Duke 2008 | ASHE | Equity and Efficiency in Health and Healthcare

The evidence is clear that providing informal care can cause adverse emotional and physical health effects on elderly spousal caregivers, at least in the short term, and even up to five years after a caregiving experience (Schultz et al., 2001). Less is known about the health effects of caregiving on adult children, who will become increasingly important sources of informal care as the baby boomer generation ages and the number of divorcées increase. This paper tests whether caregiving by adult children has adverse effects on their health. We examine both the short-run and the persistence of caregiving health effects by observing adult children over 14 years. In addition, we examine how these health effects and caregiving in general influence an adult child's own work and retirement behavior.

We exploit information about intra-family caregiving behavior, in order to control for the endogenous decision to become a caregiver. Building on the model proposed by Holmes and Van Houtven (2002), we examine the joint production of informal care within a family. The Health and Retirement survey (HRS) contains information about caregiving behavior of the adult child, the adult child's spouse, and up to four randomly drawn siblings and siblings-inlaw.

The main explanatory variables, duration of caregiving and hours of caregiving per month, are gathered on the respondent, on the respondent's spouse and on the respondent's siblings. Information on siblings helps identify our selection model into caregiving: whether a sibling works, lives nearby, lives with persons under age 18, and whether the respondent thinks a sibling better or worse off financially than himself/herself.

We also account for the endogeneity of health effects from caregiving. Death of the care recipient is used as an instrumental variable for the end of caregiving. This allows us to compare health effects of caregivers who are still caregiving, caregivers who have stopped caregiving, and non-caregivers who have experienced the death of a parent. This approach allows us to disentangle the health effects of aging or bereavement from the health effects of informal caregiving, and to control for the inherent endogeneity.

The outcome variables of interest are both self-reported and objective measures of physical and mental health. We use data from 8 waves of the HRS (1992-2006). The sample consists of respondent's with at least one parent alive in wave 1, who have at least one other sibling, who are not initially caregiving, and who are observed for at least three waves. Even with these limitations on the sample, we have approximately 1,500 observations per wave on average, or a stacked sample of almost 10,000 observations.

Finally, we examine the long-term economic effects of caregiving by examining the intensity of work (as measured by full, part-time, or no paid work) and temporary or permanent exit from the labor force. We breakdown the effects into those caused by the health effects of caregiving, and those caused through other channels of caregiving (for example, less time available for paid work during care provision). This enables us to provide estimates of the costs of caregiving that can be redressed through workplace reform, and those that can only be redressed through health expenditures or a direct payment scheme for caregivers like that instituted in Germany.