One widely reported finding in interdisciplinary socio-economic research is a reduction over the last decades of housework in partnerships (Aguiar and Hurst 2007; Coltrane 2000; Sayer 2005). Nevertheless, this development persistently fails to represent an equal division of work between the sexes. Comparing data from the German Federal Statistical Office’s 1991/1992 Time Use Survey (GTUS) with results from its 2001/2002 survey indicates that over the decade, women living in partnerships (with or without children) reduced time spent on housework substantially, whereas their partners increased theirs only marginally. Hence, despite a slight reduction in the disparity of housework division between the sexes, a great amount of inequality still exists. This is a finding that can also be observed in other countries (e.g., Baxter 2002; Gershuny 1999; Sousa-Poza et al. 2001; Sullivan 2000). Furthermore, there also appears to be some international evidence that while differences between genders are gradually getting smaller, this is primarily due to the relatively large decline in women’s time spent on household work and less on men increasing their share (Baxter 2002).

Previous research has offered some empirically based explanations for the unequal division of housework, such as a slow change in gender roles (Alvarez and Miles 2003; Gimenez et al. 2007), financial inequality between men and women (Anxo and Carlin 2004; Bech-Moen 2006; Blau and Kahn 2005; Hersch and Stratton 1994; Klaveren et al. 2008) and a combination of the two (Bittman et al. 2001; Breen and Cooke 2005; Haberkern 2005; Bonke et al. 2007). The aim of this study is to take a closer look at the change in time spent on housework in Germany over the past decade and, most notably, to try to determine the reasons for the persistent inequality in this time allocation between men and women. Besides being the first such study for Germany (and one of the few studies that analyses changes in time use using diary data), this study implements an empirical method that is not widely used, namely structural equation modelling (SEM). The main benefit of using such an approach is that interdependencies between the main determinants of housework time can be integrated into one model more easily. As our results will show, these interdependencies are important and largely neglected in current research. A further advantage of SEM is that it allows us not only to include conventional determinants in the analysis but also to procure important information usually not observed in conventional datasets.

The paper will focus on two main questions: first, how can the changes in the amount of housework carried out by men and women be explained? For this purpose data from the 1991/1992 GTUS are compared with those from the 2001/2002 GTUS so that changes in housework over this time period can be attributed to changes in the effects of some of these determinants. These findings provide the background for the second research question: where do we observe gender differences in the determinants of the allocation of time to housework and why is the unequal division of housework so persistent? By comparing the identical estimated models of men and women, the determinants of housework that have different effects for both sexes can be isolated.

The paper is organized as follows. Following a brief review of literature in “Previous Research”, “Data and Methodology” provides relevant information about the GTUS and SEM methodology. “Results” then presents first the descriptive findings followed by the comparative results for time use and the estimated models of housework for men and women. “Summary and Policy Implications” concludes and delineates policy implications.

Previous Research

Of the many recent studies addressing the phenomenon of the unequal intra-household division of work, the most pertinent to this present investigation are those that find gender-specific characteristics to be housework determinants, those that point to economic characteristics and those that cite both aspects. Other possible additional determinants are discussed at the end of the section.

Alvarez and Miles (2003), as well as Gimenez et al. (2007), identify an existing relationship between housework performed and gender aspects. For example, in analysing the Spanish TUS 1991, Alvarez and Miles (2003) explain the unequal division of housework by unobservable characteristics that are assumedly gender-specific differences rather than by observable spousal characteristics. Likewise, in their comparison of 15 European countries, Gimenez et al. (2007) find that the degree of specialisation between spouses is higher when gender-related norms are rather traditional, which accounts especially for Southern Europe.

