Will dispersed housing programmes reduce social problems in the US?

In recent years, US policy-makers have given increasing emphasis to geographically dispersing recipients of housing subsidies, based on the assumption that residence in concentrated poverty neighbour hoods abets socially dysfunctional behaviours. The paper assesses this assumption, both theoretically and through a metaanalysis of extant empirical studies. It demonstrates how only modest differences in the functional relationship between spatially concentrated poverty and resultant socially problematic behaviours will radically affect conclusions about the desirability of housing dispersal programmes. Dispersal will only lead to a net reduction in problem behaviours in society as a whole if the relationship between neighbourhood poverty rate and individual propensity to engage in problem behaviours is characterised by a positive threshold or by an increasing marginal impact. Three types of empirical studies are reviewed in an attempt to ascertain the state of knowledge regarding the nature of this functional relationship: (1) case studies of participants in dispersed housing programmes; (2) statistical studies of property value impacts of dispersed housing programmes; and (3) statistical studies of the neighbourhood correlates of the behaviour of individuals. Meta-analysis concludes that the evidence is thin and contradictory. Thus, the US now faces the unenviable situation of having adopted a major housing strategy with only a shred of evidence to suggest what effect it might have on aggregate social problems.