Policymakers often lack reliable information on the likely impacts of policies intended to promote upward mobility. Few studies attempt or are able to estimate the effects of a policy intervention in childhood on outcomes decades later. Even randomized controlled trials examining a single intervention generally are uninformative as to whether one policy administered to the same target population is more or less effective than another, whether the timing of interventions matters, or whether multiple interventions—concurrent or sequential—have notably bigger impacts than single ones. The Brookings Social Genome Model is a new dynamic microsimulation model being developed to answer these questions. The paper discusses the challenges facing any effort to model mobility processes and simulate effects of early interventions on outcomes at middle age. These include the unavailability of satisfactory data sets covering birth through middle adulthood, as well as a number of difficulties stemming from this basic problem. No less daunting are modeling challenges around parsimony, endogeneity, and heterogeneity.  We describe our approach to these challenges and present some illustrative results on the effect of several policies aimed at disadvantaged children.