Short run responses to the new 50%
income tax rate in the UK
James Browne and David Phillips
Institute for Fiscal Studies
© Institute for Fiscal Studies
What’s coming up
• Policy background
• Bunching at new tax thresholds
• Changes in aggregate incomes around the time the tax rate was
introduced
© Institute for Fiscal Studies
Policy background (1)
• As in many other OECD countries, top rate of income tax cut
during 1980s
– Fell from 83% to 40% for earned income, 98% to 40% for unearned
income
© Institute for Fiscal Studies
Top tax rates and income shares, 1978–2003
1979: top
rate reduced
from 83% to
60%
80%
18%
1988: top
rate reduced
from 60% to
40%
70%
16%
14%
60%
12%
50%
10%
40%
8%
30%
6%
20%
4%
10%
2%
0%
1978
0%
1983
1988
1993
1998
2003
5%-1% marginal tax rate
Top 1% marginal tax rate
5%-1% income share (right axis)
Top 1% income share (right axis)
Sources: Income share from Atkinson (2007), marginal tax rates from Brewer, Saez and Shephard (2010).
© Institute for Fiscal Studies
Income share
Marginal effective tax rate
90%
Policy background (1)
• As in many other OECD countries, top rate of income tax cut
during 1980s
– Fell from 83% to 40% for earned, 98% to 40% for unearned income
– Reduced marginal tax rate for top 1%, but not next-richest 4%
• Top income shares also increased during this period
– Particularly for top 1%, but also next-richest 4%
– But a gradual change – no big jumps in incomes of richest 1%
following tax cuts in 1979 and 1988
– (Changes not pre-announced, so no timing responses)
• Using next-richest 4% as a comparison group, Brewer, Saez and
Shephard (2010) use a difference-in-difference methodology to
estimate a taxable income elasticity of 0.46
– Implies a marginal income tax rate of 50% would take overall MTR
beyond top of Laffer curve
© Institute for Fiscal Studies
Policy background (2)
• Tax rises for high income individuals introduced in 2010–11
– Withdraw tax-free allowance above £100k (affects ~1.5% of adults)
– 50% income tax rate above £150k (affects less than 1% of adults)
– Changes announced more than a year in advance, giving individuals
opportunities to bring forward income to avoid higher rates
• 50% rate reduced to 45% in 2013–14
– Reaction to big behavioural response in first year of the policy
– Announced at very end of 2011–12 tax year, though much
speculation beforehand
© Institute for Fiscal Studies
Income tax schedule, 2009–10 and 2010–11
Marginal income tax rate
70%
60%
50%
40%
30%
2009-10
20%
2010-11
10%
0%
£0
© Institute for Fiscal Studies
£50,000
£100,000
£150,000
Annual taxable income
£200,000
Responses to the new tax rates (1)
• Theory would lead us to expect that individuals bunch at convex
kink points in the tax system
• In fact, we observe little bunching at the new kink points at £100k
and £150k...
© Institute for Fiscal Studies
Bunching at new tax thresholds
20
18
16
Density
14
12
10
8
6
4
2
Annual taxable income
© Institute for Fiscal Studies
Source: Authors’ calculations using 2010–11 Survey of Personal Incomes
£160,000
£155,000
£150,000
£145,000
£140,000
£135,000
£130,000
£125,000
£120,000
£115,000
£110,000
£105,000
£100,000
£95,000
£90,000
0
Less bunching than at (long-standing) threshold
for 40% income tax rate
3000
2500
Density
2000
1500
1000
500
Distance from higher rate threshold
© Institute for Fiscal Studies
Source: Authors’ calculations using 2003–04 to 2007–08 Survey of
Personal Incomes
£10,000
£8,000
£6,000
£4,000
£2,000
£0
-£2,000
-£4,000
-£6,000
-£8,000
-£10,000
0
Most bunching among directors of closely-held
incorporated businesses
Timing response rather than real income response?
12
Not self employed
10
Self-employed, unincorporated
Density
8
Director of closely-held company
6
4
2
Annual taxable income
© Institute for Fiscal Studies
Source: Authors’ calculations using 2010–11 Survey of Personal Incomes
£160,000
£155,000
£150,000
£145,000
£140,000
£135,000
£130,000
£125,000
£120,000
£115,000
£110,000
£105,000
£100,000
£95,000
£90,000
0
Responses to the new tax rates (2)
• Aggregate figures reveal significant income shifting to 2009–10
from future years
© Institute for Fiscal Studies
250
£115k to £145k
200
£150k+
150
100
50
Source: Survey of Personal Incomes for all years except 2008–09 (scaled
SA302 data), and 2011– 2 (scaled SA302 data with PAYE addition)
© Institute for Fiscal Studies
2011-12
2010-11
2009-10
2008-09
2007-08
2006-07
2005-06
2004-05
2003-04
2002-03
2001-02
2000-01
1999-2000
1998-99
1997-98
1996-97
0
1995-96
Total income (2000-01 =100)
Total incomes of affected group and group with
slightly lower income, 1999–2000 to 2010–11
Responses to the new tax rates (2)
• Aggregate figures reveal significant income shifting to 2009–10
from future years
– Not just a timing response though
– HMRC (2012) use difference-in-difference approach to estimate
response to introduction of higher rate using lower-income group as a
comparison group
– Estimate how much of income brought forward was from 2010–11
by examining trends among those with ‘stable’ incomes from
different sources
© Institute for Fiscal Studies
HMRC methodology
© Institute for Fiscal Studies
Source: HMRC (2012), ‘The Exchequer effect of the 50 per cent
additional rate of income tax’.
