J Popul Econ (2004) 17:369–383

DOI 10.1007/s00148-003-0157-y

The impact of cash-benefit reform on parents’ labour

force participation
Ghazala Naz
Department of Economics, University of Bergen, 5007 Bergen, Norway
(Fax: +47-55-959439; e-mail:[email protected])

Received: 6 May 2002/Accepted: 21 February 2003

Abstract. This paper evaluates the impact of a recent Norwegian family-

policy reform. The reform provides benefits of up to NOK 3,000 (approxi-
mately e 400) per month to families with one- to three-year-old children, who
do not utilize state-subsidized day-care centres. We investigate the reform’s
effect on parents’ labour force participation. We find that, on average, the
reform reduced women’s labour force participation and increased the spe-
cialization of work between couples. We find that the effect of the reform
depends on women’s schooling. Specifically, the labour force participation of
highly educated mothers fell by more than that of mothers with less education.

JEL classification: I38, J16, J22

Key words: Public policy, cash benefits, specialization

1. Introduction

Many empirical studies support the notion that the presence of young
children typically increases the value of women’s time at home (see Gronau

I am grateful to Associate Professor Espen Bratberg and Professor Alf Erling for their valuable
comments. My special thanks to Kjell Vaage, Arild Aakvik and Afsane Bjorvatn for helpful
suggestions. This paper was presented in a seminar at the Department of Economics, University
of Bergen, Norway. I am thankful to the participants of the seminar for their positive feedback.
The paper has also benefited from the reports given by three anonymous referees. I am also
thankful to Professor Heather Joshi for helpful remarks on the paper. Responsible editor:
Christoph M. Schmidt.
370 G. Naz

1973; Leibowitz and Klerman 1995; Nakamura and Nakamura 1994), which
reduces their labour-market activity. Female labour force participation rates
in Scandinavia are the highest among the developed nations (see Table A in
the Appendix). In Norway, in 1997, the labour force participation rate of
married and cohabiting mothers was 75% if the youngest child was less than
three years old, and 83% if the child was from three to six years old
(Statistics Norway 1998). Empirical studies in Scandinavia and in other
Western countries suggest that high-quality public child care encourages
labour-market activity by women with pre-school children (see, for
instance, Anderson and Levine 2000; Gustafsson and Stafford 1992; Hofferth
1999; Hofferth and Collins 2000; Jenkins and Symons 2001; Kreyenfeld and
Hank 2000; Lundholm and Ohlsson 1998; Kravdal 1996; Van and Siegers
1996).
Blomquist and Christiansen (1995) and Bergstrom and Blomquist (1996)
argue that owing to the income-tax wedge between home production and
market work, too little market work is provided. They claim that public
provision of day care that stimulates female labour force participation may
improve the performance of the economy as a whole. Duncan and Giles
(1996) also discuss several efficiency-based arguments for subsidizing care.
However, they assert that the strongest arguments are based on distributional
concerns. They argue that child-care subsidies may minimize employment
gaps for those with child-care responsibilities relative to those without, and
could substantially improve women’s relative economic status by providing
them with income that is not dependent on their spouse.
In the spring of 1998, the Norwegian government decided to introduce
cash benefits of up to NOK 3,000 (approximately e 400) per month to those
parents with one- to three-year-old children who did not utilize state-sub-
sidized day-care facilities. This amount is roughly equivalent to the state
subsidy per child given to day-care centres. Parents who utilize some, but not
all, day-care facilities are entitled to receive a proportionally lower cash
benefit (see Table B in the Appendix). The cash benefits are neither taxable
nor tested against the parents’ labour-market participation or income. It is
therefore quite possible for both parents to work while receiving benefits.
Nonetheless, these benefits increase the costs of child-care centres for parents.
The empirical literature suggests that high child-care costs have a negative
effect on married women’s labour force participation (see for instance, Blau
and Robins 1988; Chiuri 2000; Klerman and Leibowitz 1990; Leibowitz et al.
1992; Powell 1997; Ribar 1992). Thus, the cash benefit reform may reduce
parents’ – especially mothers’ – labour force participation in Norway. Eval-
uating the effect of the reform on parents’ working behaviour is the main
objective of this paper. Two dimensions of parents’ working behaviour are
evaluated: specialization and total market intensity. Lundberg and Rose
(1999) define market intensity as the sum of the working hours of a couple,
and define specialization as the husband’s working hours minus the wife’s
working hours. We use these definitions in this paper.
The literature on household economics suggests that a large part of the
surplus in a marriage is generated by specialization. If one spouse has a
comparative advantage in the provision of home-time to commodity pro-
duction, while the other has a comparative advantage in earning the income
that is used to purchase market inputs, then family utility will be maximized
by specialization in the time allocation of the couple (see Becker 1973, 1985;
The impact of cash-benefit reform on parents’ labour force participation 371

