VENTANAS, PUERTAS DE ALUMINIO Y TABIQUERIA VIDRIADA

Raising a child is connected with considerably high financial costs for the household. Not only due to the cost of raising the child itself, but additionally due to the loss of one full income, at least for a certain amount of time. Our analyses on mothers’

employment showed that fathers with an non-employed partner increased their working hours. Beside the reason that specialisation between the partners enables him to invest more time in the labour market, this result could be an indicator that fathers need to compensate for the partner’s income loss with increased time investments in labour

work.

We therefore explore the financial resources of the household as a potentially important determinant for fathers’ flexibility in preferred and actual work hours. To capture

financial constraints, we first control for mother’s wage and, second, we differentiate father’s wage into three groups (Table 4).

Financial factors may not just play an important role for men’s number of working

62

(Lundberg and Pollak 1994) predict that the partner with more labour force relevant resources, such as income, has more bargaining power within the relationship and can pass housework to the partner and spend more time with labour work. We analysed the effect of mother’s wage, divided in four groups (no wage, low wage, medium wage and

high wage). Interestingly, her wage did not show significant effects, neither for his work hours, housework hours, nor his work hour preferences. Thus, her contribution to the household income does not seem to affect his responsibility as main provider. Mother’s income is strongly correlated with her employment status in our dataset, and

because it shows no significant effects it is not included in our analyses to avoid distortion.6

In combination with the previous analysis (Table 3), these results indicate that the father’s involvement in household and labour work are more determined by the couple’s time restrictions rather than partners’ relative power due to labour market

relevant resources. This supports the results of Schober (2013) who analysed the changes in women’s and men’s housework or paid work time from before childbirth to

two years after the first birth. She did not find significant effects either for the mother’s wage on the father’s work hours in her analysis based on 16 waves of the BHPS.

Another relevant finding of her study is that British mothers seem to need a minimum level of earnings to be able to return to the labour market, probably due to the very high childcare costs in the United Kingdom. This result indicates that low-income households in the UK are less flexible with respect to changing traditional arrangements, because childcare costs prevent mothers’ returns to the labour market.

This assumption is supported by a recent report by the Family and Childcare Trust (2015), which revealed that increasing nursery costs in many cases make it cheaper for

63

one parent to stay at home than for both parents to be employed and pay for childcare. After childbirth, the man’s income is usually the main income of the family, even when mothers return to work. It may thus provide a good measure of families’ possibilities

to substitute or outsource household and care work.

For our analyses in Table 4, we divided men’s wages into three groups: (a) the low wage group has less than 60 per cent of the median wage, (b) the medium wage group is between 40 per cent less and 40 per cent more than the median wage, and (c) has more than 40 per cent than the median wage, on condition that the female partner’s wage does not exceed the average wage of women. Women’s high wages would remove fathers’ financial constraints to be the main provider and distort the results.

We expect that men with low wages who become fathers increase their work hours, or at least have a wish to increase their working hours, to compensate for income losses in the household. A medium wage makes it possible to provide financially for a whole family, despite mothers reduced labour work. We thus expect a negative effect of medium wages on his paid work and a positive effect on his housework hours due to the increased time investments which are necessary to raise a child. A relatively high wage could have contradictory effects on his working hours. On the one hand it enables him to provide for the family even when he reduces working hours to spend more time with them.

64

Table 4: Interaction Effect of First Child’s Birth (Child’s Age) and Wage Groups on Work Hour Preferences, Actual Working Hours and Housework Hours of Men in the UK.

Linear FE Model Correlated RE Logit Model Linear RE Model Correlated RE Logit Model Total working hours (+overtime) Work more than 48 hours (yes/no) Total housework hours Wish to reduce work hours Wish to increase work hours M1 M2 M3 M4 M5

Age of first child No child in HH (ref.) Up to and incl. one year old

-0.35 -0.02 1.09*** 0.29* -0.25

(0.332) (0.150) (0.170) (0.138) (0.254) Between one and five

years old

-0.04 0.20 0.56*** 0.12 0.33

(0.328) (0.141) (0.150) (0.118) (0.200)

Five years old and older -0.54 0.09 0.38* 0.00 0.44+

(0.460) (0.185) (0.184) (0.138) (0.236) Men’s hourly wage

Medium wage group (>60% <140% of median, ref)

Low wage group 3.82*** 1.09*** 0.22 0.37** 0.50*

(<60% of median) (0.302) (0.120) (0.164) (0.131) (0.205)

High wage group -2.19*** -0.34*** -0.07 0.12 -0.03

(>140% of median) (0.235) (0.096) (0.117) (0.092) (0.177) Interaction: age child*

men’s wage

medium wage OR no child (ref.)

