Poverty and household spending in Britain Mike Brewer Alissa Goodman Andrew Leicester Institute for Fiscal Studies 17th May 2006 Income and Expenditure Poverty Andrew Leicester Motivation • “… if we don’t raise the standard of living of the poorest people in Britain we will have failed as a government.” Tony Blair, 1997 • Poverty debate has focused on income as financial measure of living standards • Expenditure provides alternative / complementary measure Measuring living standards • Typical focus on income • Correlation with welfare indicators • Government targets • Government policy • Expenditure may have advantages • Well-being depends on what we consume, not what money we bring in • Saving and borrowing (smoothing) Smoothing out living standards • Short-term variability of incomes • Unemployment / Illness • Bonuses / Windfalls • Self-employed have more volatile incomes • Variability across life-cycle • Student loans • Pensioners run down accumulated assets • Maybe spending reflects longer-term inequalities Measuring living standards • Typical focus on income • Correlation with welfare indicators • Government targets • Government policy • Expenditure may have advantages • Well-being depends on what we consume, not potential consumption • Saving and borrowing (smoothing) • Incomes for poorest households badly measured? • But also disadvantages • Measurement problems • Expenditure versus consumption (e.g. durables) Key questions • What would poverty story of recent years be had focus been on expenditure? • Why might the two tell us different things? • Does giving poor people more money translate into higher expenditure, and on what? Measuring poverty • Measure expenditure poverty in same way as income poverty in annual Households Below Average Income publication • Define household as ‘poor’ if its income or spending below 60% of contemporary median (“relative poverty”) Data • Wish to obtain best available data on household incomes and expenditures • Construct measures which are as conceptually similar as possible • Expenditure • Family Expenditure Survey / Expenditure and Food Survey 1974 – 2002/3 • Weekly household spending, equivalised, real terms • No housing (rent, mortgage, local tax) • Income • Households Below Average Income 1961 – 2004/5 (net income after housing costs) • All reported at weekly household level, equivalised and in real terms Poverty lines (2002/3) Weekly Income Weekly Spending Childless couple £173 £158 Single £95 £87 Couple + 8-year-old £213 £194 Couple + 1- and 3-year-old £216 £197 Lone parent + 8-year-old £135 £123 Lone parent + 1- and 3-year-old £138 £126 Growth rates 8% Income 1979 - 1990 Average annual growth rate 7% 6% Spending 1979 - 1990 5% 4% 3% 2% 1% 0% -1% -2% Poorest 2 3 4 5 Decile 6 7 8 9 Richest Growth rates 8% Income 1979 - 1990 Spending 1979 - 1990 Income 1996/7 - 2002/3 Spending 1996/7 - 2002/3 Average annual growth rate 7% 6% 5% 4% 3% 2% 1% 0% -1% -2% Poorest 2 3 4 5 Decile 6 7 8 9 Richest Poverty rates 2002/3 Overall Self-employed Jobseeker Pensioner couple Single pensioner Income Spending poverty rate (%) poverty rate (%) Poverty rates 2002/3 Overall Self-employed Jobseeker Pensioner couple Single pensioner Income Spending poverty rate (%) poverty rate (%) 21.8 22.2 Poverty rates 2002/3 Income Spending poverty rate (%) poverty rate (%) Overall 21.8 22.2 Self-employed 22.6 12.6 Jobseeker 70.1 50.4 Pensioner couple Single pensioner Poverty rates 2002/3 Income Spending poverty rate (%) poverty rate (%) Overall 21.8 22.2 Self-employed 22.6 12.6 Jobseeker 70.1 50.4 Pensioner couple 23.1 29.3 Single pensioner 19.7 42.1 Poverty rates, 1961 – 2004/5 30 25 Income Expenditure Poverty rate (%) 20 15 10 5 0 1961 1964 1967 1970 1973 1976 1979 1982 1985 1988 1991 1994 1997 2000 2003 Child poverty, 1961 – 2004/5 35 Poverty rate (%) 30 Income Expenditure 25 20 15 10 5 0 1961 1965 1969 1973 1977 1981 1985 1989 1993 1997 2001 Pensioner poverty, 1961 – 2004/5 45 40 35 Poverty rate (%) 30 25 20 15 10 Income Expenditure 5 0 1961 1964 1967 1970 1973 1976 1979 1982 1985 1988 1991 1994 1997 2000 Non-pensioner, no-children poverty, 1961 – 2004/5 20 18 Poverty rate (%) 16 Income Expenditure 14 12 10 8 6 4 2 0 1961 1965 1969 1973 1977 1981 1985 1989 1993 1997 2001 Poverty trends • Rose sharply in later 1980s, particularly for income • Income poverty rates stablised in early 1990s and fell in latter 1990s as policy measures enacted • Expenditure poverty risen continually • Recent falls in child income poverty not matched by falls in child expenditure poverty • Pensioner income poverty pro-cyclical except recently • pensioners now less likely to be income poor than non-pensioners • Pensioner spending poverty much more stable, higher and not falling • Other groups have seen rises on both measures Why do trends differ? • Some possibilities: • Low income and low spenders are different people • Only half of those income poor are also spending poor • Increased means-tested