Weiss 1997). There is ample evidence for specialization within the household.
Married men work longer hours in the labour market and have substantially
higher wages than single men. Married women have lower wages and work
more at home than unmarried women (see Gronau 1986; Daniel 1992;
Korenman and Neumark 1992). Owing to both social and biological factors,
many women have a comparative advantage in the production and care of
children (see Becker 1991; Lundberg and Rose 1999). We therefore suggest
that the cash benefit reform particularly increases the value of mothers’ time
at home. Whether it follows from this that specialization increases in
households, is an empirical question that is addressed in this paper. Lundberg
and Rose (1999) emphasize that child-care costs also affect parents’ total
market intensity. They find that in the United States, specialization increases
and the total market intensity of parents decreases following the birth of a
child. This motivates us to evaluate the effect of the cash benefit reform on
Norwegian couples’ total market intensity.
The impact of cash benefits on specialization and market intensity may
differ between groups of individuals according to their characteristics.
Becker describes human capital as an important determinant of specializa-
tion in households. Human capital analysis assumes that schooling raises
earnings and productivity in the market sector, which establishes a positive
association between females’ education and their labour force participation.
This in turn leads to a lower level of specialization between couples in which
wives have more schooling relative to couples in which wives have less
schooling. As mothers’ schooling is one of the important determinants of
specialization and of parents’ total market intensity, we investigate whether
the effect of the reform depends on mothers’ schooling. Nilsen et al. (2000)
find that women with more education, experience, and income are sig-
nificantly less likely to quit jobs in Norway. We investigate whether women
with a higher level of schooling were less motivated to reduce their hours of
paid work relative to those with a lower level of schooling, following the
cash benefit reform.
A number of empirical studies in Norway have compared women’s labour
force participation before and after the cash-benefit reform of 1998. Hellevik
(2000) and Rønsen (2001) report that mothers shifted from full-time work to
part-time work following the introduction of cash benefits. Langset et al.
(2000) estimate that the labour supply of working women was reduced most
in the health sector. These studies use cross-section data for women before
and after the reform. Langset et al. (2000) and Hellevik (2000) use only
descriptive-statistic techniques, while Rønsen (2001) runs multinomial
regressions on the following five alternatives: full-time work/subsidized care;
full-time work/non-subsidized care; part-time work/subsidized care; part-time
work/non-subsidized care; not working. A shortcoming of these studies is
that they discuss the labour force participation of the women who receive the
cash benefits without considering the problem of sample selection bias. In the
econometric literature, a wide range of non-experimental estimators has been
proposed to evaluate the unbiased effects of social programs (see Heckman
and Robb 1985; Moffitt 1991; Heckman et al. 1999). This paper addresses the
selection-bias problem and evaluates the reform using a difference in differ-
ence estimator.
The rest of the paper is organized as follows. Section 2 briefly reviews
Norwegian public policies directed towards families with children. Section 3
372 G. Naz

reports the data sources and explains our sample. Section 4 specifies the
econometric model used to evaluate the effect of the reform. Section 5 dis-
cusses the effect of the cash-benefit reform on specialization and the total
market intensity of couples. Section 6 concludes the paper.

2. Family policies and transfers to parents

An important feature of Scandinavian welfare states is the transfer of

resources to families with children. Other than enhancing equal opportunities
for children, the aim of these policies have had been to promote female labour
force participation (see Risa 1998). This section briefly reviews Norwegian
transfer policies directed towards families.

2.1. Maternity leave

Since 1993, working parents have been entitled to receive 52 weeks leave with
80% wage compensation, or 42 weeks leave with 100% wage compensation,
in connection with childbirth. The leave is funded by the state. The mother
must take nine weeks leave, three weeks before birth and six weeks after
delivery, while the father must take four weeks leave or forfeit it. Parents can
share the remaining time for paid parental leave however they wish. If the
female is not attached to the labour market, neither spouse can claim paid
leave, and the family receives a fixed amount of NOK 32,138 (approximately
e 4,000).

2.2. Child allowance

All residents in Norway with children under 16 years of age receive an

allowance of about NOK 900 (approximately e 115) per child per month.
This allowance is neither taxable nor means tested.

2.3. Taxation of families

Working couples are taxed independently in Norway. This tax system

provides strong incentives for female labour-market participation given that
tax brackets are quite progressive.

