GDP and the value of family caretaking: how much does Europe care?

Abstract

This study estimates the size and value of unpaid family caretaking activities at a European level. While at a country level several studies are available, a comprehensive evaluation for Europe as a whole was missing so far, mainly due to data limitations. This article fills this gap using a method that merges the information of the European Survey of Income and Living Conditions (EU-SILC) with the Harmonised European Time Use Surveys (HETUS). Monetary values of unpaid family domestic work and unpaid family childcare work are obtained applying both the opportunity cost and the market replacement approaches. For Europe as a whole, the total value of these activities ranges between 17% and 31.6% of the EU Gross Domestic Product (GDP), depending on the applied methodology. The national values of these activities are discussed and an interpretation of the country and gender differentials in family caretaking costs is given in terms of differences in culture, economic development and welfare state.

Keywords:

• unpaid domestic work
• family care work
• gender inequality
• opportunity cost
• market replacement approach

JEL Classification::

• D19
• J16
• R20
• J22

Acknowledgements

This article grew out of a study for the European Parliament conducted by the authors on behalf of the Fondazione Giacomo Brodolini in Rome. It was presented in 2009 at the XXIV National Conference of Labour Economics in Sassari, and in 2010 at the conference of the International Association for Time Use Research in Paris, and at the CEPR-IZA European Summer Symposium in Labour Economics in Munich, benefiting from the suggestions of their participants. The authors also would like to thank two anonymous referees for their comments and suggestions. The usual disclaimer applies.

Notes

¹ Several works are based on the experiences of individual researchers, such as Gronau (1973), Nordhaus and Tobin (1973), Kendrick (1979), Eisner et al. (1982), Ironmonger (1994), Goldschmidt-Clermont and Pagnossin-Aligisakis (1999), and others by initiatives of several international institutions, such as OECD (1992, 1999); ILO (1995, 1996); Swiebel (1999); EUROSTAT (2000, 2003); Trewin (2000); Holloway et al. (2002).

² In the national economic accounts, the economic activities are categorized as (i) SNA production activities; (ii) non-SNA production activities (food preparation, childcare, adult care, making and care of textiles, upkeep of dwelling and surroundings, repairs and maintenance of dwelling and of household equipment, household management and shopping, gardening and pet care) and unpaid work for the community; (iii) noneconomic activities, sometimes called personal activities (physiological and recreational activities and self-education).

³ Also see Iazaro et al. (2004) for an analysis of unpaid caring work in Spain and Alesina and Ichino (2010) for Italy.

⁴ See Prodromos-Ioannis (2005) for an analysis on the use of time of British women.

⁵ Since 2006, all municipalities had to offer a childcare guarantee until the child is 6-month old. In fact, in Denmark there is a tendency to see childcare as much as an offer to the children in their development as democratic citizens, as an offer to the parents to have their children cared for while they are working.

⁶ HETUS is a collection of national time use surveys recorded in different years: Belgium (2005–2006), Bulgaria (2001–2002), Estonia (1999–2000), Finland (1999–2000), France (1998–1999), Germany (2001–2002), Italy (2002–2003), Latvia (2003), Lithuania (2003), Norway (2000–2001), Poland (2003–2004), Slovenia (2000–2001), Spain (2002–2003), Sweden (2000–2001), the UK (2000–2001). For this reason whenever a figure or a table is based on HETUS, the reference year is not reported. See http://www.tus.scb.se.

⁷ According to HETUS data, the individual average time of market work and domestic work amount to around 200 minutes per day each (Table 2), leisure amounts to 321 minutes per day and other activities, which include sleeping, amount to more than 700 minutes per day.

⁸ The choice of the technique to be applied is often related to the kind of information that is matched, but in the present case there was not a real option of choice. In fact, direct access to HETUS micro data was not allowed. The reason is that HETUS is a harmonized collection of independent national time-use surveys, and to have access to the whole dataset, single agreements with each statistical institute involved would have been needed. The only viable solution was to use the on-line application that can generate personalized tables of average times devoted to a broad range of activities.

⁹ Domestic work and childcare work are the only two family caretaking activities that are accounted for in this analysis, since the HETUS source does not allow distinguishing care of the elderly from other activities.

¹⁰ More details about this step are presented in Appendix A.

¹¹ This lower bound is chosen for consistency with HETUS tables, which include people aged 20–74.

¹² In the literature concerning the comparison between the two approaches, the OC value of unpaid work is usually larger than the Market Replacement (MR) value because the average market wage is generally higher than the ‘domestic worker’ wage used as general wage to apply to domestic work. This is so because the reference population performing this work is, for comparative purposes, of the same size. On the other hand, with EGMR the value of unpaid work supplied by retirees is also taken into account.

¹³ We could have chosen the average wage, but as the distribution of wages is skewed towards high salaries, the average wage is always higher that the median wage, implying a possible overestimation of the values of UFDW and UFCW.

¹⁴ For UFDW, HETUS detailed information is also available (like time devoted to ironing, washing, cleaning the house and so on), but it would not match with any occupational ISCO-88 code other than 91.

¹⁵ Since code 23 includes university professors, code 33 (teaching assistant professionals) could have been chosen, in order to avoid overestimating the average parental teaching ability. However, code 33 too has few observations in the EU-SILC survey (for instance, Finland only have six observations, Greece 10), thus it would produce unreliable results.

¹⁶ A similar example of the use of OC for this purpose is provided in Davis and You (2010), who study the consequences on obesity of the increase in the opportunity cost of time of food preparation at home.

¹⁷ The wage of a professional childcare worker is usually higher than that of a domestic worker, allowing to evaluate more accurately the value of UFCW.

¹⁸ This is a very broad approximation, though, and a more precise indicator of the provision of services by the state should be used to avoid uncertainty about what are the real factors influencing the time devoted to unpaid activities. This is, however beyond the scope of this article and could be an issue to investigate in future work.

¹⁹ Under this respect, a more detailed analysis accounting for Esping-Andersen's classification of welfare state regimes could lead to interesting results (Esping-Andersen, 1990).

²⁰ It would be possible to produce several other analyses using data from Table 2. These are left out of this article for the sake of brevity and may be object of future work.

²¹ Italy, Portugal, Latvia and Greece had to be excluded from this comparison because of the use of net wages instead of gross wages, like in the rest of countries.

²² In general, smaller values of wages are expected for nonworkers due to the selection bias. People who are out of the labour market are indeed less productive, on average, with respect to actual workers, leading to lower predicted wages for the unemployed.