Education and entitlement to
household income.
A gendered longitudinal analysis
of British couples
Jerome De Henau and Susan Himmelweit
IAFFE annual conference, 22-24 July 2010, Buenos Aires
Aims
 Understanding determinants of satisfaction with household
income
 Indicator of access to household resources?
 Indicator of material (subjective) well-being?
 Investigating gender specialisation as source of efficiency
(gendered?)
 Educational homogamy and (gender) equality
 More similar gender role attitudes
 Reduced gain from specialisation
2
Motivation
 Long record of research on income inequality – also gender
income inequality
 Inside household, long record of research attempts to identify
how men and women in couples share their resources
 Also long record of research on financial satisfaction as proxy
for subjective (material) well-being (SMWB)
 Most analyses are carried out at individual level even for
couples
 Very little account of gender
 Investigating those missing links
3
Research on satisfaction
 Ferrer-i-Carbonell and van Praag (2003):
 income satisfaction inequality: if IS is equal for two
individuals with different characteristics then their welfare
is equal
 Analysis at individual level: implicit assumption that hh
income is equally (or at least fairly) shared between
partners
 Kristoffersen (2010):
 Measures of satisfaction represent valid accounts of
respondents’ evaluation of their situation (e.g. financial
situation / income)
Interpersonal comparability
Possibly interval / difference validity
4
Research on division of work
 Sigle-Rushton (2010):
 Weakly rejects Becker’s specialisation (efficiency gains in
marriage) – against many studies who found evidence of it
(e.g. Blau et al. 2000)
 Men’s unpaid work reduced probability of divorce
 But need better account of perceptions
 Can man’s and woman’s tasks have different economic
values for partners?
 What does it tell us about gains from specialisation /
individual access/command to income?
5
Access to income
 Satisfaction with household income is influenced by:
 Income, expenditure (costs), assets and debt
 Efficiency gains (from specialisation), economies of scale
(household composition)
 Degree of security of resources (risk sharing / capacity of
mobilising new resources through employment, savings,
etc.)
 Personal access to income for individual purposes
(including altruistic use)
 Unobserved fixed characteristics (personality, degree of
adaptation to adverse events)
 Other (potentially observable) variable characteristics
(importance of other aspects such as leisure time, social
life, etc., expectations of change and past change)
6
Data
 British Household Panel Survey (1996-2007)
 Stable couples of working age with or without children
 Household information (real annual household income,
number and age of children, home ownership)
 Individual information for each partner (potential wage, earns
> 75% of total earnings, weekly hours of housework, working
part-time, inactive, unemployed, long term disabled,
expectations of future financial situation, change in financial
situation since last year, whether saves each month, overall
satisfaction with life)
 Four educational groups (different gender role attitudes,
expectations and opportunities) – high/low educ.
7
Descriptive results
Sample size (obs. / groups)
Man's satisfaction with hh income
Woman's satisfaction with hh income
Real annual hh income (£)
Man 75-100% of hh earnings
Woman 75-100% of hh earnings
Man working full-time
Man working part-time
Man econ. inactive
Man unemployed
Man long term disabled
Woman working full-time
Woman working part-time
Woman econ. inactive
Woman unemployed
Woman long term disabled
Man's hours of housework (week)
Woman's hours of housework (week)
Educational group
both low man high / man low / both high
woman
woman
low
high
4244 / 822 2817 / 478 2086 / 373 4691 / 758
4.27
4.52
4.60
4.85
4.39
4.65
4.58
4.87
27,120
35,343
35,488
46,655
45%
49%
29%
34%
6%
3%
7%
4%
82.5%
92.0%
92.1%
92.5%
2.6%
2.2%
2.7%
4.3%
2.5%
1.2%
1.3%
1.5%
6.3%
1.9%
1.8%
1.2%
6.1%
2.6%
2.1%
0.5%
36.3%
42.7%
59.3%
56.0%
31.8%
31.5%
26.3%
28.8%
25.2%
21.9%