2.4. Single parents

Single parents can get transfers for living expenses until their children reach
the age of eight. During the first three years of the child’s life, transfers are
paid with earnings restrictions, but no further restrictions apply. If transfers
commence after the child’s third birthday, the single parent must be available
for employment suggested by the labour-market authorities, or be enrolled in
formal education that may enhance future employment opportunities. The
period of entitlement to lone-parenthood transfers may be increased by two
years if necessary for the completion of formal education.
The impact of cash-benefit reform on parents’ labour force participation 373

2.5. Subsidized day care

The costs of public and private day-care centres are shared by the state, the
municipalities, and the parents. Parental payments for day care vary
according to local rules and may be means tested. Knudsen (1998) estimated
cost-sharing arrangements in day-care centres for 1998. The average cost of
full-time public day care was NOK 11,500 (approximately e 1,400) per
month per child for one- to three-year-olds and NOK 6,700 (approximately
e 800) per month for three- to six-year-olds. The costs were slightly lower in
private day care. Average parental payments without means-tested reductions
were NOK 3,600 (approximately e 450) per month for all age groups.
The main reason for introducing the cash-benefit reform in 1998 was to
offer an alternative to subsidized day-care to parents by giving them freedom
of choice in child rearing. The cash-benefit reform is quite innovative inter-
nationally. It was first introduced in Finland in 1985. The Finnish and
Norwegian reforms are quite similar. The main component of both programs
is a flat-rate basic payment, but the Finnish system also has a means-tested
component and a siblings supplement. Unlike in Finland, the cash benefits in
Norway are not taxable.

3. Data and sample

The sources of our data are living-standard surveys conducted in April–May

1998 and 1999. The 1998 survey was carried out six months before the
introduction of the cash-benefit reforms, and includes 2,436 households. The
1999 survey was conducted six months after the reform and includes those
mothers in 1998 who still had a pre-school child as well as new mothers who
had given birth to a child between the surveys. The sample in 1999 included
3,334 households.
Missing variables reduce our data to 98% and 96% of the households
included in the surveys of 1998 and 1999 respectively. In the 1998 survey, 7%
of respondents are single parents, 30% are cohabitants, and 63% are in
married couples. In the 1999 survey, 9% are single parents, 30% are coha-
bitants, and 61% are in married couples. We exclude the single parents, but
include cohabitants.1 We also exclude women without a child and those with
children that are less than one year old.
Our final sample comprises 1,797 and 2, 741 households from the surveys of
1998 and 1999 respectively. We define parents with a child aged from one to
three, and parents with a child aged from three to six, as non-participants.2
Using the surveys of 1998 and 1999, we form a pseudo panel of the two groups.
A genuine panel tracks the same individuals over time, whereas a pseudo panel
tracks cohorts/groups over time. A cohort/group in a pseudo panel may very
well comprise different sets of individuals in each period. In a pseudo panel,
estimators are based on cohort/group means rather than on individual ob-
servations. Moreover, measurement error varies over time in a pseudo panel
because it is averaged over a different number of individuals at different points
in time. To minimize measurement-error variance, the individuals in each co-
hort/group should be as homogenous as possible (see Deaton 1985 for details of
pseudo panels). To get groups that are as homogeneous as possible, we define all
parents with a child aged from one to three included the 1999 survey (whether
374 G. Naz

they choose to receive the cash benefits or not) as participants. We do so because

we cannot identify those parents with a child aged from one to three who are
included in the survey of 1998 survey and who would choose not to receive cash
benefits were they given the opportunity.
The total number of observations in our sample is 4,538, of which 2,712
are for couples with children aged from one to three. The rest are for couples
with children aged from three to six. Some descriptive statistics for these two
groups are given in Table 1.
Table 1 shows that specialization between parents increases, and house-
holds’ total market intensity decreases, following the reform for participants,
whereas for non-participants, specialization between parents decreases, and
households’ total market intensity increases. It is quite likely that the direc-
tion of change in specialization and market intensity would have been the
same for participants and non-participants had the reform of 1998 not been
introduced. Thus, we use the difference in difference estimator, which is
explained in the next section, to evaluate the effects of the reforms. Table 1
also illustrates that wives with a university education have higher working
hours than women without a university education. Consequently, total
market intensity is higher and specialization is lower for wives with a uni-
versity education.
In almost 8% of the households in our sample, wives work more hours
than their husbands. We define one of our dependent variables, speciali-
zation, as the difference between the husband’s working hours and the
wife’s working hours. This is negative if the wife’s working hours are
higher than her husband’s, and therefore may result in underestimation of
the level of specialization. Thus, we also run regressions that exclude
couples in which the wife’s working hours are higher than her husband’s.
Since these regressions do not change our results much, they are not re-
ported.

