2749
Views
744
Downloads |
Does going beyond income make a difference? Income vs. equivalent income in the EU over 2007-2011
Marko Ledić* 
Article | Year: 2020 | Pages: 423 - 462 | Volume: 44 | Issue: 4 Received: March 20, 2020 | Accepted: June 30, 2020 | Published online: December 1, 2020
|
FULL ARTICLE
FIGURES & DATA
REFERENCES
CROSSMARK POLICY
METRICS
LICENCING
PDF
|
|
2007
|
2011
|
Difference
|
2007 and 2011 pooled
|
|
|
Est.
|
SE
|
Est.
|
SE
|
Diff.
|
p-value
|
Est.
|
SE
|
|
lnY (β)
|
0.358
|
[0.031]***
|
0.296
|
[0.025]***
|
-0.062
|
0.124
|
0.324
|
[0.020]***
|
|
UNE (γ1)
|
-0.661
|
[0.084]***
|
-0.690
|
[0.056]***
|
-0.029
|
0.773
|
-0.678
|
[0.047]***
|
|
HEAvbad
(γ2)
|
-2.264
|
[0.150]***
|
-2.313
|
[0.111]***
|
-0.049
|
1.000
|
-2.285
|
[0.089]***
|
|
HEAbad
(γ3)
|
-1.645
|
[0.071]***
|
-1.614
|
[0.060]***
|
0.031
|
1.000
|
-1.622
|
[0.046]***
|
|
HEAfair
(γ4)
|
-0.896
|
[0.050]***
|
-0.849
|
[0.042]***
|
0.047
|
0.474
|
-0.865
|
[0.032]***
|
|
HEAgood
(γ5)
|
-0.457
|
[0.042]***
|
-0.443
|
[0.035]***
|
0.014
|
0.803
|
-0.446
|
[0.027]***
|
|
HOUspace
(γ6)
|
-0.251
|
[0.042]***
|
-0.299
|
[0.036]***
|
-0.048
|
0.382
|
-0.282
|
[0.027]***
|
|
HOUrot
(γ7)
|
-0.330
|
[0.052]***
|
-0.317
|
[0.048]***
|
0.013
|
0.844
|
-0.317
|
[0.035]***
|
|
HOUdamp
(γ8)
|
-0.245
|
[0.048]***
|
-0.181
|
[0.042]***
|
0.064
|
0.318
|
-0.205
|
[0.032]***
|
|
HOUbathtoil
(γ9)
|
-0.219
|
[0.067]**
|
-0.054
|
[0.064]
|
0.165
|
0.076
|
-0.143
|
[0.046]**
|
|
CRImoder
(γ10)
|
-0.126
|
[0.034]***
|
-0.131
|
[0.029]***
|
-0.005
|
0.924
|
-0.125
|
[0.022]***
|
|
CRImajor
(γ11)
|
-0.202
|
[0.053]***
|
-0.232
|
[0.063]***
|
-0.030
|
0.721
|
-0.202
|
[0.039]***
|
|
ENVwater
(γ12)
|
-0.160
|
[0.037]***
|
-0.147
|
[0.035]***
|
0.013
|
0.803
|
-0.153
|
[0.025]***
|
|
ENVair
(γ13)
|
-0.087
|
[0.034]*
|
-0.116
|
[0.032]***
|
-0.029
|
0.534
|
-0.097
|
[0.023]***
|
|
male (π1)
|
-0.123
|
[0.030]***
|
-0.067
|
[0.025]**
|
0.056
|
0.151
|
-0.087
|
[0.019]***
|
|
age (π2)
|
-0.059
|
[0.006]***
|
-0.053
|
[0.005]***
|
0.006
|
0.415
|
-0.056
|
[0.004]***
|
|
age2
(π3)
|
0.641
|
[0.057]***
|
0.584
|
[0.045]***
|
-0.057
|
0.432
|
0.616
|
[0.035]***
|
|
edulow
(π4)
|
-0.212
|
[0.056]***
|
0.015
|
[0.051]
|
0.227
|
0.003
|
-0.094
|
[0.038]*
|
|
eduhigh
(π5)
|
0.020
|
[0.037]
|
0.015
|
[0.029]
|
-0.005
|
0.921
|
0.020
|
[0.023]
|
|
marrcoh
(π6)
|
0.353
|
[0.057]***
|
0.277
|
[0.043]***
|
-0.076
|
0.286
|
0.295
|
[0.034]***
|
|
divsep
(π7)
|
-0.309
|
[0.071]***
|
-0.231
|
[0.054]***
|
0.078
|
0.380
|
-0.263
|
[0.043]***
|
|
widow (π8)
|
-0.173
|
[0.077]*
|
-0.170
|
[0.059]**
|
0.003
|
0.975
|
-0.182
|
[0.047]***
|
|
urban (π9)
|
-0.048
|
[0.037]
|
-0.050
|
[0.030]+
|
-0.002
|
0.963
|
-0.062
|
[0.023]**
|
|
child (π10)
|
0.144
|
[0.048]**
|
0.146
|
[0.037]***
|
0.002
|
0.970
|
0.158
|
[0.029]***
|
|
trust (π11)
|
0.286
|
[0.011]***
|
0.222
|
[0.009]***
|
-0.064
|
0.000
|
0.247
|
[0.007]***
|
|
year2011
(π12)
|
|
|
|
|
|
|
0.110
|
[0.020]***
|
|
pseudo
R2
|
0.106
|
|
0.077
|
|
|
|
0.087
|
|
|
log
likelih.
