White grandparents who take on primary responsibility for raising a grandchild face a higher risk of dying than peers who never take on that role — and the increased risk shows up no matter how long the caregiving lasts. That is the central finding of a new U.S. Census Bureau working paper that tracked a weighted 64,027,000 adults age 50 and older for two decades, linking 2000 census long-form records directly to Social Security Administration death records through 2019. For the roughly 2.1 million grandparents nationwide who are primarily responsible for a grandchild’s basic needs, that unevenness is the real story.
What sets this study apart from earlier work on grandparent caregiving is the data itself. Instead of relying on surveys where people self-report their health years after the fact, the Census team matched individual respondents from the 2000 census long-form questionnaire directly to their eventual Social Security death records, then followed that linked population for nineteen years.
The racial split nobody expected
Among White coresident grandparents, the study found consistently elevated mortality risk compared with White adults who did not live with or care for a grandchild — and that risk held steady whether the caregiving arrangement was brief or lasted for years. Asian coresident grandparents showed the opposite pattern: a measurable protective effect, with lower mortality than their non-caregiving peers, again regardless of how long the arrangement continued. The peer-reviewed version of the research frames this as evidence that the stress of caregiving does not translate into the same biological cost across every population.
Hispanic and Black grandparents fell somewhere between those two extremes, but the timing mattered. Hispanic grandparents saw elevated mortality risk emerge specifically among long-term primary caregivers. Black grandparents showed the reverse timing effect: lower mortality risk appeared among short-term caregivers, a pattern that did not hold once the arrangement stretched on.
Why this matters beyond the numbers
Grandfamilies are not a fringe phenomenon. Census Bureau survey data from the 2017-2021 American Community Survey found that about 6.7 million people age 30 and older lived with their grandchildren in 2021, and roughly a third of grandparents in those households were the ones responsible for the child’s care. The arrangement is also geographically lopsided — Southern states, Alaska and Hawaii run above the national average for grandparent-headed households, while the Northeast and Midwest run below it.
What the Census team is not able to say, at least not in the abstract released so far, is exactly why White grandparents face a flat, duration-independent risk while Hispanic grandparents only see it emerge over time. Researchers who study caregiving stress have long pointed to a mix of financial strain, disrupted retirement plans, social isolation and the physical demands of raising a child later in life.
The takeaway is not that grandparents should avoid stepping in when a family needs them. It is that the households absorbing this responsibility — millions of them, concentrated in specific regions and communities — deserve support systems built around the fact that caregiving has a measurable physical cost, and that cost does not land the same way twice.
The study’s reliance on death-record linkage rather than survey self-report is part of what makes its findings harder to dismiss than earlier research on the topic. Self-reported health surveys are vulnerable to a well-documented bias: people under chronic stress often underreport how much it’s affecting them, particularly caregivers who may not want to characterize a role they’ve chosen, or felt obligated to take on, as harmful. By tracking an actual outcome — whether and roughly when someone died — the Census team sidestepped that reporting problem entirely, even though it means the research can only describe mortality risk in aggregate, not explain the specific biological or social mechanisms driving it for any individual family.
What researchers who study caregiving broadly agree on is that the underlying stressors are real and measurable even without a mortality study: disrupted retirement savings, reduced social contact outside the household, and the physical toll of caring for young children later in life all show up consistently in caregiving research across different populations. This new Census data adds a harder, more consequential outcome to that existing picture rather than introducing an entirely new finding.
