It must be heartbreaking for a child to see a parent go to jail for some crime. But is it necessarily all bad for how the kid’s life turns out in the long run?

Not according to a paper in the latest American Economic Review  by Samuel Norris, Matthew Pecenco, and Jeffrey Weaver, “The Effects of Parental and Sibling Incarceration: Evidence from Ohio.” (Ungated version here.)

Many earlier studies found that kids with incarcerated parents had worse long-term life outcomes -  things like education, jobs, and health. However, most of these studies took the straightforward but wrong approach of simply comparing life outcomes for kids who had parents in prison with those who didn’t.  The problem is that the two groups are also likely to differ systematically in other ways. For example, children with imprisoned parents tend to come from poorer households. Poverty also worsens life outcomes, possibly swamping the specific effect of parental incarceration.

A statistically superior but morally awful and practically impossible approach would be to run a randomized experiment: take a large group of kids with criminal parents, flip a coin to decide whose parents go to jail, and come back after, say, fifteen years to check how the lives of the two groups had turned out.  In an interview with the “Probable Causation” podcast, Jeffrey Weaver explains how the new AER paper was nonetheless able to approximate the theoretical ideal using clever statistics and a lot of hard work to gather obscure data. It did this by exploiting the fact that some US states assign criminal cases to judges randomly, and judges vary in the severity of the sentences they hand down. The random variation in the severity of punishments allows the paper to pin down (“identify”) the effect of parental imprisonment on children’s lives.

Perhaps surprisingly, the study finds significant positive impacts in at least a couple of ways. First, parental incarceration reduces the probability of a child being charged with a crime by age 25 by 6.6 percentage points, around one-fifth. Second, it also helps children achieve a significantly better socio-economic status, as measured by the kind of neighborhood they live in as adults. The reasons for these positive effects could include the benefits of a better, more stable family structure when kids go to live with relatives and of being separated from toxic, “criminogenic” parents. There could also be a deterrent effect from seeing the parent in prison.

The impact of parental incarceration on reducing child crime is much bigger and more statistically significant when the parent is black than white. For example, there is an 8.8 percentage point estimated reduction in the probability of the child being charged with a crime by age 25 when the parent is black, compared to a 1.6 percentage point cut for whites. This finding may be one among the many reasons why blacks oppose reducing the police presence in their neighborhoods by margins of over 80%.