|
-34,000
|
|
-47,000
|
|
|
|
-82,000
|
|
|
chi-2
|
5,783
|
|
5,983
|
|
|
|
11,000
|
|
|
N
|
18,899
|
|
25,117
|
|
|
|
44,016
|
|
childless, surveyed in 2011, from Austria, non-unemployed, with all non-income dimensions at the best levels. Weights are used in all estimations.
Note: Example of interpretation: 0.1 above the bar labelled HEAvbad, means that is a person is in very bad, rather than very good, health (i.e., if HEAvbad = 1) and has all the other nonincome dimensions at their reference (best) levels, her equivalent income is 0.1 percent of her income. The figure is calculated as: 100exp(φ2). Other figures can be calculated and interpreted in the same way.
Note: The first (second) number in parentheses next to a country label is the country’s rank by the average income (equivalent income) or, on the bottom panel, by the respective growth rates.
Note: The first (second) number in parentheses next to a country label is the country’s rank by the Atkinson index for income (equivalent income) or, on the bottom panel, by the respective growth rates.
 Note: Decomposition is based on equation (9). Y, UNE, HEA, HOU, CRI and ENV refer to the contributions of income, unemployment, health, housing, crime and environment, respectively.
|
|
Percent of variance due to
variation in
|
|
|
ln μ
|
ln(1-A1)
|
Δln μ
|
Δln(1-A1)
|
ln μ*
|
ln(1-A1*)
|
Δln μ*
|
Δln(1-A1*)
|
|
Variance
of:
|
|
lnW (2007)
|
94.5
|
5.5
|
-
|
-
|
-
|
-
|
-
|
-
|
|
lnW (2011)
|
93.5
|
6.5
|
-
|
-
|
-
|
-
|
-
|
-
|
|
ΔlnW
|
-
|
-
|
83.8
|
16.2
|
-
|
-
|
-
|
-
|
|
lnW* (2007)
|
-
|
-
|
-
|
-
|
75.6
|
24.4
|
-
|
-
|
|
lnW* (2011)
|
-
|
-
|
-
|
-
|
71.2
|
28.8
|
-
|
-
|
|
ΔlnW*
|
|
|
|
|
|
|
81.2
|
18.8
|
|
Contribution to var(lnW*)=100 or var(ΔlnW*)=100 of variation in:
|
2007
|
2011
|
Change
2007-2011
|
|
1. C(W)
or C(ΔW)
|
39.3
|
45.2
|
12.7
|
|
1a. C(μ) or C(Δμ)
|
36.7
|
42.3
|
10.3
|
|
1b. C(A1) or C(ΔA1)
|
2.6
|
2.9
|
2.4
|
|
2. C(nonY)
or C(ΔnonY)
|
60.7
|
54.8
|
87.3
|
|
2a. C(UNE) or C(ΔUNE)
|
1.6
|
3.9
|
2.2
|
|
2b. C(HEA) or C(ΔHEA)
|
31.4
|
31.1
|
41.7
|
|
2c. C(HOU) or C(ΔHOU)
|
14.7
|
12.9
|
9.7
|
|
2d. C(CRI) or C(ΔCRI)
|
3.8
|
0.8
|
13.5
|
|
2e. C(ENV) or C(ΔENV)
|
9.2
|
6.1
|
20.2
|
Note: Decomposition based on the version of equation (12) representing changes over 2007-2011 (see text).
