Abstract Research on neighborhood structural conditions like concentrated disadvantage and crime largely focuses on between-neighborhood differences; for example, places with more disadvantage are expected to experience higher homicide rates. However, empirical research often does not consider within-neighborhood dynamics of structural stability and change. Furthermore, several recent studies have found cross-sectional associations between structural variables and crime outcomes can vary significantly across units, violating a key assumption of global modeling strategies. The current work explores if and how historical changes in disadvantage influence neighborhood collective efficacy and homicide rates, net of the level of disadvantage at a given time point. Collective efficacy theoretically mediates the relationship between conditions and crime, and is hypothesized to be the mechanism through which structural change influences homicide rates. It is also hypothesized that spatial variation in cross-sectional associations between disadvantage and social outcomes can be explained by accounting for within-neighborhood changes in disadvantage. Using a sample of Chicago neighborhoods and ordinary least squares and geographically weighted regression models, I find that within-neighborhood changes in disadvantage significantly predict neighborhood collective efficacy, though the effects of this change on homicide rates are not completely mediated by collective efficacy. Within-neighborhood change completely accounts for spatial variation in cross-sectional associations, offering one explanation of prior research findings. Within-neighborhood structural changes appear to disrupt collective efficacy and contribute to higher homicide rates than predicted by the level of disadvantage alone.