New Study Finds Lasting Effects of Divorce on Kids

The divorce debate is alive and well in America. If you google “divorce” and limit results to big magazines and media, you’ll find these in the top results: “Divorce Can Be an Act of Radical Self-Love.” “Women are divorcing — and finally finding happiness.” “In Defense of Divorce.” And just last Friday, “Divorce is a Gift” from The New York Times. 

Characteristically, these articles map out the divorce journeys of their respective authors as they grapple with their personal identities. Each article intends to make divorce a little less taboo and perhaps ease the concerns of would-be divorcees. But what about the kids in these families? 

The relationship between divorce and kids has always been a hot button in the discourse. That kids are affected by divorce is quite clear. But while the immediate effects of divorce go mostly undisputed, many doubt that divorce has a lasting impact into adulthood. Whatever discrepancies that exist between children from intact families and children with divorced parents can be explained, the argument goes, by factors other than or surrounding the divorce. It may be that bad outcomes are merely symptoms of an unhappy home, and not the divorce itself. Or it may be that divorcees are more prone to conflict for idiosyncratic or genetic reasons, which carries on to kids.

Comparing outcomes between kids from intact and divorced homes is relatively simple. The difficulty is establishing causation. To this end, researchers need a large enough sample, over a sufficient length of time, and a way to control for differences between families. A new working paper from Andrew C. Johnston, Maggie R. Jones, and Nolan G. Pope achieves this goal. Using tax records for over 5 million children born between 1988 and 1993, Johnston and his coauthors follow the lives of children whose parents divorced, measuring income, child mortality, teen birth rates, incarceration, and college residency. 

What makes this study particularly robust is its use of one million sibling groups whose parents divorced. By comparing siblings, the authors can see how divorce affects, say, a 10-year-old versus an 18-year-old within the same family. This, we will see, is extremely useful.

Furthermore, the study examines three distinctive effects the come during and immediately after divorce for families: declines in household income, declines in neighborhood quality, and increased distances between non-resident parents. To the degree these explain negative outcomes associated with divorce, like decreased adult income of the child, we have strong evidence that divorce itself is a cause, and not just associated with, such outcomes. That is to say, the negative outcomes associated with divorce are not merely from underlying household issues, but from the act of separation itself.

So what did the authors find? Let’s begin with the immediate effects of a family breakup, starting with income. Leading up to a divorce, average income is between $90,000 to $100,000 for the family. But following divorce, household income falls to $42,000, less than half. While household incomes generally recover in the period following divorce, they remain about 30% below their pre-divorce level even after a decade. This income loss has a number of downstream effects. While the decline in household income overstates a decline in resources available for the child, additional financial pressure is evident, which in part is due to a loss of economies of scale. Whereas a married couple needs just one house with the relevant furnishings, a divorced couple needs two. Perhaps as a consequence of the extra income needed after a divorce, the authors find an increase in working hours among parents. Specifically, fathers work 16% more hours, and mothers work 8% more hours after the family splits up. Fathers' work hours continue to rise over the decade after divorce. This, we can reasonably conclude, means less time and flexibility for the kids.

Far from being a mere change in legal status, Johnston, Jones, and Pope demonstrate that divorce has a tangible negative impact on factors relevant to child outcomes.

With a divorce also comes a move. Parents split, and at least one parent needs to find a new home. The authors found that 35% of children change addresses in the year of the divorce, a rate nearly three times higher than pre-divorce. The median distance between non-resident parents (in the vast majority of cases, the dad) is 4 miles in the year of the divorce, with the average distance between parents at over 100 miles. This suggests that a substantial portion of divorces result in extreme distance between parent and child. The elevated move-rate persists even through the decade following a divorce, with the median distance between parents growing to 10 miles, and the average growing to over 200 miles.

And when we consider the decline in household income, it is no surprise that children generally move to lower-quality neighborhoods, often with their mothers. Indeed, the authors found that neighborhood household incomes fall 7% in the years following a divorce. The financial pressure caused by going from one house to two means that parents often must sacrifice on neighborhood quality, representing a real decline in standard of living.

Far from being a mere change in legal status, Johnston, Jones, and Pope demonstrate that divorce has a tangible negative impact on factors relevant to child outcomes. These effects, for the most part, are felt immediately and continue to be felt year-after-year.

How does this show up in child outcomes? This is where the paper really stands out. For years, causal skeptics of divorce have argued that most child-divorce outcome research is flawed. The studies, they say, don’t account for the fact that the kinds of families that end up divorcing are different from those that stick together. Divorce is then merely a symptom of an underlying dysfunction, and not necessarily a cause of worse child outcomes. Johnston, Jones, and Pope account for this by comparing siblings within families that experience divorce. The hypothesis: younger siblings more exposed to divorce will have worse outcomes than older siblings with less exposure. 

Consider income. Those who experience an early-childhood divorce earn about $2,500 less at age 25 than those whose parents divorced at age 25. This represents a 9% decrease relative to average earnings at that age. By 27, they earn about 13% less, suggesting that the effect of early-childhood divorce on future income persists with age. The authors remark that this is comparable to “losing a year of education” or “moving to a one-standard-deviation lower-quality neighborhood for all of childhood.”

Teen birth rates also rise for children of divorce. Before a divorce, teen birth rates hover around 7 births per 1,000 girls annually, and dip right at the divorce. But following a divorce, teen birth rates climb, rising to 13 teen births per 1,000 girls annually. Child mortality also increases: the authors note that following a divorce, there is a “sharp and persistent” increase in child mortality of 10 to 15 additional deaths per 100,000 children per year. 

To quote the study:

These results reveal substantial effects of divorce on children’s outcomes. The absence of pre-trends in both outcomes supports a causal interpretation. The magnitude of the effects—a 35 to 55 percent increase in mortality and up to a 63 percent increase in teen births—underscores how divorce can dramatically reshape children’s outcomes, potentially through changes in resources, supervision, and family dynamics.

It is quite clear, then, that parental divorce is associated with negative outcomes for children, even within families. This undermines the view that divorce effects are simply a result of some innate or idiosyncratic factor like genetics. 

But what of the view that bad outcomes are from underlying problems, and not from the divorce itself? As shown above, the stark increases in outcomes post-divorce like mortality and teen births (and incarceration, which is discussed in the paper) lend evidence to divorce being the cause. Consider also the effects that follow from divorce of decreasing household income, lower-quality neighborhoods, and decreasing parental proximity. The authors find that these three factors explain between 25% to 60% of the negative child outcomes associated with divorce in adult income, child mortality, teen birth rates, college residence, and incarceration. This mediation analysis tells us that the immediate effects of divorce play a role, further supporting a causal story. 

Is the divorce debate finally settled? To a degree. Johnston and his coauthors give strong evidence that exposure to divorce is at least a partial cause of the negative outcomes for children following parental divorce. As Patrick T. Brown notes in his newsletter, some critics may argue that negative shocks around divorce (e.g. job loss) are the real drivers of bad outcomes. The authors preemptively counter this line by noting that divorces generally come about from long-standing issues, and that they don’t see any evidence of shocks prior to divorce. 

In any case, there is still more to study: changes in household income, neighborhood quality, and parental proximity do not completely explain the outcomes they measured. That leaves the door open to further debate about the degree of causation. In addition, these outcomes are just a fraction of the ways children can be affected by divorce. 

With all that said, this new study provides further evidence that lasting marriages matter, especially for child well-being.

Grant Bailey is a research associate at the Institute for Family Studies.