I NFORMAL CARE FORMALCARE AND DAUGHTERS EMPLOYMENT CONDITIONS

I NFORMAL CARE, FORMALCARE AND DAUGHTERS EMPLOYMENT CONDITIONS : EVIDENCEFROM E UROPEAN COUNTRIES Journée de la Chaire Santé Paris, March 30, 2018 E R I C B O N S A N G( U N I V E R S I T É P A R I S- D A U P H I N-E P S L ) J O A N C O S T A- F O N T ( L S E )

Gender employment gaps and long-term care 2 I NTRODUCTION • With the ageing of the population, there is growing attention on what influences the use of informal care and formal care among older individuals. • Informal care is the primary type of long-term care for elderly individuals in need of care (Bettio and Verashchagina, 2010; Rodrigues et al. , 2013; Arno et al. , 1999). • So far, informal care from the adult children, in particular daughters, is the most common form of long-term care for older parents. • However, this important source of care might shrink due to the decrease in family size, increased divorce and more particularly as more daughters join the labour force, the focus of this paper.

Gender employment gaps and long-term care 3 I NTRODUCTION • The increasing labour force attachment of daughters is likely to translate into higher opportunity cost for providing care to older parents and to decrease the supply of informal care. • A limited number of studies have investigated the effect of labour market participation on the provision of informal care (Carmichael, Charles and Hulme, 2010; Nizalova 2012). • Results suggest a negative effect of employment on informal care.

Gender employment gaps and long-term care 4 I NTRODUCTION • At the same time, we may expect that it also affects the demand formal care. • Indeed, several studies show that changes in the supply of informal care influences the use of formal care (Van Houtven and Norton, 2004, 2008; Bolin et al. 2008; Bonsang, 2009). • As a consequence, we expect that the current trend toward closing the gender employment gap should affect the demand for long-term care.

Gender employment gaps and long-term care 5 T HIS PAPER • This paper investigates the effect of the gender composition of children alongside the gender-specific employment rate across European countries. • Given women have historically been the primary provider of home production, the gender composition of the children influences the probability of receiving care (Horrowitz, 1985). • However, our contribution to the literature lies in exploring whether gender differences in local labour markets on long-term care arrangements is related to the effect of daughters on caregiving arrangements.

Gender employment gaps and long-term care 6 T HIS PAPER • Using the Survey of Health, Ageing and Retirement in Europe (SHARE), we estimate the effect of the gender composition of the children on the use of informal and formal care of older parents • We allow this effect to be heterogeneous according to variations in the aggregate gender employment gap (at the country and year level) of the children. • Our model controls for unobserved heterogeneity at the country-year level by including country-year fixed effects. • The identification strategy relies on the assumption that the gender of the children is randomly distributed according to major sources of heterogeneity such as health, preferences, and abilities.

Gender employment gaps and long-term care 7 C ONTRIBUTION • First, we add to the early literature on the effect of employment on informal caregiving (Mentzakis et al. , 2009; Carmichael et al. , 2010) by considering labour market conditions on the decision to provide care. • Second, we contribute to the literature investigating the interactions between informal and formal care by providing another source of exogenous variation in the receipt of informal care that exploits the fact that the heterogeneous effect of the gender of children according to the gender employment gap (Bolin 2008; Van Houtven and Norton 2004; Bonsang, 2009). • While previous literature highlight that daughters are more likely to provide care to their older parents, we actually show that this effect varies across country and that it is highly associated with the gender employment gap.

Gender employment gaps and long-term care 8 C ONTRIBUTION • Moreover, we show that the gender composition of the children also affects the use of formal care among older individuals. • This effect varies according to the gender employment gap: older individuals with daughters are less likely to use formal care, especially paid domestic help, but only when the gender employment gap is larger. This result suggests that informal care is a substitute to formal care.

Gender employment gaps and long-term care E MPIRICALSTRATEGY 9

Gender employment gaps and long-term care 10 D ATA: T HE SAMPLE • Pooled data from waves 1, 2, 5 and 6 of the Survey of Health, Ageing and Retirement in Europe (SHARE). • 19 European countries: Austria, Belgium, Croatia, Czech Republic, Denmark, Estonia France, Germany, Greece, Ireland, Israel, Italy, Luxembourg, the Netherlands, Poland, Portugal, Slovenia, Spain, Sweden, and Switzerland. • Sample selection: • • individuals aged 65 or above who have between one to four children whose children are between 20 and 64 year-old only keep individuals who do not live with a partner or a spouse remove individuals who live with one of their children Remove individuals living in a nursing home The final analytical sample includes 18, 169 observations.

Gender employment gaps and long-term care 11 D ATA: I NFORMALCARE • Informal care is measured as the sum of care and assistance received by the respondents from all their children • Derived from three questions that describe the relationship with the caregiver (if any), the frequency (daily, weekly, monthly or less often) of the care received and the average number of hours per day/week/month /year respectively. • However, from the wave 4 onwards, the average number of hours was not asked anymore • On average, a child providing daily (resp. weekly, monthly, yearly) care to his parent provides 120 (resp. 25, 7, 2) hours of care per month. • We use this method to impute the amount of informal care received from the children for all waves.