|
Countries
|
Year
|
|
2007
|
2011
|
|
N
|
%
of total N
|
N
|
%
of total N
|
|
AT
|
534
|
2.8
|
729
|
2.9
|
|
BE
|
713
|
3.8
|
734
|
2.9
|
|
BG
|
586
|
3.1
|
691
|
2.8
|
|
CY
|
683
|
3.6
|
586
|
2.3
|
|
CZ
|
749
|
4.0
|
708
|
2.8
|
|
DE
|
1,364
|
7.2
|
2,312
|
9.2
|
|
DK
|
772
|
4.1
|
841
|
3.3
|
|
EE
|
740
|
3.9
|
702
|
2.8
|
|
EL
|
702
|
3.7
|
658
|
2.6
|
|
ES
|
451
|
2.4
|
901
|
3.6
|
|
FI
|
774
|
4.1
|
852
|
3.4
|
|
FR
|
1,126
|
6.0
|
1,889
|
7.5
|
|
HU
|
686
|
3.6
|
690
|
2.7
|
|
IE
|
403
|
2.1
|
807
|
3.2
|
|
IT
|
484
|
2.6
|
1,407
|
5.6
|
|
LT
|
749
|
4.0
|
897
|
3.6
|
|
LU
|
484
|
2.6
|
650
|
2.6
|
|
LV
|
591
|
3.1
|
756
|
3.0
|
|
MT
|
528
|
2.8
|
474
|
1.9
|
|
NL
|
771
|
4.1
|
811
|
3.2
|
|
PL
|
936
|
5.0
|
1,632
|
6.5
|
|
PT
|
376
|
2.0
|
544
|
2.2
|
|
RO
|
638
|
3.4
|
1,148
|
4.6
|
|
SE
|
900
|
4.8
|
820
|
3.3
|
|
SI
|
650
|
3.4
|
623
|
2.5
|
|
SK
|
731
|
3.9
|
656
|
2.6
|
|
UK
|
778
|
4.1
|
1,599
|
6.4
|
|
18,899
|
100.0
|
25,117
|
100,0
|
|
Variable name
in life
satisfaction model
|
Description
|
Reference (best) category for non-income dimensions
|
2007
|
2011
|
|
Mean
|
SD
|
Mean
|
SD
|
|
S
|
life
satisfaction; integer scale from 1 (very dissatisfied) to 10 (very satisfied)
|
-
|
7.06
|
2.08
|
7.05
|
2.12
|
|
lnY
|
logarithm
of household disposable income equivalized by the OECD equivalence scale
|
-
|
9.99
|
0.71
|
9.81
|
0.71
|
|
UNE
|
1 if
unemployed; 0 otherwise
|
“non-unemployed”:
(self)employed, retired, students, other inactive
|
0.05
|
0.22
|
0.08
|
0.27
|
|
HEAvbad
|
1 if
self-assessed health (SAH) very bad; 0 otherwise
|
SAH
very good
|
0.02
|
0.13
|
0.02
|
0.15
|
|
HEAbad
|
1 if SAH
bad; 0 otherwise
|
SAH
very good
|
0.08
|
0.27
|
0.08
|
0.27
|
|
HEAfair
|
1 if SAH
fair; 0 otherwise
|
SAH
very good
|
0.27
|
0.44
|
0.28
|
0.45
|
|
HEAgood
|
1 if SAH
good; 0 otherwise
|
SAH
very good
|
0.41
|
0.49
|
0.4
|
0.49
|
|
HOUspace
|
1 if
there is shortage of space; 0 otherwise
|
there
is no shortage of space in dwelling
|
0.19
|
0.39
|
0.16
|
0.37
|
|
HOUrotten
|
1 if
dwelling has rotten parts; 0 otherwise
|
there
are no rotten parts in dwelling
|
0.13
|
0.33
|
0.10
|
0.30
|
|
HOUdamp
|
1 if
there are damps or leaks in dwelling; 0 otherwise
|
there
are no damps or leaks in dwelling
|
0.15
|
0.35
|
0.14
|
0.35
|
|
HOUbathtoil
|
1 if
there is no bath or toilet in dwelling; 0 otherwise
|
there
are both bath and toilet in dwelling
|
0.07
|
0.25
|
0.06
|
0.23
|
|
CRImoder
|
1 if
moderate problems with crime or vandalism in the neighbourhood
|
no
problems with crime or vandalism in neighbourhood
|
0.33
|
0.47
|
0.26
|
0.44
|
|
CRImajor
|
1 if
major problems with crime or vandalism in neighbourhood.
