This paper estimates the relationship between initial village inequality and subsequent household income growth for a large sample of households in rural China. Using a rich longitudinal survey spanning the years 1987-2002, and controlling for an array of household and village characteristics, the paper finds that households located in higher inequality villages experienced significantly lower income growth through the 1990s. However, local inequality s predictive power and effects are significantly diminished by the end of the sample. The paper exploits several advantages of the household-level data to explore hypotheses that shed light on the channels by which inequality affects growth. Biases due to aggregation and heterogeneity of returns to own-resources, previously suggested as candidate explanations for the relationship, are both ruled out. Instead, the evidence points to unobserved village institutions at the time of economic reforms that were associated with household access to higher income activities as the source of the link between inequality and growth. The empirical analysis addresses a number of pertinent econometric issues including measurement error and attrition, but underscores others that are likely to be intractable for all investigations of the inequality-growth relationship.