Labour markets, social transfers and child poverty

Transcription

1 Labour markets, social transfers and child poverty Bruce Bradbury, Markus Jäntti and Lena Lindahl and

2 Objectives o Both earnings and social transfers are important determinants of the living standards of disadvantaged children and their families o How do the relative disposable incomes of the most disadvantaged fifth of children vary across nations? o What is the role of transfers, wages and hours of work in contributing to this? o How does this vary across the economic cycle?

3 Previous work o Our previous work on child poverty in the 1990s (B&J, 2001): Variations in market income explained more of the cross-national variation in child disadvantage than transfers o Chen and Corak (2008) (data for 1990s) and Bitler, Hoynes and Kuka (2014), Chzhen (2014) (recent crisis) Transfers often do not increase sufficiently (or at all) to mitigate the impacts of recessions. o Whiteford and Adema (2007): Low joblessness does not necessarily imply low poverty English-speaking: Higher benefits more joblessness Not the case for Nordic countries o Wage inequality Strong cross-national correlation with poverty, but Weak impact of minimum wage increases on poverty Wages transfers?

4 Methods o We examine the living standards of disadvantaged children, both in real terms and relative to the norm in their society o Dependent variable Mean income of the bottom fifth: Mean disposable income of the one-fifth of children in families with the lowest disposable income Real (PPP 2013 USD) Relative: Relative to the overall median income in their country. All incomes equivalised Scale = (Nadults+0.75*Nchildren)^0.8

5 Average incomes and poverty rates o Relative income of the bottom fifth is strongly correlated with relative child poverty o Real income of the bottom fifth is strongly correlated with (the logit of) poverty based on the US poverty line (r = in both cases) o For cross-national comparisons, we focus on relative income Indication of the degree of exclusion experienced by the poor. Relative to national capacity to intervene. o When looking at trends over time, both relative and real measures are relevant

6 LIS income data and decomposition o Initial decomposition M: Market income (wage, self-employment, capital, private transfer income) T: Net social transfers (Transfers minus income taxes and social insurance contributions) M + T = D (Disposable income) o Adjustments Income components truncated at zero Tax liabilities truncated at 90% of gross income o Problems Italy, Greece. Some transfers paid through the wage system => M too high and T too low. Less important: Belgium, France, Russia Slovenia. Taxes deducted from M => M too low, T too high

7 OECD wage inequality data o Better source of wage inequality data than LIS We use the OECD wage inequality data to decompose the LIS earning income inequality into that due to wage rates and that due to hours differences o Limitations Some cross-national definitional variations A cross-check with a sub-set of LIS data gives us some confidence that these are not substantial OECD wages are for all workers not just parents

8 Decomposition identity (all incomes relative to overall median income) D tc = M tc + T tc M tc = K tc + E tc E tc = w tc h tc E Q3tc where D tc = mean disposable income of bottom fifth M tc = mean market income of bottom fifth T tc = mean net social transfers of bottom fifth K tc = mean capital and private transfer income of bottom fifth E tc = mean earnings of bottom fifth E Q3tc = mean earnings of middle fifth w tc = w 10tc w50tc, wage rate of 10 th relative to 50 th percentiles h tc = E tc E Q3tc, (imputed) relative hours of bottom to middle w 50tc w 10tc

9 Difference from cross-national mean D tc D = T tc T + K tc K + (E tc E) T tc T + K tc K + E w tc w w + h tc h h + E Q3tc E Q3 E Q3

10 Relative poverty and mean relative income of bottom-fifth

11 Children in the bottom fifth: Average market income (M), net social transfers (T) and disposable income (D=M+T): All relative to median disposable income

12 Children in the bottom fifth: Average market income (M), net social transfers (T) and disposable income (D=M+T): Real 2013 USD

13 Trends since the 1990s: 2013 USD

14 Trends since the 1990s, income relative to median

15 Income components relative to median disposable income (countries with >1m children and non-missing w) D T M K E w h E Q3 Denmark Finland Sweden Norway Netherlands Ireland Czech Republic United Kingdom Germany South Korea Japan France Australia Poland Canada Greece United States Israel Italy Spain Mean Correlation with D Identities: D = T + M; D = T + K + E; D = T + K + w.h.e Q3

16 Decomposition summary Note:: Right-hand panel is a decomposition of E in left-hand panel

17 Conclusions: Cross-national variation o Market incomes and net transfers explain similar proportions of the cross-national variation in the relative living standards of the bottom fifth of children o Strong negative correlation between the two. Only partly due to variations across the cycle Labour supply issues? Transfer vs service policies? o English-speaking countries have similar (less than average) market incomes, but with substantial variation in transfers Relatively low employment hours in all, plus Relatively low wages in the US and Canada o Nordic countries have similar disposable income, but with variations in the market-transfer income mix

18 Conclusions: Changes over the last 2 decades o Relative income Most common: increases in transfers compensating for decreases in market income Especially when considered relative to the fall in median income o Real income trends after the 2008 financial crisis Dramatic decrease in Spain Substantial in Ireland Small decline in Australia, the US, the UK, Germany and the Netherlands Increase in Norway, Finland, Canada and Denmark