This is a first effort to quantify the contribution of different factors in explaining racial difference in low birthweight rate (LBW). Mother's health, child characteristics, prenatal care, socioeconomic status (SES), and the socioeconomic and healthcare environment of mother's community are important inputs into the birthweight production function, and a vast literature has delved into obtaining causal estimates of their effect on infant health. What is unknown is how much of the racial gap in LBW is explained by all these inputs together. We apply a nonlinear extension of the Oaxaca–Blinder method proposed by Fairlie to decompose this gap into the portion explained by differences in observed characteristics and the portion that remains unexplained. Data are obtained from several sources in order to capture as many observables as possible, although the primary data source is the Natality Detail Files. Results show that of the 6.8 percentage point racial gap in LBW, only 0.9–1.9 points are explained by white–black differences in endowments across those measures, and of those endowments, most of the gap in LBW is explained by the differences in SES. The unexplained difference is attributed to racial differences in the returns to or the marginal product of investments in infant health. Copyright © 2011 John Wiley & Sons, Ltd.