EUROPEAN DATA CENTER FOR WORK and WELFARE

EUROPEAN DATA CENTERFOR

WORK and WELFARE

your direct access to

EUROPEAN COMPARATIVE DATA

- opinions -- indicators and statistics -- policies and institutions -

WWW.EDACWOWE.EU

Institutions or Market Forces?

Employment insecurity among European workers in a period of financial crisis

Heejung ChungWim van Oorschot

Seminar "The financial crisis, welfare state challenges and new forms of risk management"

Aalborg, DK2-3 March, 2011

BACKGROUND

The flexicurity agenda of the European Commission

- The balance between flexibility and security is key to success for the future of the European social-economy

- A belief in the power of institutional arrangements to regulate the flexicurity balance, i.e. flexicurity policies do matter

The financial crisis

- Does economic depression lead to a larger imbalance between overall flexibility and overall security irrespective of flexicurity policies?

- Or, how strong are institutions against market forces: do they make a difference in times of economic depression?

- What is the gain of flexicurity policies, i.e. of a European way of organizing the social-economy, when the economy is down? And what are the costs?

Interesting questions, but…

ANALYTIC PROBLEMS and SOLUTIONS

Limits- No data about the actual balance between flexibility and security in countries- Let alone such data before, during and after the financial crisis

Our partial contribution- Focus on the security aspect of flexicurity: subjectively experienced employment security by workers- Are employment security differences between European workers influenced more by institutional or by economic differences between the countries they live in?

RESEARCH QUESTION

What is the relative contribution of institutional and economic context factors to the explanation of differences in feelings of employment insecurity among workers in European countries?

HYPOTHESES

Workers will be less likely to perceive employment insecurity in countries with:

Institutional factors1. stronger active labour market policies2. stronger passive lmp (unemployment benefits)3. stronger EPL (workers with permanent contracts only)

Economic market factors4. higher average unemployment rate 2004/7

5. lower average employment rate 2004/7

6. higher ue increase in mid 2008 to mid 20097. lower average GDP growth 2004/8

8. lower GDP growth 2008-2009

HYPOTHESES

Composition effects (individual level control variables)

Employment insecurity higher among:

- women, lower educated, with income from benefit, with minority background, without in-job training experience, with previous unemployment experience, with health problems, with no partner in paid work, with children, not being a union member, not having a permanent contract, with part-time contract, having no influence at work, working long hours, working in smaller firms, working in the private sector

DATA and METHODS

Data- European Social Survey 2008, wave 4- N=22 countries- people in employment < 65 years of age: N=20.809

Dependent variable“How likely is it that during the next 12 months you will be unemployed and looking for work for at least four consecutive weeks?” 1= likely + very likely, 0=else

Independent context variables ALMP and PLMP, 2008: national spending as % GDP: EurostatEPL: OECD, 2008UE rates, employment rates, GDP growth: Eurostat

Analysis random intercept, multi-level logistic regression

FINDINGS

0%

10%

20%

30%

40%

50%

60%

70%

80%

NO NL FI DK SE GB DE BE CY SK SI FR ave PL PT ES HU GR

RO CZ EE BG LV

Perceived employment insecurity by country % of employed people

FINDINGS

Percentage of variation in Y at country level

Total (ICC; empty- or null-model): 17%

Composition effects: 2%

Other context factors: 15%

Now, which other context factors play a role?

FINDINGS

Model B SE R2 a

0-1 PLMP expenditure/unemployment rate -4.005 1.018 42.0

0-2 ALMP expenditure/unemployment rate -6.751 1.921 36.5

0-3 LMP expenditure/unemployment rate (ALMP+PLMP) -2.622 0.671 41.7

0-4 EPL for regular workers 0.166 0.266 2.2

0-5 EPL for temporary workers 0.131 0.157 3.9

0-6 Unemployment rate average 0.100 0.060 11.3

0-7 Employment rate average -0.072 0.022 32.2

0-8 Change in unemployment 2008 - 2009 0.155 0.049 31.8

0-9 GDP growth rate average 0.251 0.090 26.6

0-10 GDP growth rate for 2009 -0.121 0.034 36.8

‘Bi-variate’ effects of context vars on perceived employment insecurity(controlled for composition effects)

B in italics: P < .10 , B in bold: P< .01a Percentage explained variance at country level

FINDINGS

But what about multi-variate effects?

Due to country N of 22 we cannot include all context vars in a single model, so three steps:

1. Model with institutional vars2. Model with economic vars3. Model combining vars with significant effects in 1 and 2

FINDINGS

Multi-variate effects of institutional context variables(controlled for composition effects)

B in italics: P < .10 , B in bold: P< .01a Percentage explained variance at country level

Employment Insecurity

Model 1

B SE

PLMP exp as a % of GDP

ALMP exp as a % of GDP

LMP exp as a % of GDP (= ALMP+PLMP: multi-collinearity) -2.091 0.535

EPL for regular workers -0.041 0.204

EPL for temporary workers 0.113 0.117

EPLreg* permanent 0.184 0.086

variance level 2 0.190 0.068

R2 a 50.1%

N level 2 18

N level 1 15508

FINDINGS

Multi-variate effects of market context variables(controlled for composition effects)

B in italics: P < .10 , B in bold: P< .01a Percentage explained variance at country level

Employment InsecurityModel 2

B SE

Employment average -0.077 0.014

Unemployment change 08-09

GDP growth average

GDP growth 08-09 -0.128 0.023

variance level 2 0.156 0.050

R2 (from model2)a 73.5%

N level 2 22

N level 1 17936

FINDINGS

Multi-variate effects of institutional and market context variables(controlled for composition effects)

B in italics: P < .10 , B in bold: P< .01a Percentage explained variance at country level

Employment Insecurity

Model 3

B SE

LMP exp as a % of GDP -0.148 0.092

EPL for regular workers

EPLreg* permanent

Employment rate average -0.067 0.015

GDP growth rate for 2009 -0.116 0.023

variance level 2 0.139 0.045

R2 (from model2)a 76.5%

N level 2 22

N level 1 17936

CONCLUSIONS

1. Differences in employment insecurity between individual workers depend for about 15% on the country they live in.

2. Workers are more insecure in countries with longer-term poorer labour markets and in countries with a stronger recent economic downturn

3. Differences in spending on active or passive labour market policies, as well as differences in EPL, do not impact much on feelings of employment insecurity

DISCUSSION

1. It may be that the effect of GDP Growth 2009 is time specific (related to this financial crisis)

2. Policies may have an indirect effect on insecurity, through their effect on employment , on human capital of workers, etc.

3. Policies may have an impact on objective measures of employment security

4. Market forces or policies? Need for longitudinal data