Shifting the Beveridge curve: What affects labour market matching?

• Published date: 01 June 2018
• DOI: http://doi.org/10.1111/ilr.12046
• Date: 01 June 2018

International Labour Review, Vol. 157 (2018), No. 2

Copyright © The authors 2018

Journal compilation © International Labour Organization 2018

*

International Monetary Fund, Fiscal Affairs Department, Washington, DC, emails: ebova@

imf.org; jjalles@imf.org. ** International Monetary Fund, Strategy, Policy and Review Depart-

ment, Washington, DC, email: ckolerus@imf.org. The authors would like to thank Vítor Gaspar, Bene-

dict Clements, Romain Duval, Christian Ebeke, Robert Sierhej, Rima Turk, Li Zeng, Helge Berger

and participants at the IMF Fiscal Affairs Department Seminar Series for their useful comments

and suggestions. They would also like to thank Ethan Alt for his excellent research assistance. Any

remaining errors are the authors’ sole responsibility, and the views expressed herein do not neces-

sarily reect those of the IMF or its member countries.

Responsibility for opinions expressed in signed articles rests solely with their authors, and

publication does not constitute an endorsement by the ILO.

Shifting the Beveridge curve:

What affects labour market matching?

Elva BOVA,* João TOVAR JALLES* and Christina KOLERUS**

Abstract. This article explores short-run determinants of the matching between la-

bour demand and supply by identifying shifts in the Beveridge curves for 12 OECD

countries between 200 0Q1 and 2013Q4. Using three complementary methodologies

(visual examination, cointegration techniques and non-linear estimations), we nd

that labour force growth and employment protection legislation reduce the likeli-

hood of outward shifts, and the higher the share of employees with intermediate

levels of education and the long-term unemployment, the more difcult the matching

process. Active labour market policies (such as incentives for start-ups or job-sharing

programmes) could facilitate matching, while passive policies (unemployment bene-

ts or labour taxation) make matching signicantly more difcult.

While much of the increase in unemployment since the global nancial

crisis has been attributed to cyclical factors (Kugler, 2014), mismatches

between labour demand and labour supply have become more relevant in

comparison with pre-crisis years. The evidence reveals that positive signs of

recovery, such as an increase in the advertised number of vacancies beginning

in the last quarter of 2009, coexist with stubbornly high levels of unemploy-

ment, thus suggesting signicant mismatches in labour supply and demand (g-

ure 1). More importantly, long-term unemployment has increased markedly in

the aftermath of the global nancial crisis. As OECD data show, the number

of unemployed rose by almost 50 per cent between 2007 and 2013, with the

number of long-term unemployed increasing by more than 80 per cent. Given

the path dependency associated with long-term unemployment, addressing la-

bour market mismatches is becoming an even more urgent task.

International Labour Review268

Improving the efciency of labour market matching requires policies

that go beyond those aimed merely at stimulating aggregate demand, given

that frictional unemployment originates from institutional inefciencies, from

skill gaps between the market forces of supply and demand and from any fac-

tor that dissuades jobseekers from accepting a job or that makes employers

choosier in their job selection process.

Frictions and mismatches in the labour market can be captured by

the so-called Beveridge curve, which relates vacancies to the number of un-

employed. An economic slowdown, during which the job destruction process

is more volatile than the job creation process (Mortensen and Pissarides, 1994),

would lead to a downward movement along the curve, corresponding to fewer

vacancies and higher unemployment. A recovery, on the other hand, would

trigger an upward ride. Some policies, such as the short-time working scheme

applied by several countries during the global nancial crisis, might prevent

or attenuate an increase in unemployment. These types of policies could even

lead to an inward shift of the curve if vacancies are decreasing at a given un-

employment level due to a slowdown in economic activity. Conversely, during

periods of jobless recoveries, the Beveridge curve would show an outward shift

as vacancies remain constant and unemployment increases.

This article assesses the role played by policies and institutions in affect-

ing shifts of the Beveridge curves for a sample of 12OECD countries over the

Figure 1. Average unemployment and vacancies in the great recession

(percentage of labour force)

Source: OECD.

4.0

5.0

6.0

7.0

8.0

9.0

3.0

2.0

1.0

0

Unemployment (percentage)

2007Q1

0.4

0.6

0.8

1.0

1.2

1.4

0.2

0

Vacancies

2007Q3

2008Q1

2008Q3

2009Q1

2009Q3

2010Q1

2010Q3

2011Q1

2011Q3

2012Q1

2012Q3

2013Q1

2013Q3

Unemployment

Vacancies

Years, quarters

Shifting the Beveridge curve: What affects labour market matching? 269

2000 Q1–2013Q4 period. In an initial step, we detect shifts in each country’s Bev-

eridge curve and determine the magnitude and direction of those shifts accord-

ing to three methodologies: visual examination, cointegration techniques and

non-linear estimation. In a second step, using a panel probit model, we assess

several factors that could have an inuence on the probability of these shifts, to

better understand which policies and institutions can affect the efciency of la-

bour market matching. Our main ndings can be summarized as follows:

• Shifts. Of the 12 OECD countries that are examined, ten exhibit a shift in

their respective Beveridge curve. We identify seven countries with outward

shifts (i.e. deteriorations of labour market matching), which in many cases

took place at the onset of the global nancial crisis, and two countries with

inward shifts. One country features both an outward and an inward shift.

• Labour market structure. We nd strong and robust evidence throughout

our specications that labour force growth reduces the likelihood of an

outward shift of the Beveridge curve. Also, higher labour market protec-

tion makes outward shifts less likely and is thus negatively associated with

frictional unemployment. We furthermore nd that the higher the share

of employees with intermediate education in the labour force, the more

likely an outward shift.

• Categories of unemployment. We nd robust evidence that the larger the

share of long-term unemployed with respect to the total number of un-

employed, the greater the frictions. This is possibly due to outdated skill

sets or employer bias against this group. Our preferred specication of the

model also provides evidence that matching is more difcult the larger the

share of female jobseekers and young jobseekers, while matching is easier

the larger the share of elderly workers, possibly due to greater experience.

• Policies. Tax and expenditure policies can play a role in reducing frictional

unemployment. We nd that higher social security contributions and, more

generally, the tax wedge are more likely to shift the Beveridge curve out-

ward and have a detrimental impact on matching, especially at higher levels

of income. As expected, a similar effect is found for higher unemploy-

ment benets, given that they lower the urgency to nd a job. Conversely,

spending on active labour market programmes has a positive impact on

reducing frictions, particularly when such programmes are aimed at pro-

viding incentives for start-ups and promoting job sharing programmes.

• Interactions. Our results show that, during the 2008 global shock, un-

employment benets had a stronger negative impact on frictional unemploy-

ment, while the impact of both long-term unemployment and low and

intermediate levels of education was weaker. There was no statistically

signicant change in the impact of the determinants identied above, in-

cluding testing for complementarities across labour market institutions,

thus conrming the nding by Bassanini and Duval (2009).

•

Labour market reforms and aggregate demand. Our results also show that

labour tax reforms and changes in unemployment benets may have an