Designing Public Housing: Evidence from Cities in Brazil

Policy Context

Public housing is one of the most widespread policies in cities across the globe, providing subsidized homes to the poor in both developed and developing countries (del Pero et al. 2016). To maximize the number of new homes, policymakers typically build subsidized units far from city centers where land costs are lower. By focusing on building at scale, this common cost-driven policy design overlooks the importance of local amenities for the socioeconomic outcomes of vulnerable adults and children (Chyn and Katz 2021). Existing evidence shows that public housing programs isolate beneficiaries from jobs, effectively moving them away from opportunity (Banhardt et al. 2017, Doria et al. 2023, Franklin 2020). Housing assistance initiatives under alternative designs are scarce, and so is the evidence on whether they could improve the socioeconomic perspectives of beneficiaries and induce intergenerational mobility.

In this paper, researchers provide novel evidence on alternative designs for public housing. Using data for hundreds of thousands of beneficiaries in more than 400 cities in Brazil, researchers examine how differences in the local implementation of the same program drive the impacts of public housing on economic outcomes of low-income adults and their children. Researchers focus on three policy parameters: (i) the characteristics of the neighborhoods where beneficiaries move to, (ii) the income profile of the targeted families, and (iii) the scale of the program at the local level.

Study Design

Researchers divide the analysis into two parts. First, they estimate the nationwide average causal effects of the public housing program. They do that by implementing a staggered, matched event study design. Researchers match families who received a home in the first half of the program (2010-2014) with a control group of families who received their homes in the second half of the program (2015-2018). Using rich data from admin records of the program, the Brazilian Welfare Registry and matched employer-employee records, researchers run event studies at the individual level to recover the average effects of the policy on housing quality, labor market outcomes of adult beneficiaries, and educational attainment and labor market outcomes of their children.

Second, researchers investigate how the impacts of public housing vary with the 3 policy dimensions of interest. They recover the causal effect of public housing for more than 400 cities by estimating the event studies described above for each city separately. Combining Empirical Bayes shrinkage and a Bayesian Hierarchical model, we then estimate how much of the effects of the program at the city level are driven by: (i) the characteristics of the receiving neighborhoods where beneficiaries move to, such as distance to city center, average income and education of residents, and local availability of public goods, (ii) the position of the average family of beneficiaries in the city income distribution, and (iii) the share of families that receive housing in a given city.

Results and Policy Lessons

The analysis reveals contrasting effects of subsidized housing on adults and children. Adult beneficiaries of Brazil’s Minha Casa Minha Vida program experience modest declines in formal employment following receipt of subsidized housing. These effects are not driven by spatial isolation or reduced access to labor markets. Instead, the evidence is consistent with income effects, suggesting that improved housing conditions slightly reduce labor supply among adults without substantially affecting wages or long-term employment trajectories.

In contrast, the impacts on children are large and transformative. Earlier exposure to subsidized housing during childhood substantially increases educational attainment and improves labor market outcomes in early adulthood. Children who receive housing assistance at younger ages are significantly more likely to complete high school, enroll in college, and hold formal jobs as adults. These gains translate into higher earnings and stronger attachment to the formal labor market.

The benefits for children are concentrated among those from the most disadvantaged households and are closely linked to improved neighborhood conditions, particularly proximity to higher-quality public schools. Preliminary education data further support this mechanism, indicating that access to better educational environments plays a central role in driving long-run gains. Overall, the findings suggest that while housing assistance has limited short-run labor effects for adults, it can generate substantial intergenerational mobility when it improves the environments in which children grow up.