|
no
problems with crime or vandalism in neighbourhood
|
0.14
|
0.34
|
0.06
|
0.23
|
|
ENVwat
|
1 if any
problems with drinking water in neighbourhood
|
no
problems with drinking water in neighbourhood
|
0.36
|
0.48
|
0.22
|
0.41
|
|
ENVair
|
1 if
any problems with air in neighbourhood
|
no
problems with air in neighbourhood
|
0.44
|
0.50
|
0.26
|
0.44
|
|
male
|
1 if
male; 0 if female
|
-
|
0.49
|
0.50
|
0.49
|
0.50
|
|
age
|
age in
years
|
-
|
46.96
|
17.27
|
48.21
|
17.78
|
|
edulow
|
1 if
low education (primary or less); 0 otherwise; reference: intermediate educ.
(secondary)
|
-
|
0.10
|
0.31
|
0.10
|
0.30
|
|
eduhigh
|
1 if
high education (tertiary); 0 otherwise; reference: intermediate educ.
(secondary)
|
-
|
0.21
|
0.41
|
0.25
|
0.43
|
|
marrcoh
|
1 if
married or in cohabitation; 0 otherwise; reference: single
|
-
|
0.67
|
0.47
|
0.54
|
0.5
|
|
divsep
|
1 if
divorced or separated; 0 otherwise; reference: single
|
-
|
0.08
|
0.27
|
0.13
|
0.33
|
|
widow
|
1 if
widowed; 0 otherwise; reference: single
|
-
|
0.09
|
0.28
|
0.12
|
0.32
|
|
urban
|
1 if
living in urban area; 0 otherwise
|
-
|
0.24
|
0.43
|
0.25
|
0.43
|
|
child
|
1 if
person has a child; 0 otherwise
|
-
|
0.75
|
0.43
|
0.69
|
0.46
|
|
trust
|
average
of trust in people, legal system and police; 1-10 scale; higher value means
higher trust
|
-
|
5.54
|
1.96
|
5.30
|
1.95
|
|
Countries
|
Year
|
|
2007
|
2011
|
|
N
|
%
of total N
|
N
|
%
of total N
|
|
AT
|
534
|
2.8
|
729
|
2.9
|
|
BE
|
713
|
3.8
|
734
|
2.9
|
|
BG
|
586
|
3.1
|
691
|
2.8
|
|
CY
|
683
|
3.6
|
586
|
2.3
|
|
CZ
|
749
|
4.0
|
708
|
2.8
|
|
DE
|
1364
|
7.2
|
2312
|
9.2
|
|
DK
|
772
|
4.1
|
841
|
3.3
|
|
EE
|
740
|
3.9
|
702
|
2.8
|
|
EL
|
702
|
3.7
|
658
|
2.6
|
|
ES
|
451
|
2.4
|
901
|
3.6
|
|
FI
|
774
|
4.1
|
852
|
3.4
|
|
FR
|
1126
|
6.0
|
1889
|
7.5
|
|
HU
|
686
|
3.6
|
690
|
2.7
|
|
IE
|
403
|
2.1
|
807
|
3.2
|
|
IT
|
484
|
2.6
|
1407
|
5.6
|
|
LT
|
749
|
4.0
|
897
|
3.6
|
|
LU
|
484
|
2.6
|
650
|
2.6
|
|
LV
|
591
|
3.1
|
756
|
3.0
|
|
MT
|
528
|
2.8
|
474
|
1.9
|
|
NL
|
771
|
4.1
|
811
|
3.2
|
|
PL
|
936
|
5.0
|
1632
|
6.5
|
|
PT
|
376
|
2.0
|
544
|
2.2
|
|
RO
|
638
|
3.4
|
1148
|
4.6
|
|
SE
|
900
|
4.8
|
820
|
3.3
|
|
SI
|
650
|
3.4
|
623
|
2.5
|
|
SK
|
731
|
3.9
|
656
|
2.6
|
|
UK
|
778
|
4.1
|
1599
|
6.4
|
|
18899
|
100.0
|
25117
|
100,0
|
Table 1Ordered logit estimation of life satisfaction model DISPLAY Table
Figure 1Individual counterfactual equivalent income DISPLAY Figure
Figure 2Average incomes and equivalent incomes DISPLAY Figure
Figure 3Income and equivalent income inequality DISPLAY Figure
Figure 4Decomposition of variance of logarithms of individual equivalent incomes DISPLAY Figure
Table 2Decomposition of cross-country variance of levels and changes in social welfare DISPLAY Table
Table 3Variance decomposition of equivalent income-based social welfare DISPLAY Table
Figure 5Decomposition of growth in equivalent income-based social welfare DISPLAY Figure
Table A1Number of observations per country and year DISPLAY Table
Table A2Variable definitions and summary statistics DISPLAY Table
Figure A1Gini coefficients for original and rescaled incomes DISPLAY Figure
Figure A2Decomposition of the difference between the Gini coefficient for original and rescaled incomes DISPLAY Figure
* The authors would like to thank the two independent reviewers for their most useful suggestions and comments.
