The first rows of each panel present results for the probit equation for maids services. Both coefficient estimates and analytic marginal effects are reported. These marginal effects are calculated for married couples with sample average opportunity costs, sample average ages, and one teenage child. All other covariates are assumed to take a value of zero. These covariate values generate a predicted probability of maid service of 3.6% in the UK and of 2.1% in France. The price effects indicate that in both the UK and France, maids services is negatively related to its own price, positively related to both partners’ opportunity costs of time, and positively related to the presence of non‐labour income. In the UK, higher electricity prices are associated with a higher probability of hiring a maid, but the effect is not statistically significant (estimates are shown in Appendix Table A1). The opportunity cost of the wife is significant at the 1% level in both countries. A 10% increase in her opportunity cost of time is associated with about a 1 percentage point increase in the probability of hiring a maid in the UK, and a 0.6 percentage point increase in the probability of hiring a maid in France. These marginal effects are substantial given that the base probability of hiring a maid is only 3.6% to 2.1%. The opportunity cost of the husband has a positive effect in both countries that is significant at the 5% level. A 10% increase in his predicted wage is associated with a 0.8 percentage point increase in the probability of hiring a maid in the UK, and a 0.3 percentage point increase in France. Thus the impact of the wife’s opportunity cost is both more statistically significant and larger in magnitude than that of the husband, particularly in France, where women spend more time on these tasks. Since women on average earn less than men, the predicted gender differential following an absolute rather than relative change in opportunity cost is even greater. Receipt of non‐labour income is significantly related to the probability of hiring a maid in both countries, having a marginal impact of approximately 2 percentage points in both countries. The wage of the maid has a negative impact as expected, but though its marginal effect is rather large (a 10% increase in the wage of the maid reduces the probability of having a maid by 2 percentage points in the UK and 1 percentage point in France), its effect is not precisely estimated. The cross‐country similarities are striking.
The next rows present results relating to the probability of having a dishwasher. As with maids services, in each country higher opportunity costs are associated with a higher probability of having a dishwasher. In this case, the marginal impact of his predicted wage is greater than the marginal impact of her predicted wage—with the difference being particularly large in the UK perhaps because British men spend more time on housework. Thus in the UK (France), a 10% increase in his opportunity cost is associated with a 7.6% (4.0%) higher probability of having a dishwasher, while a 10% increase in her opportunity cost of time is associated with only a 1.9% (3.1%) higher probability of having a dishwasher (the base level is 46% in the UK and 63% in France). Overall, her value of time seems to have a greater impact on the decision to purchase maid services particularly in France where she spends more time on these activities, while his value of time has a greater impact on the probability of having household appliances particularly in the UK where he spends more time on these activities. The price of maid services is negatively related to having a dishwasher, significantly so in France. The receipt of non‐labour income has a positive effect in both countries, but this effect is statistically significant only in France. Electricity prices in the UK are negatively associated with having a dishwasher, but not significantly so.
Strong cross‐country similarities persist in the partners’ housework time equations. Looking first at the opportunity costs, results in the first column indicate that his predicted wage has no consistent association with household time use. Only in the case of her weekday time in France is the association individually statistically significant—and in this case positive indicates that as his opportunity cost rises, she spends more time on housework. In both countries, a joint test of the statistical significance of his predicted wages in the four household time‐use equations indicates there is no significant relation (p‐value 0.41 in the UK and 0.39 in France).
By contrast, her opportunity cost has a quite consistent and statistically significant association with housework time. Higher predicted wages for the wife are consistently associated with more housework time by the husband, significantly so in three of four cases. A 10% increase in her predicted wage leads to a 0.3–0.7 minute increase in his reported weekday time, and a 2.1 minute increase in his reported weekend time. These magnitudes translate to about a 5% and a 10% increase in his average weekday and weekend housework time. In three of four cases, an increase in her predicted wage is associated with a statistically significant reduction in her housework time. In the UK, a 10% increase in her wage is associated with a 3 minute decrease in her weekday housework time—a decrease of almost 4% compared to the sample average. In France, a 10% increase in her wage is associated with a decrease of over 7 minutes in her weekday time and about 4 minutes in her weekend time, differences of between 7% and 3% of sample means. The larger magnitude of the results in France is likely attributable in part to the greater amount of time that French women spend on housework. More generally, in both countries the estimates indicate that when her opportunity cost increases, he spends more time and she spends less time on housework. Thus men and women are not complements but likely substitutes in production.