on the size of the autocorrelation
coefficient. It is tempting to interpret the persistence of these shocks as the “credibility” that a reform
has. In fact, there can be cases in which a not-so-credible government would implement a reform which
is widely believed to be short-lived. This would be similar — although not equivalent — to rational
agents knowing that the value of the institutions would go back to its original value in, say, about four
quarters.9
Therefore, this exercise suggests that reduced form regressions of unemployment on labour
market institutions that do not control for some kind of indicator of government credibility, would suffer
of severe problems of omitted variable bias.
5 Sources of fluctuations in the Euro area
This section attempts to shed light on the main sources of output, unemployment and inflation fluctuation
in the Euro area. In the model used here, fluctuations can be due to two kinds of different shocks. The
first set of shocks includes standard macroeconomic shocks: productivity, interest rate, price markup,
and to government consumption. We will refer to them simply as “structural” shocks. The second set
includes the policy shocks, i.e. shocks to labour market institutions: ALMP, bargaining power, firing
costs, income and payroll taxes, and unemployment benefits. Whether policy shocks generate substantial
fluctuations in the model is important because it would help in: i) understanding whether the secular
increase of European unemployment was due to excessive labour market regulation, as opposed to the
“Anglo-Saxon” experience; ii) assessing the importance of labour market regulation for inflation and output
volatility. To this extent, we estimate the model on the Euro area. If it is true that the labour market
experiences differed substantially within Europe, the largest Euro area economies - Germany, France,
Italy and Spain - all experienced high and persistent unemployment. At the same time, they constitute
the largest countries in the area. Therefore, with some abuse of terminology, we will use the Euro area as
an approximation of what is generally called “Europe” in the literature on unemployment.
5.1 Estimation: Data and priors
The estimation strategy consists in keeping the parameters determining the steady state as in Section
3, and estimating those governing shocks and nominal side of the model. Thus, there are twenty three
parameters to be estimated. Twenty of them characterize the shock processes. These are the AR
coefficients of the four structural shocks (productivity, interest rate, price markup and government
consumption), those of the six labour market institutions shocks (firing costs, payroll and income taxes,
the replacement rate, matching efficiency, and bargaining power), the standard deviations of all shocks.
The remaining three parameters are those of the Taylor rule (policy inertia and reaction to inflation),
and the Calvo price parameter.
We use the following variables for the estimation: i) real GDP per working age persons; ii) the employment
rate; iii) the short term nominal interest rate; iv) the year on year inflation rate; v) the real wage; vi) the
replacement rate; vii) payroll taxes; viii) income taxes; ix) an index of EPL. Data for i) to v) are take
from the AWM dataset10
, while data on institutions are computed from the CEP-OECD institutions data
set11
. Due to limitations of the institution data, our sample goes from 1970Q4, to 2003Q4, leaving
9 This would be the case with an AR coefficient equal to 0.5
10 http://www.eabcn.org/area-wide-model
11 http://eprints.lse.ac.uk/19789/
12 ILO Research Paper No. 12
Figure 2: Euro area labour market institutions series
2A: Original series
2B: Smooth series
Note: The smooth trend in Figure 2A is obtained with an HP-Filter with smoothing parameter equal to 100. In
Figure 2B. a HP-Filter with smoothing parameter equal 1600 is used. Source: CEP-OECD institutions dataset.
us with 133 observations. Data series i) and v) are HP-filtered with smoothing parameter equal to 1600,
while ii), iii) and iv) are demeaned. Data on labour market institutions has a few drawbacks. First of all,
they are provided only for individual countries and not as an aggregate for the Euro area. Therefore, we
compute artificial series for vi) to ix) by computing averages weighted by the size of the labour force
in the largest four Euro countries, Germany, France, Italy and Spain. Another issue is that data on
institutions have annual frequency, while the model is calibrated to quarterly frequency. To deal with
this problem, for each institution we extend the annual series to quarterly frequency by keeping the value
at the beginning of the year constant for each of the corresponding four quarters. Then, we filter the
series with an arbitrary smoothing parameter (equal to 100) until obtaining a smooth trend as shown in
Labour market reforms in the Euro area: A DSGE approach 13
Figure 2A. Since these series display fluctuations that are much larger than those of the other series, it
seems natural to treat them as non-stationary. For this reason we then apply again the HP-Filter with
smoothing parameter 1600 to isolate the business cycle frequencies. The result is shown in Figure 2B.
