The Effects of Income Support Settings on Incentives to Work. Nicolas Hérault, Guyonne Kalb and Justin van de Ven

Transcription

1 The Effects of Income Support Settings on Incentives to Work Nicolas Hérault, Guyonne Kalb and Justin van de Ven

2 Objectives of research Key research question: What relationships are described by survey data between benefits policy changes and employment outcomes? Main objective is to improve the existing evidence base for setting the policy levers of income support allowances => Go beyond studies focused on specific policy reforms Focus is on labour force participation and hours of work given employment

3 Analytical approach Reduced form regression analysis of the relationships between income support policies and employment outcomes Innovation: Policy parameters tailored to the individual using a microsimulation model Exploit multiple policy variations observed over a 17-year period in Australia Separate regressions are estimated for six demographic groups: sole parents, single males, single females, couple men, couple women without kids and couple women with kids Effect of policy parameters on hours worked by wage rate quintile (to account for heterogeneous effects along the income distribution)

4 Average Relationship Between Market and Net Incomes, 1999 and 2007, singles Source: Herault & Azpitarte (ER 2015), Authors calculations based on MITTS and SIHC data

5 Policy parameters Three generic variables have been constructed to capture the key payment features of means-tested benefits for each individual: 1. Benefit rate: value of eligible transfer income if the individual doesn t work 2. Withdrawal threshold: the income the individual can earn before reaching the end of the withdrawal-free area 3. First withdrawal rate: the marginal effective tax rate when this first exceeds 10% 4. Average effective tax rate at 40 hours of work We also include dummy variables to account for discrete changes in activity requirements: 1996/97: tightening of activity requirements 2002/03: Mutual Obligations and Work for the Dole 2006/07; reduction in youngest child age for Parenting Payment

6 Control variables We also control for: Age, kids (by age band), education, State & region (rural or urban), ESB Non-employment income and spouse s employment status Macro conditions: unemployment rate and level of the minimum wage Those on DSP and DVA pensions are excluded

7 Labour force participation rates by age and year MEN

8 Labour force participation rates by age and year WOMEN

9 Summary statistics for population aged 25 to 54: income support parameters couple women no children single parents single men single women couple men no children with children net eligible transfer ($2009 per week) mean percentile percentile percentile % with positive net benefit income withdrawal threshold % with positive threshold of those with positive thresholds: mean percentile percentile percentile withdrawal rate mean percentile percentile percentile average effective tax rate at 40 hours mean percentile percentile percentile Source: authors' calculations based on SIH data and imputed taxes and benefits

10 Labour force participation Estimated marginal effects (in ppt) of income support variables on the probability of labour force participation for individuals aged 25 to 54 Marg. Eff. Std. Err. Marg. Eff. Std. Err. Marg. Eff. Std. Err. single parents single men single women Number of observations Net eligibility for transfers (a) -0.06*** *** *** Withdrawal-free threshold (a) $1 to $30/week 11.52*** *** *** $30 to $50/week 7.81* *** *** $50 to $80/week *** *** > $80/week First withdrawal rate (1 to 100) *** *** Av. Eff. Tax Rate at 40hrs (1 to 100) -0.20*** *** *** married women married men no children with children Number of observations Net eligibility for transfers (a) 0.003** ** *** Withdrawal-free threshold (a) $1 to $30/week 3.94*** *** *** $30 to $50/week 2.31*** *** ** $50 to $80/week -4.19*** > $80/week *** *** *** First withdrawal rate (1 to 100) -0.03*** *** Av. Eff. Tax Rate at 40hrs (1 to 100) -0.02*** *** Post 1997/8 dummy (b) *** ** Post 2002/3 dummy (b) (c) Post 2007/8 dummy (b) (d) *** (a) Expressed in 2009 dollars per week

11 Estimated coefficients for tobit regressions of hours of employment (age 25 to 54) single single single couple couple women parents men women men no child w. children net eligibility for transfers (g) 1 quintile 1 (h) * 0.040*** 0.034*** *** ** 2 quintile 2 (h) *** 0.039*** ** 3 quintile 3 (h) *** 0.053*** *** 4 quintile 4 (h) *** 0.029*** 0.025** *** 5 quintile 5 (h) ** 0.034*** 0.020** * ***

12 Estimated coefficients for tobit regressions of hours of employment (age 25 to 54) single single single couple couple women parents men women men no child w. children net withdrawal eligibility free for threshold transfers (g) 1 quintile 1 (h) * *** *** 0.042*** *** *** * ** 2 quintile 2 (h) *** *** *** *** quintile 3 (h) *** *** *** *** quintile 4 (h) *** *** ** * *** 0.009** 5 quintile 5 (h) ** 0.044*** ** 0.020* * ** 0.016** *** 0.005