4. Econometric specification

We are interested in evaluating the impact of the cash-benefit reform of 1998

on the specialization and total market intensity of Norwegian households.
For this purpose, we use a difference in difference estimator, which is defined
and specified in this section. Let Y be the outcome variables, specialization
and market intensity, and consider the following definitions:
Yit
¼ level of outcome variable for couple i who have a one- to three-year-
old child at time t if the cash-benefit reform had not been introduced;
Yit

¼ level of outcome variable for couple i who have a one- to three-
year-old child at time t after the cash-benefit reform was introduced.
The difference between these two outcomes is the effect of the reform,
denoted by a:

a ¼ Yit

 Yit
ð1Þ

Our aim is to obtain an estimator of a, the effect of the reform. The difficulty
is that we cannot observe Yit
directly for couples with a one- to three-year-old
The impact of cash-benefit reform on parents’ labour force participation 375

Table 1. Descriptive statistics for parents with children 1–3 and 3–6

Variable Mean

1–3 3–6

Before After Before After

Wife’s education 2.33 2.29 2.30 2.24

If Wife’s education university level 3.14 3.13 3.11 3.19
If Wife’s education lower than university 1.80 1.79 1.81 1.81
Husband’s education 2.42 2.45 2.40 2.42
If Wife’s education university level 2.90 2.92 2.97 2.99
If Wife’s education lower than university 2.12 2.17 2.06 2.14
Wife’s working hours (per week) 24.44 23.73 24.47 26.53
If Wife’s education university level 30.40 28.37 29.10 31.34
If Wife’s education lower than university 20.55 20.93 21.65 24.19
Husband’s working hours (per week) 40.88 41.27 40.80 40.75
If Wife’s education university level 41.29 41.32 41.26 40.51
If Wife’s education lower than university 40.62 41.23 40.54 40.87
Specialization 16.44 17.54 16.34 14.22
If Wife’s education university level 10.88 12.95 12.16 9.17
If Wife’s education lower than university 20.08 20.29 18.89 16.68
Market intensity 65.32 64.99 65.28 67.28
If Wife’s education university level 71.69 69.70 70.36 71.85
If Wife’s education lower than university 61.17 62.17 62.19 65.06
Number of children 1.60 1.53 1.46 1.27
If Wife’s education university level 1.66 1.59 1.43 1.30
If Wife’s education lower than university 1.57 1.49 1.50 1.27
Wife’s age 30.68 32.06 33.10 33.11
If Wife’s education university level 32.28 33.67 34.53 34.55
If Wife’s education lower than university 29.64 31.09 32.22 32.42
Dummy if wife in the health sector 0.23 0.21 0.20 0.20
If Wife’s education university level 0.29 0.28 0.27 0.27
If Wife’s education lower than university 0.18 0.18 0.15 0.16
Amount of cash benefits 1485
If Wife’s education university level 1327
If Wife’s education lower than university 1585
# of observations 1128 1584 669 1157
If Wife’s education university level 445 594 253 379
If Wife’s education lower than university 683 990 416 778

Education is a categorical variable from 1 to 4. 1 is below high school, 2 is college education, 3 is

bachelors and 4 is masters degree. In the paper we define levels 1 and 2 as lower than university
education and levels 3 and 4 as university education.

child because cash benefits have already been introduced. However, we know
the value of Yit
for non-participants. In our sample, parents with a child aged
from three to six are defined as non-participants. Hence, the estimator of the
reform’s effect can be obtained by estimating the difference between Yit

and
Yit
for participants and non-participants, respectively. This estimator is:
 
a ¼ ðY
it

d ¼ 1Þ  ðY
it
d ¼ 0Þ;
~ ð2Þ
376 G. Naz

where d ¼ 1 if couples have a child aged from one to three, and d ¼ 0 if

couples have a child aged from three to six. Y
it

d ¼ 1 and Y
it
d ¼ 0 are the
corresponding average values of Yit.
However, the estimator defined in (2) is likely to be biased because:
 
Y
it
d ¼ 1 6¼ Y
it
d ¼ 0: ð3Þ

Even if the cash benefits had not been introduced, the level of outcome variable
for parents with a child aged from one to three would probably differ from that
of parents with a child aged from three to six. This is mainly because parents’
labour force participation may vary with the age of their children. We can solve
this potential selection-bias problem by using the cohort data from the
pre-reform to the post-reform period. The following difference in difference
estimator can then be used to estimate the effect of the reform:





a ¼ ðY
it

 Y
it1
^ d ¼ 1Þ  ðY
it
 Y
it1 d ¼ 0Þ; ð4Þ



where Y
it

 Y
it1 d ¼ 1 is the change in Y between  the pre-reform and post-


reform periods for participants, and Y
it
 Y
it1 d ¼ 0 is the corresponding
change in Y for non-participants.
The estimator in (4) requires that the change for participants and non-
participants between the pre-reform and post-reform periods would have
been the same in the absence of the reform, that is:


 


ðY
it
 Y
it1 d ¼ 1Þ ¼ ðY
it
 Y
it1 d ¼ 0Þ; ð5Þ

where t-1 is the pre-reform period and t is the post-reform period.