As expected, the business cycle fluctuations of labour market institutions are of an order much smaller
than the other series. The bottom part of Figure 2 shows fluctuations ranging from 3% to less than 1%.
Fluctuations at the extreme of the sample are somewhat larger than in the middle, with replacement
rate showing the wider changes. As for the choice of priors, we follow Gertler, Sala and Trigari (2008) in
specifying i) beta distributions with mean 0.5 and standard deviation 0.15 for all AR coefficients, and
ii) inverse-gamma distributions with mean 0.15 and standard deviation 0.15 for all shocks’ standard
deviation. Following the logic of the same authors, we specify a iii) beta distribution for the Calvo
parameter with mean 0.88 and standard deviation 0.15; iv) a normal distribution with mean 3.15 and
standard deviation 0.15 for the inflation coefficient of the Taylor rule, and v) a beta distribution with
mean 0.7 and standard deviation 0.15 for the policy inertia parameter.
5.2 Estimation results
Table 2 summarizes priors and posterior means of the estimated parameters. The data favor a specification
with very low policy inertia but a high autocorrelation coefficient for the interest rate shock. The Calvo
parameter is substantially higher than what is found in Thomas and Zanetti (2009), implying that firms
re-optimize once every three quarters. All structural shocks are quite persistent, with a standard deviation
substantially higher than the priors. Among the structural shocks, government consumption has the
highest standard deviation. This is quite common as the government spending shock acts as a residual in
the resource constraint and it captures movements in the trade balance that are not included into the
model. labour market institution-shocks are also very persistent, exception made for the bargaining power
shock that has an AR coefficient similar to the prior, and the matching efficiency shock. This should
not be surprising as these two shocks are free to adjust and they are not constrained by the data on
institutions, which are very persistent in nature. The matching efficiency shock has the highest variance
for the same reason. Interestingly, the bargaining power shock - usually accounting for most of wage
fluctuation, is very low. In fact, in this kind of models, bargaining power shocks are always very large
because they need to account for the excessive variation of the model-wage with respect to the data. We
interpret this fact as follows: the inclusion of labour market institution-shocks helps standard search and
matching model to reproduce the volatility of wages.
To assess the behavior of the model following “standard” structural shocks (productivity, interest rate,
markup and government spending), figures 3 to 6 display the impulse response functions of several key
macroeconomic variables. According to the analysis, all expansionary shocks increase output, employment
and real wages. However, unemployment and vacancies are positively correlated that is, the model
generates an upward-sloped Beveridge curve. As a consequence, labour market tightness is countercyclical.
This is because during good times reservation productivity decreases, and so does average worker
productivity. Therefore, firms have less incentives to hire, generating the positive relationship between
vacancies and unemployment. A positive productivity shocks generates, in accordance with the literature,
decreasing marginal costs. Job creation and job destruction rates tend to move in the same direction
over the business cycle and to be roughly of the same order of magnitude. Job destruction however
tends to reach its extreme value with a lag of about three to four quarters, whereas job creation does so
immediately on impact.
