Abstract
maids services, household appliances and housework time are key inputs to domestic production. This study uses data from the UK and France to estimate the effects of resource prices on the demand for these inputs. We conclude that higher opportunity costs of time increase the likelihood of having maid services and appliances. Women’s time costs are also positively related to his housework time and negatively related to hers. Finally, maid service appears to be a closer substitute for housework time on weekend days than weekdays, suggesting smaller labour supply effects than anticipated by earlier literature.

Introduction
As is the case with most goods, domestic services are produced using a combination of labour and capital resources. The labour may be purchased in the market by hiring a maid or provided by household members at the cost of foregone time for employment or leisure. The capital resources consist of household appliances that reduce the amount of time necessary to complete domestic work. Maids, household appliances and household time inputs to domestic work all contribute to wellbeing by producing a pleasant home environment. Though output in this sector is not measured, inputs have been, and they indicate that these services are of substantial value. The US Economic Census indicates that in 2002, receipts for residential housekeeping services totalled 2.4 billion dollars, while sales of household appliances totalled over 14 billion dollars. Our calculations from the 2003 American Time Use Survey indicate that on average, couple households reported spending 28 hours per week on housework, which is about 44% of the time that these households devoted to market employment. Here we add to the literature in this area by estimating the effects of resource prices on the demand for all inputs to domestic production—maids, appliances and individual provision of housework. Measures of both market prices and individual opportunity costs of time are incorporated in the model, using rich data drawn from household and industry surveys in France and the United Kingdom (UK). By using data from more than one country, we gauge the degree to which our results are robust and might be generalizable.

Aguiar and Hurst (2005) were in the vanguard in this area of research with their work addressing meal preparation. They recognize that inputs to food production include not just food (modelled by food expenditures) but also the time spent shopping and preparing meals. By analysing both these inputs to food production, they resolve a long‐standing puzzle—the observation that food expenditures decline substantially following retirement. What they find is that these declining expenditures are offset by increased time spent shopping and preparing meals—suggesting that time and money are substitutes in food production. Stancanelli and van Soest (2012) added to this literature by allowing the retirement status of both partners to affect the hours of home production of both partners. Hamermesh (2007), using time diary and expenditure survey data from 1985 and 2003 in the USA, examines this mix of inputs for the working age population. He estimates linear input demand equations, specifying as inputs the raw food materials that make up meals and also the time devoted to buying food, preparing meals, eating them and cleaning up afterwards. His key explanatory variables are income and the husband’s and wife’s value of time. He concludes that income has a positive effect on all inputs, while the opportunity cost of time is negatively related to the time inputs. These key studies of household food production illustrate the importance of taking both market and household inputs into account in modelling the home production sector.

Most of the scant research to date on the domestic help industry (Suen 1994; Cortés and Pan 2013; Cortés and Tessada 2011—using microdata on foreign maids) and on home appliances (Cavalcanti and Tavares 2008; Coen‐Pirani et al. 2010) focuses on the positive impact these alternative inputs to domestic production have on women’s labour supply. Cortés and Tessada (2011) find, combining information from different data sources, that the greater availability of maids services, instrumented with migration flows, has increased the labour supply of high‐earning women by between 4 and 20 minutes per week. Their estimates suggest a modest decrease of about 7 minutes per week in housework. None of these studies addresses the role of men as either potential providers or consumers of household services.

Related research focuses on the impact of opportunity costs on the time inputs to domestic services. Suen (1994) reports that women’s predicted wages are significantly positively related to the probability of hiring domestic servants in Hong Kong. Cohen (1998) finds similar results in the USA, but also finds a weak positive relation to men’s earnings. The more extensive literature relating earnings to couples’ time use (see, for example, Hersch and Stratton 1994; Friedberg and Webb 2007) typically finds a negative relation between own opportunity costs and housework time.

We are not aware of any earlier study that has taken as comprehensive an approach to analysing the inputs to domestic work in couple households as we do here. While previous work has related the availability of maids services and appliances to female labour supply, we analyse the link between maid services, appliances, individual housework time and resource prices. The time allocated by each partner in couple households to housework is valued using potential wages, as is often done in the labour supply literature. The prices of maid services and of electricity are constructed using regional measures. By simultaneously modelling the demand for maid services, appliances and individual housework time as a function of these prices, we can shed some light on the degree of substitutability among these inputs. Furthermore, our data distinguish between weekend and weekday days, allowing us to estimate demand equations across different days of the week. Because many people in our surveys do not work for pay in the market on the weekend, housework done on weekend days is more likely to be carried out at the cost of foregone leisure time than reduced labour supply hours, a potentially relevant issue that has been overlooked in the earlier literature.