To calculate ^ a, we estimate the following equation:

Y ¼ b0 þ b1 d1 þ b2 d2 þ b3 d1 d2 þ l; ð6Þ

where d1 ¼ 1 if parents have a child aged from one to three, d1=0 if parents
have a child aged from three to six, d2 ¼ 1 if the time period is after the
reform, and d2 ¼ 0 if the time period is before the reform.
Substituting (6) into (4) yields:

a ¼ ðb0 þ b1 þ b2 þ b3  b0  b1 Þ  ðb0 þ b2  b0 Þ ¼ b3 :
^ ð7Þ

To determine whether the effect of the cash-benefit reform depends on the

education level of women, we estimate the following equation:

Y ¼ b0 þ b1 d1 þ b2 d2 þ b3 d1 d2 þ c0 xi þ c1 d1 x þ c2 d2 x þ c3 d1 d2 x þ v; ð8Þ

where x ¼ 1 if the education level is high and x ¼ 0 if the education level is

low. The estimator of the effect of the reform for highly educated women is
equal to b3 þ c3 .
An alternative estimation method is to stratify the whole sample with
respect to women’s education levels and then estimate Eq. (6) for each stra-
tified sample. We run regressions by stratifying our sample in addition to
estimating Eq. (8) but get the same results. Although the standard errors are
The impact of cash-benefit reform on parents’ labour force participation 377

different, they do not change the significance of the coefficients (so the results
are not reported).
To control for other variables, we interact each variable with the same
variables with which we interact the education-level variable and estimate the
following equation:

Y ¼ b0 þ b1 d1 þ b2 d2 þ b3 d1 d2 þ c0 x þ c1 d1 x þ c2 d2 x þ c3 d1 d2 x
þ d1 d1 y þ d2 d2 y þ d3 d1 d2 y þ g; ð9Þ

where y is the vector of control variables.

5. Results and discussion

The main objective of our research is to evaluate the impact of the reform of
1998 on the specialization and market intensity of Norwegian couples. For
this purpose, we estimate Eq. (6) for each outcome. Table 2 reports the post-
reform effects of the cash-benefit reform.
We suggested in Sect. 1 that the reform of 1998 may have increased spe-
cialization among working parents. Our results confirm this suggestion for
our sample. We find that after the cash-benefit reform, couples with children
aged from one to three increased their specialization by 3.28 hours and
reduced their market intensity by 2.42 hours per week.
Our definitions of specialization and total market intensity are based on
the working hours of husbands and wives. Therefore, it is worth noting
changes in individuals’ working hours following the reform. Table 2 illus-
trates that wives did less paid work and husbands did more following the
reform. The increase in husbands’ paid work is insignificant. We find that
87% of the increased specialization and 85% of the decrease in the total
market intensity of couples is due to the reduction in working hours by
women. Our results are in line with the findings of Rønsen (2001) and
Hellevik (2000) that Norwegian women did less paid work following the
cash-benefit reforms.
The impact of any social program may differ between groups of
individuals according to their characteristics. Here, we examine the impact of
mothers’ schooling on the effects of the reform.
Household economic models predict a positive association between
females’ education and their paid work, which leads to a lower level of spe-
cialization between couples. Figure 1 illustrates the relationship between

Table 2. Effect of cash benefits on specialisation and market intensity

Variables b0 t ratio b3 t ratio

Specialization 16.18 21.97 3.28 2.75

Market intensity 65.63 87.89 )2.42 )2.00
Wife’s working hours 24.73 41.65 )2.85 )2.97
Husband’s working hours 40.90 92.02 0.43 0.60
Number of observations 4538

b0 ” constant; b3 ” reform’s effect.

378 G. Naz

Fig 1. Education is a categorical variable from 1 to 4. 1 is below high school, 2 is college

education, 3 is bachelors and 4 is masters degree. In the paper we define levels 1 and 2 as lower
than university education and levels 3 and 4 as university education

Fig 2. Education is a categorical variable from 1 to 4: 1 denotes below high school, 2 denotes
college education, 3 denotes bachelors degree, and 4 denotes masters degree In this paper, we
characterize 1 and 2 as below university-level education and levels 3 and 4 as university-level
education

education and average working hours for the women in our sample who have
one- to three-year-old children. Figure 2 illustrates the relationship between
the wife’s level of education and the level of specialization for couples with
one- to three-year-old children. As expected, there is a positive relationship
between a woman’s education level and her labour force participation, and a
negative relationship between a woman’s education level and the couples’
specialization in the pre-reform and post-reform periods. Nevertheless,
average working hours for highly educated women are lower and the level of
specialization is higher in the post-reform period than in the pre-reform
period.
To determine whether the effect on specialization and total market
intensity of the cash-benefit reform depends on the education level of women,
we ran regressions that interact our independent variables with dummies for
mothers’ education. The results are reported in Table 3.
Table 3. Difference in the effect of cash benefits with respect to wife’s education level

