This study examines the effects of changes in family structure on children's economic well-being. An initial shift-share analysis indicates that, had the proportion of children living in female-headed families remained constant since 1970, the 1998 child poverty rate would have been 4.4 percentage points lower than its actual 1998 level of 18.3 percent. We then use the March 1999 Current Population Survey to conduct a second analysis in which marriages are simulated between single mothers and demographically-similar unrelated males. The microsimulation analysis addresses some of the shortcomings of the shift-share approach by allowing us to account for the possibility of a shortage of marriageable men, to control for unobservable differences between married men and women and their unmarried counterparts, and to measure directly the effects of increases in marriage on the economic well-being of children. Results from the microsimulation analysis suggest that, had the proportion of children living in female-headed families remained constant since 1970, the child poverty rate would have been 3.5 percentage points lower than its actual 1998 level. Among children whose mothers participated in a simulated marriage, the poverty rate would have fallen by almost two-thirds.