Other studies accentuate the important role of the individual and/or partner’s economic characteristics (e.g., Hersch and Stratton 1994; Anxo and Carlin 2004; Blau and Kahn 2005; Bech-Moen 2006; Klaveren et al. 2008). First, the analysis of Hersch and Stratton (1994) of the 1979–1987 Panel Study of Income Dynamics (PSID) points to the important role of economic determinants for the housework hours performed. The larger the income differences between spouses, the greater the degree of specialisation, the more time women spend on housework. In contrast, the analysis of Anxo and Carlin (2004) of French TUS from 1999 suggests that in households where women are employed, not only is less housework done in general but women also show a positive cross-wage elasticity, implying that a higher wage for females results in a higher share of housework done by their male partners. However, Blau and Kahn (2005a) find a decrease in wage elasticities for women that resemble that for men, implying that the housework amount for American women may not be as sensitive to wage changes as it once was. A recent study by Klaveren et al. (2008) using the 2004 British Household Panel Survey also identifies a tendency for the working behaviour of women to become more equal to that of men. Nevertheless, as shown by the analysis of Bech-Moen (2006) of Norwegian TUS (1971–2000) and the American PSID (1971–1997), wages still have a major impact on housework. Particularly interesting is the analysis of Connelly and Kimmel (2007) that, based on data from the 2003/2004 American TUS, shows no relationship between an individual’s household time and its partner’s wage or partner’s labour market status.

Several authors identify both gender and economic aspects of housework determinants (e.g., Bittman et al. 2001; Breen and Cooke 2005; Haberkern 2005; Bonke et al. 2007). Among these, Bittman et al. (2001) use 1992 Australian TUS to show the important but not exclusive effects of economic characteristics on housework. Like Anxo and Carlin (2004), they find lower levels of housework for women with higher levels of employment, even though their partner’s housework time does not increase simultaneously. Even when market work and wages are similar for both spouses, women still perform more housework than their male counterparts, leading the authors to postulate a gender aspect. Based on an analysis of the International Social Survey Programme data from 22 countries, Breen and Cooke (2005) argue that the unequal share of housework between the sexes will persist as long as women’s economic situation does not equal that of men and as long as men do not change their gender ideology. Both Haberkern (2005) and Bonke et al. (2007), using 2001/2002 GTUS data and a comparison between 2001 Danish TUS and 1992–1994 American National Survey of Family and Households data, respectively, confirm that the degree of specialisation within couples is dependent on economic determinants, restrictions on time and gender-related norms.

The study by Bonke et al. (2007), which focuses on the impact of children on housework, concludes that Danish couples do not specialise as much as American couples, a finding they explain by child care seeming to be more of a market good in Denmark. Not only does this finding imply that dimensions other than economic and gender-specific factors may play a role in time spent on housework but it also accounts for the substitutability between home-produced and market goods. For instance, Knowles (2007) explains the reduction of female housework over the last decades in terms of the lowered prices of and enhancements to household appliances and their resultant wider distribution, whereas Gørtz (2007) states that despite easier access to market substitutes for household goods, women might experience extra benefits from goods they produce themselves.

Overall, the literature suggests that the determinants of housework have both economic and gender aspects, while current research on the sexual division of housework reveals that spousal effects have become less important.

Data and Methodology

The German Time Use Data

We use data from the German Federal Statistical Office’s 2001/2002 GTUS which covers a representative sample of 5,400 households that includes over 12,000 individuals and 36,000 time diary days. The GTUS not only provides information about the everyday life activities of diverse population groups but also demographic data like age, sex, employment status, the level of education and economic status of both the individual and the household. This current analysis concentrates on couples aged 20 to 59, meaning that 2,497 couples remain in the analysis. For reasons of comparison, the analysis also includes data from the 1991/1992 GTUS for individuals with corresponding characteristics to produce a sample of 3,871 couples.

The variable of interest, namely time spent on household activities, is constructed with the information from 3 days of the time-use diaries in which the average time per day is calculated by weighting weekdays and weekends, respectively.Footnote 1

Structural Equation Model

In this study, SEM is used—a common econometric approach in psychology and marketing, but not often encountered in sociology and economics. The main advantage of using this technique is that it allows us to model interactions among determinants. Furthermore, and as will become clearer below, SEM allows us to procure important information usually not observed in conventional datasets. We implement a partial least squares (PLS) model, also known as path modelling, which was developed by Wold (1982, 1985). Its minimal requirements for residual distributions and measurement scales make it very robust. In contrast to the better-known covariance SEM approach, it is a variance-based, iterative analysis based on multivariate regression that applies the least-squares algorithm (Fornell and Cha 1994). One limitation of the method, however, is that standard errors must be calculated via re-sampling procedures such as bootstrapping (Efron and Gong 1983).