Variables b0 c0 b3 c3 b3 + c3 d3

Specialization
No control variables 18.73 (20.30) )6.57 ()4.44) 2.43 (1.64) 2.63 (1.09) 5.06 (2.68)
With control variables 19.35 (13.34) )6.81 ()4.09) 3.63 (1.65) 2.00 (0.74) 5.63 (1.75)
If husband’s education university 2.29 (0.87)
# of children > 2 )0.15 ()0.03)
Works in the health sector )1.19 ()0.41)
Wife’s age > 30 )2.65 ()1.04)
Market intensity
No control variables 62.63 (66.94) 7.72 (5.14) )1.90 ()1.26) )1.59 ()0.65) )3.49 ()1.82)
With control variables 61.55 (41.85) 6.88 (4.08) )4.60 ()2.05) )2.82 ()1.04) )7.42 ()2.27)
Husband’s education university 2.33 (0.87)
# of children > 2 5.62 (0.94)
Works in the health sector 2.53 (0.86)
Wife’s age > 30 1.62 (0.63)
Wife’s working hours
No control variables 21.95 (29.87) 7.15 (6.06) )2.17 ()1.83) )2.11 ()1.10) )4.28 ()2.84)
With control variables 21.10 (18.28) 6.85 (5.17) )4.12 ()2.34) )2.41 ()1.13) )6.53 ()2.54)
Husband’s education university 0.02 (0.01)
# of children > 2 2.89 (0.61)
Works in the health sector 1.87 (0.81)
Wife’s age > 30 2.14 (1.05)
Husband’s working hours
The impact of cash-benefit reform on parents’ labour force participation

No control variables 40.68 (71.57) 0.58 (0.63) 0.27 (0.29) 0.52 (1.35) 0.79 (0.68)
With control variables 40.45 (45.16) 0.04 (0.04) )0.48 ()0.35) )0.41 ()0.25) )0.89 ()0.45)
Husband’s education university 2.31 (1.42)
Children > 2 2.73 (0.75)
Works in the health sector 0.67 (0.38)
Wife’s age > 30 )0.52 ()0.04)

t ratio in parenthesis; b0 ” constant for low education level; b0+c0 ” constant for high education level; b3 ” reform’s effect on low educated; c3 ” difference in the
effect of the b3+c3 ” reform for low and high educated; reform’s effect on high educated; d3 ” coefficients for control variables.
379
380 G. Naz

We find that the reform affects households in which the wife lacks a university
education as well as households in which the wife has more schooling. The
difference between the reform’s effects for the two types of household are
statistically insignificant, but the reform’s effect is larger if the wife has a
university education.
Households in which the wife does not have a university education spe-
cialized by 2.43 hours more following the reform, whereas those in which the
wife was university educated specialized by 5.06 hours more. The decrease in
the total market intensity of couples is lower, but not significantly, if the wife
does not have a university education, whereas total market intensity falls
by 3.49 hours per week in households in which the wife has a higher level of
schooling.
Table 3 reports changes in individuals’ working hours for the two groups
of women. Changes in specialization and market intensity are largely
explained by lower working hours by women, since the changes in
husbands’ working hours are insignificant for both groups. Women without
a university education do 2.17 hours less paid work, whereas the corre-
sponding reduction for university-educated women is 4.28 hours. Control-
ling for age, industry, number of children, and the husband’s education does
not change the basic pattern of the results, though we get larger coefficients
for the outcome varialbes. All control variables are statistically insignificant.
F-tests show that the control variables are also jointly insignificant (results
not shown).
A greater effect of the reform in households in which the wife has a
university education relative to households in which the wife has less
schooling requires explanation. We suggest that one of the reasons why the
effect of the reform differs for more- and less-educated wives is a difference in
their respective use of child-care facilities. The use of subsidized day care is
more common among higher-educated mothers (see Hellevik and Koren
2000). It is quite likely that less-educated working women who were not using
subsidized day care before the reform made other arrangements (such as
using husbands, relatives or private nannies to take care of the children
during working hours). Since the cash-benefit reform did not raise the price of
non-subsidized child care, it would have had a smaller effect on the labour
force participation of less-educated mothers’ than on that of highly educated
mothers.
Comparing with findings from Finland suggests that favourable and
flexible labour-market conditions might be important in explaining the effects
of cash benefits on less- and higher-educated women in Norway. The un-
employment rate in Norway is around 3%, whereas in Finland it was below
3% in the late 1980s, rose sharply in the 1990s and is currently above 9%
(OECD 2002; Statistics Finland 2002). Studies on Finnish data suggest that
the education level of mothers did not affect the choice to receive cash benefits
during the late 1980s, but during the 1990s, mothers who kept their jobs, who
were often highly educated, were much less likely to use child-care cash
benefits in Finland (see Rissanen and Knudsen 2001). These findings from
Finland suggest that the low rate of unemployment in Norway may explain
why the cash-benefit reform affected households in which the wife did not
have a university education as well as households in which the wife had a
higher level of education.
The impact of cash-benefit reform on parents’ labour force participation 381

