Comparisons and contrasts of the impact of the crisis on euro area labour markets Report prepared by ad hoc ESCB Team Robert Anderton, Team Chairman Deputy Head of Surveillance Division (SUR)

Report coordinators/editors: V. Jarvis, B. Szörfi (OAD) Presentation prepared by V. Jarvis, M. Périnet and B. Szörfi (OAD), on behalf of the ESCB Team.

Presentation to OECD 18 March 2015

Overview of Rubric

the presentation

1

Labour market dynamics and unemployment over the crisis

2

Labour market slack and structural unemployment

3

Wage adjustment

4

Concluding remarks: summary of main findings and policy implications

2

www.ecb.europa.eu ©

A Rubric larger, longer-lasting impact on employment of the crisis compared to the past. Both recession phases of crisis had strong impacts on euro area labour market, despite a much milder downturn second time around… Euro area employment across recessions

Euro area employment, GDP and unemployment since 1999

(index T=100 at cyclical peak in GDP; intervals are quarters)

(year-on-year growth rates; % of labour force) GDP

1980 Q1

1992 Q1

2008 Q1

employment (lhs)

unemployment rate (rhs) 6

102

13

4 100

12

2 11 0 10

98 -2

9 -4 96 8

-6

-8 1999

94 T-8

T-4

T

T+4

T+8

T+12

T+16

T+20

T+24

Source: Eurostat and ESCB calculations.

7 2001

2003

2005

2007

2009

2011

2013

Source: Eurostat and ESCB calculations. 3

www.ecb.europa.eu ©

Differential impacts of Rubric

the two recessions on euro area labour markets

… with stronger impacts across the stressed euro area countries. Evolution of unemployment rates across euro area countries 2008

2011

2013

30 25 20 15 10 5 0 euro area

CY

SI

IE

IT

GR

PT

ES

NL

stressed countries

AT

LU

EE

MT

FI

BE

FR

DE

LV

SK

other euro area economies

Sources: Eurostat and ESCB calculations. Note: within groups, countries are ordered according to their unemployment rate level in 2008. 4

www.ecb.europa.eu ©

Labour Rubricmarket flows: Job separation rates rose, but job-finding rates declined much more…

Flows from unemployment to employment over the crisis

Flows out of employment into unemployment over the crisis

(job finding probabilities)

(job separation probabilities) 8

35 2005Q1-2008Q1

7

2005Q1-2008Q1

2008Q2-2011Q2

30 2011Q3-2013Q1

2008Q2-2011Q2 2011Q3-2013Q1

6 25 5 20 4 15 3 10

2

5

1

0

0 EA11

EE

IE

GR

ES

FR

IT

NL

AT

PT

SI

SK

FI

EA11

EE

IE

GR

ES

FR

IT

NL

AT

PT

SI

SK

FI

Source: Eurostat (EU Labour Force Survey microdata) and ESCB calculations. Notes: EA11 excludes Portugal for which data are not available before 2008. Separation probabilities computed as percentage of the labour force each quarter. Separations include voluntary quits. Job finding rates are computed as a share of those unemployed in the previous quarter. 5

www.ecb.europa.eu ©

Job losses heavily concentrated by country, sector, skills and age Rubric Employment losses by sectors and worker groups (percentage change in employment between 2008 and 2013)

by sectors

by educational attainment

by age

Source: Eurostat and ESCB calculations. 6

www.ecb.europa.eu ©

Were young people “pushed out” of Rubric

employment by older workers?

Increased employment of older workers also reflects rising participation of this group following pension reforms in many countries…

But micro analysis shows no negative impact on young persons from higher employment propensities of older workers …perhaps even complementary?

Euro area participation rates by age groups

Estimation of the potential displacement effect of young workers by older workers.

(annual change in the participation rate, in percentage points) total

15-24

25-54

55-64

3

Dependent Variable Probability of being employed for a young person

Dependent Variable Probability of being unemployed for a young person

0.017

-0.0751*

Probability of young persons entering employment

Probability of young persons exiting employment

0.0393

0.0258

2 Employment rate (55-64)

1

0

-1 Change in employment rate (55-64) -2 2007

2008

2009

2010

2011

2012

2013

2014

Sources: Eurostat and ESCB calculations. 7

Notes: Eurostat and ESCB calculations. Besides the aggregate employment rate of senior workers, the other regressors are gender, age, education, aggregate employment rate for the 40-49 age group (expressed in changes for entry/exit), constant, fixed effects: region, year, country, year*country. Sample period: 2006-2012. www.ecb.europa.eu © * Indicates statistically significant at 10% level.