Responses to the new tax rates (2)
• Aggregate figures reveal significant income shifting to 2009–10
from future years
– Not just a timing response though
– HMRC (2012) use difference-in-difference approach to estimate
response to introduction of higher rate using lower-income group as a
comparison group
– Estimate how much of income brought forward was from 2010–11
by examining trends among those with ‘stable’ incomes from
different sources
• HMRC estimate that 50% rate would raise £1.1 billion a year more
than 40% rate in long run; cut to 45% would cost £0.1 billion
– But much uncertainty, particularly around how much of increase in
2009–10 brought forward from 2010–11 rather than future years
© Institute for Fiscal Studies
What has changed since initial analysis?
• Since HMRC did analysis, now have fuller data for 2010-11 and
data for 2011-12
• Full data for 2010-11 show incomes fell more in that year than the
early data suggested, but especially for the control group
– Would suggest less behavioural response than initially estimated
• Data for 2011-12 does not show a big bounce back in top
incomes; strongly suggests response was not all forestalling
– Indeed, bounce back somewhat less than HMRC’s estimates would
suggest, which suggests more of what response there was may have
been “long term”
• At this stage HMRC estimates remain best estimates
– But these factors going in different directions mean clear need for
further research (which we are undertaking in coming months)
© Institute for Fiscal Studies
250
£115k to £145k
200
£150k+
150
100
50
Source: Survey of Personal Incomes for all years except 2008–09 (scaled
SA302 data), and 2011– 2 (scaled SA302 data with PAYE addition)
© Institute for Fiscal Studies
2011-12
2010-11
2009-10
2008-09
2007-08
2006-07
2005-06
2004-05
2003-04
2002-03
2001-02
2000-01
1999-2000
1998-99
1997-98
1996-97
0
1995-96
Total income (2000-01 =100)
Total incomes of affected group and group with
slightly lower income, 1999–2000 to 2011–12
How did taxpayers respond to the higher tax
rate?
• Clear evidence of timing response for dividends
– Easiest type of income to shift between periods
– Some bounce back in 2011–12, but remains well below pre-reform
level
– Delaying income in 2011–12 because anticipate cut in tax rate?
• Employment income much less responsive to higher tax rate
© Institute for Fiscal Studies
350
Employment income
300
Dividend income
250
Other income
200
150
100
2011-12
2010-11
2009-10
2008-09
2007-08
2006-07
2005-06
2004-05
2003-04
2002-03
50
2001-02
Total income from income source
(2001-02 =100)
Trends in different income sources for group
affected by 50% rate, 2001–02 to 2011–12
Source: SA302 (income tax return) data. Note that more individuals had to file
tax returns in 2010–11, slightly understating income falls in that year.
© Institute for Fiscal Studies
How did taxpayers respond to the higher tax
rate?
• Clear evidence of timing response for dividends
– Easiest type of income to shift between periods
– Some bounce back in 2011–12, but remains well below pre-reform
level
– Delaying income because anticipate cut in tax rate?
• Employment income much less responsive to higher tax rate
• Deductions disclosed on the tax return did not seem responsive to
new tax rate
– We observe charitable donations, losses and some pension
contributions (mainly for the self-employed)
– Perhaps because of reduced annual limit on pension contributions?
– Contrary to usual result from the literature that most reduction in
taxable income the result of higher deductions
– But are other deductions we don’t observe, in particular most pension
contributions for employees
© Institute for Fiscal Studies
450
Employment income
400
Dividend income
350
Other income
300
Deductions
250
200
150
100
2011-12
2010-11
2009-10
2008-09
2007-08
2006-07
2005-06
2004-05
2003-04
2002-03
50
2001-02
Total income from income source
(2001-02 =100)
Trends in different income sources for group
affected by 50% rate, 2001–02 to 2011–12
Source: SA302 (income tax return) data. Note that more individuals had to file
tax returns in 2010–11, slightly understating income falls in that year.
© Institute for Fiscal Studies
Conclusion
• UK introduced a 50% income tax rate on incomes above £150,000
in 2010
– Affects less than 1% of adults
– Short-lived experiment: rate reduced to 45% three years later
• Individuals did not bunch at the new kink points introduced in
2010–11
– Might increase as awareness of optimal response to kink points
increases
• Clear evidence of timing effects, particularly for dividend income
– Though employment income much less affected
– Not the whole story though: incomes didn’t bounce back in 2011–12
• Key challenge is to disentangle timing response from permanent
response
– Looking at this as part of ongoing work in this area
© Institute for Fiscal Studies
Optional Extra Slide
© Institute for Fiscal Studies
200
150
100
100k to £150k, HMRC
150k+, HMRC
50
Source: HMRC analysis of the 50% tax rate
© Institute for Fiscal Studies
2010-11
2009-10
2008-09
2007-08
2006-07
2005-06
2004-05
2003-04
2002-03
0
2001-02
Total income (2000-01 =100)
HMRC trends in income for 100-150k, and 150k
plus groups