© Institute for Fiscal Studies
What is happening to inequality?
Paul Johnson
September 13, 2012
What is actually happening to inequality
• The standard analysis
– What HBAI tells us
– The components of inequality
• The complications
– Under reporting of income– Under reporting of income
– Other measures of income and consumption
• The role of government
– Importance of tax and transfers
– How different measures give different answers
– Going beyond transfers
• Conclusions
© Institute for Fiscal Studies
Median household £419 a week in 2010-11
1,000
1,500
2,000
2,500
Household income (£ per
week)
50th percentile:£419
© Institute for Fiscal Studies
0
500
1,000
10 20 30 40 50 60 70 80 90
Household income (£ per
week)
Percentile point
Source: Figure 3.1 of Living Standards, Poverty and Inequality: 2012
90th centile income about 4 times 10th centile
1,000
1,500
2,000
2,500
Household income (£ per
week) 90th percentile:
£84650th percentile:£419
© Institute for Fiscal Studies
0
500
1,000
10 20 30 40 50 60 70 80 90
Household income (£ per
week)
Percentile point
10th percentile:£216
Source: Figure 3.1 of Living Standards, Poverty and Inequality: 2012
Income growth 1996-97 to 2009-10 equalising over middle 90%
1%
2%
3%
4%
Average annual income gain
Income growth: 1996–97 to 2009–10
© Institute for Fiscal Studies
-2%
-1%
0%10 20 30 40 50 60 70 80 90
Average annual income gain
Percentile point
Source: Figure 3.5 of Living Standards, Poverty and Inequality: 2012
Income falls in 2010-11 bigger at top end
1%
2%
3%
4%
Average annual income gain Income growth: 1996–97 to 2010–11
Income growth: 1996–97 to 2009–10
© Institute for Fiscal Studies
-2%
-1%
0%10 20 30 40 50 60 70 80 90
Average annual income gain
Percentile point
Source: Figure 3.5 of Living Standards, Poverty and Inequality: 2012
In stark contrast to story since 1979
1%
2%
3%
4%
Average annual income gain
Income growth: 1979 to 1996–97
Income growth: 1996–97 to 2010–11
Income growth: 1996–97 to 2009–10
© Institute for Fiscal Studies
-2%
-1%
0%10 20 30 40 50 60 70 80 90
Average annual income gain
Percentile point
Source: Figure 3.5 of Living Standards, Poverty and Inequality: 2012
Gini up sharply to 1990, modestly 1996-2009
0.30
0.35
0.40
Gini coefficient
1979 to 1990 Gini increased from 0.25 to 0.34
1996-97 to 2009-10 Gini rose from 0.33 to 0.36
© Institute for Fiscal Studies
0.25
1979
1982
1984
1986
1988
1990
1992
1994–95
1996–97
1998–99
2000–01
2002–03
2004–05
2006–07
2008–09
2010–11
Gini coefficient
Source: Figure 3.7 of Living Standards, Poverty and Inequality: 2012
With sharp fall in 2010-11
0.30
0.35
0.40
Gini coefficient
© Institute for Fiscal Studies
0.25
1979
1982
1984
1986
1988
1990
1992
1994–95
1996–97
1998–99
2000–01
2002–03
2004–05
2006–07
2008–09
2010–11
Gini coefficient
• Gini fell from 0.36 to 0.34 in 2010-11
• Largest one year fall since at least 1961
• Returns Gini to below its 1997-98 level
Source: Figure 3.7 of Living Standards, Poverty and Inequality: 2012
Inequality ratios: 1979 to 2010-11 (GB)
4.0
4.5
5.0
5.5
6.0
Ratio
99/50 ratio fell to 5.1
(still higher than 2008-09)
© Institute for Fiscal Studies
3.0
3.5
99/50 ratio
Source: Figure 3.9a of Living Standards, Poverty and Inequality: 2012
At the bottom of the distribution
• Pensioners have done very well
© Institute for Fiscal Studies
Relative pensioner poverty lowest since 1984
25%30%35%40%45%
Number of pensioners living in poverty now:
2.0 million (BHC)
0%5%
10%15%20%
1962 1970 1978 1986 1994 2002 2010
Pensioners AHC Pensioners BHCSource: HBAI Data (FES and FRS)
1.7 million (AHC)
© Institute for Fiscal Studies
At the bottom of the distribution
• Pensioners have done very well
• Families with children on an improving trend
© Institute for Fiscal Studies
15%
20%
25%
30%
While relative BHC child poverty rates are back at mid
1980s levels
Lowest since 1984
0%
5%
10%
15%
1961 1968 1975 1982 1989 1996 2003 2010
© Institute for Fiscal Studies
Notes: Poverty line is 60% of median income. Years up to and including 1992 are calendar years; thereafter, years refer to financial years. Incomes are measured before housing costs have been deducted. Source: Authors’ calculations using Family Resources Survey and Family Expenditure Survey.
