Differences in Characteristics or Differences in Risk: Decomposing the Black-White Difference in Nonmarital Fertility

Daniel A. Powers, University of Texas at Austin

Multivariate decomposition is used to partition the black-white difference in out-of-wedlock fertility rates into compositional and return-to-risk components. About 45% of the observed difference in rates can be attributed to racial differences in characteristics, with the remaining portion attributed to racial differences in the effects of these characteristics and to differences in the baseline hazard. A detailed decomposition is carried out, thus allowing an assessment of the contribution of each model predictor to the race difference in risk. We assess the relative contribution of socioeconomic background and family structure to components of the nonmarital fertility differential. Family background and family structure variables and the effects of these factors differ markedly by race. We find that family background variables together explain about 70% of the compositional differential and about 55% of the return-to-risk differential, with family structure variables making relatively little contribution to either component.