Lower intergenerational income mobility for blacks is a likely cause behind the persistent inter-racial gap in economic status in the US. However, few studies have analyzed black-white differences in intergenerational income mobility and the factors that determine these differences. This is largely due to the absence of appropriate methodological tools. We develop nonparametric methods to estimate the effects of covariates on two measures of mobility. We first consider the traditional transition probability of movement across income quantiles. We then introduce a new measure of upward mobility which is the probability that an adult child's relative position exceeds that of the parents. Conducting statistical inference on these mobility measures and the effects of covariates on them requires nontrivial modifications of standard nonparametric regression theory since the dependent variables are nonsmooth functions of marginal quantiles or relative ranks. Using NLSY data, we document that blacks experience much less upward mobility across generations than whites. Applying our new methodological tools, we find that most of this gap can be accounted for by differences in cognitive skills during adolescence.