13 Estimated coefficients for tobit regressions of hours of employment (age 25 to 54) single single single couple couple women parents men women men no child w. children net first eligibility withdrawal for rate transfers (1 to 100) (g) 1 quintile 1 (h) * *** *** * *** 0.029** * ** 2 quintile 2 (h) *** 0.037** 0.051*** *** quintile 3 (h) *** *** *** quintile 4 (h) *** *** 0.043** 0.025** *** quintile 5 (h) ** *** 0.045** 0.020* ** *** 0.014

14 Estimated coefficients for tobit regressions of hours of employment (age 25 to 54) single single single couple couple women parents men women men no child w. children Average net eligibility effective for transfers tax rate at (g) 40 hours (1 to 100) 1 quintile 11 (h) (h) * *** *** *** * 0.028** ** 2 quintile quintile 22 (h) (h) *** 0.047*** *** *** 3 quintile 3 (h) *** 0.051*** *** quintile 3 (h) ** * 4 quintile 4 (h) *** 0.029*** 0.025** *** quintile 4 (h) *** 0.089* *** *** 5 quintile 5 (h) ** 0.044*** 0.020* ** *** quintile 5 (h) ** *** ** ** ***

15 Summary and conclusions Benefit rates (maximum eligible benefits): Limited effect (in magnitude) on LFP Negative for sole parents and couple women Positive for singles and couple men Limited and largely negative effect on hours worked for sole parents and couple members Large positive associations with hours worked for singles that extends throughout the earnings distribution

16 Summary and conclusions Withdrawal-free threshold: Labour force participation Sizeable effects Expected positive association (particularly large for singles and sole parents) but only up to $50/week Negative effect beyond for all groups except for sole parents Hours worked: Most individuals with wage rates in the lower half of the distribution tend to work fewer hours when subject to higher first withdrawal threshold Significant association only for singles Effects diminish at high incomes

17 Summary and conclusions Withdrawal rate: Negative effect on labour force participation Large for singles and couple women with no kids Only weak associations with hours worked Average effective tax rate at 40 hours of work Systematic negative effect on LFP Particularly large for singles and sole parents Negative associations with hours worked Effects tend to increase at higher wages

18 Summary and conclusions Activity requirements: Evidence of positive effects of 1996/97 tightening on LFP 2002/03 reforms (Mutual obligation and Work for the Dole) only associated with increased LFP for sole parents 2006/07 reform of Parenting Payment associated with reduction of LFP => grandfathering dominates Largely non-significant associations with hours worked

19 The Effects of Income Support Settings on Incentives to Work Nicolas Hérault, Guyonne Kalb and Justin van de Ven

20 Summary statistics for population aged 25 to 54: individual characteristics couple women single parents single men single women couple men no children with children sample (observations) 4,989 8,676 6,344 24,827 10,092 18,739 age band of individual number of dependent children aged aged aged aged aged born in Australia arrived in Australia within last 15 years full-time study part-time study university graduate other post-school qualifcations receipt of investment income receipt of child support employed spouse non-employment income ($2009 pw) (b) umemployment rate (%) (c)

21 Estimated coefficients for tobit regressions of hours of employment (age 25 to 54) single single single couple couple women parents men women men no child w. children net eligibility for transfers (g) 1 quintile 1 (h) * 0.040*** 0.034*** *** ** 2 quintile 2 (h) *** 0.039*** ** 3 quintile 3 (h) *** 0.053*** *** 4 quintile 4 (h) *** 0.029*** 0.025** *** 5 quintile 5 (h) ** 0.034*** 0.020** * *** withdrawal free threshold (g) quintile 1 (h) *** *** quintile 2 (h) *** quintile 3 (h) *** quintile 4 (h) * ** quintile 5 (h) ** * 0.016** first withdrawal rate (1 to 100) quintile 1 (h) * 0.029** quintile 2 (h) ** quintile 3 (h) quintile 4 (h) ** quintile 5 (h) ** Average effective tax rate at 40 hours (1 to 100) quintile 1 (h) ** quintile 2 (h) *** quintile 3 (h) ** * quintile 4 (h) *** 0.089* *** *** quintile 5 (h) ** *** ** ** ***

22 Estimated coefficients for tobit regressions of hours of employment (age 25 to 54) couple couple women men no child w. children post 1997/8 and positive transfer (i) quintile 1 (h) quintile 2 (h) quintile 3 (h) quintile 4 (h) * quintile 5 (h) post 2002/3 and positive transfer (i) quintile 1 (h) quintile 2 (h) quintile 3 (h) * quintile 4 (h) quintile 5 (h) post 2007/8 and positive transfer (i) quintile 1 (h) quintile 2 (h) quintile 3 (h) quintile 4 (h) quintile 5 (h)