This is our tenth blog post for our Job Market Paper Series blog for 2024-2025.
Joaquin A. Urrego is a Postdoctoral Scholar at the Haas School of Business at the University of California Berkeley. His fields are Urban Economics, Development, Economics of Crime, and Real Estate. You can find his JMP here.
Slums have existed for decades, and public policies aiming to close them have been key components in many developing countries’ agendas. While slums offer affordable housing to low-income families, they are characterized by substandard housing quality, limited access to urban amenities, and unsanitary conditions. Research on slum dynamics and the effects of different public policies is limited but has become more relevant as urbanization continues to accelerate in cities across developing countries (Alves, 2021; Gechter and Tsivanidis, 2023; Harari and Wong, 2024; Marx et al., 2013; Rojas-Ampuero and Carrera, 2023). But why is it so difficult to study slums, and even more challenging to close them? A significant issue is that we barely know what happens inside a slum. Limited data availability has restricted our understanding of informal settlements and the effectiveness of slum renewal policies.
We aim to bridge this gap by focusing on three analyses: (i) studying the location and evolution of all slums in Chile over the last two decades; (ii) estimating the direct effects of two policies implemented since 2011—in-situ upgrading and population relocation—on slums’ physical and sociodemographic characteristics; and (iii) identifying spillover effects of these policies on nearby formal neighborhoods in terms of physical and sociodemographic characteristics and crime outcomes. Our analysis focuses on areas that have been occupied by slums at some point rather than on the household level. This approach highlights policymakers’ concerns about urban renewal and nearby formal residents’ interest in spillover effects.
Chile presents an exceptional case study due to the proliferation of slums, the simultaneous implementation of different policies, and the availability of granular data. Chile defines a slum as an area that hosts eight or more families living in dwellings close to each other, lacks property rights, and lacks access to at least one basic service (water, electricity, or sanitation). In the last decade, the number of slums in Chile has increased by 66%, and the number of households living in slums has more than doubled. Since 2011, the government has spent over $400 million on two main policies:
- In-situ upgrading aims to transform slum areas into more formal neighborhoods by providing essential services and, in some cases, constructing new, improved houses.
- Population relocation aims to clear slum areas by moving households into formal neighborhoods outside the slum perimeter.
[Extensive] Data Efforts
To overcome the data scarcity, we collected over 90,000 satellite images and employed machine-learning algorithms to extract building footprints and monitor changes over time. This approach provided insights into the physical characteristics of slums, such as housing quality, density, and infrastructure, allowing us to create a detailed panel of slum areas. We combined this data with geocoded population censuses to measure sociodemographic characteristics.
Our efforts extended beyond satellite imagery. We used geocoded building permits from the National Statistics Institute to identify building activity and neighborhood development, gathered land tax records from the Chilean Internal Revenue System to measure housing investment, and included geocoded crime reports from the Crime Prevention Subsecretary. Land tax records and crime reports are available only outside slum areas, highlighting the contrast between the information available in formal neighborhoods versus informal settlements.
Location and Evolution of Slums
Our results show that slums are disproportionately located on the city periphery, but their specific locations are not random; they are strategically situated near low-skilled job opportunities. This proximity suggests residents prioritize access to employment over other amenities when choosing where to settle (Celhay and Undurraga, 2022). Comparing slums with a random sample of non-slum areas, both are at similar distances from clusters of firms in low-skill industries. However, slums are approximately 200 meters farther from the nearest school and 100 meters farther from the nearest bus stop.
Our dynamic analysis yields two additional insights. First, Chilean population censuses in 2002 and 2017 suggest that slums are initially populated by males, with families arriving later. Second, most slums within cities are settlements absorbed by urban expansion. When slums are established, they overwhelmingly form at the city’s edge. Over time, slums are “absorbed” into city boundaries as urban areas grow. Figure 1 illustrates where slums are situated in the Viña del Mar/Valparaiso area, alongside the urban footprint in 1993 and 2020. Slums are primarily positioned at the city’s borders. Moreover, new slums—those established in the last 10 years—are mostly found in the expansion areas of the urban footprint.
Direct Effects of Slum Renewal Policies
We use Synthetic Difference-in-Differences (SDiD) and Two-Way Fixed Effects models to estimate the causal effects of each slum renewal policy on slums’ physical and sociodemographic characteristics. The SDiD methodology combines Synthetic Controls with DiD (Arkhangelsky et al., 2021). For each cohort of treated slums, SDiD finds a synthetic control group from never-treated slums that matches the pre-treatment trends of the treated group.
We find that in-situ upgrading is more effective than population relocation in improving slum conditions. Specifically, in-situ upgrading leads to significant enhancements in housing quality, such as larger, more regular, and better-distributed buildings within the slum. It also stimulates neighborhood transformation by increasing the share of paved streets and attracting formal housing investments –more building permits– (Gonzalez-Navarro and Quintana-Domeque, 2016). Regarding sociodemographics, we observe higher-educated individuals moving into these newly formed neighborhoods. In contrast, population relocation shows no evidence of improved housing quality, housing development, or changes in socioeconomic composition within slums.
Spillovers of Slum Renewal Policies
In-situ upgrading also generates significant positive spillover effects on nearby formal neighborhoods. Areas within 200 meters of in-situ upgraded slums experience neighborhood development and increased housing investment, as evidenced by more approved building permits and a surge in building renovations. These neighborhoods also witness improvements in sociodemographic characteristics, with a greater presence of residents of higher socioeconomic status. Most of these effects persist up to 500 meters from the slum border. For areas near population relocation slums, we do not find similar effects. In fact, we do not see changes in residents’ composition even after a significant population decline (>15%) in formal neighborhoods within 500 meters of the slum.
Importantly, we find a substantial decline in property and violent crimes in areas adjacent to in-situ upgraded slums. After five years of intervention, formal neighborhoods within 200 meters of in-situ upgraded slums have 5.1 fewer property crimes and 2.6 fewer violent crimes per km² (see Table 1). We find no significant effects for areas near population relocation slums. This improvement in safety can be attributed to several factors, including changes in socioeconomic composition, enhanced state presence due to infrastructure investments (roads and new construction), and increased private security measures from housing renovations. The influx of higher socioeconomic status residents and increased employment opportunities may also contribute to reduced crime rates.
Conclusion
Our research underscores the importance of in-situ upgrading as a more effective and efficient slum policy compared to population relocation. By improving conditions within slums and fostering positive spillovers in surrounding neighborhoods, in-situ upgrading addresses both the immediate needs of slum residents and broader urban development goals. It is also less expensive, with the cost per household being two-thirds that of population relocation.
For policymakers, these findings highlight the need for comprehensive strategies that not only mitigate the adverse effects of existing slums but also limit the creation of new ones. Expanding affordable housing options and implementing inclusive urban planning can accommodate the needs of low-income populations, reducing the pressure on informal settlements.
Despite our extensive data collection and analysis, areas requiring further research remain. Notably, in the case of the population relocation strategy, we have focused on the geographical location of the slum rather than on the potential effects on relocated households and how the disruption of community ties affects them (Rojas-Ampuero and Carrera, 2023). Additionally, exploring the relationship between local labor markets and slum formation can inform policies that address both housing and employment challenges faced by low-income households.
Feature image is aerial photography from the Slum Bosque Hermoso in the municipality of Lampa near Santiago. It is taken from a photojournalism article from El Pais here.
