4 Conclusion
We estimate a model of input demands for domestic work, analysing the use of maids services, the availability of a dishwasher and the time that partners spend on weekend and weekday days performing routine and generally disliked household chores. All the input data derive from the same source, allowing us to model all these equations simultaneously and thus capture correlations in unobservable factors affecting these demands. Of particular interest is the sensitivity of input demand to resource prices. These prices are captured by the regional price of maid service, regional electricity prices, and the opportunity costs of time for the partners. We use data from both France and the UK.

Our results are remarkably similar across countries, suggesting that our findings may be representative for highly developed economies. The results indicate that the decision to purchase maids services is negatively related to the price of maid service, and positively related to electricity prices, each partner’s opportunity cost and the availability of non‐labour income. Both maid and electricity prices vary little from region to region, resulting in some lack of precision. Women’s wages have the strongest marginal effect, with a 10% increase in her wages associated with a 1 percentage point increase in the probability of hiring a maid in the UK and a 0.6 percentage point increase in France. The association with men’s wages is less precisely measured and of a smaller magnitude, particularly in France. That the time freed by hiring a maid could alter the time devoted to activities other than household production, such as childcare or gardening, is something that we do not study in this paper, but it suggests that work expanding this analysis to incorporate other time uses would be worth exploring.

Each partner’s opportunity cost of time is likewise highly positively associated with access to a dishwasher. In this case, the resource is more sensitive to men’s than to women’s value of time. Women’s but not men’s opportunity cost of time is significantly associated with household time inputs. In households where she has a higher opportunity cost, she reports less time on these housework tasks while he reports more, suggesting that higher wages may be associated with negotiation power or that women earning higher wages may choose partners who do more housework! That cohabiting women in France and women cohabiting with men who have more earnings power in the UK spend less time on housework is also suggestive that power may be important. Cross‐sectional analysis does not allow us to distinguish between these alternatives, but the issue is certainly deserving of future work.

Earlier literature suggested a large impact of maids and appliances on female labour supply. While we find evidence in both countries that, all else equal, in households more likely to have maids services women report spending less time on housework on weekdays, we find a much more direct association between maids services and weekend time use. Where maid service costs more, both men and women report more time doing housework on weekend days, as might be expected if maids are disproportionately likely to engage in housework tasks that can be put off or delayed. To the extent that maids provide a better substitute for weekend than weekday housework, the impact of maid services on both men’s and women’s leisure time is likely to be more substantial than its impact on their labour supply.

We thank the anonymous referees, Daniel Hamermesh, Arthur Lewbel, Margaret McElroy, John Pencavel, Robert Pollak, Arthur van Soest, Frank Stafford, and participants at the 2011 SOLE Conference in Vancouver, the 2010 IATUR conference in Paris, and the 2010 Society of Government Economists meeting in Washington, DC, as well as seminar participants at Aarhus University, Zaragoza University and Cergy Pontoise University, for helpful comments. All errors are ours.