benefits are targeted at low income households, not low spending households • Suggests spending rises due to benefit increases may be reflected higher up the income distribution • Changes in savings behaviour • Low income households may not be spending all their new income • Would happen if uncertainty over how permanent income changes are likely to be • Middle income households maybe reducing savings or increasing borrowing by more than poorer households Conclusions • Living standards have risen whether measured by income or spending • Income growth particularly strong for poorer people • Expenditure growth strongest for higher spenders • Relative position of poor improved if measured on incomes, worsened if measured on spending • Shows up in increased expenditure poverty rate even as income poverty declined • Reasons for different trends not yet clear • Expenditure poverty ought to be monitored alongside other indicators The effect of increased benefit entitlements on pensioners’ spending Mike Brewer Motivation • State benefits for 60+ risen under Labour, yet spending poverty of pensioners little changed • Have extra benefits improved pensioners’ living standards? • Related work • Meyer and Sullivan (2004) [US data, lone parents] • Gregg, Waldfogel and Washbrook (2004, 2006) [UK data, families with children] • Blow, Walker and Zhu (2005), [UK data, families with children] • Munro, Walker and Zhu (ongoing) [UK data, winter fuel allowance] Outline • • • • Policy changes affecting pensioners Method and data Results Conclusions Main benefit changes affecting pensioners since 1997 • Rise in basic state pension (April 2001 & 2002) • Increases in means-tested benefits (from April 1999) and introduction of pension credit (from 2003) • Equalisation of pensioner premia in meanstested benefits (by 2001) • Winter fuel allowance (from 1999) Changes to benefit entitlement for single pensioners (1996=1) 1.3 1.25 "Poor", 60-74 1.2 "Poor", 75-79 1.15 1.1 "Poor", 80+ or on disability benefits 1.05 1 "Rich" 19 96 19 97 19 98 19 99 20 00 20 01 20 02 0.95 Graph shows maximum entitlement to IS or BSP for single pensioner Overview of method • Compare (changes in) spending of pensioners affected by rise in benefits to pensioners not affected • Attribute difference to policy. • Called “conditional difference-in-differences”. Difference-in-differences: data • FES/EFS from 1996/7 to 2002/3 • Single adults born before April 1936 • Aged 60+ in 1996, 66+ in 2002 (pseudo-panel) • 3,056 “poor” pensioners (entitled to a meanstested benefit under 1996/7 system) • 1,281 “young” & 1,775 “old” • 1,778 “rich” pensioners (not entitled to a meanstested benefit under 2002/3 system) • Some pensioners omitted entirely (neither “poor” nor “rich”) Changes in benefit entitlements, income and spending, 1996/7-2002/3 35% 30% Rich 25% Young poor 20% Middle poor Old Poor 15% 10% 5% 0% Benefit entitlement Income Spending Difference-in-differences: overview • • • • • • Compare spending before and after rise in means-tested benefits (April 1999) Rich pensioners tell us about general trends affecting pensioners: B-A Poor pensioners tell us about general trends and impact of policy: D-C. Difference tells us about impact of policy: (D-C) – (B-A) • Assumes “common trends” Control for various factors (regression-adjusted DiD) Also compare “young” and “old” “poor” pensioners Mean spend 4/96 4/99 to to 3/99 3/03 Rich A B Poor C D Impact of benefit changes on pensioners Impact of policy on log( . ) Income April 1999 MIG (Poor vs Rich) 0.112 *** April 2001 eq’n (Young vs Old) 0.016 Spending (nonhousing) Spending on basics 0.096 *** 0.097 * 0.030 *** -0.008 Spending on non-basics 0.149 *** 0.184 *** * = significant @ 10% *** significant @ 1% Impact of benefit changes on pensioners Impact of policy on log( . ) Income April 1999 MIG (Poor vs Rich) 0.112 *** April 2001 eq’n (Young vs Old) 0.016 Spending (nonhousing) Spending on basics 0.096 *** 0.097 * 0.030 *** -0.008 Spending on non-basics 0.149 *** 0.184 *** * = significant @ 10% *** significant @ 1% Impact of benefit changes on pensioners Impact of policy on log( . ) Income April 1999 MIG (Poor vs Rich) 0.112 *** April 2001 eq’n (Young vs Old) 0.016 Spending (nonhousing) Spending on basics 0.096 *** 0.097 * 0.030 *** -0.008 Spending on non-basics 0.149 *** 0.184 *** * = significant @ 10% *** significant @ 1% Conclusions • Pensioners look poorer when assessed using spending than income • Recent rises in means-tested benefit for pensioners were translated into higher spending • Results rely on untested “common trends” assumption: evidence stronger for introduction of MIG than equalisation of agerelated premia Summing up • Living standards have risen whether measured by income or spending • Increased expenditure poverty rate since 1997 even as income poverty declined • Reasons for different trends not yet clear • Recent rises in means-tested benefit for pensioners were translated into higher spending