14 ILO Research Paper No. 12
Table 2: Priors and posteriors of the estimated parameters
Parameter Prior (mean, sd) Posterior mean 90% Interval
Policy inertia ρin beta (0.7, 0.15) 0.158 [ 0.089, 0.221 ]
Response to inflation rπ normal (3.15, 0.15) 3.460 [ 3.199, 3.725 ]
Calvo price λ beta (0.88, 0.15) 0.690 [ 0.595, 0.811 ]
AR coefficient
Unemployment benefits ρb beta (0.5, 0.15) 0.936 [ 0.909, 0.968 ]
Firing costs ρf beta (0.5, 0.15) 0.943 [ 0.918, 0.969 ]
Payroll taxes ρτf beta (0.5, 0.15) 0.949 [ 0.929, 0.974 ]
Income taxes ρτw beta (0.5, 0.15) 0.931 [ 0.904, 0.961 ]
Bargaining power ρη beta (0.5, 0.15) 0.503 [ 0.220, 0.755 ]
ALMP ρalmp beta (0.5, 0.15) 0.872 [ 0.829, 0.909 ]
Productivity ρa beta (0.5, 0.15) 0.883 [ 0.828, 0.936 ]
Interest rate ρi beta (0.5, 0.15) 0.962 [ 0.938, 0.995 ]
Markup ρµ beta (0.5, 0.15) 0.937 [ 0.905, 0.968 ]
Government spending ρg beta (0.5, 0.15) 0.942 [ 0.905, 0.977 ]
Standard deviation
Unemployment benefits σb inv. gamma (0.15, 0.15) 0.425 [ 0.385, 0.472 ]
Firing costs σf inv. gamma (0.15, 0.15) 0.169 [ 0.152, 0.187 ]
Bargaining power ση inv. gamma (0.15, 0.15) 0.129 [ 0.047, 0.228 ]
Payroll taxes στf inv. gamma (0.15, 0.15) 0.124 [ 0.110, 0.135 ]
Income taxes στw inv. gamma (0.15, 0.15) 0.101 [ 0.091, 0.111 ]
ALMP σalmp inv. gamma (0.15, 0.15) 1.258 [ 0.158, 1.408 ]
Productivity σa inv. gamma (0.15, 0.15) 0.357 [ 0.320, 0.390 ]
Interest rate σi inv. gamma (0.15, 0.15) 0.667 [ 0.566, 0.739 ]
Markup σµ inv. gamma (0.15, 0.15) 0.286 [ 0.245, 0.321 ]
Government spending σg inv. gamma (0.15, 0.15) 6.967 [ 3.279, 13.084 ]
Labour market reforms in the Euro area: A DSGE approach 15
Figure 3: Impulse response functions of selected variables: one standard deviation productivity
shock
0 10 20 30 40
−4
−3
−2
−1
0
Unemployment
Vacancies
0 10 20 30 40
0
0.2
0.4
0.6
0.8
Output
Employment
0 10 20 30 40
−4
−3
−2
−1
0
Job creation
Job destruction
0 10 20 30 40
−0.4
−0.2
0
0.2
0.4
Reservation prod.
Real wage
0 10 20 30 40
−0.12
−0.1
−8 · 10−2
−6 · 10−2
−4 · 10−2
−2 · 10−2
0
Inflation
Marginal cost
0 10 20 30 40
−2
−1.5
−1
−0.5
0
Tightness
16 ILO Research Paper No. 12
Figure 4: Impulse response functions of selected variables: one standard deviation interest rate
shock
0 10 20 30 40
0
0.5
1
1.5
Unemployment
Vacancies
0 10 20 30 40
−0.14
−0.12
−0.1
−8 · 10−2
−6 · 10−2
−4 · 10−2
−2 · 10−2
Output
Employment
0 10 20 30 40
0
0.5
1
1.5
Job creation
Job destruction
0 10 20 30 40
−0.1
−5 · 10−2
0
5 · 10−2
0.1
0.15
Reservation prod.
Real wage
0 10 20 30 40
−0.4
−0.3
−0.2
−0.1
0
0 10 20 30 40
0
0.2
0.4
0.6
0.8
Tightness
Labour market reforms in the Euro area: A DSGE approach 17
Figure 5: Impulse response functions of selected variables: one standard deviation markup shock
0 10 20 30 40
0
1
2
3
4
Unemployment
Vacancies
0 10 20 30 40
−0.4
−0.3
−0.2
−0.1
0
Output
Employment
0 10 20 30 40
0
1
2
3
4
Job creation
Job destruction
0 10 20 30 40
−0.2
0
0.2
0.4
0.6
Reservation prod.
Real wage
0 10 20 30 40
−0.3
−0.2
−0.1
0
0.1
Inflation
Marginal cost
0 10 20 30 40
0
0.5
1
1.5
2
Tightness
18 ILO Research Paper No. 12
Figure 6: Impulse response functions of selected variables: one standard deviation government
spending shock
0 10 20 30 40
−2
−1.5
−1
−0.5
0
0.5
Unemployment
Vacancies
0 10 20 30 40
0
5 · 10−2
0.1
0.15
0.2
Output
Employment
0 10 20 30 40
−2
−1.5
−1
−0.5
0
Job creation
Job destruction
0 10 20 30 40
−0.2
−0.1
0
0.1
0.2
Reservation prod.