<=1 year * low wage -0.72 -0.03 -1.63*** -1.24** 1.18* (0.871) (0.354) (0.479) (0.459) (0.538)

1-5 years* low wage 1.32+ 0.42 0.11 -0.09 0.35

(0.766) (0.302) (0.416) (0.342) (0.467) >=5 years * low wage 1.09 1.18*** -0.34 -0.08 -0.15

(0.834) (0.339) (0.459) (0.375) (0.556) <=1 year * high wage 0.29 0.06 -0.69* -0.51* -0.37

(0.503) (0.226) (0.273) (0.223) (0.525)

1-5 years* high wage -0.09 -0.22 -0.04 -0.30+ -0.61+

(0.424) (0.183) (0.217) (0.178) (0.364) >=5 years * high wage 0.46 0.05 -0.19 0.18 -0.83*

(0.517) (0.206) (0.251) (0.195) (0.404) Observations: 20147 20147 12246 12397 12397 Couples 5653 5653 2800 2882 2882 BHPS+ UKHLS BHPS+ UKHLS BHPS+ UKHLS BHPS only BHPS only Source: BHPS 1991-2008 + UKHLS 2010-2013 +p<0.10, * p<0.05, ** p<0.01, *** p<0.001 Standard errors in parentheses. Models include men’s: age, female partner’s employment status, education, number of employees at workplace, permanent job, overtime, time travel to work, social class, sector, wave. M2, M4 and M5 include means of all time varying covariates.

65

On the other hand, he may earn enough to support the family with his income alone, reducing the necessity for the mother’s return to full-time work. In this case his income

is needed to provide financially for the family and he should not reduce working hours. Additionally, higher wages mean he has higher opportunity costs for every hour less he works.

The results of the analysis for different wage groups (Table 4) do not support our predictions. Children do not have significantly different impacts on fathers’ work hours (Model 1) or their likelihood to work very long hours (Model 2) across the three wage groups. One exception is fathers of the lower wage group who are more likely to work 48 hours or more when the child reaches school age.

The high involvement in household work of fathers with children under one year old, evident in the previous analyses, is significantly different for fathers in each of the three wage groups. It seems that it is especially very small children, who require a high level of care, that make it necessary for fathers to increase their participation in domestic work. We see a higher involvement in housework for fathers with a medium wage with children in this age group, while fathers with a lower or higher wage are less likely to invest time in the private sphere.

For low and high wage fathers, a child under one has negative effects for the wish to reduce work hours, indicating that his work hour preferences are related to his involvement in housework. Fathers with child under one in the medium wage group want to reduce their working hours, while fathers of the two other wage groups show a negative effect in their preference to reduce their involvement in the labour market. An explanation of the negative effects for children on fathers’ housework hours in the lower wage group could be connected to more rigidities in low-wage jobs which make

66

changes in working hours less likely and thus affect his available time to spend with his family. An indicator that men in jobs in lower wage groups face more rigidity is our result that these men are more likely to wish to change their working hours. However, fathers in the higher wage group might be able to react to a mismatch between working hours and domestic work duties with buying external childcare and household work. We know from our analyses in Table 3 that the female partner’s employment status is one of the main explanatory factors for men’s work hour changes after childbirth. The

exit of the female partner after childbirth, and thus the loss of one income, may affect men with lower wages more than men who earn enough to support the family alone. To be able to see whether fathers with lower wages are more affected by the mother’s

income loss, in Table 5 we calculate separate models for the three wage groups and focus on men’s total working hour changes.

Table 5: Fixed Effects Models for Men’s Total Working Hours (+overtime) for Different Wage Groups (real hourly wage (<60% of average wage =low) (>60% of average wage =high)

All men All men + interaction Low wage Medium wage High wage One child in HH (yes) -0.12 0.13 3.78* -0.40 0.19 (0.223) (0.333) (1.953) (0.427) (0.521) Employment status (woman)

employed -0.10 0.04 2.38 -0.37 0.02

(0.232) (0.270) (1.500) (0.345) (0.413) Interaction between child (yes) *

woman’s employment status

Child (yes) * employed mother - -0.01 -0.19** -0.01 -0.01

(0.010) (0.062) (0.012) (0.014) Observations 19656 19656 1767 11569 6320 Couples BHPS+ UKHLS 5528 5528 1094 3972 2116 Source: BHPS 1991-2008+ UKHLS 2010-2013, + p<0.10, * p<0.05, ** p<0.01, *** p<0.001 Standard errors in parentheses. Models include men’s: age, education, number of employees at workplace, permanent job, time travel to work, social class, sector, wave.

Observation numbers differ to the analyses in Table 2, 3, and 4 (Model1) as we exclude men’s wage groups where the female partner has a relatively high income and thus might distort the results (125 couples, 2%). Households with children that have left the household are excluded.

67

We distinguish between fathers versus non-fathers and between employed versus non- employed partners, which leave us enough observations within each group to be able to include interactions between both variables. Since we have three separate models the results may be imprecise and should be interpreted with caution.

Nevertheless, our analysis indicates that fathers with a low hourly wage (60 per cent or less than the average) and a not employed partner are more affected by the birth of a child than fathers who earn more. Men in the lower wage groups increase their working hours by over three hours when they become fathers and their partner is not employed. The employment of their partners reduces the difference significantly. We find no significant differences for men with medium or high wages. An explanation could be that these households are less dependent on the female partner’s additional income than

households with a main earner who has a relatively low wage.