Transcript Reforming family assistance in southern Europe Is there a

Microsimulation and the analysis of
poverty and inequality
an illustration
Manos Matsaganis & Panos Tsakloglou
Athens University of Economics
FBBVA Microsimulation Workshop
Madrid 15-16 November 2004
Microsimulation, poverty and inequality (1)
 MS can help analyse the effect of policy changes
 examine redistributive impact of current policies
 explore impact of reforms vs. some baseline
 answer counterfactual («what if»-type) questions
Microsimulation, poverty and inequality (2)
 MS can help produce various types of output
 analyse the impact of policy changes in terms of:
 changes in the income distribution
 poverty and inequality
 distribution of gains and losses
 i.e. winners and losers in the entire population ...
 ... or by population group (lone parents, elderly etc.)
 fiscal implications of policy changes
 excl. administrative costs
Microsimulation, poverty and inequality (3)
 MS can help calculate RRs and METRs
 replacement rates
 does work pay?
 marginal effective tax rates
 poverty trap?
 unemployment trap?
Microsimulation, poverty and inequality (4)
 MS can help disentangle separate effects
 taxes and benefits interact
 e.g. raising benefit rates will raise pre-tax incomes, so
it may move recipients to a higher tax band ...
 ... while lowering taxes will raise pre-tax incomes, so
it may cause recipients of income-tested benefits to
lose eligibility
 no other way to account (separately and jointly) for the
effect of taxes and benefits on final disposable incomes
An illustration
Matsaganis M., O’Donoghue C., Levy H., Coromaldi M.,
Mercader-Prats M., Rodrigues C.F., Toso S., Tsakloglou P.
«Child poverty and family transfers in southern Europe»
Working Paper EM 2/04
Microsimulation Unit, University of Cambridge
http://www.econ.cam.ac.uk/dae/mu/emod.htm
 1. Introduction
 2. Data and methodology
 3. Incidence of child poverty by household type
 4. Distributional impact of family transfers
 5. Simulating reforms
 6. Discussion and policy implications
The paper
 aims to assess the impact
– of family transfers
– on child poverty
– in southern Europe
• Greece
• Italy
• Spain
• Portugal
 using microsimulation
– the tax-benefit model EUROMOD
Family transfers
 definition
 all income transfers to families with children
 child benefits (non-contributory)
 family allowances (contributory)
 tax relief for dependent children
 non-cash benefits (e.g. child care) not included
Child poverty
 issue has risen to prominence
– poor children by definition “deserving”
– social costs of child poverty vs. benefits of early
intervention
 anti-poverty measures  human capital investments
 high future returns!
 political commitments
– Blair pledge to eliminate child poverty in Britain by 2020
– Commission proposal to halve child poverty in the EU by
2010 (not endorsed by the Council)
Southern Europe
 tradition of “familialism”
 strength of informal safety nets
 but are family resources adequate for intra-family redistribution?
 subsidiary role of formal safety nets
 public assistance to families meagre / not available at all
 reliance of tax benefits  what about those too poor to pay tax?
 uneven coverage  gaps in protection
 1. Introduction
 2. Data and methodology
 3. Incidence of child poverty by household type
 4. Distributional impact of family transfers
 5. Simulating reforms
 6. Discussion and policy implications
EUROMOD (1)
 what is it?
– a tax-benefit model for all 15 “old” EU members
 construction funded by EC TSER / FP5 programmes
 40 persons in 18 centres led by Cambridge University
 main objective: comparability
 what can it do?
– analyse policy changes in a comparative setting
 examine impact of current policies
 explore impact of reforms
 answer “what if?”-type questions
EUROMOD (2)
 what data does it rely upon?
 EC Household Panel (Greece / Spain / Portugal)
 Bank of Italy Survey of Household Income & Wealth
 updated to 1998
 2001 dataset in progress
 what sort of policies does it simulate?
 income taxes, social contributions
 social assistance, housing, family, unemployment and
some social insurance benefits, e.g. social pensions
 1998 & 2001 rules applied
 2003 update in progress
EUROMOD (3)
 what types of output does it produce?
– estimates of policy impact in terms of:
 income distribution
 fiscal costs
 distribution of gains and losses (winners and losers)
– calculations of:
 marginal effective tax rates
 replacement rates
EUROMOD (4)
 temporary difficulties?
