Part One: What Five Million Children Tell Us
For decades, American culture has told competing stories about divorce. One story says unhappy marriages poison children, so dissolution frees everyone to flourish. A second story says divorce harms children by fracturing the stability they need to grow. A third insists that genes and temperament dwarf family structure, making divorce a minor player in a child’s life. A new working paper from the National Bureau of Economic Research, Divorce, Family Arrangements, and Children’s Adult Outcomes, by economists Andrew C. Johnston, Maggie R. Jones, and Nolan G. Pope, settles much of the argument with data earlier researchers could only dream of.
The authors linked federal tax records with Census Bureau data to track more than five million American children born between 1988 and 1993. They followed these children into adulthood and watched what happened when their parents separated. The findings are sobering, and they deserve a wider hearing than a technical paper typically receives.
When parents divorce, household income falls by roughly half. A family that stood at the fifty-seventh percentile of income before divorce drops to the thirty-sixth. A decade later, the household has recovered only part of what it lost and still sits below its pre-divorce standing. Families move to poorer neighborhoods with seven percent lower average incomes and measurably weaker economic mobility, as documented in the Opportunity Atlas work of Raj Chetty and his colleagues. The median distance between a child and the nonresident parent climbs to five miles in the year of divorce and grows wider over the following decade, with mean distances exceeding one hundred miles.
Parents start working longer hours. Mothers increase their weekly hours by about eight percent and fathers by sixteen percent. Roughly half of divorced parents remarry within five years, bringing stepparents into the household and diffusing parental attention across new children and new relationships. Children absorb each of these shifts at once, and each carries consequences.
In the year following divorce, child mortality rises between thirty-five and fifty-five percent and remains elevated for at least a decade. Teen birth rates climb to sixty-three percent above pre-divorce levels within several years. The authors carefully rule out pre-divorce trends that might suggest these outcomes were already deteriorating. Divorce itself marks the turning point.
To estimate long-run effects, Johnston and his coauthors compared siblings within the same family who had different lengths of exposure to the same divorce. This design controls for the selection effects that have plagued earlier research, since siblings share the same parents, genes, and baseline family environment. A child who experienced parental divorce between ages zero and five earned nine percent less income at age twenty-five than a sibling exposed later in childhood. By age twenty-seven, the gap widened to thirteen percent. Early-childhood divorce increased the probability of a teen birth by seventy-three percent, incarceration by forty-three percent, and mortality by age twenty-five by thirty-five percent. Children of early divorce were also substantially less likely to be living on a college campus in their late teens and early twenties.
The authors quantify three channels through which these effects operate. Lost household income, deteriorating neighborhood quality, and growing physical distance between children and nonresident parents together account for between twenty-five and sixty percent of divorce’s impact on adult outcomes. Resource losses explain roughly ten to forty-four percent of the earnings effect. Neighborhood quality explains sixteen to twenty-nine percent. Parent proximity explains fifteen to twenty-two percent of the effects on mortality and teen birth. Much of what divorce does to children still operates through channels administrative data cannot see, including time investment, parental cooperation, and emotional availability.
Because divorce is more common among lower-income families, these effects compound across generations. Children born at the twentieth percentile of parental income are more than thirty percentage points more likely to experience divorce than children at the top. Black children face divorce rates near forty-five percent, compared with seventeen percent for Asian children. Family dissolution has become one of the quiet engines of intergenerational disadvantage.
The paper does not moralize, and readers should resist simple policy conclusions. Many marriages end for grave reasons, including abuse and irreparable conflict, and the authors note they are measuring the effect of formal divorce, not the ease or difficulty of obtaining one. What this research makes unavoidable is the recognition that divorce functions as a bundle of changes rather than a single legal event, and that the bundle presses hardest on the youngest and the poorest. The costs are real, measurable, and long-lasting. A culture that pretends otherwise is telling its children a comfortable lie.
