approach and concluded that marriage would reduce poverty among single moth- ers substantially, by about 65 percent.18
Both Lerman’s and Thomas and Sawhill’s estimates assume that getting married has no effect on men’s labor supply (and therefore male earnings). Most research on this topic, by contrast, finds that marriage leads to a modest increase in male labor supply, which would further reduce poverty rates. David Ribar did a useful survey of the literature on the impact of marriage on men’s earnings.19
Other research that seeks to analyze the impact of marriage on poverty consists of studies that conduct a “shift-share analysis,” which show what poverty rates would be if the proportion of households in different family structures remained constant over a given time period. Examples of this research include work done by Hilary Hoynes, Marianne Page, and Ann Stevens and by Rebecca Blank and David Card. These studies find that over 80 percent of poverty is related to changes in family structure, such as increases in households headed by single mothers.20
One cautionary note, however, is that these studies could overstate the impact of family structure on income because they do not account for the likelihood that, as Thomas and Sawhill say, “single-parent families possess characteristics that dispro- portionately predispose them to poverty.”21 For example, persons struggling with mental illness, substance abuse, or criminal records might find it difficult both to hold a job and to get or stay married. Nevertheless, even studies that attempt to control for these factors strongly suggest that family fragmentation negatively affects the income available to single parents and their children.
Does Family Fragmentation Increase Crime?
In addition to poverty, family fragmentation also appears to have large effects on rates of crime, according to three separate bodies of literature.
For example, research that considers entire communities has found a strong associ- ation between the percent of single-parent households and crime rates. In one case, Robert O’Brien and Jean Stockard found that increases in the proportion of adoles- cents born outside of marriage were linked to significant increases in homicide arrest rates for fifteen to nineteen year olds.22
A second large body of literature—investigations of individual families using vari- ables, such as parent-child relationships or mothers’ education levels—finds that a child raised outside of an intact marriage is more likely to commit crimes as a teen and young adult. In one study Cynthia Harper and Sara McLanahan control for a large number of demographic and other characteristics and find that boys reared in single-mother households and cohabitating (or “living together”) households are