Figure 1 outlines the model estimated in this paper. Time spent on housework is assumed to depend on the wage, aspiration level for market goods and household goods.Footnote 2 The aspiration for the consumption of market goods describes a minimum level of market goods and services an individual aspires to consume (see hereto Deaton and Muellbauer 1980; Simon 1978). The concept of household goods is basically defined as the consumption of goods and services that are mainly produced for the household itself and that are consumed by household members. Figure 1 also shows that the aspiration level of market goods as well as the consumed household goods depends on wages.

Fig. 1
figure 1

A structural equation model

Full size image

It is important to note that not all of these determinants can be measured directly as they are not included in the data set. With SEM, it is, however, possible to indirectly measure these concepts via different sets of observed variables. These measurements are described in the following section.

Definition of Variables

Housework

Housework is defined in the classical way and thus, includes all activities for the household that could have been done by a third party (Reid 1934). It incorporates activities such as the preparation of meals, cleaning, gardening, pet care, laundry, handicrafts, shopping, childcare, and care for adults in need (see Zick et al. 2008). Housework is recorded according to this definition in the diary. We computed housework in minutes per day by weighting the mean by weekdays and weekend days based on the assumption of different behaviour in each.

Household Goods

Household goods are assumed to be marketable; hence, it would be possible to either sell or buy them. This distinguishes household goods from leisure activities (Gronau and Hamermesh 2006). There are many examples for possible market substitutes of household goods: meals can be consumed at home or in a restaurant, parents could look after their children or buy child care. As in most datasets, the GTUS offers no output data for household goods, but the literature hints at several adequate indicators (e.g., Fitzgerald and Wicks 1990; Goldschmidt-Clermont 1982; Säntti et al. 1982). Drawing on these, this study integrates three different dimensions into the measurement of household goods. Thus, with the aid of SEM, we develop a construct that emulates the household goods consumed by an individual.

The first most important dimension of household goods consumption is the number and age of children living within the household (Künzler et al. 2001; Säntti et al. 1982). As different investments must be made in housework based on the number and age of children, children are divided into the following subgroups: those aged 0–1, those aged 2–6 and those aged 7–15. Also included in this dimension of housework is the number of adults in need of care (Kettschau et al. 2004).

Second, household goods produced by others—most notably by an individual’s partner—are assumed to be another important indicator of the level of consumption of household goods (Kettschau et al. 2004). An especially good indicator for household goods produced by the partner is his or her housework time (in average minutes per day). As housework may also be done by non-household members, the models include unpaid external housework (quantity of received help within the last week) and whether or not external unpaid childcare is used.

The third important dimension of household goods is housing characteristics. Specifically, a large living space (in 10 m2) implies high household consumption. Living in a house (as opposed to an apartment) and the possession of certain domestic appliances (number of dishwashers and laundry dryers) are also assumed to imply higher levels of consumption of household goods (Künzler et al. 2001).

Aspiration for Market Goods

The (non-observable) aspiration levels for market goods can be interpreted as preferences for the consumption of market goods that are influenced by earlier consumption levels and/or by the social environment (Ajzen and Fishbein 2000; Fehr and Götte 2005). Thus, measuring the aspiration for market goods is by definition rather explorative, and we are not aware that it has been done before. Nevertheless, consumer theory suggests three dimensions relating to this behaviour: individual characteristics, environmental parameters, and psychological processes (e.g., Blackwell et al. 2001; Solomon et al. 2006).