– no simulation of benefits in kind (publicly provided
services)
– full tax compliance (no tax evasion)
– no targeting errors (100% take up / no leakage)
 given sufficient effort (and funding), all of the above
may be fully or partly amenable to treatment
EUROMOD (5)
 structural weaknesses?
 static microsimulation
 no behavioural responses
 e.g. labour supply
 consumer demand etc.
 nonetheless:
 dynamic models sensitive to assumptions
 in the short term, static MS may be sufficient
 1. Introduction
 2. Data and methodology
 3. Incidence of child poverty by household type
 4. Distributional impact of family transfers
 5. Simulating reforms
 6. Discussion and policy implications
Table 2a Child poverty rates
Greece
Italy
Spain
Portugal
couple with 1 child 0-17
10.0
13.0
12.5
11.3
couple with 2 children 0-17
14.9
20.7
15.6
16.1
couple with 3+ children 0-17
12.0
37.0
33.8
35.7
lone parent: all children 0-17
42.2
49.0
45.2
56.6
lone parent: 1+ child 18+
30.0
67.7
41.6
34.4
couple with 1+ child 18+
18.3
36.0
30.2
21.2
other hh types with children
28.5
27.6
19.5
27.4
all households with children
17.0
26.5
21.6
23.1
all households
20.5
20.7
18.3
22.0
Table 2b Contribution to aggregate child poverty
Greece
Italy
Spain
Portugal
8.6
7.8
8.4
7.9
39.3
30.7
29.6
21.4
couple with 3+ children 0-17
8.7
24.2
19.9
19.6
lone parent: all children 0-17
7.2
5.2
5.4
13.5
lone parent: 1+ child 18+
2.1
3.0
3.1
2.0
couple with 1+ child 18+
10.3
20.0
24.7
13.7
other hh types with children
23.8
9.3
9.0
21.9
100.0
100.0
100.0
100.0
couple with 1 child 0-17
couple with 2 children 0-17
all households with children
 1. Introduction
 2. Data and methodology
 3. Incidence of child poverty by household type
 4. Distributional impact of family transfers
 5. Simulating reforms
 6. Discussion and policy implications
Table 7a Estimated value of cash benefits
family size
1 child
2 children
3 children
4+
children
income
Greece
Italy
Spain
Portugal
low
88
250
36
191
middle
71
267
16
188
high
114
41
5
185
low
76
640
124
233
middle
72
390
9
202
high
131
29
2
182
low
319
569
165
227
middle
313
664
24
225
high
172
93
0
217
low
327
177
203
254
middle
457
895
95
308
high
275
47
36
249
Table 7b Estimated value of tax relief
family size
1 child
2 children
3 children
4+
children
income
Greece
Italy
Spain
Portugal
low
67
188
114
80
middle
68
202
138
87
high
84
203
126
89
low
27
120
60
49
middle
71
165
158
86
high
79
159
176
96
low
11
85
42
38
middle
83
147
185
76
high
108
161
210
85
low
47
71
0
24
middle
88
129
113
24
high
132
163
210
80
Graph 1 Concentration curves
Greece
Italy
100
100
cash benefits
cash benefits
80
tax relief
60
40
all family transfers
20
disposable income
(net of family
transfers)
line of perfect
equality
0
0
20
40
60
80
100
C umulative population share
(by equivalent income)
Share
Share
80
tax relief
60
40
all family transfers
20
disposable income
(net of family
transfers)
line of perfect
equality
0
0
20
40
60
80
100
C umulative population share
(by equivalent income)
Spain
Portugal
100
100
cash benefits
cash benefits
80
80
all family transfers
40
20
0
0
20
40
60
80
C umulative population share
(by equivalent income)
100
disposable income
(net of family
transfers)
line of perfect
equality
Share
Share
tax relief
60
tax relief
60
all family transfers
40
20
0
0
20
40
60
80
C umulative population share
(by equivalent income)
100
disposable income
(net of family
transfers)
line of perfect
equality
Table 11a Impact on child poverty rates
Greece
Italy
Spain
Portugal
couple with 1 child 0-17
0.0
17.6
2.7
5.9
couple with 2 children 0-17
3.3
25.6
6.4
17.6
couple with 3+ children 0-17
32.0
23.4
7.8
34.6
lone parent: all children 0-17
4.1
14.1
4.4
9.8
lone parent: 1+ child 18+
0.0
0.0
2.3
4.0
couple with 1+ child 18+
5.5
10.1
5.7
14.4
other hh types with children
9.4
8.1
7.5
24.1
all households with children
8.1
19.0
7.3
20.8
Percentage reduction in the number of poor children due to current policies
Table 11b Impact on child poverty gaps
Greece
Italy
Spain
Portugal
couple with 1 child 0-17
1.9
14.4
8.5
16.4
couple with 2 children 0-17
4.8
31.6
12.0
29.6
couple with 3+ children 0-17
34.0
35.7
18.1
58.0
lone parent: all children 0-17
16.8
