Skip to content

How to evaluate area based initiatives: Urban renewal policies and neighbourhood dynamics

arrow down
simone-hutsch-8p6pdSkA-6A-unsplash

Why is it relevant to study this topic?

Residential segregation has become a striking feature among many cities in the US, the UK and Europe. Policies aimed at deconcentrating poverty and minority population groups are widespread in OECD countries. Often known as income-mixing policies, these initiatives can fall into two broad categories. The first includes policies that encourage poor households to move into more affluent neighbourhoods, for example Section 8 and Moving to Opportunity in the US. The second considers policies that encourage richer households to move into less affluent neighbourhoods — for example Hope VI and CDBG in the US or URBAN projects in the EU — by improving their infrastructure. Examples of this latter category include housing rehabilitation, improving public facilities, and streetscape improvements. Income-mixing policies are controversial since, if successful, they can cause gentrification. They also are costly and their effectiveness is unclear.

What is the aim of this study?

This study was conducted to shed light on the effectiveness of income-mixing policies that invest in deprived neighbourhoods in order to deconcentrate poverty and immigration. It evaluates the effects on population dynamics at the neighbourhood level for a prominent place-based policy implemented in the Spanish region of Catalonia that falls into the second category of income-mixing policies.

What was the programme and what did it aim to do?

The policy started in 2004 and consisted of large, geographically concentrated investments that improved public spaces and facilities in the targeted neighbourhoods. [1] One objective of the policy was to deconcentrate poverty and immigration in a period when the share of immigrant population was rising rapidly, especially in low income urban neighbourhoods. The intervention was implemented through annual calls for proposals, with an annual budget of €99 million between 2004 and 2010. However, the degree of execution was low among those projects accepted in the 2008-2010 funding calls, mainly due to the fall in public sector revenues. For this reason, the study focuses on the 39 interventions corresponding to the 2004-2007 calls, with an average investment of €3,065 per inhabitant.

What were the programme eligibility criteria?

The selection process consisted of two rounds. In the first round, a deprivation index was calculated for each application made, using 20 socioeconomic indicators of neighbourhood characteristics. Neighbourhoods scoring above a given threshold entered the second round, where projects were ranked according to a final score that partly depended on the deprivation index. The pool of control neighbourhoods comprised the rejected projects as well as the projects accepted in the 2008-2010 calls that were never executed. The study used all the neighbourhood indicators used by policymakers when selecting the treated neighbourhoods in order to get robust treatment estimates.

What’s the evaluation challenge?

As the most deprived neighbourhoods were treated first, the projects accepted during the initial calls differ significantly from the control neighbourhoods. Moreover, these differences translate into differential pre-treatment trends in neighbourhood population dynamics.

What did the evaluation do?

The study solved this by including linearly control for pre-treatment trends in standard difference-in-differences estimators. Also, it combined the difference-in-differences with a matching approach (the Oaxaca-Blinder estimator) which is suitable for this setting. The results obtained with the two alternative approaches are qualitatively similar.

How good was the evaluation?

According to our scoring guide, the difference-in-differences method receives a maximum of three out of five on the Maryland Scientific Methods Scale (SMS). This is because although it does well to control for observable differences (e.g. neighbourhood indicators) between treated and non-treated neighbourhoods, it is unable to control for unobserved differences (e.g. migration shocks) that change with time. The study ensured that the control group was as similar as possible to the treated neighbourhoods by considering only future treated and rejected neighbourhoods — within the pool of control neighbourhoods — and controlling for the neighbourhood indicators used to target the policy in the regression analysis. Furthermore, there is a clearly defined treatment date. Therefore, we score the study three out of five on the SMS.

What did the evaluation find?

The graph below summarises the findings. In each chart, the solid line represents the evolution of each variable for the treated sample, while the dotted and dashed lines represent the analogous evolution for the control sample and the non-treated sample, respectively. The effect of the policy is shown by the difference between the treated group (solid line) and the control group (dotted line), after its introduction in 2004. Focusing on the top chart, which considers the percentage of natives as the outcome variables, the study does not find a systematic difference between the treated (solid line) and the control (dotted line) samples. The same observation holds for the two other outcomes: the percentage of non-EU15 immigrants (middle chart) and the percentage of EU15 immigrants (bottom chart).

What can we learn from this?

The average results showed in the previous graph provide evidence that this policy had no effect on the population dynamics of the intervention areas, indicating that substantial investment in public spaces and facilities is insufficient to attract native and/or high income households to deprived neighbourhoods. One notable exception seems to be the urban renewal projects carried out in the historic districts of Barcelona — the one large metropolitan area in the region. Specifically, the findings indicate that, here, the policy may have accelerated the ongoing process of gentrification that the city centre is experiencing.

Fig_1_URP_Case_Study_Graph

Reference

González-Pampillón, N., Jofre-Monseny, J., and Viladecans-Marsal, E. (2019). “Can urban renewal policies reverse neighbourhood ethnic dynamics? Journal of Economic Geography (forthcoming).

[1] For example, the Santa Caterina & Sant Pere project was a €15 million intervention carried out in Barcelona´s city centre. It involved restoring squares, public buildings, and providing equipment for collective use in these public spaces.