6. Conclusion

The main aim of this research has been to evaluate the effect of the recent cash-
benefit reform on the specialization and individual working-hours of
Norwegian couples. We found that, on average, women did less paid work
following the reform. This led to an increase in specialization between
husbands and wives, and a decrease in market intensity in households.
Husbands’ working hours did not increase significantly. These findings suggest
that the cash-benefit reform clearly departs from the strand of Norwegian
family policies, which increase women’s labour-market participation.
Another objective of this paper was to determine whether the effect of cash
benefits depended on the mother’s education. We found that women with a
low or high level of education reduced their working hours following the
reform, which led to an increase in specialization between husbands and
wives, and we found that this effect was stronger for highly educated mothers.
We found that the cash-benefit reform reduced paid work by women,
particularly by highly educated women. This reduction may have led to
efficiency losses of the kind described by Blomquist and Christiansen (1995)
and Bergstrom and Blomquist (1996) mentioned earlier. However, a full
analysis of the efficiency and distributional effects of the cash-benefit reform
requires a richer data set than the one used in this paper.

Appendix

Table A. Females labour force participation aged 15-64 in the year 2000

Countries Percentage in the labor force

Belgium 56.6
Denamrk 75.9
Finland 61.7
France 65.7
Netherlands 65.7
Norway 76.5
Sweden 76.4
United Kingdom 68.9
United States 70.8

Source: OECD Employment outlook.

Table B. Distribution of cash benefits by the use of subsidized care

Time per week in the day care Cash benefits per month

No day care 3000

1–8 hours 2400
9–6 hours 1800
17–24 hours 1200
25–32 hours 600
> 32 hours 0

Source: Social Office, Norway.

382 G. Naz

Endnotes
1
Cohabitants with one or more children probably behave like married couples and are also
treated as married couples as far as public policies are concerned. Running separate regressions
for cohabitants and married couples gave us the same results
2
Children aged from one to three are between 12 and 35 months old, whereas children aged from
three to six are between 36 and 71 months old.

References
Anderson PM, Levine PB (2000) Child Care and Mother’s Employment Decision. In: Card DE,
Blank RM (eds) Finding Jobs: Work and Welfare Reform. Russel Saga Foundation, New
York
Becker GS (1973) A Theory of Marriage. Journal of Political Economy 81 (4):813–846
Becker GS (1985) Human Capital, Effort, and the Sexual Division of Labor. Journal of Labor
Economics 3 (1–2):33–58
Becker GS (1991): A Treatise on the Family. Harvard University Press, Cambridge, MA
Bergstrom T, Blomquist S (1996) The Political Economy of Subsidised Day Care. European
Journal of Political Economy 12:433–457
Blau D, Robins P (1988) Child-Care Costs and Family Labor Supply. Review of Economics and
Statistics 70 (3):374–381
Blomquist S, Christiansen V (1995) Public Provision of Private Goods as a Redistributive Device
in an Optimum Income Model. Scandinavian Journal of Economics 97:547–567
Chiuri MC (2000) Quality and Demand of Child Care and Female Labour Supply in Italy.
Labour 14 (1):97–118
Daniel K (1992) Does Marriage Make Men More Productive? Working Paper No 92-2,
Economic Research Center, NORC
Deaton A (1985) Panel Data from Time Series of Cross Sections. Journal of Econometrics
30:109–126
Duncan A, Giles C (1996) Should We Subsidise Pre-School Childcare, and If So, How? Fiscal
Studies 17 (3):39–61
Gustafsson S, Stafford F (1992) Child Care Subsidies and Labor Supply in Sweden. Journal of
Human Resources 27:204–230
Gronau R (1973) The Effect of Children on the Housewife’s Value of Time. Journal of Political
Economy 81:168–199
Gronau R (1986) Home-Productiona – Survey. In: Ashenfelter O, Layards R (eds) Handbook of
Labor Economics. North-Holland, Amsterdam
Heckman JJ, Lalonde RJ, Smith JA (1999) The Economics and Econometrics of Active Labor
Market Programs. In: Ashenfelter O, Card D (eds) Handbook of Labor Economics. North-
Holland, Amsterdam
Heckman JJ, Robb R (1985) Alternative Methods for Evaluating the Impact of Interventions:
An Overview. Journal of Econometrics 30:239–267
Hellevik T (2000) Småbarnsforeldres yrkesdeltakels og valg av barnetilsyn før og etter
kontantstøttens innføring (Families’ labour force participation and childcare choice before
and after the introduction of cash benefit reform) NOVA-Rapport 2/00 Norwegian Social
Research. Oslo (in Norwegian)
Hellevik T, Koren C (2000) Parents Between Work and Care: Effects of the new cash benefit
scheme. Nova Skriftseria 7/00 Norwegian Social Research, Oslo
Hofferth SL (1999) Child Care, Maternal Employment and Public Policy. Annals of the American
Academy of Political and Social Science 563:20–38
Hofferth SL, Collins, N (2000) Child Care and Employment Turnover. Population Research and
Policy Review 19 (4):357–395
Jenkins S, Symons EJ (2001) Child Care Costs and Lone Mother’s Employment Rates: UK
Evidence. Manchester School 69 (2):121–147
Klerman JA, Leibowitz A (1990) Child Care and Women’s Return to Work after Childbirth.
American Economic Review 80 (2):284–288
The impact of cash-benefit reform on parents’ labour force participation 383