One of the most important aspects related to the individual characteristics is economic resources (Mitchell and Walsh 2004). Hence, the models incorporate own net income (in categories) and occupation (dummies for civil servants and white collar workers). However, because time resources constitute another individual aspect, two variables that measure aspects of an individual’s satisfaction with time use in leisure and market work are added in separately.Footnote 3 It is assumed that a higher job burden can be compensated by an increased aspiration for market goods (Blackwell et al. 2001), although the reverse is expected for leisure.

Aspiration for market goods is supposedly influenced by the social environment (Kroeber-Riehl and Weinberg 2003). As children aged 16–25 are said to influence parental consumption behaviour (Blackwell et al. 2001), a variable that depicts the number of children living in the household in this age group is incorporated into the models. Likewise, because car ownership ensures mobility and increases both the radius and the freedom of action, it is chosen to represent external environmental parameters.

Whereas little information is available about psychological processes, communicative equipment within the household—for example, telephones, internet connections and computers—assists in gathering information which, in turn, may affect aspirations levels (Mitchell and Walsh 2004). We, therefore, include three variables that capture the presence and quantity of such equipment in the household.

Wage

The wage data in the GTUS have two drawbacks: first, due to the introduction of the EURO in 2001, comparisons of wages between the 1991/1992 and 2001/2002 surveys is difficult. Second, precise wage information is only available for about 40% of all respondents. We, therefore, develop a construct that emulates wages. A first measurement is the relative hourly wage rates measured in quartiles. Furthermore, we incorporate the type of contract (full-time, part-time, marginally part-time or no employment), the age of the respondent (and a quadratic age term to account for non-linear effects), and three variables for the level of education.

Descriptive statistics as well as quality measures for these models are presented in Tables 4 and 5 in Appendix. Note that the exact same measurement models were used for both time periods and both men and women in order to allow closer comparison of time and gender differences.

Results

As outlined earlier, the study’s goal is twofold: to understand why individual household members, both male and female, change their amount of housework time and why gender differences in household time-use behaviour exist. This section first traces descriptively the development of housework time and then outlines a chronological comparison of the PLS models for men and women. Finally, it identifies the differences in time-use behaviour between the sexes.

Descriptive Analysis of Time Use

Between 1991/1992 and 2001/2002, men slightly increased the time they spent on housework by 6 min/day, but women reduced the amount of time for this activity by an enormous 42 min/day. One important aspect of the above time-use shift is that, over this decade, couples reduced their housework by more than 30 min/day, which is in line with findings for other countries such as the USA (see Bianchi et al. 2000).

Yet, even though women reduced and men increased their time spent on housework, the split is still unequal. In 2001/2002, among all couples, 66% of all housework was done by women, which equates to about 5 h and 20 min a day (SD 138.03), leaving about 2 h and 45 min a day (SD 115.84) to partners. Irrespective of household type and despite the convergence in the amount of time men and women spent on housework, a gender gap is still evident.

Results for the Chronological Change in Housework

Answering the research questions begins with a chronological comparison of the models for men and women. To this end, the 2001/2002 model is presented first and then compared with the 1991/1992 model.

In general, the 2001/2002 (1991/1992) model explains 35% (37%) of the amount of housework done by women in partnerships. The explained variance of housework for men is much smaller than that observed for women, about 15% (12%), which is a common finding (e.g., Sousa-Poza et al. 2001).

Women’s Housework in 2001/2002 and 1991/1992

Figure 2 shows the standardized path coefficients for women. The first coefficients in each path are for women in 2001/2002; those in brackets are for women in 1991/1992. The significance is tested by a bootstrapping re-sampling procedure, and the results are given in Table 5 in Appendix. All paths presented are highly significant.