Target
At the bottom of the distribution
• Pensioners have done very well
• Families with children on an improving trend
• But the working age childless have not done well
© Institute for Fiscal Studies
Relative poverty among working-age adults without children on continuing upward trend
10%12%14%16%18%20%
0%2%4%6%8%
10%
1962 1970 1978 1986 1994 2002 2010
60% of AHC Median 60% of BHC Median
Source: HBAI Data (FES and FRS)
© Institute for Fiscal Studies
Absolute poverty among working-age adults without children similar to 1970s levels...
20%25%30%35%40%
0%5%
10%15%20%
1962 1970 1978 1986 1994 2002 2010
AHC BHC
Source: HBAI Data (FES and FRS)
© Institute for Fiscal Studies
1.5%
2.0%
2.5%
Aver
age
annu
al re
al in
com
e ga
in
Children
Individuals in childless households
Families with children have done better than the childless across the
distribution since 1998-99
0.0%
0.5%
1.0%
10 20 30 40 50 60 70 80 90
Aver
age
annu
al re
al in
com
e ga
in
Percentile point of household income
Notes: Changes in income at the 1st
, 2nd
, 98th
and 99th
percentiles are not shown on this graph due to very high levels of statistical uncertainty. Incomes have been measured before housing costs have been deducted. Source: Authors’ calculations using Family Resources Survey, 1998-99 and 2010-11.
Growth in median income
So the story seems to be
• Big increase in inequality since 1979
• Modest increase during 2000s driven by the tails
• Childless working age done particularly badly
• But inequality growth is overwhelmingly within • But inequality growth is overwhelmingly within identifiable groups
© Institute for Fiscal Studies
200
250
300
350
Ineq
ualit
y (lo
g va
rianc
e x
1000
)
Employment StatusHealth
Ethnic group
Education
Age
Income inequality – factor decomposition
0
50
100
150
200
1968 1972 1976 1980 1984 1988 1992 1996 2000 2004
Ineq
ualit
y (lo
g va
rianc
e x
1000
)
Age
Gender
Household typeRegion
Residual
Source: HBAI Data (FES and FRS) and authors’ analysis
But is that the whole story?
• Two good reasons for thinking we are (increasingly) underestimating living standards lower down the distribution
– First: important elements of benefit spending appear to be very poorly captured in the data
© Institute for Fiscal Studies
Only half of tax credit and two thirds of Income Support spending captured in the data
Coverage Spend (£m/yr)
Retirement pension 95% 66,480
“Other” 52% 27,970
Working and child tax
credits 50% 21,270
Rent rebates and
Notes: based on Barnard (2011) analysis of LCFS 2009 and 2010
allowances 83% 18,930
Income support &
pension credit 68% 16,580
Child benefit 96% 11,880
Incapacity benefit 74% 6,670
Maternity/Statutory
maternity pay 119% 1,900
Jobseekers allowance 80% 1,200
War pensions 33% 1,020
Student support 236% 970
But is that the whole story?
• Two good reasons for thinking we are (increasingly) underestimating living standards lower down the distribution
– First: important elements of benefit spending appear to be very poorly captured in the data
– Second: incomes at the bottom of the distribution – Second: incomes at the bottom of the distribution don’t match consumption
© Institute for Fiscal Studies
Those with the lowest cash incomes do not have the lowest cash outlays...
£280
£350
£420
£490
Median Expenditure
Notes: LCFS 2009; Great Britain only
£0
£70
£140
£210
£0 £100 £200 £300 £400 £500
Median Expenditure
Income
Median expenditure
...but those with the lowest cash outlays do have the lowest cash income
£280
£350
£420
£490
Median Expenditure
Notes: LCFS 2009; Great Britain only
£0
£70
£140
£210
£0 £100 £200 £300 £400 £500
Median Expenditure
Income
Median expenditure
Median income
But is that the whole story?
• Two good reasons for thinking we are (increasingly) underestimating living standards lower down the distribution
– First: important elements of benefit spending appear to be very poorly captured in the data
– Second: incomes at the bottom of the distribution – Second: incomes at the bottom of the distribution don’t match consumption
• In addition
– Inequality in consumption has grown much less quickly than inequality in income
© Institute for Fiscal Studies
Remember the rising Gini
0.32
0.34
0.36
0.38
HBAI Inc
0.20
0.22
0.24
0.26
0.28
0.30
1978 1982 1986 1990 1994 1998 2002 2006 2010
Doesn’t look so dramatic when you look at consumption
0.32
0.34
0.36
0.38
HBAI Inc
Consumption
0.20
0.22
0.24
0.26
0.28
0.30
1978 1982 1986 1990 1994 1998 2002 2006 2010
n
While a broader measure of income matters little
0.32
0.34
0.36
0.38HBAI Inc
Broad Inc
Consumption
0.20
0.22
0.24
0.26
0.28
0.30
1978 1982 1986 1990 1994 1998 2002 2006 2010
n
But is that the whole story?