Gender employment gaps and long-term care D ATA: I NFORMALCARE 12

Gender employment gaps and long-term care 13 D ATA: F ORMAL CARE • We consider the utilization of two types of formal home care: paid domestic help and nursing care at home. • Respondents were asked to report whether they received either paid or professional home help, which they could not perform due to health problem, or professional or paid nursing or personal care during the twelve months preceding the survey. • Our analysis is restricted to the use of formal care at the extensive margin given the intensity is not asked from wave 5.

Gender employment gaps and long-term care D ATA: F ORMAL CARE 14

Gender employment gaps and long-term care D ATA: F ORMAL CARE 15

Gender employment gaps and long-term care 16 D ATA: G ENDER GAP IN EMPLOYMENTRATE • The employment rate of daughters by country and wave is calculated by taking the sum of all daughters in employment divided by the sum of all daughters. • The same calculation is made for sons. • The gender employment gap among children is defined as the difference in the employment rate of sons and the employment rate of daughters at the country and wave level. • A sensitivity analysis will use an alternative measure of the gender employment gap by using the measures of employment rates of men and women aged 20 to 64 from Eurostat. • Correlation coefficient between the measure from SHARE and from Eurostat is equal to 0. 898.

Gender employment gaps and long-term care D ATA: G ENDER GAP IN EMPLOYMENTRATE 17

Gender employment gaps and long-term care 18 D ATA: C ONTROLVARIABLES • We include age and age squared to take into account for the fact that the need of care increases with age • We also include the number of limitations with the activities of daily living and the number of limitations with the instrumental activities of daily living. • The level of education is included and classified according to the ISED-1997 classification. • We also control for the gender of the parent and for the number of children.

Gender employment gaps and long-term care M AIN R ESULTS 19

Gender employment gaps and long-term care R ESULTSBY GENDER 20

Gender employment gaps and long-term care G RAPHICALREPRESENTATION OF THE RESULTSFOR INFORMALCARE 21

Gender employment gaps and long-term care G RAPHICALREPRESENTATION OF THE RESULTSFOR FORMALCARE ( KIND OF FORMALCARE) 22 A LL

Gender employment gaps and long-term care R ESULTS: S EPARATINGEMPLOYMENTRATES OF MEN AND WOMEN 23

Gender employment gaps and long-term care 24 R ESULTS: S AMPLE AND SPECIFICATIONCHECKS • Robust to alternative specification where the number of daughters is included instead of the proportion of daughters. • Robust to the use of the gender employment gap obtained from Eurostat for individuals aged 20 -64. • Robust to the inclusion of parents living with their children (either by not • Robust to restricting the sample to individuals reporting being limited in activities because of a health problem. • Robust to the larger sample including individuals living with a partner. imputing informal care for co-resident children, or by arbitrarily imputing 30 additional hours of informal care per month for co-residing children).

Gender employment gaps and long-term care R ESULTSBY GENDERFOR THE SAMPLE WITH A PARTNER /SPOUSE INCLUDINGPARENTSLIVING 25

Gender employment gaps and long-term care R ESULTSFOR MOTHERS CHILDREN ACCORDINGTO THE DISTANCEFROM THE 26

Gender employment gaps and long-term care 27 R ESULTS: S AMPLE AND SPECIFICATIONCHECKS • Robust to alternative specification where the number of daughters is included instead of the proportion of daughters. • Robust to the use of the gender employment gap obtained from Eurostat for individuals aged 20 -64. • Robust to the inclusion of parents living with their children (either by not • Robust to the larger sample including individuals living with a partner. • Robust to an alternative definition of the gender employment gap that take into account part-time workers. imputing informal care for co-resident children, or by arbitrarily imputing 30 additional hours of informal care per month for co-residing children)

Gender employment gaps and long-term care R ESULTSBY GENDER PART-TIME WORK WITH GENDERGAP TAKINGINTO ACCOUNT 28

Gender employment gaps and long-term care W HAT WOULDHAPPENIF WE CLOSETHE GENDEREMPLOYMENTGAP IN ALL COUNTRIES ? 29

Gender employment gaps and long-term care 30 C ONCLUSIONS • We find that the effect of daughters on informal care is associated with the gender specific labour market conditions. • That is, the effect of daughters on the supply of care is highly significant in countries with large gender labour market gaps but insignificant in the absence of such a gap. • We show that it is also associated with changes in the use of formal care. • So far, informal care from the adult children, in particular daughters, is by far the most common form of long-term care for older parents. However, our results suggest that this important source of care might shrink as more daughters join the labour force. • It also suggests that the current trend toward closing the gender gap in employment rate is likely to affect the demand for long-term care and longterm care expenditures.

Gender employment gaps and long-term care A LTERNATIVEINTERPRETATIONS AND • 31 L IMITATIONS Alternative interpretations for informal care? • The results may be due to reverse causality. However, literature seem to show the effect of informal care on employment is small. • The results may be due to confounding factors. Social norms where the role of women is to take care of the non-market activities. • The wording of the questions formal care changed from wave 5. However, separate analysis provide consistent results. • Proportions of daughters exogenous? • Robust to the use of the sex of the first kid