Education Effects of Unemployment and the Transition to Parenthood in Germany and the UK

Education Effects of Unemployment and the Transition to Parenthood in Germany and the UK Christian Schmitt Rostock University & DIW Berlin Education and Reproduction in Low Fertility Settings Wittgenstein Centre, 2-4 Dec. 2015, Vienna Christian Schmitt Education effects of unemployment 1

Study Focus Impact of unemployment on the transition to first-birth risk across educational groups in Germany and the UK How does educational attainment moderate the impact of unemployment? Gender specific similar or opposing effects? Universal effect or cross country heterogeneity differences due to cultural influence and social policy settings? Christian Schmitt Education effects of unemployment 2

Cultural & Institutional Orientation Germany (continental conservative): LM: Strong state and union role Generous unemployment benefits Low labor-market turnover High incidence of long-term UE. Generous & lasting leave Affordable childcare with low levels of acceptance of ECEC Traditional breadwinner model Strong norms of maternal care Encouragement of mothers to retreat from the labour market & strong male breadwinner model UK (anglo-american liberal): Labour market widely deregulated Limited & means tested UE. support High labor-market turnover Low incidence of long-term UE. Need to pursue a dual earner model High cost of a widely privatized childcare system Modified breadwinner model Strong norms of maternal care Mothers as carer and adult workers == double burden & a challenged male breadwinner 3

How does unemployment affect family formation plans? Opportunity costs of parenthood (Mincer) Increased likelihood to start a family (women/higher educated) Human capital hypothesis (Mincer, Becker) Avoidance of human capital depreciation (women/higher educ.) Breadwinner/homemaker hypothesis UE undermining the (male) breadwinner role (Hochschild 1989), (primarily men) Compensation hypothesis Compensation of loss in social esteem/status (Lindenberg), family form. as uncertainty reduction (Friedmann et al. 1994), (primarily women, lower educated) 4 4

Data & Methods Cohorts 1964-1981 Time of analysis 1991-2008 Semi-balanced Panel Piecewise-constant exponential hazard model Process time starts with 16th year of life (month 180) Database Germany: n DE = 7.789 Socio-economic panel study (SOEP) Events DE = 2.636 Database UK: n UK = 4.413 British Household panel study (BHPS) Events UK = 1.595 5

Data & Methods Figure 1: Kaplan Meier survival estimates of transition to first birth UK and Germany (cohorts 1955 to 1985) 1.00 1.00 0.75 0.75 0.50 proportion 0.50 0.25 0.25 0.00 0.00 10 20 30 40 Analysis time: age in months (displayed in years) 10 20 30 40 Analysis time: age in months (displayed in years) Women Men Women Men UK n = 6.620 (3.093 events) Germany n = 14.366 (7.280 events) Source: SOEP and BHPS 1991-2008, biographical data, own calculations 6

Data & Methods.008.006.004 Hazard-Rates of first birth decisions: Women in cohorts 1964-1981 Approximation of a normally distributed baseline hazard by a piecewise-constant exponential model.002 0 10 20 25 30 35 40 50 Years of age (estimates based on time in months) UK Germany Model specification: X Z t t 1 X Z t t 2 3 X Z t t 4 X Z t t exp 192; 252 1 1 1 exp 253;312 2 2 2 t exp X Z t t 313;396 3 3 3 exp 397; 456 4 4 4 exp 457;540 5 5 5 5 7 7

List of covariates Choice Model: Preferences for having children / having good job Activity status, incl. unemployment duration Educational attainment Type of relationship Net individual income & benefits Controls (year, location of residence, etc.) Interaction terms of UE. and education t n-m Choice t n birth Focus of Analysis 8 8

Results First-birth decision hazards ( exp(b) ) Educational Attainment Comprehensive schooling Germany Germany UK UK 1.01 1.11 1.25 + 1.24 O-Level degree 1 1 1 1 A-Level degree 0.62*** 0.81** 0.90 0.77** Academic degree 1.19* 1.04 0.94 0.74** n of cases/events 4161 / 1193 3628 / 1443 2366 / 745 2047 / 850 Sources: SOEP & BHPS 1991-2008; piecewise constant exponential hazard p < 0.10 ( + ), p < 0.05 (*), p < 0.01 (**), p < 0.00 9 9

Results First-birth decision hazards ( exp(b) ) Activity Status Germany Germany UK UK Full-time emply. 1 1 1 1 Part-time emply. 0.48** 0.97 0.48 + 3.90*** In education 0.50*** 0.29*** 0.17*** 0.29*** Unemployed 0.81 1.40** 0.65 3.83*** Frictional UE 1 0.81 1.41*** 0.67 3.92*** UE > 3 months 1 0.73** 1.51*** 0.83 3.13*** 1) based on separate model estimates Sources: SOEP & BHPS 1991-2008; piecewise constant exponential hazard p < 0.10 ( + ), p < 0.05 (*), p < 0.01 (**), p < 0.00 10 10

Results First-birth decision hazards ( exp(b) ) Interaction terms Germany Germany UK UK Log-UE duration *low edu. Log-UE duration *A-level Log-UE duration *academic degree Interaction sets based on separate model estimates 1.01+ 1.00 1.03*** 1.04*** 0.96 1.01 1.01 1.03*** 1.02 0.99 0.94 1.02 + UE*low edu. 1.20 1.28** 1.95*** 2.77*** UE*A-level 0.57 1.17 (+) 1.10 2.79*** UE *academic 1.52 1.16 (+) 0.23 1.90* Sources: SOEP & BHPS 1991-2008; piecewise constant exponential hazard p < 0.10 ( + ), p < 0.05 (*), p < 0.01 (**), p < 0.00 14 14