12.1%
13.0%
2.0%
1.7%
1.4%
1.3%
4.6%
2.3%
0.9%
0.9%
5.4
4.8
5.4
5.4
18.4
17.7
14.0
13.7
All
13998 / 2473
4.57
4.63
36,586
40%
5%
89.2%
3.1%
1.7%
3.0%
2.9%
47.7%
29.9%
18.5%
1.6%
2.3%
5.3
16.0
8
Descriptive results
both low
Man saves each month
Woman saves each month
Man's fin. sit. expected to improve
Man's fin. sit. expected to worsen
Man's fin. sit. not expected to change
Man's fin. sit. uncertain change
Woman's fin. sit. expected to improve
Woman's fin. sit. expected to worsen
Woman's fin. sit. not expected to change
Woman's fin. sit. uncertain change
Man's fin. sit. has improved
Man's fin. sit. has worsened
Man's fin. sit. hasn't changed
Man's fin. sit. has improved
Man's fin. sit. has worsened
Man's fin. sit. hasn't changed
Man's overall satisfaction with life
Woman's overall satisfaction with life
37.0%
36.6%
32.0%
7.2%
57.0%
3.8%
28.4%
5.7%
62.2%
3.6%
29.2%
22.7%
48.1%
29.1%
20.0%
50.9%
5.16
5.21
Educational group
man high / man low /
woman
woman
low
high
50.9%
45.3%
44.1%
48.9%
36.8%
42.0%
8.2%
5.9%
52.4%
49.6%
2.6%
2.5%
32.2%
37.3%
6.8%
8.1%
57.9%
52.4%
3.2%
2.2%
35.8%
38.7%
22.3%
19.0%
41.9%
42.3%
31.1%
37.6%
21.7%
21.2%
47.2%
41.2%
5.18
5.29
5.31
5.34
All
both high
58.1%
53.9%
40.3%
8.2%
49.8%
1.7%
35.6%
9.0%
53.6%
1.8%
42.0%
20.0%
38.1%
39.3%
21.4%
39.3%
5.31
5.33
48.1%
45.7%
37.3%
7.5%
52.5%
2.7%
32.9%
7.4%
57.0%
2.7%
36.2%
21.1%
42.6%
34.2%
21.0%
44.8%
5.24
5.29
9
Model specification
 Dep. Var.: individual satisfaction with hh income
 Linear fixed effect regression (Ferrer-i-Carbonell and Frijters
2004), for each partner and jointly with sex inter.
 Comparison of effects of own and partner’s characteristics on
individual satisfaction according to gender; we expect:
 If specialisation gains (and following gender norms), man’s
employment and woman unpaid work to be valued by both
(and symmetrically, man’s unpaid work and woman’s fulltime work to reduce income satisfaction of both partners)
 greater specialisation in low educated couples (traditional
GRA and more scarce resources)
 Higher individual expectations with own education
 Effects of variables presented in terms of the equivalent %
change in hh income
10
Selected overall effects (all sig. at 95%)
-100% -75%
-50%
-25%
0%
25%
50%
75%
100%
125%
150%
175%
200%
225%
250%
Man
No. children 0-4y **
Woman
Weekly hours of
housework (+10) **
75-100% of hh
earnings ***
Working part-time
**
Dif f betw een man's and
w oman's satis. signif .
* at 0.10
** at 0.05
*** at 0.01
Econ. inactive **
Unemployed ***
Fin. sit. has
improved
Fin. sit. has
worsened
Partner's weekly hours of
housework (+10)
Partner working
part-time
Partner econ. inactive *
Partner unemployed ***
Partner's fin. sit.
has worsened **
Partner's fin. sit.
has improved *
11
By level of education in hh (sig.>90%)
-100%
Nchi 0-4y
Both low
M high, F low **
-75%
-50%
-25%
0%
25%
50%
75%
100%
M
F
M low, F high *
Ptn >75%
hh
>75% hh
earnings earnings
Ptn
housewk
(+10)
Housewk (+10)
Both high
Both low **
M low, F high
Both high
Both low
M high, F low
Both low
191%
M low, F high ***
336%
M low, F high
Both high *
12
By level of education in hh (sig.>90%)
-100%
-50%
-25%
0%
25%
50%
75%
100%
Part-time
Ptn
parttime
Both low *
M high, F low
Both low
Ptn
ec.
inact. Ec. inact.
-75%
M
F
Both high *
Both low *
M low, F high
Both high
Both low *
Unempl.
Both low ***
M high, F low **
M low, F high ***
Ptn unempl.
Both high ***
Both low ***
M high, F low ***
M low, F high **
Both high **
13
Main results
 Gender effects quite striking:
 Men don’t actually care about their partner’s employment
of housework situation
 Both men and women value man’s employment
considerably more than woman’s (and so dislike male
unpaid work, inactivity or unemployment)
 A bit less for higher educated partners (but still present)
 Children remain a woman’s burden (remaining effect
seems to be time-use related as costs of young children
taken into account in financial situation change), except in
low educated couples (cost for both)
 However, women value their higher share of earnings
(especially when man is low educated) and value change in
their own situation more than in their partner’s
14