• Two good reasons for thinking we are (increasingly) underestimating living standards lower down the distribution
– First: important elements of benefit spending appear to be very poorly captured in the data
– Second: incomes at the bottom of the distribution – Second: incomes at the bottom of the distribution don’t match consumption
• In addition
– Inequality in consumption has grown much less quickly than inequality in income
– Relative performance of different groups is highly sensitive to measure of income/consumption used
© Institute for Fiscal Studies
Relative poverty rate, HBAI income (<60% of median household income)
0.30
0.35
0.40
0.45Children
Working-age
0.05
0.10
0.15
0.20
0.25
1978 1982 1986 1990 1994 1998 2002 2006 2010
Relative poverty rate, broad income (<60% of median household income)
0.30
0.35
0.40
0.45Children
Working-age
0.05
0.10
0.15
0.20
0.25
1978 1982 1986 1990 1994 1998 2002 2006 2010
Relative poverty rate, consumption (<60% of median household income)
0.30
0.35
0.40
0.45Children
Working-age
0.05
0.10
0.15
0.20
0.25
1978 1982 1986 1990 1994 1998 2002 2006 2010
Relative poverty rate by age and time, HBAI income
0.20
0.25
0.30
1978-1982
2003-
0.00
0.05
0.10
0.15
25 30 35 40 45 50 55 60 65 70 75Age
2003-
Relative poverty rate by age and time, broad income
0.20
0.25
0.30
1978-1982
2003-
0.00
0.05
0.10
0.15
25 30 35 40 45 50 55 60 65 70 75Age
2003-
So overall
• There may have been more a reduction in inequality in living standards in the 2000s than HBAI data suggests
• Once you account for a broader definition of income pensioners have not only done relatively better than other groups but are now better than other groups but are now significantly less likely to be poor than working age people with or without children
© Institute for Fiscal Studies
What about the role of government
• Of course the tax and transfer system is important
• 2000s saw an additional £30 billion a year spent on benefits and tax credits
– Mostly focussed on pensioners and families with childrenchildren
© Institute for Fiscal Studies
Effects of tax benefit changes 1997-2010
5%
10%
15%
20%
25%
30%
Loss as a percentage of net income
© Institute for Fiscal Studies
-15%
-10%
-5%
0%
Poorest 2 3 4 5 6 7 8 9 Richest All
Loss as a percentage of net income
Income decile group
Families with children Pensioners Working-age without children All
Couple, 3 children, no
work
Lone parent, 1 child, no
work
Lone parent, 1 child, part-time
work
Change in BHC relative
child poverty rate in UK
1999-00 + + + -
2000-01 + + + -
2001-02 + + + -
2002-03 + - + -
2003-04 + + + -
A very close correspondence between movements in benefit levels and child poverty levels
2003-04
2004-05 + + + -
2005-06 - - - +
2006-07 - - - +
2007-08 - - - +
2008-09 + + + -
2009-10 + + + -
2010-11 + + + -
© Institute for Fiscal Studies
Notes: Ignores Housing Benefit, Council Tax Benefit and the value of free school meals. The working lone parent earns an amount that is below the personal income tax allowance and the primary threshold for National Insurance contributions.Source: Authors’ calculations using TAXBEN, the IFS tax and benefit micro-simulation model.