Nevertheless, the crisis resulted in a strong rise in youth unemployment. Rubric … as a consequence in part of “last in, first out” practices, lower human capital, and a higher propensity of temporary contracts among the young…particularly in the stressed countries…. …….but unemployment ratios give less dramatic picture… Unemployment rate (total and youth) in euro area, stressed countries and other euro area economies

Unemployment rate and unemployment ratio (as % of youth labour force and youth population), 2013

euro area - total

euro area - youth

stressed countries - youth

other countries- youth

unemployment rate

unemployment ratio

70

50 45

60

40

50

35 30

40

25

30

20 15

20

10

10

5 0

0 07Q1

08Q1

09Q1

10Q1

11Q1

12Q1

13Q1

14Q1

EA

Source: Eurostat and ESCB calculations. Note: Youth refer to the 15-24 age group.

GR ES IT CY PT SK IE FR BE LV SI FI EE LU MT NL AT DE

Sources: Eurostat and ESCB calculations. Note: the unemployment ratio is defined as the share of youth unemployed over the youth population. 8

www.ecb.europa.eu ©

Overview of Rubric

the presentation

1

Labour market dynamics and unemployment over the crisis

2

Labour market slack and structural unemployment

3

Wage adjustment

4

Concluding remarks: summary of main findings and policy implications

9

www.ecb.europa.eu ©

Discouragement and underemployment have increased considerably... Rubric Significant rise in: (1) discouraged workers and (2) underemployment (involuntary part-time workers who would like to work more hours). Implies labour market slack is considerably greater than suggested by headline unemployment rate… Broader measures of economic slack for the euro area (annual averages) PTER as % active labour force

(headline) unemployment rate

Discouraged as % active labour force

unemployment rate + part time for economic reasons unemployment rate + part time for economic reasons + discouraged

12 25

10

20

8

15

6

10

4

5

2

0

0 2008

2009

2010

2011

2012

2008

2013

2009

2010

2011

2012

2013

Source: Eurostat, EU Labour force survey and ECB calculations Notes: PTER refers to those working "part time for economic reasons". Active labour force includes discouraged workers, where appropriate.

10

www.ecb.europa.eu ©

Long-term unemployment has also risen and the Beveridge curve shows an outward shift Rubric Long term unemployment (>12m) has grown substantially and the euro area Beveridge curve shows a marked outward shift (statistically significant in econometric analysis)..… Movements in the euro area Beveridge curve

Long term unemployment in stressed and other euro area countries

(as % of labour force; limits to production from shortage of labour)

(as a percentage of total unemployment) 60

1

euro area

01Q1

stressed economies 55

12

0.9

other euro area countries

10 0.8

08Q1

50

45

Labour shortages

0.7 0.6 0.5

40

8 12Q1

02Q1

6

07Q1

14Q1 11Q1

0.4

35

13Q1

4

03Q1 04Q1

0.3 30 0.2

09Q1

2

10Q1

25

0.1 0

20 07 Q1

0 08Q1

09Q1

10Q1

11Q1

12Q1

13Q1

7

14 Q1

Source: Eurostat and ESCB calculations Notes: Long-term unemployment is defined as those without jobs for 12 months or more.

11

8

9

10

11 12 Unemployment rate

13

Source: Eurostat and ESCB calculations. Notes: vacancy developments approximated by DG-ECFIN series of labour shortages as reported by manufacturing employers. www.ecb.europa.eu ©

… due to increasing skill mismatch Rubric Skill Mismatch Index (SMI) suggests notable increase in skill mismatch over the course of the crisis, particularly in stressed economies Skill mismatch across stressed and other euro area countries

Skill mismatch in the euro area labour market (difference between skill distributions between employed workers vs. all labour force participants) euro area

country

(difference between skill distributions between employed workers vs. all labour force participants)

region

stressed countries

other euro area economies

0.16%

0.10%

0.14% 0.08%

0.12% 0.10%

0.06%

0.08% 0.04%

0.06% 0.04%

0.02%

0.02% 0.00% 2001

2003

2005

2007

2009

0.00%

2011

1998

2000

2002

2004

2006

2008

2010

2012

Source: Eurostat and ESCB calculations. . 12

www.ecb.europa.eu ©

Migration trends over the crisis helped to reduce labour market pressures…. Rubric Migration trends helped to alleviate occupational bottlenecks and demographic pressures (eg, declining national populations). Population growth and contributions, other euro area countries

Population growth and contributions, stressed countries (annual percentage growth)

(annual percentage growth) immigration growth nationals growth population growth

immigration growth nationals growth population growth

2.0%

0.6%

1.5%

0.4%

1.0% 0.2% 0.5% 0.0% 0.0% -0.2% -0.5% -0.4%

-1.0%

-0.6%

-1.5% 2006

2007

2008

2009

2010

2011

2012

2006

2013

2007

2008

2009

2010

2011

2012

2013

Source: Eurostat and ESCB calculations.

