Economist has traditionally been skeptical about the potential effects of place-based policies. Some authors (see, e.g., Overman and Nathan, 2013, and Kline and Moretti, 2014) argued that place-based policies are an imperfect solution to deal with social problems, for many reasons. First, because these policies create incentives to work, invest and live in less productive of hospitable areas (see Busso, Gregory and Kline, 2014), shifting activity from good to bad places and perpetuating the problems of lagging areas (see Jofre-Monseny, 2014). Furthermore, it is not clear that shifting resources from one place to other would deliver aggregate benefits. Unless there is spatial heterogeneity in agglomeration economies, this kind of policy would lead to a zero sum game or even worse raising distortions among areas. Second, assuming a low level of housing supply constraints and highly mobile individuals, it could lead to the unintended effects of increasing house prices in the targeted area and then increasing the cost of living in those areas making difficult to improve living conditions. In this context, this paper studies the effect of a place-based policy carried out in Catalonia on the neighborhood composition. This program is defined as urban integral interventions aimed at improving the quality of life and/or the economic performance of small areas. Interestingly, the policy is not only meant for reducing social problems and improving housing condition but also at improving neighborhood amenities. In the context of a segregation model with amenities, neighborhoods affected by this program could be described as evolving towards full immigrant equilibrium. Depending on the number of immigrants already in the neighborhood, preferences of both groups, and the magnitude of the shift, we could observe a reversal in the share of immigrants in the targeted areas after this urban intervention. Therefore, in this context, we empirically test whether there is scope for reversal after this policy intervention. To test this hypothesis, we use a difference-in-difference evaluation methodology. The novelty of this intervention is that treated areas were selected based on a first round score constructed with socio-economic and urban indicators, and a second round score which depends on the first round score, objective features of the proposal, and the quality of the project. In that sense, our control group consists of areas that were rejected in the second round, areas initially rejected in the second round but accepted in future calls, and future applicants. Alternatively, we use control areas outside Catalonia. To find these control units, we exploit the fact that we have the information necessary to replicate the first round score in other areas of the country.