29.4
14.4
39.3
lone parent: 1+ child 18+
16.7
14.7
10.0
32.6
couple with 1+ child 18+
10.4
22.3
9.1
21.0
other hh types with children
14.0
16.1
9.7
31.2
all households with children
11.4
28.2
12.1
36.7
Percentage reduction in the aggregate child income gap due to current policies
Table 12 Redistributive impact of family transfers
Greece
Italy
Spain
Portugal
before any family transfers
18.5
32.7
23.4
29.2
after tax relief
18.2
31.9
22.2
28.6
after cash benefits
17.4
27.7
23.2
23.7
after all family transfers
17.0
26.5
21.6
23.1
before any family transfers
3.8
6.5
4.3
2.5
after tax relief
3.7
6.2
4.3
2.4
after cash benefits
3.3
4.7
3.8
1.4
after all family transfers
3.3
4.5
3.7
1.4
child poverty rate
FGT index ( = 2)
The FGT index attaches greater weight to larger income gaps
Figure 1 Target efficiency of social transfers
PRE
A / A+B+C
PGE
post-transfer disposable
income
A / A+D
B
poverty line
C
D
A
pre-transfer disposable income
households ranked by income
Table 13 Target efficiency of family transfers
PRE
PGE
=1
=2
=3
21.1
4.7
7.3
10.2
8.2
1.0
1.0
1.1
ITALY - cash benefits
55.2
15.2
18.1
18.8
ITALY - tax relief
26.2
4.5
4.7
4.5
SPAIN - cash benefits
80.5
4.9
8.4
11.1
SPAIN - tax relief
12.3
1.6
0.7
0.3
PORTUGAL - cash benefits
38.7
20.7
31.5
39.6
PORTUGAL - tax relief
13.6
2.6
2.9
3.4
GREECE - cash benefits
GREECE - tax relief
Greater values of  indicate greater concern for those with larger income gaps
 1. Introduction
 2. Data and methodology
 3. Incidence of child poverty by household type
 4. Distributional impact of family transfers
 5. Simulating reforms
 6. Discussion and policy implications
Reforms (1)
 reforming family transfers
 abolish all current policies
 introduce universal child benefits
 what if ...?
Reforms (2)
 why universal child benefits?
 obvious solution to the problem of gaps in coverage
 ... though controversial
 good for illustration
 easy to explain
 simple to implement
Reforms (3)
 which universal child benefits?
 various issues involved
 benefit level
 variation by age
 variation by no. of children etc.
 5 variations of UCB reform
2 “artificial” universal child benefits
 poverty-neutral / budget-neutral
3 “actually existing” universal child benefits
 British / Danish / Swedish schemes ...
... adjusted in terms of average male f-t earnings
Table 14 Simulated reforms
Greece
Italy
Spain
Portugal
Reform I: poverty neutral UCB
204
912
196
350
Reform II: budget neutral UCB
197
582
135
284
Reform III: British CB
eldest
401
676
555
369
other children
326
551
452
300
Reform IV: Danish CB 0-3 yrs
594
1 003
824
547
4-7 yrs
540
912
749
498
8-17 yrs
422
711
584
388
Reform V: Swedish CB 1st-2nd
354
597
490
326
3rd
448
756
621
413
4th
637
1 074
883
586
5th+
707
1 194
981
651
10 253
17 300
14 212
9 441
annual earnings male f-t employee
Table 15a Impact of simulated reforms
(I): child poverty rates
Greece
Italy
Spain
Portugal
baseline: current policy
17.0
26.5
21.6
23.1
Reform I poverty neutral UCB
17.0
26.5
21.9
23.1
Reform II budget neutral UCB
17.1
28.4
22.5
25.4
Reform III British CB
16.0
28.1
18.9
23.5
Reform IV Danish CB
15.5
27.1
17.9
19.6
Reform V Swedish CB
15.9
28.1
18.9
23.2
child poverty rate
Table 15b Impact of simulated reforms
(II): intensity of child poverty
Greece
Italy
Spain
Portugal
baseline: current policy
3.3
4.5
3.7
1.4
Reform I poverty neutral UCB
3.2
3.8
3.7
1.2
Reform II budget neutral UCB
3.2
4.7
3.9
1.3
Reform III British CB
2.8
4.6
3.0
1.3
Reform IV Danish CB
2.6
4.2
2.6
1.0
Reform V Swedish CB
2.8
4.5
2.9
1.2
FGT index (α=2)
Table 17 Winners vs. losers
reform II (budget neutral CB)
Greece
Italy
80
40
30
20
winners
losers
10
0
percent
percent
50
60
winners
losers
40
20
0
1
2
3
4
5
6
7
8
9 10
1
2
3
income decile
4
winners
losers
40
20
0
4
5
6
7
income decile
8
9 10
percent
percent
60
3
7
8
9 10
Portugal
80
2
6
income decile
Spain
1
5
60
50
40
30
20
10
0
winners
losers
1
2
3
4
5
6
7
income decile
8
9 10
Table 19 Winners vs. losers
reform IV (Danish UCB)
Italy
60
50
40
30
20
10
0
winners
losers
1
2
3
4
5
6
7
8
9
percent
percent
Greece
60
50
40
30
20
10
0
winners
losers
1
10
2
3
4
6
7
8
9
10
income decile
income decile
Spain
Portugal
70
60
50
40
30
20
10
0
80
winners