Fig. 2
figure 2

Path model for women, 2001/2002 (1991/1992)

Full size image

In both time periods, a positive relationship between consumption level of household goods and housework time proves to be true. As shown by the strong path coefficients of 0.320 in 2001/2002 and 0.342 in 1991/1992, the effect of household goods—measured by children, housing characteristics, and social networks—is highly significant. However, once these two coefficients are compared, the importance of household goods for housework time drops significantly (see Table 1). One possible explanation for this drop is an increased productivity with regard to the production of household goods by women. This can be attributed to the increased availability of technical appliances at reasonable prices in households, i.e., in 2001/2002, households had a larger number of time-saving appliances than in 1991/1992 (Statistisches Bundesamt 2006). At the same time, fewer household goods are consumed, which is reasonable given the demographic changes shown by the GTUS data. Specifically, within this decade, the number of children aged 0–15 decreased dramatically. Ignoring the effect of wages at this stage, all these aspects might explain the reduced impact of household goods on the time allocated to housework.

Table 1 Comparison of path coefficients for women, 2001/2002 and 1991/1992
Full size table

The relationship between aspiration for market goods and housework is negative, with a significant effect of −0.063 for 2001/2002, meaning that the higher the aspiration level for market goods, the less housework will be done. This is reasonable as an increase in the aspiration level implies an increase in the marginal utility of market goods consumption. An increase in market goods consumption might occur through the partner or through more own market work. In the latter case, less time remains to be allocated to housework and leisure time. However, once this figure is compared to the impact of −0.123 for the aspiration for market goods in 1991/1992, its size diminishes substantially. One suggested explanation for such reduced housework is lower prices for adequate market substitutes (Albanesi and Olivetti 2007), which is indeed true for Germany (see Statistisches Bundesamt 2008).

When predicting the relationship between wages and housework, the common negative effect can be observed (−0.355 in 2001/2002 and −0.229 in 1991/1992). This finding is not surprising given the high opportunity costs associated with (time-intensive) housework. This will also affect the prices for household goods and for their market substitutes. This makes it reasonable also to investigate the interplay between wages and household goods consumption as well as between wage and the aspiration for market goods. The impact of wages on household goods consumption is significantly negative with −0.206 in 2001/2002 and −0.177 in 1991/1992. The higher the wage, the more expensive the housework needed to produce household goods, which is why, all else being equal, household goods become rather expensive. Interestingly, wages have gained in importance over time.

The same tendency emerges for the impact of wages on the aspiration for market goods consumption. There is a highly significant positive impact for both sets of survey data, but the path coefficient becomes stronger in 2001/2002 (0.844 vs. 0.819 in 1991/1992). While household goods are more expensive with increasing wages, market goods become relatively cheaper. Hence, the relative marginal utility shifts in favour of the consumption of (and hence aspiration for) market goods, which results in a substitution of household goods with market goods. When comparing the wage impact on the aspiration for market goods chronologically, we once again observe that the role of wages has gained in importance.

It is, therefore, not surprising that the impact of wages increases in opposite directions for the two variables: negatively for household goods and positively for the aspiration for market goods. Thus, one has to distinguish between the direct effect of wages on housework time and the indirect effects that result from the effect that wages have on household goods consumption and the aspiration for market goods. The total wage effect clearly increases between 1991/1992 and 2001/2002.

Based on these findings, the reduction in housework can be primarily attributed to the gained importance of wages, irrespective of whether the effects are direct or indirect.

Men’s Housework in 2001/2002 and 1991/1992

The models for men’s housework output are estimated in the same manner as those for women’s to allow comparison between the sexes. The effects are presented in Figure 3, and the bootstrapping results are given in Table 6 in Appendix. Interestingly, for men, not every path comes out to be significant and additional deviations from the results of women are observable.

Fig. 3
figure 3

Path model for men, 2001/2002 (1991/1992)

Full size image

As in the models for women, the consumption of household goods significantly influences the amount of housework done by men. Nonetheless, a comparison of the path coefficients—approximately 0.178 in 2001/2002 and 0.187 in 1991/1992—suggests that men’s behaviour with regard to this variable did not really change (see Table 2). The effect of the aspiration for market goods consumption on housework has diminished significantly between 1991/1992 and 2001/2002. Whereas in 1991/1992 we in fact observe a positive effect, in 2001/2002 no significant correlation is obtained. As regards wages, the relationship between this variable and housework is very strong—with a path coefficient of −0.330 and −0.315, respectively—and supports the assumption of a normal labour supply. Furthermore, no changes are observable over time, and surprisingly, wages, in no way, impact the consumption of household goods in either 2001/2002 or 1991/1992. Wages influence the aspiration for market goods positively and quite strongly, and again, there is no difference between 2001/2002 and 1991/1992.