Knudsen K (1998) Knotantstøtten: en alternative økonomisk vinkling. (Economic Aspects of

Cash Benefits) Sosialøkonomer 6 (in Norwegian)
Korenman S, Neumark D (1992) Marriage, Motherhood, and Wages. Journal of Human
Resources 27:233–255
Kravdal Ø (1996) How the Local Supply Day-Care Centres influences Fertility in Norway: A
Parity Specific Approach. Population Research and Policy Review 15:201–218
Kreyenfeld M, Hank K (2000) Does the Availability of Childcare Influence the Employment of
Mothers? Findings from West Germany. Population Research and Policy Review 19
(4):317–337
Langset B, Lian B, Thoresen TO (2000) Kontantstøtten – hva har skjedd med yrekesdeltakelsen?
(Cash benefits – what happened to labour force participation?) Økonomisk Analyse 3 (in
Norwegian)
Leibowitz A, Klerman JA, Waita LJ (1992) Employment of New Mothers and Child Care
Choice: Differences by Children’s Age. Journal of Human Resources 27 (1):112–133
Leibowitz A, Klerman JA (1995) Explaining Changes in Married Mothers’ Employment Over
Time. Demography 32 (3):365–378
Lundberg S, Rose E (1999) The Determinants of Specialization Within Marriage. Mimeo,
University of Washington, Seattle
Lundholm M, Ohlsson H (1998) Wages, Taxes and Publicly Provided Daycare. Journal of
Population Economics 11 (2):185–204
Moffitt R (1991) Program Evaluation with Nonexperimental Data. Evaluation Review 15
(3):291–314
Nakamura A, Nakamura M (1994) Predicting Female Labour Supply: Effects of Children and
Recent Work Experience. Journal of Human Resources 29 (2):304–327
Nilsen A, Risa AE, Torstensen A (2000) Transitions From Employment Among Young
Norwegian Workers. Journal of Population Economics 13:21–24
OECD 2002 Employment Outlook. OECD, Paris
Powell LM (1997) The Impact of Childcare Cost on the Labour Supply of Married Mothers:
Evidence from Canada, Canadian Journal of Economics 30 (3):577–594
Ribar DC (1992) Child Care and Labor Supply of Married Women. Journal of Human Resources
27:134–165
Risa, AE (1998) Family Policy in Norway. Working Paper 22, Centre for Economic Studies in
Social Insurance. University of Bergen, Bergen
Rissanen T, Knudsen C (2001) The Child Home Care Allowance and Women’s Labour Force
Participation in Finland, 1985–1998. Nova Skriftseria 6/01 Norwegian Social Research, Oslo
Rønsen M (2001) Market Work, Child Care and the Division of Household Labour. Statistics
Norway Report 2001/3, Oslo
Statistics Finland (2002) Labour Force Statistics. Statistics Finland, Helsinki
Statistics Norway (1998) Labour Force Statistics. Statistics Norway, Oslo
Van DL, Siegers JJ (1996) The Effect of Supply of Subsidised Day Care Facilities on Female
Labour Supply. Labour 10 (3):559–582
Weiss Y (1997) The formation and dissolution of families: Why marry? Who marries whom? And
what happens upon divorce? In: Rosenzweig MR, Stark O (eds) Handbook of Population and
Family Economics. North-Holland, Amsterdam