losers
percent
percent
5
60
winners
40
losers
20
0
1
2
3
4
5
6
7
income decile
8
9
10
1
2
3
4
5
6
7
income decile
8
9
10
Table 21 Fiscal effects of simulated reforms
Greece
Italy
Spain
Portugal
baseline: current policy
0.9
1.3
0.5
1.5
Reform I poverty neutral UCB
0.9
2.2
0.7
1.9
Reform II budget neutral UCB
0.9
1.3
0.5
1.5
Reform III British CB
1.6
1.4
1.8
1.8
Reform IV Danish CB
2.1
1.8
2.4
2.3
Reform V Swedish CB
1.7
1.4
1.8
1.8
child poverty rate
 1. Introduction
 2. Data and methodology
 3. Incidence of child poverty by household type
 4. Distributional impact of family transfers
 5. Simulating reforms
 6. Discussion and policy implications
Concluding remarks (1)
 performance of family transfers modest
 many poor families are ineligible for assistance (GR / IT)
 ... or receive low benefits (SP / PT)
 non-refundable tax credits exclude poor families by
design
Concluding remarks (2)
 performance of UCBs disappointing?
 replacing current policies by UCBs would not reduce
the number of poor children by much – and could
even increase it!
 impact of UCBs weak or negative where current
policies provide substantial benefits to a considerable
subset of the low-income population (as in Italy)
Concluding remarks (3)
 headcount poverty too severe a test for UCBs
 where existing policies leave coverage gaps, those
currently ineligible for assistance will be better off
under a UCB even when they remain below the
poverty line
 headcount poverty rates cannot capture such
improvements
 bringing in the FGT index (implying more concern for
those at the bottom of the income distribution) does
more justice to the anti-poverty impact of UCBs
Concluding remarks (4)
 not all UCBs are the same
 the Danish UCB, paying higher amounts to younger
children, emerges ahead of the others in terms of
generosity and anti-poverty effectiveness
 The British and the Swedish UCBs, though different in
structure (the former paying a higher rate to the
eldest child, the latter rising in value with family size)
have similar effects on child poverty and fiscal costs
Concluding remarks (5)
 playing field uneven
 100% take up a reasonable assumption for universal
benefits – not so for currrent income-tested policies
 in favour of universal child benefits
 low administrative costs
 no stigma
 no adverse labour incentives (i.e. no poverty traps)
Concluding remarks (6)
 the case for UCB wider
 horizontal redistribution
 from single tax payers to families with children
 children a (partly) public good
 social citizenship
 access to certain benefits can be a citizen right
 political economy considerations
 narrowly targeted programmes at risk of backlash
 support for universal programmes broader-based
Concluding remarks (7)
 a basic trade off at work
 more generous UCBs are more effective but costlier
 but: current spending on family transfers far too low
 hard to reduce poverty via internal reallocations alone
Concluding remarks (8)
 way forward?
 judiciously combine
a universal (even if low) income base ...
 to address the problem of gaps in coverage
 to achieve horizontal redistribution
... with more targeted (non-categorical) interventions
 to direct extra resources to families in need
 to improve effectiveness of both type I and II
Concluding remarks (9)
 bring in services ...
 there is more to fighting child poverty than cash benefits
alone
 universal access to affordable, good-quality family
services a high priority
 a child care guarantee a promising route out of poverty
Concluding remarks (10)
 ... but avoid the other extreme
 services a complement of cash benefits, not a substitute
 proper design of income transfers still relevant!
Concluding remarks (11)
 the role of microsimulation: counterfactual
 “what if”-type questions
 involve interactions between benefits and taxes ...
 ... that are difficult to predict
 only tax-benefit models
 can provide full answers to counterfactual questions
 in spite of their limitations!
Concluding remarks (12)
 the role of microsimulation: aid to policy making
 a valuable tool to decision making
 “informing” policy dilemmas
 promoting a more rational and dispassionate debate on
tax-benefit reform