In his majority opinion in United States v. Windsor (2013), Justice Anthony Kennedy offered a novel argument for invalidating the federal refusal to recognize same-sex marriages. The Defense of Marriage Act, Kennedy explained, humiliates children raised by same-sex couples. Many regulatory schemes, whether civil or criminal, have spillover effects onto the children of regulated subjects. Judicial acknowledgement of those effects in the same-sex marriage context, however, contrasts strikingly with the judicial incuriosity about parallel effects in other domains, particularly in the criminal law.

Since the 1970s, the United States has experienced explosive growth in incarceration. Mass incarceration unequally affects different racial groups. Black men lacking a high school diploma are more than five times more likely to be incarcerated, for example, than similarly situated white men (p. 15). One in four black children in the 1990 cohort experienced paternal incarceration (p. 41). Does mass incarceration humiliate, or inflict worse harms, on children? And does the effect differ by racial group in ways that influence larger patterns of racial inequality?

The answers are not obvious. The incarcerated population is drawn from a relatively narrow slice of the left-tail of the income distribution. Its direct effect on inequality is tightly bounded. Western, for example, estimated the effect of incarceration on the gap between black and white earnings to be around three percent (Reference WesternWestern 2006). Further, parental incarceration might have no effect if it is caused by the same underlying factors that cause other undesirable childhood outcomes (e.g., poverty, health problems, and homelessness). Alternatively, parental incarceration might enhance the welfare of children by removing a source of violent risk and criminality from the household.

Documenting the causal effect of parental incarceration on childhood outcomes raises nettlesome methodological problems: Randomized experiments are out of bounds. Few large-scale data-sets contain extensive longitudinal evidence. Existing data presents significant identification problems. The central contribution of Children of the Prison Boom is the application of sophisticated econometric methodologies to a range of national and city-level data to generate deeply sobering evidence of incarceration's devastating effect on children, and in particular African-American children. Focusing on the causal effects of (largely paternal) incarceration on children's mental health, behavioral outcomes, homelessness, and infant mortality, Wakefield and Wildeman also persuasively demonstrate that mass incarceration is a significant causal pathway for the intergenerational reproduction of racial inequality.

To generate estimates of parental incarceration's causal effects, Wakefield and Wildeman rely on three large longitudinal data-gathering efforts: the twenty-city Fragile Families and Child Wellbeing study, the single-city Project on Human Development in Chicago Neighborhoods, and the multistate Pregnancy Risk Assessment Monitoring Systems. Longitudinal data allows them to deploy propensity-score matching, within-person change models, and synthetic regressions to estimate the average effect of parental incarceration on outcomes of interest. In respect to each independent variable of interest, Wakefield and Wildeman apply different methods and present several model specifications. By generating estimates of causal effects that are robust to both method and model specification, they generate powerful evidence of causality. Nonspecialist readers are supplied sufficient information on methods and their limits to make informed judgments about the validity of the book's findings.

The central finding of Children of the Prison Boom is that parental incarceration worsens childhood health and well-being outcomes. Two headline findings underscore the importance of Wakefield and Wildeman's research. First, recent parental incarceration is associated with a 49% increase in infant mortality in rigorous specifications (p. 108). This is greater than the increase linked to maternal smoking. This, moreover, is an average effect. Because the 13% of women who have experienced abuse do not face an increased risk to their infants, the average effect where abuse is absent may be even larger. More generally, they find that the adverse effects of parental incarceration are concentrated among children whose parents were arrested for nonviolent crimes. Second, paternal incarceration doubles the risk of child homelessness. Again, decomposition of the data by race is even more revealing: Parental incarceration has no effect on homelessness in white and Latino families. Its effect is solely concentrated on African-American children.

In contrast to Western's findings about incarceration's direct effect on racial inequality, Wakefield and Wildeman estimate large black-white gaps in incarceration's effect on children, in particular respecting the risk of homelessness. Because black-white disparities in incarceration has increased since the beginning of mass incarceration in the 1970s, the inequality-related effect on children varies by age cohort, with the intergenerational transmission of inequality growing larger with each cohort from the 1970s onward (p. 146). As much concern as mass incarceration provokes today, therefore, Wakefield and Wildeman suggest that its largest effects are still to come.

Wakefield and Wildeman provide exemplary large-n analyses, using the best available methods, and a diverse range of quantitative data. Quite reasonably, they limit their ambit in accordance with their methodological expertise. This leaves two gaps for other work to fill. First, while drawing on qualitative data to supplement their central quantitative conclusion, Wakefield and Wildeman do not offer thick accounts of how paternal incarceration changes intrafamilial dynamics and neighborhood culture. Here, for example, Goffman's (2014) recent study of a Philadelphia neighborhood provides a useful methodological counterpoint, which is more suggestive of the phenomenology and cultural effects of parental incarceration.

Second, Wakefield and Wildeman have little to say about the policy considerations of their analysis. They touch briefly on early childhood intervention policies, drug treatment programs, and reforms to the probation system, but only in passing. It is surely to be hoped that their pathmarking empirical work catalyzes among social scientists wider engagement with and attention to previously unappreciated spillover effects from mass incarceration, effects that promise to cast a long shadow. It is also to be hoped that legal scholars and judges too recognize that the collateral costs of laws on children is not an artifact of the same-sex marriage context, but rather a pervasive and deeply troubling aspect of our wider criminal regulatory state.