Table 2 Comparison of path coefficients for men, 2001/2002 and 1991/1992
Full size table

As stated, wages are the most important determinant of housework for men. However, the results for the direct and two indirect effects of wage show no significant change between 2001/2002 and 1991/1992. In 2001/2002, for example, the total effect of wages equals the direct effect, which can be attributed to the missing influence, on the one hand, of wages on household goods and, on the other, of aspiration to market goods on housework. Nonetheless, an increased total effect of wages is observable, whereas hardly any other change appears in men’s housework-related behaviour.

Results for the Division of Work Between Couples

Table 3 presents the comparative results of the 2001/2002 PLS models for men and women. As the comparisons make clear, the effect of household goods on housework for men is much smaller than for women.Footnote 4 Moreover, we reveal a weak but important difference—significant at the 10% level—between the effects of the aspiration for market goods on housework between the sexes. The aspiration for market goods influences women’s housework but does not affect that of men.

Table 3 Gender-specific comparison of path coefficients for 2001/2002
Full size table

The results for the direct effect of wages on housework, in contrast, are similar for both sexes in 2001/2002, and no difference can be observed. This finding is similar to the observation of Blau and Kahn (2005) that women’s wage labour supply elasticities adapt to men’s. In our analysis, we observe the same development for housework time. Blau and Kahn (2005) offer several explanations for converging labour supply elasticities, including the fact that women’s increased participation in the labour market or the rising divorce rate may make women’s labour supply less sensitive to their own wages. To a certain extent, this may also apply to housework time.

Nevertheless, effects of wages on household goods and on aspiration for market goods are significantly different for men and women, i.e., in these two cases women tend to react more sensitively than their partners to changes in wages. There are, however, no changes detected over time: differences between the sexes remain rather stable. Hence, given that the total effect of wages on housework is significantly higher for women than for men, the findings of this study highlight that some indirect effects of wages on housework, little addressed in research to date, must still be considered.

In all, the results point to a large distinction between the sexes in time-use patterns, although a comparison over time shows that these differences have been reduced. Such a reduction cannot, however, be clearly attributed to the changing behaviour of men; rather, women appear to adjust their housework behaviour to that of their partners. In addition, even though the impact of men’s household goods consumption and aspiration for market goods on housework did not change between 1991/1992 and 2001/2002, a depreciating effect was found for women. In terms of household goods, as shown in other studies, better appliances seemingly increase the productivity of women’s household time (e.g., Knowles 2007). Yet, surprisingly, this development does not apply to men; the time invested in the production of household goods has not changed. However, women may adapt their behaviour with regard to the consumption of household goods and market goods to that of men, which has implications for the housework performed by women. Empirically, there is some evidence in favour of substituting household goods with market equivalents; for example, in the fields of nutrition or childcare (e.g., Albanesi and Olivetti 2007; Bonke et al. 2007; Gørtz 2007; Ribar 1995).

Summary and Policy Implications

This study analyses data from the German Time Use Surveys in an attempt to explain the rather unequal division of housework between couples. It also implements SEM in order to not only integrate conventional determinants in the analysis but also to procure important information usually not available in most datasets, such as the consumption of household goods or a preference for market goods consumption. The analysis also uses this methodology in order to replicate other results and generate new insights by considering determinants simultaneously.