Real wage
0 10 20 30 40
0
5 · 10−2
0.1
0.15
0.2
Inflation
Marginal cost
0 10 20 30 40
−1
−0.8
−0.6
−0.4
−0.2
0
Tightness
5.3 Decomposing the volatility of unemployment, output and inflation in
the Euro area
Unemployment increased in almost all OECD countries during the 1970s, but then it decreased in some
countries (like the “USA”), while it remained high in others (such as Germany, France, Italy and Spain,
that we call the “Euro area”). Why did this happen? Such a question generated a heated debate which
divided scholars into two main factions: on one hand there are those argue that the Euro area was hit
by several and persistent shocks during the 1980s and 1990s, while the US was not. The other faction
believes that it was change in labour market regulation in the Euro area that increased the structural
level of unemployment, as opposed to the US where flexible labour market institutions made prices
absorbing the adverse shocks and let unemployment decrease to its equilibrium level.12
This section uses
variance decomposition techniques to shed light on the relative merit of these two views. One of the
interesting features of treating labour market institutions as stochastic processes is precisely that: to be
able to distinguish between standard macroeconomic shocks and changes in regulation to account for the
historical variation of unemployment.
12 See for example Nickell (2003) and Saint Paul (2004)
Labour market reforms in the Euro area: A DSGE approach 19
The results are summarized in the top panel of Table 3. The first row presents the unconditional
variance decomposition of unemployment over the sample (1970Q4 - 2003Q4). It says that only 30 %
of total unemployment fluctuation was due to changes in labour market institutions, in particular to
unemployment benefits and changes in the bargaining power of workers. The 70% left is due to standard
macro shocks with productivity and interest rate shocks being the most important. Inspecting the
conditional variance decomposition at one quarter, one year and five years ahead, it is evident how in
the short run, changes in unemployment are mostly due to demand factors, captured by the interest
rate shock. Some role is played by changes in matching efficiency — the effectiveness of ALMPs, and
firing costs. As the horizon increases, supply factors gain importance, together with unemployment
benefits and the bargaining shock. Since the bargaining power shocks is very important to explain wage
dynamics, one can interpret this result as showing the role of wage rigidity in preventing the adjustment
of prices following a productivity shock, a mechanism already proposed by Bruno and Sachs (1985). It
is remarkable how the results obtained with this exercise are in line with those in Nickell and Layard
(1999), who operate in a very different framework.
The second issue under examination is whether changes in labour market regulation had a significant
impact on inflation and output volatility. This exercise is motivated by the fact that labour market
institutions create frictions in the labour market affecting marginal costs and thus the pricing decisions
of firms. As a consequence, they might contribute to the variability of output and inflation. Thomas
and Zanetti (2009) performed a similar exercise - although using different techniques - but assessing
only the impact of changes in firing costs and unemployment benefits. Since there is not theoretical
reason to believe that other institutions should not be as important, we conduct the exercise including a
more comprehensive set of institutions. From the central and bottom panel of Table 3, it is immediately
evident that inflation and output volatility is almost entirely due to macro shocks, rather than changes in
labour market regulation. The variance of inflation is explained by cost push factors, represented by the
markup shock, and to a less extent by changes in government consumption. This is true unconditionally
and conditional on every time period. The variance of output is explained by shocks to interest rate and
productivity. Bargaining power shocks play a minor role only, especially in the long run. That firing
costs and unemployment benefits do not significantly contribute to the variance of inflation was already
showed in Thomas and Zanetti (2009). However, this exercise shows that the even other important
institutions such as ALMP, labour taxes and institutions affecting the bargaining power of workers are
equally unimportant and that to a large extent, the results extend to output volatility.
20 ILO Research Paper No. 12
Table3:Unconditionalandconditional(1,4and20periodsahead)variancedecompositionofunemployment,inflationandoutput.
