1.5 An increase in payment generosity
1.5.1 At the threat points
In this section I examine how the policy change affects the bargaining position of the par- ents. As described in Section 1.2, the relative bargaining position in the current set-up is determined by the relative well-being of the parents in the case of a disagreement where they are assumed to act non-cooperatively according to a non-cooperative Nash bargaining game. With the estimated parameters I can simulate also the non-cooperative (unobserved) outcomes before and after the policy change. If one parent gains in utility relative to the other parent as a result of the policy, the policy has changed the bargaining positions of the parents. Such an effect may have important consequences for the well-being within the household. As described in Section 1.2 and in the illustration of Figure 1.3, when the bar- gaining power of one parent increases relative to the other parent the preferences of the first parent will be weighted higher in the cooperative solution. This in turn will change the resource distribution within the household and this parent would be relatively better off.
Table 1.8 shows the outcomes for the non-cooperative equilibrium. In the first column we find the baseline predictions, in the second column the predictions after the change in policy, and the third column shows the percentage change from baseline. A first thing to note is that at the baseline both the mother and the father work on average much more than in the cooperative solution. This is a result of the nature of the consumption good. As this is private and can only be paid for by private income the parent him/herself needs to work to consume. Moreover, in the set-up of the non-cooperative game the parents do not take the well-being of the other parent into account when they make their decisions. As such they will spend more time on the goods from which they derive utility, not taking into account the effects of this decision on the utility of the other parent. As a result of this time
40A non-linear effect in time spent with the child may explain why Carneiro et. al. (2009) find a positive
effect of maternity leave on child outcomes, while the other natural experiment studies do not (see Baker and Milligan 2008, 2010, 2011; Dustmann and Schnberg, 2008; Liu and Nordstrm Skans, 2010 and Rasmussen, 2010).
allocation, the child development is lower for non-cooperating couples. The consumption of the mother is much lower than the consumption of the father which is mainly due to her working less and paying all the costs for non-parental childcare. As we can see, the childcare usage is fairly high.
Table 1.8: Changes in the case of no cooperation
Outcome Baseline New policy Percentage change Total hours mother 1077.72 950.27 -10.49
Total hours father 2458.19 2456.55 -0.06 Child development 5.45 5.55 1.91 Consumption father 18324.07 18309.48 -0.06 Consumption mother 6380.83 6721.95 19.73 Child care 1123.07 992.41 31.51 Maternity leave paymentc 2016.66 3162.34 67.18
Note that the hours of work of the mother are conditional on mother working under the original policy, that the non-parental childcare variables are conditional on using non-parental care under the original policy, and that the maternity leave payment change is calculated only for mothers who had payments in the first period.
From the second and third columns we can see that with the current change in policy the mother would, in the case of disagreement, decrease her working time about 10 percent on average. The father would also decrease his working time because of the complementarity in child development production, but the decrease is very small. This move has a negligible negative effect on his consumption. The average maternity leave payments would increase from about 1026 to about 3162 and the mother increases her consumption of about 20 percent. Due to the decrease in working time of the mother the child development increases by around 1.91 percent from the baseline prediction.
If we compare how much better off the parents are after the policy change we first note that both parents are better off. Measured in consumption units the mother is 830 better off and the father is 1350 better off. However, even if it seems like the father may have gained in utility relative to the mother, one cannot draw this conclusion from the consumption values. Due to the decreasing marginal utility of consumption and the fact that the father consumes much more than the mother under the original policy he needs a bigger amount of consumption than the mother to gain one unit of utility. Another way to look at this is to compare the relative utility of the parents before and after the policy change. The average relative utility of the father to the mother at the baseline is 0.5429 and after the policy change it is 0.5425. Thus mother seems to have improved her bargaining position slightly
relative to the father on average. However, it may be interesting to study this result closer over types. As we noted in Table 1.5, some mothers are estimated to have positive utility from work (63 percent) and one may expect that the results might look different because of this. The relative utility is shown in Table 1.9 and we can see that the high type mothers have a lower bargaining position after the policy change, whereas the low type mothers have a higher bargaining position. As hinted at, this is due to the high type mothers decreasing their labor supply while the fathers can enjoy more child development without sacrificing any consumption or labor supply. Therefore, the father increases his utility relative to the mother in a household where the mother is of a high type.
Table 1.9: Relative utility father to mother no cooperation Mother type
l h Total Baseline 0.5603 0.5327 0.5429 New policy 0.5572 0.5339 0.5425