A crucial issue for this study is to ascertain to what extent the associations between family fragmentation and these negative outcomes are causal. There are of course powerful selection effects into marriage, divorce, and unwed childbearing, and some portion of the negative outcomes for children in nonmarital families are caused by habits, traits, circumstances, and disadvantages among adults that may also lead to divorce and nonmarital childbearing.12
For example, a dating couple facing an unexpected pregnancy may choose not to marry because the man is unemployed. Depending on how one looks at it, the out-of-wedlock birth may be said to result from the father’s low-earnings or the mother and child’s poverty may be said to result from the out-of-wedlock birth. Untangling “what causes what” is a challenge faced by many researchers who study the family.
How Much Does Marriage Reduce Poverty?
Researchers respond to this challenge by using a variety of methods to control for unobserved selection effects (that is, to account for other factors that could be explaining the finding) and to tease out causal relationships (that is, to untangle “what causes what”). In this case, the idea that family fragmentation contributes to child poverty has been studied extensively and is widely accepted.13 Marriage can help to reduce poverty because there are two potential wage earners in the home, because of economies of scale in the household, and possibly also because of changes in habits, values, and mores that may occur when two people marry.14
In addition, there is recent, intriguing research that uses naturally occurring evi- dence to examine whether family fragmentation causes poverty. Elizabeth Ananat and Guy Michaels, for example, use an unusual predictor of whether a married couple will stay married (the predictor is whether their firstborn child is a male, since research has shown that divorce is less likely when this is the case). With this predictor they are able to study married couples who do and do not divorce and conclude that “divorce significantly increases the odds that a woman with children is poor.”15 Their analysis suggests that almost all of the increase in poverty observed among divorced mothers is caused by the divorce. Less than 1 percent of these women and children live in poverty if their first marriage is intact, while more than 24 percent of divorced women with children are living in poverty.16
Another area of research uses national data to simulate changes in family structure. For example, Robert Lerman used the Current Population Survey (CPS) and simu- lated “plausible” marriages by matching single mothers to single males who were the same race and were similar in age and education levels. He found that if these theoretical marriages occurred they would reduce poverty by 80 percent among these single-mother households.17 Adam Thomas and Isabel Sawhill used a similar