Unempl.PRODINRATEMARKUPGCONSALMPPAYROLLINCOMEFIRINGBARGAINUBENEF
-34.2716.506.136.130.413.281.090.0024.957.25
10.0070.140.006.538.435.670.009.230.000.00
444.7720.213.084.660.252.610.740.0018.265.44
2036.1417.344.985.830.373.231.020.0024.047.03
InflationPRODINRATEMARKUPGCONSALMPPAYROLLINCOMEFIRINGBARGAINUBENEF
-0.630.2683.9415.130.000.000.000.010.020.01
13.281.2974.7920.270.000.030.020.060.200.07
41.690.6677.5619.950.000.010.010.020.070.02
200.780.3281.8417.010.000.000.000.010.030.01
OutputPRODINRATEMARKUPGCONSALMPPAYROLLINCOMEFIRINGBARGAINUBENEF
-37.6839.992.652.740.071.531.100.0010.843.39
142.8846.670.901.560.020.790.570.024.971.63
442.4843.531.181.790.031.070.690.016.982.24
2038.5340.782.132.560.061.481.020.0010.233.22
Labour market reforms in the Euro area: A DSGE approach 21
6 Conclusions
This paper builds a DSGE model with a rich characterization of the labour market and models explicitly
key labour market institutions. The model is used to simulate the short and long run outcomes of
several labour market reforms. This environment allows to study the general equilibrium response of
unemployment to truly exogenous changes in regulation, a difficult endeavor to achieve with reduced form
regressions. Firing costs and labour taxes increase output and decrease unemployment in the long run,
while unemployment benefits and higher bargaining power of workers have the opposite effect. In the
short run, whether the change in policy is expected to last, or to be in place for a few periods only makes
a large difference in terms of unemployment. After estimating the model on the Euro area, a variance
decomposition reveals that most of the historical volatility of unemployment, output and inflation is
due to structural shocks, rather than changes in regulation. This evidence suggests that the rise and
persistence of European unemployment was due to macroeconomic shocks, rather than to a progressive
tightening of labour market regulation. In addition, according to the results, central banks should not
worry too much about labour market regulation.
22 ILO Research Paper No. 12
References
Andolfatto, D. 1996. “Business cycles and labour-market search”, in The American Economic Review,
Vol. 86, No. 1, pp. 112–132.
Barnichon, R. 2010. “Building a composite Help-Wanted index”, in Economics Letters, Vol. 109, No. 3,
pp. 175–178.
Bentolila, S.; Bertola, G. 1990. “Firing costs and labour demand: how bad is Eurosclerosis?”, in The
Review of Economic Studies, Vol. 57, No. 3, pp. 381–402.
Blanchard, O.J. 2007. “A review of Richard Layard, Stephen Nickell, and Richard Jackman’s unemployment:
macroeconomic performance and the labour market”, in Journal of Economic Literature, Vol.
45, No. 2, pp. 410–418.
Blanchard, O.J.; Nordhaus, W.D.; Phelps, E.S. 1997. “The medium run”, in Brookings Papers on
Economic Activity, No. 2, pp. 89–158.
Blanchard, O.J.; Wolfers, J. 2000. “The role of shocks and institutions in the rise of European unemployment:
the aggregate evidence”, in The Economic Journal, Vol. 110, No. 462, pp. 1–33.
Bruno, M.; Sachs, J.D. 1985. Economics of worldwide stagflation (Cambridge, MA, Harvard University
Press)
Calvo, G.A. 1983. “Staggered prices in a utility-maximizing framework”, in Journal of Monetary
Economics, Vol. 12, No. 3, pp. 383–398.
Christiano, L.J.; Eichenbaum, M.; Evans, C.L. 2005. “Nominal rigidities and the dynamic effects of a
shock to monetary policy”, in Journal of Political Economy, Vol. 113, No. 1, pp. 1–45.
Christoffel, K.; Kuester, K.; Linzert, T. 2009. “The role of labour markets for euro area monetary policy”,
in European Economic Review, Vol. 53, No. 8, pp. 908–936.
Cooley, T.F.; Quadrini, V. 1999. “A neoclassical model of the Phillips curve relation”, in Journal of
Monetary Economics, Vol. 44, No. 2, pp. 165–193.
Davis, S.J.; Faberman R.J.; Haltiwanger, J. 2006. “The flow approach to labour markets: New data
sources and micro-macro links”, in Journal of Economic Perspectives, Vol. 20, No. 3, pp. 3–26.
Davis, S.J.; Haltiwanger J.; Schuh, S. 1996. “Small business and job creation: Dissecting the myth and
reassessing the facts”, in Small Business Economics, Vol. 8, No. 4, pp. 297–315.
Den Haan, W.J.; Ramey, G.; Watson, J. 2000. “Job destruction and the experiences of displaced workers”,
in Carnegie-Rochester Conference Series on Public Policy, Vol. 52, No. 1, pp. 87–128.
Galı, J.; Gertler, M. 1999. “Inflation dynamics: A structural econometric analysis”, in Journal of
Monetary Economics, Vol. 44, No. 2, pp. 195–222.
