The cost of caregiving is hard to measure. Here’s a way to do it well.

Current methods may be substantially overestimating caregivers’ lost wages.

The number of people living with dementia is increasing quickly as the world population ages. Because, in many cases, people living with dementia eventually need help doing everyday activities, the resources devoted to dementia care extend beyond the medical system to include a substantial amount of unpaid care provided by family and friends. Lower birth rates in the US mean that this care will be shared among fewer family members in the future, likely leading to greater demands on each caregiver’s time.  

The time and energy invested in providing unpaid care can impact caregivers in many ways, both tangible and intangible. Tangible impacts include out-of-pocket costs and lost, or forgone, wages. The intangible impacts of caregiving can be positive but also include stress, depression, and a loss of social opportunities. To effectively support caregivers, we need to measure these impacts carefully. Forgone wages of caregivers are a particularly important metric to track since giving up paid employment can have a huge financial impact on caregivers and their families. National estimates of forgone wages for dementia caregivers assume that all caregivers are impacted in the same way. While this method leads to eye-popping values for caregivers’ forgone wages, it covers up important differences among different subgroups of caregivers.  

How are forgone wages currently estimated?

Estimates of the amount of unpaid care provided to people living with dementia are based on surveys that typically have a limited amount of information about unpaid caregivers. In a widely cited (and recently updated) study, Hurd and colleagues estimated that forgone wages of those caring for a loved one with dementia were $159 billion. Because the researchers used a national survey that did not ask caregivers directly about their lost wages, they had to answer two questions: First, how many of the caregiving hours would have been spent earning money instead? Second, what is the value of each of those hours? To answer the first question, Hurd and colleagues started with an assumption that every hour of care could have been spent working. They then adjusted the hours down to account for the fact that not all caregivers would have been in the workforce. To make this adjustment, they used census data to estimate the probability that someone would be working (based on their age, sex, and education). If only 50% of people in a demographic group were in the labor force, caregiving hours for that group were cut in half. To answer the second question, the value of each caregiving hour was specified as the average wage for the caregiver’s demographic group. The average wages were also estimated from census data. While these assumptions are reasonable, this method would be overestimating the forgone wages if caregivers were able to juggle some caregiving around their jobs. Alternatively, forgone wages could be underestimated if a caregiver reached a tipping point at which they exited the labor force despite providing fewer than 40 hours of care per week.  

How many caregivers give up paid employment?

To examine in more detail how caregivers’ wages are affected by their caregiving responsibilities, we turned to data from GERAS-US, a study of the costs of Alzheimer’s disease conducted by Eli Lilly. (Eli Lilly has made the data from GERAS-US available via the AD Workbench, a data-sharing platform created by the Alzheimer’s Disease Data Initiative.) GERAS-US was designed to track participants with mild cognitive impairment or mild dementia over time, so the baseline dataset is not representative of all levels of dementia severity.  

The GERAS-US study asked caregivers several questions to determine how caregiving had impacted their work. Caregivers in the study were asked if they were employed, and if they were not, whether they had quit work due to caregiving responsibilities. Those who were employed were asked if they had cut back on work hours due to caregiving responsibilities, and if so, by how many hours per week. We found that most caregivers in this dataset were either not working (so caregiving didn’t impact formal employment) or were able to combine caregiving and employment without a loss in work hours. Only 12% of caregivers cut back on work hours or quit their job due to caregiving responsibilities. 

How do self-reported data affect forgone wage estimates?

Using carers’ self-reported hours of lost work, we compared two methods of calculating forgone wages. For both methods, an average wage for each carer’s demographic group (age, sex, and education level) was estimated from census data since caregivers were not asked about their wages.  

First, we followed the methodology of Hurd and colleagues, essentially acting as if we had no information about caregivers’ work impacts. We used the probability that each caregiver would be in the workforce, based on their demographics, and scaled down their caregiving hours accordingly. We then valued the remaining hours at the average wage for that demographic group. This led to an estimate of $461 in weekly forgone wages for the average caregiver.

Our second calculation of forgone wages used the carers’ self-reported lost hours of work due to caregiving responsibilities. Using self-reported lost work hours led to a much smaller estimate of $53 in weekly forgone wages for the average caregiver.

Why are better estimates needed?  

When caregivers reduce their work hours or drop out of the labor force completely, there is a tangible economic impact to the caregivers and to society. Producing estimates based on assumptions rather than real-world data obscures changes in caregivers’ ability to combine paid work and caregiving. Data that enable researchers and policymakers to track caregiver work impacts regionally and over time are critical for evaluating any plan to support dementia caregivers. For example, a policy that makes it easier for a caregiver to stay in their job while providing care (e.g., additional family leave) may not affect the number of care hours they provide. In this case, the policy would not lead to any difference in the estimated value of forgone wages for dementia caregivers under the previously described assumptions.  

Conclusion

These findings suggest that current methods of estimating the forgone wages of caregivers are substantially overestimating the value of forgone wages. While lost employment is only one of many impacts from providing unpaid care, it is an important one. Cutting back on or quitting work can lead to financial hardship, higher stress levels, and long-term effects on earnings. Accurate tracking of forgone wages for those providing unpaid care is a crucial step in finding the best ways to support caregivers.