The analysis shows that an important determinant of time spent on housework is (for both men and women) wages followed by the consumption of household goods. On the one hand, for women, the magnitude of nearly every determinant changed in the decade analysed in this paper: wages gained and household goods consumption as well as the aspiration for market goods consumption lost importance. These changes in the determinants can, to a large extent, explain the observed decrease in time spent on housework by women. Whereas these changes can be partly explained by better technology (i.e., better household appliances), women also tend to consume fewer household goods, which can be accounted for by the decreased number of childrenFootnote 5 and/or by higher substitution rates of household goods with market goods. On the other hand, men’s behaviour has remained remarkably constant in the decade analysed in this paper. Furthermore, it can be observed that women have changed their housework-related behaviour to better match that of men. In spite of this convergence in the housework-related behavioural patterns of men and women, the unequal division of housework persists because there are still large and significant differences in the determinants between the sexes.

As pointed out by the United Nations, assuring more gender equality, reduced welfare recipients and less poverty is to a large extent determined by a society’s ability to depart from the traditional breadwinner model (United Nations 2007). In highly specialized households that adhere to this traditional model, individual members face a higher risk of poverty and welfare dependence whenever major life events occur, such as, for example, unemployment spells, divorce or the death of the breadwinner. Although female employment rates have increased substantially over the past decade, the traditional breadwinner model is still quite firmly entrenched in Germany—or for that matter in most industrialized countries. The increase in female employment rates has not been matched by an adjustment of men’s time allocation, and as our results indicate, it is the women who have changed their behaviour and distribute their time in a way that is starting to resemble that of men. One consequence of this change is that less time is being spent on housework.

The diminishing importance of housework and the increased relevance of market substitutes implies that, for example, food provisioning or child care are no longer solely unpaid tasks preformed within households, but increasingly an activity that falls into the domain of markets and the public sector. Needless to say, such a development calls for public awareness and, in most countries, institutional adjustments. As women’s labour participation rises, then there is a need for appropriate market goods that can replace household goods. These substitutes should be affordable, available, and (especially in the case of child care) of appropriate quality. Hence, amended and new products and services are sold on the markets—entailing new challenges for consumer policy. Here, we offer three examples in which consumer policy plays a central role when looking at the gender inequality of the distribution of housework within a household.

  1. 1.

    Children

As shown in our study, women’s time allocation is becoming more similar to that of men. One major factor contributing to the overall decrease in time invested in housework is the decrease in the number of children per household. Recently implemented policy tools such as a new parental allowance scheme called Elterngeld were meant to counteract this trend. Elterngeld compensates parents for foregone earnings that arise while looking after a newborn child. The goal of the Elterngeld is twofold: first, this policy instrument should enable higher educated couples to rear children while pursuing their career and, second, it is supposed to motivate fathers to spend more time with their children. The maximum of 14 months can only be reached when both parents take on a certain minimum share of parental leave. This policy thus not only tries to reduce the costs of having children but also tries to get more men involved in housework, i.e., to break away from the traditional breadwinner model.

  1. 2.

    Child care

Families with both partners employed have to rely on high-quality child care including well-educated child carers, nutritious meals or sufficient opportunities for physical activity. Herbst and Tekin (2008) show some consequences of low quality child care, e.g., children in lower-quality care institutions perform worse in maths and reading tests as well as of teachers’ assessment of behavioural problems. The OECD (2008) finds a growing income inequality between lower and higher socioeconomic status families in Germany. Well-trained and equipped child carers could reduce such inequalities. Indeed, a number of recent studies suggest that a child care subsidy is an effective policy tool not only for increasing the labour force participation of mothers but also to increase equal opportunities (e.g., Blau and Tekin 2007).

  1. 3.

    Convenience food

Regarding other market substitutes for household goods, Cawley and Liu (2007) find, for example, that, in comparison to non-employed mothers, employed mothers are more likely to buy prepared food than to cook. The food industry recognized this development and offers more convenience food than in the past (Statistisches Bundesamt 2008). The increased offer of such highly processed foods leads to unhealthier nutrition due to its inferior nutritional value and high energy content (Jeffery and Utter 2003). One emerging problem is that consumers experience difficulties in evaluating food quality. Hence, one often-discussed policy tool is food labelling. Another policy issue is the security of food.