Eldercare hours, work hours and perceived filial obligations

ABSTRACT

In this paper, we take a fresh look at the magnitude of the trade-off between caring informally for a parent and paid work. We adopt a simultaneous approach with a primary focus on how hours of care are influenced by hours of work rather than the other way round. We also investigate the role that filial obligations play in choices of caring versus working. Using the SHARE data (2004 and 2006) we find that the elasticity of informal care hours in response to working hours is between −0.17 in the caregivers sample and −0.19 in the women-only caregivers sample; small but not negligible. Moreover, we find that a 10%increase in the index measuring the strength of filial obligations increases weekly hours of care by about two and a half hours.

KEYWORDS:

• Informal care
• caregiver
• culture
• endogeneity
• SHARE

JEL CLASSIFICATION:

• I11
• I12
• D10
• J22

Disclosure statement

No potential conflict of interest was reported by the authors.

Correction Statement

This article has been republished with minor changes. These changes do not impact the academic content of the article.

Notes

¹ Spiess and Schneider (2003), Bolin, Lindgren, and Lundborg (2008) and Kotsadam (2011) all dismiss the importance of endogeneity, based on testing or evidence from the literature. In contrast, the evidence offered by Crespo and Mira (2014), Michaud, Heitmueller, and Nazarov (2010), Heitmueller (2007) and Casado-Marín, García-Gómez, and López-Nicolás (2010) lends support to the assumption of endogeneity, respectively, for 11 SHARE countries, the UK and Spain. To compare additional contributions on this point, see Moussa (2019).

² Fernández (2008, 1232) is critical of Guiso,Sapienza and Zingales’ choice of using religion as an instrument in cross-country analysis, but concedes that it may perform better with individual-level data, like ours.

³ The definition of the target population is in SHARE Release Guide 7.0.0: 11 (http://www.share-project.org/data-documentation.html; accessed 3 April 2019). Only one age-eligible member plus his/her partner/spouse is interviewed per household.

⁴ The unweighted country samples are more balanced than the size of the respective populations: with the exception of Austria and Switzerland, overall country samples range between roughly one and two thousand observations. Since we were interested in exploiting between-country diversity we did not use country weights for the estimates reported here. This was aimed at not giving undue leverage to the large countries in the sample. However, we did ascertain that weighting did not appreciably change the results.

⁵ Vectors ${\mathrm{X}}_1,{\mathrm{X}}_2$ in Equations (1) and (2) can coincide without prejudicing identification.

⁶ Detailed results are available from the authors.

⁷ Testing differences in mean values between two groups avoids the problems of computing correlations when one of the variables is binary, as in this case.

⁸ Instances of respondents reporting 24 h of care or more are sufficiently rare in our sample, about 0.5% of total care hours in the caregivers sample and less than 1% in the overall sample. While reporting more than 24 h is a classic example of measurement error, reporting as many as 24 h may reflect the fact that, if needed, carers develop strategies to combine their own sleep with monitoring the elderly patient at night, and some feel that their sleeping time should be counted towards care time.

⁹ The incidence of adult children also caring for someone in the household is very limited in our caregivers sample, less than 10%.

¹⁰ The SHARE survey does not provide information on formal care received by the parents of the respondents, i.e. our care receivers.

¹¹ Switzerland was added to the Nordic countries cluster on the basis of two indicators pointing in the same direction: high coverage of formal supplies as opposed to cash benefits (which tend to support informal care); low weekly hours of informal external care (Rodrigues, Huber, and Lamura 2012: figure 7.1 and 5.4). According to the original taxonomy, and with exclusive reference to elderly care, Belgium could be placed in group 1 or group 3. We eventually chose group 1 because hours of informal external care were low in the country, in line with the fact that services were mostly given in kind (Ibidem: figures 5.4 and 7.2).

¹² In order to avoid discarding observations, missing values were generally set at zero, with the exception of the Family Obligations Score, where missing values were replaced by mean values. An indicator taking value 1 for missing values was added at the estimation stage to control for possible distortion due to imputation. Unless stated otherwise, the descriptive statistics in Table 1 were computed excluding missing values whenever the latter were set at zero.