Unemployment and economic crisis: stress testing the welfare systems in Europe Mariña Fernandez, Francesco Figari, Holly Sutherland and Alberto Tumino Università dell Insubria and ISER University of Essex EUROMOD RESEARCH WORKSHOP 13 th October 2011
Some facts Increase in unemployment rate during the crisis (+3% between 2007 and 2010 in OECD countries) Crisis affects unemployment even after GDP recovers ES IT UK EU27 2008 0.9-1.3-0.1 0.5 GDP Growth rate 2009-3.7-5.2-4.9-4.2 2010-0.1 1.3 1.3 1.8 2008 11.3 6.7 5.6 7.1 Unemployment rate 2009 18 7.8 7.5 9 2010 20.1 8.4 7.8 9.6
Stress testing the welfare system Current indicators might have serious difficulties in capturing the effects of the crisis (Nolan 2009, Aaberge et al. 2000) Timeliness of micro data To assess the extent to which current benefits provide an effective safety net for the individuals and families affected by the crisis and to assess the cost to the government (Atkinson 2009; Figari et al. 2010, 2011) Stress test is a way to assess the extent to which welfare systems in the EU cushion the effects of loss of market income offering relative protection preventing to falling below an absolute income threshold
Approach Identification of covariates associated with unemployment (country specific LFS micro data) and matching LFS-EUROMOD dataset in order to identify in the latter those more likely to become unemployed during the crisis Use EUROMOD to: Take into account predicted changes in market income due to unemployment Calculate new tax liabilities and benefit entitlements; and hence new household disposable income Analysis We focus on the situation of people most likely to become unemployed, not the distribution as a whole. Analysis at individual level. It is not a forecasting exercise!
Coverage of this analysis Countries: ES, IT, UK Data Spain: Italy: UK: EM: 2007 EU-SILC Eurostat UBD, 2006 incomes updated to 2009 LFS: Encuesta de Pobolaciòn Activa EM: 2007 IT-SILC, 2006 incomes updated to 2009 LFS: Rilevazione sulle Forze di Lavoro EM: 2008/9 Family Resources Survey LFS: UK Labour Force Survey Policy year: 2009 (Analysis to be extended to BE, EE and SE)
Unemployment associated covariates and matching Unemployment transition between 2009 and 2010 (LFS data) Select people currently in work in EUROMOD database with similar characteristics of the new unemployed as observed in LFS Coarsened exact matching (Iacus, King and Porro, 2011) Matching based on determinants of unemployment (age, gender, education, citizenship, economic activity, industry, region of residence) plus number of household members (adult and children), number of employed people in the household The exact matching, based on coarsened characteristics, allows to respect their joint distribution. Matching quality overall imbalance ranges from.37 to.55 showing a good overall balance of observed characteristics. Means of observed characteristics in the two groups (LFS vs EM) are not statistically different for most of the observed characteristic
Unemployment risk Spain Italy UK Male 0.007-0.069*** 0.235*** Age 25-34 -0.351*** -0.224*** -0.222*** 35-44 -0.577*** -0.470*** -0.317*** 45-54 -0.733*** -0.558*** -0.364*** 55+ -1.022*** -0.740*** -0.456*** Education Upper secondary -0.223*** -0.048* -0.084*** Tertiary -0.258*** -0.027-0.196*** National -0.353*** -0.220*** -0.077* Self employed -0.602*** -0.249*** -0.114*** Secot of activity Industry -0.190*** 0.200*** 0.465*** Services -0.366*** 0.013 0.382** Constant -0.364*** -1.220*** -1.934*** No. Observations 68041 55647 44300 Pseudo R 2 0.1365 0.0815 0.0374
Unemployment welfare systems ES Scheme Duration Subject to Tax / SICs UB Insurance 60-70 % previous earnings 4 months to 2 years yes / yes IT UB Insurance 40-60% previous earnings, ceilings8 to 12 months yes / no UK UB Insurance Flat-rate ( 61 to 76 pw) 6 months yes / no
Unemployment welfare systems ES Scheme Duration Subject to Tax / SICs UB Insurance 60-70 % previous earnings 4 months to 2 years yes / yes UB Assistance Means tested at individual level 6 to 18 months yes / no IT UB Insurance 40-60% previous earnings, ceilings 8 to 12 months yes / no UK UB Insurance Flat-rate ( 61 to 76 pw) 6 months yes / no UB Assistance Means tested at family level Unlimited no / no
Unemployment welfare systems ES Scheme Duration Subject to Tax / SICs UB Insurance 60-70 % previous earnings 4 months to 2 years yes / yes UB Assistance Means tested at individual level 6 to 18 months yes / no Social Assistance Regional schemes, means tested at family level IT UB Insurance 40-60% previous earnings, ceilings 8 to 12 months yes / no UK UB Insurance Flat-rate ( 61 to 76 pw) 6 months yes / no UB Assistance Means tested at family level Unlimited no / no Social Assistance Means tested at family level Unlimited
Indicators Relative Welfare Resilience Relative Welfare Resilience Indicator: ratio of household disposable income after and before the shock (i.e. Equivalent to Net Replacement Rate) Welfare Compensation rate: ratio of difference in net benefits and difference in original income before and after the shock Absolute Welfare Resilience Risk of falling below a low absolute income threshold (poverty line. Equivalent to change in poverty status with fixed threshold)
Average Relative Welfare Resilience Indicator Spain Italy UK All with UBs 0.753 0.647 0.576 without UBs 0.480 0.423 0.552 Sole earner households with UBs 0.706 0.501 0.500 without UBs 0.434 0.175 0.476
Average RWRI by tax-benefit component
Average RWRI by baseline income quintile (with UB) 1 0.8 Spain 1 0.8 Italy 0.6 0.6 0.4 0.4 0.2 0.2 0 0-0.2-0.2-0.4 Bottom 2nd 3rd 4th Top -0.4 Bottom 2nd 3rd 4th Top 1 0.8 0.6 UK Social Assistance Unemployment benefits 0.4 Other benefits 0.2 Taxes and contributions 0-0.2-0.4 Bottom 2nd 3rd 4th Top Original income RWRI
Average Welfare Compensation rate Spain Italy UK All with UBs 0.658 0.541 0.425 without UBs 0.267 0.253 0.381 Sole earner households with UBs 0.684 0.539 0.508 without UBs 0.360 0.235 0.480
Average Welfare Compensation rate by baseline income quintile (with UB) 0.8 Spain 0.8 Italy 0.6 0.6 0.4 0.4 0.2 0.2 0 Bottom 2nd 3rd 4th Top 0 Bottom 2nd 3rd 4th Top 0.8 UK Social Assistance 0.6 Unemployment benefits 0.4 Other benefits 0.2 Taxes and contributions 0 Bottom 2nd 3rd 4th Top Compensation rate
Absolute resilience (with UBs) Are the new unemployed protected from falling below the poverty line? 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 protected at risk poor in work 0.1 0 All Sole earners All Sole earners All Sole earners Spain Italy UK
Absolute resilience (without UBs) Are the new unemployed protected from falling below the poverty line? 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 protected at risk poor in work 0.1 0 All Sole earners All Sole earners All Sole earners Spain Italy UK
Conclusions Differences across countries in the impact of economic downturn on unemployment risk and in the protection offered by the welfare systems Other household income play a key role in protection. Net public support flat across the distribution (ES and IT) and focused at the bottom (UK) Welfare state has a complementary role, more remarkable in ES (with UBs) and the UK (for the poorest; without UBs)