Gertler, M.; Sala, L.;Trigari A. 2008. “An estimated monetary DSGE model with unemployment and
staggered nominal wage bargaining”, in Journal of Money, Credit and Banking, Vol. 40, No. 8, pp.
1713–1764.
Labour market reforms in the Euro area: A DSGE approach 23
Gal´ı, J.; Smets, F.; Wouters, R. 2011. “Unemployment in an estimated new Keynesian model”, in
National Bureau of Economic Research Macroeconomics Annual, Vol. 26, No. 1, pp. 329-360.
Hagedorn, M.; Manovskii, I. 2008. “The cyclical behavior of equilibrium unemployment and vacancies
revisited”, in The American Economic Review, Vol. 98, No. 4, pp. 1692–1706.
Hall, R.E. (2005). “Employment fluctuations with equilibrium wage stickiness”, in The American
Economic Review, Vol. 95, No. 1, pp. 50–65.
Krause, M.U.; Lopez-Salido, D.J.; Lubik, T.A. 2008a. “Inflation dynamics with search frictions: A
structural econometric analysis”, in Journal of Monetary Economics, Vol. 55, No. 5, pp. 892–916.
Krause, M.U.; Lopez-Salido, D.J.; Lubik, T.A. 2008b. “Do search frictions matter for inflation dynamics?”,
in European Economic Review, Vol. 52, No. 8, pp. 1464–1479.
Krause, M.; Lemke, W. 2006. Optimal monetary policy response to distortionary tax changes, Computing
in Economics and Finance No. 306 (Society for Computational Economics).
Ireland, P.N. 2004. “A method for taking models to the data”. in Journal of Economic Dynamics and
Control, Vol. 28, No. 6, pp. 1205–1226.
Layard, R.; Nickell, S.; Jackman, R. 2005. Unemployment: Macroeconomic performance and the labour
market (Cambridge, MA, Oxford University Press).
Lazear, E.P. 1990. “Job security provisions and employment”, in The Quarterly Journal of Economics,
Vol. 105, No. 3, pp. 699–726.
Merz, M. 1995. “Search in the labour market and the real business cycle”, in Journal of Monetary
Economics, Vol. 36, No. 2, pp. 269–300.
Nickell, W. 2006. The CEP-OECD institutions data set (1960-2004),CEP Discussion Paper No. 759
(London, Centre for Economic Performance).
Nickell, S. (2003). “Labour market institutions and unemployment in OECD countries”, in CESifo DICE
Report Vol. 1, No. 2, pp.13–26.
Nickell, S.; Layard, R. 1999. “Labour market institutions and economic performance”, in O. Ashenfelter
and D. Card (eds): Handbook of labour economics Vol.3, Part C (Amsterdam, Elsevier Sciences B.V.),
pp. 3029–3084.
Pescatori, A.; Tasci, M. 2011. Search frictions and the labour wedge, IMF Working Paper WP/11/117
(Washington, DC).
Saint-Paul, G. 2004. “Why are European countries diverging in their unemployment experience?”, in
Journal of Economic Perspectives, Vol. 18, No. 4, pp. 49–68.
Shimer, R. 2005. “The cyclical behavior of equilibrium unemployment and vacancies”, The American
Economic Review, Vol. 95, No. 1, pp. 25–49.
Smets, F.; Wouters, R. (2007). Shocks and frictions in US business cycles: A Bayesian DSGE approach,
ECB Working Paper Series No. 722 (Frankfurt am Main, European Central Bank).
24 ILO Research Paper No. 12
Thomas, Carlos; Francesco Zanetti (2009). “labour market reform and price stability: An application to
the Euro Area”, in Journal of Monetary Economics, Vol. 56, No. 6, pp. 885–899.
Trigari, A. 2009. “Equilibrium unemployment, job flows, and inflation dynamics”, in Journal of Money,
Credit and Banking, Vol. 41, No. 1, pp. 1–33.
Van Ours, J.C. 2002. The locking-in effect of subsidized jobs, Discussion Paper No. 527 (Bonn, Institute
for the Study of Labor (IZA)).
Zanetti, F. 2011.“Labour market institutions and aggregate fluctuations in a search and matching model”,
in European Economic Review, Vol. 55, No. 5, pp. 644–658.