Source: Eurostat and ESCB calculations. 13

www.ecb.europa.eu ©

Strong increase in structural unemployment Rubric Structural unemployment estimates have risen substantially …..consistent with rising long-term unemployment and higher mismatch. Structural unemployment rate in the euro area

EC

IMF

OECD

25

actual unemployment rate

11

10

9

8

7 2000

the

crisis…

Evolution of the structural unemployment rate between 2008 and 2013

structural unemployment rate in 2013

12

since

ES

GR

20

PT

CY

15

IE

LV IT EEFR

SI

10

FI BE NL

5

LU AT

SK

euro area

DE MT

0 2002

2004

2006

2008

2010

0

2012

Source: EC, Eurostat, IMF, OECD and ESCB calculations.

5 10 structural unemployment rate in 2008

15

Source: European Commission Notes: the dark line is a trend line 14

www.ecb.europa.eu ©

Overview of Rubric

the presentation

1

Labour market dynamics and unemployment over the crisis

2

Labour market slack and structural unemployment

3

Wage adjustment

4

Concluding remarks: summary of main findings and policy implications

15

www.ecb.europa.eu ©

Response Rubric of wages to unemployment increased… While wage response was low over the first phase of the crisis, there has been some increase over the second phase, implying a ‘steeper’ Phillips curve compared to first phase... (annual percentage point changes; annual percentage changes) pre-crisis (2005-07)

Great recession phase

second recession phase

pre-crisis (2005-2007)

Great recession phase

second recession phase

change in compensation per employee

12

8

4

0 change in unemployment rate

-4

-8 -5

0

5

Source: Eurostat and ESCB calculations. 16

www.ecb.europa.eu ©

Downward Rubric wage rigidities declined over the crisis Rolling regression of the downward rigidity parameter (D*U)

Panel estimates of euro area wage equation (elasticities)

D*U

(1)

(2)

U

-0.465***

-0.411***

Δprod

0.782***

0.754***

ΔCPI

0.779***

0.777***

D*U

0.105***

0.194***

95% confidence interval

0.3

0.25

0.2

0.15

D*U*Trend

-0.008**

Constant

4.089

3.677

0.1

Observations R2

874 0.75

874 0.751

0.05

14

14

1.428 0.74

1.424 0.741

Nb of euro area countries SER Adj-R2

Sources: Eurostat and ESCB calculations. Notes: Dependent variable - nominal compensation per hour; sample period maximum data range 1991Q4-2013Q4.; unbalanced panel, fixed effects; Long-run parameters ; D*U*Trend begins in 2008Q1.

0 08Q1

09Q1

10Q1

11Q1

12Q1

13Q1

Sources: Eurostat and ESCB calculations.

D*U: lower wage response to unemployment during downturns (confirms downward wage rigidities). D*U*Trend: wages become more responsive to unemployment as crisis becomes more protracted (downward wage rigidities become weaker: reforms,17public sector pay constraint, threshold effects?). www.ecb.europa.eu ©

Overview of Rubric

the presentation

1

Labour market dynamics and unemployment over the crisis

2

Labour market slack and structural unemployment

3

Wage adjustment

4

Concluding remarks: summary of main findings and policy implications

18

www.ecb.europa.eu ©

Rubric Concluding remarks – summary of findings and policy implications • The crisis had uneven impacts on EA labour markets –

lower GDP fall in second EA recession, but almost as many job losses, mostly in stressed economies

• Labour market flows suggest sharp rise in unemployment in 2nd recession due to marked decline in job-finding probabilities

• Young and less-skilled most heavily employment continued to increase…

affected,

while

older

workers’

... but older workers did not displace younger workers More emphasis on reducing labour market dualism – focusing on reducing average adjustment costs - and enhanced ALMPs would help increase transitions to employment

• In wider labour market, discouragement and underemployment have risen implying considerably more slack than suggested by headline unemployment figures.

• Long-term unemployment and structural mismatch have increased, leading to strong rises in structural unemployment, particularly in stressed economies – –

more emphasis on increasing mobility will help alleviate high local unemployment. ALMPs are needed to help the displaced retrain for changing sectoral composition of employment.

• Downward wage rigidities reduced as crisis persisted –

but more wage and employment flexibility – and greater wage differentiation - required to accelerate adjustment/reallocation and reduce structural unemployment.

19

www.ecb.europa.eu ©

One final reminder from the US unemployment rate of Rubric

the need for reforms

Upward shocks to euro area unemployment tend to persist (rigidities), while reversed in the US (flexibility) Unemployment rate in the euro area and the United States 13 euro area

United States

trend (euro area)

trend (United States)

11

9

7

5

3 1980

1985

1990

1995

2000

2005

2010

Source: OECD, Area Wide Model Database (AWM, ECB Working Paper Series No 42) 20

www.ecb.europa.eu ©