Entitlements grew faster than relative poverty line (median income) Entitlements grew more slowly than relative poverty line (median income)
What about the role of government
• Of course the tax and transfer system is important
• 2000s saw an additional £30 billion a year spent on benefits and tax credits
– Mostly focussed on pensioners and families with childrenchildren
• This had a big effect on inequality
© Institute for Fiscal Studies
Actual income changes 1997-2009 favoured deciles 2 to 4 (and 10)
30.0%
35.0%
40.0%
45.0%
overall % change
© Institute for Fiscal Studies
0.0%
5.0%
10.0%
15.0%
20.0%
25.0%
1 2 3 4 5 6 7 8 9 10
overall % change
Very heavily driven by tax and benefit changes
30.0%
40.0%
50.0%
© Institute for Fiscal Studies
-10.0%
0.0%
10.0%
20.0%
1 2 3 4 5 6 7 8 9 10
overall % change
of which, taxben
Without those changes pattern is much less progressive
30.0%
40.0%
50.0%
© Institute for Fiscal Studies
-10.0%
0.0%
10.0%
20.0%
1 2 3 4 5 6 7 8 9 10
overall % change
of which, taxben
implied change w/out taxben
I MITIGATED THE INCREASE IN THE GAP BETWEEN RICH AND
POOR, LEAVING THE GAP SMALLER THAN IT OTHERWISE
WOULD HAVE BEEN (UNDER RPI UPRATING OF THE 1996-97
TAX AND BENEFIT SYSTEM)TAX AND BENEFIT SYSTEM)
What about the role of government
• Of course the tax and transfer system is important
• 2000s saw an additional £30 billion a year spent on benefits and tax credits
– Mostly focussed on pensioners and families with childrenchildren
• This had a big effect on inequality
• Other spending is also equalising
– At least in the sense that cost of service provided is a higher proportion of income of poor than of rich
© Institute for Fiscal Studies
Benefits in kind matter a lot
© Institute for Fiscal Studies
What about the role of government
• Of course the tax and transfer system is important
• 2000s saw an additional £30 billion a year spent on benefits and tax credits
– Mostly focussed on pensioners and families with childrenchildren
• This had a big effect on inequality
• Other spending is also equalising
– At least in the sense that cost of service provided is a higher proportion of income of poor than of rich
• And spending on the welfare state has become much more important over time
© Institute for Fiscal Studies
Public spending in 2010-11
29%
5%
6%
2%
3%1%
10% Social protection
Personal social services
Health
Education
Transport
Defence
© Institute for Fiscal Studies
4%
18%13%
3%
6%
5%
Public order and safety
Gross debt interest
Housing
TIEEE
AFF
Other
Public spending in 1978-79
23%
2%6%
9%
1%
9%Social security
Personal social services
Health
Education
Transport
Defence
© Institute for Fiscal Studies
10%
12%
4%10%
4%
10%
Law, order & protection
Gross debt interest
Housing
TIEEE
AFFF
Other
What about the role of government
• Of course the tax and transfer system is important
• 2000s saw an additional £30 billion a year spent on benefits and tax credits
– Mostly focussed on pensioners and families with childrenchildren
• This had a big effect on inequality
• Other spending is also equalising
– At least in the sense that cost of service provided is a higher proportion of income of poor than of rich
• And spending on the welfare state has become much more important over time
• A lot of that spending now more skewed to those on low incomes© Institute for Fiscal Studies
Funding focused more deprived schools
3,000
4,000
Premium (2010-11
Primary Secondary
© Institute for Fiscal Studies
0
1,000
2,000
2005 2006 2007 2008 2009 2010Implicit FSM Premium (2010
Prices)
Notes: Implicit FSM Premium calculated as the extra funding associated with one extra pupil eligible for FSM, holding other pupil and school characteristics constant.
Sources: For a full list of sources please see Table 2.2. (link)
Though we don’t know how much value that adds
60%
70%
80%
90%
100%
£5,000
£6,000
£7,000
© Institute for Fiscal Studies
0%
10%
20%
30%
40%
50%
60%
£0
£1,000
£2,000
£3,000
£4,000
Deciles of school deprivation (% FSM)
Spending per pupil (left axis)
What about the role of government
• Of course the tax and transfer system is important
• 2000s saw an additional £30 billion a year spent on benefits and tax credits
– Mostly focussed on pensioners and families with childrenchildren
• This had a big effect on inequality
• Other spending is also equalising
– At least in the sense that cost of service provided is a higher proportion of income of poor than of rich
• And spending on the welfare state has become much more important over time
• A lot of that spending now more skewed to those on low incomes
• And just as with measuring living standards
© Institute for Fiscal Studies
VAT reform: effects by income
£0
£2
£4
£6
£8
4%
5%
6%
7%
8%
% rise in non-housing expenditure
% rise in income
cash gain/loss (£/week, RH axis)
© Institute for Fiscal Studies
-£8
-£6
-£4
-£2
£0
0%
1%
2%
3%
4%
Poorest 2 3 4 5 6 7 8 9 Richest
Income Decile Group
VAT reform: effects by expenditure
£2
£4
£6
£8
5%
6%
7%
8%
% rise in non-housing expenditure
% rise in income
cash gain/loss (£/week, RH axis)
© Institute for Fiscal Studies
-£8
-£6
-£4
-£2
£0
£2
0%
1%
2%
3%
4%
5%
Poorest 2 3 4 5 6 7 8 9 Richest
Expenditure Decile Group
To conclude
• The HBAI story tells us what has happened to household incomes over time
– But only on a particular measure
– Using a broader measure of income changes the picture between groups
– And inequality in measured consumption has grown – And inequality in measured consumption has grown much less
• Redistribution through the welfare system over recent years has been very important
– Though very costly and being partially reversed
• And don’t forget the wider role of the state in redistributing
– Nor the costs (as well as the benefits) inherent in that redistribution
© Institute for Fiscal Studies