1 Countries (abbreviations) are: Austria (AT), Belgium (BE), Bulgaria (BG), Cyprus (CY), Czech Republic (CZ), Germany (DE), Denmark (DK), Estonia (EE), Greece (EL), Spain (ES), Finland (FI), France (FR), Hungary (HU), Ireland (IE), Italy (IT), Lithuania (LT), Luxembourg (LU), Latvia (LV), Malta (MT), Netherlands (NL), Poland (PL), Portugal (PT), Romania (RO), Sweden (SE), Slovenia (SI), Slovakia (SK), United Kingdom (UK). The abbreviations will be used in figures throughout the paper.
2 Of the current EU member states, only Croatia is left out, as it was not yet a member state during in this period.
3 Research has shown that the answers to these two types of question in existing surveys are highly correlated (Clark,  2016), more than the conceptual distinction would suggest.
4 Using the terminology from Decancq, Fleurbaey and Schokkaert ( 2015a; 2015b) and Decancq and Schokkaert ( 2016).
5 Decancq and Schokkaert ( 2016) argue that interpreting changes in equivalent income over time becomes difficult if one wants to retain the heterogeneity of preferences.
7 Based on the OECD equivalence scale: the first adult is counted as 1, the remaining adults as 0.5 each, and children as 0.3 each, where children are household members aged 0-13 years.
8 European Union Statistics on Income and Living Conditions – the official source of distributional statistics for the European Union of Eurostat.
9 The rescaling procedure is described in Online-Appendix. We also provide a comparison of inequality measures for incomes before and after rescaling and, in addition, analyse how much incomes are distorted by the recaling.
10 According to the Chow test.
11 Unlike our model, Decancq and Schokkaert's ( 2016) model of life satisfaction allows for preference heterogeneity and non-linearity of income and health effects on life satisfaction. The figures we report are based on preferences of their reference group, but those for other groups do not differ very much, and thus generally imply strong suffering from deprivations in non-income dimensions as well. The model is estimated on the 2008 and 2010 European Social Survey data for 18 countries. The reason for choosing this paper for comparison is that other papers estimate preferences for single countries and thus may not be comparable (Decancq, Fleurbaey and Schokkaert, 2015a; 2017; Decancq and Neumann, 2016; Decancq, Schokkaert and Zuluaga, 2016). However, even the estimates from these papers imply very high suffering from deprivations in non-income dimensions, of the order of magnitude estimated in the present paper.
12 For details, see the data section and Online Appendix.
13 The Atkinson index for equivalent income should not be confused with Tsui’s ( 1995) generalisation of this inequality measure to the multidimensional context.
14 We checked the correlation between the variance of logarithms of income/equivalent income and the Atkinson index that we used. The correlations are quite high. In the case of income, the level (rank) correlation is 0.96 (0.96) in 2007 and 0.76 (0.89) in 2011. In the case of equivalent income, the level (rank) correlation is 0.93 (0.92) in 2007 and 0.95 (0.95) in 2011.
15 We divide equation (9) by var(ln Y*) and multiply it by 100 to get the contributions add up to 100.
16 This decomposition is shown below in this section; see equation (12).
17 Precisely, net improvements, since not all non-income dimensions necessarily improved.
|
|
December, 2020
IV/2020
|
