Modeling Gentrification Dynamics: An Agent-Based Approach to Segregation and Public Housing Policies

Gentrification is a pressing and controversial phenomenon that is reshaping thousands of cities around the world. Although it offers the possibility of revitalizing degraded urban areas, its potential to displace long-term low-income residents poses a major threat. Many aspects of this phenomenon are still widely debated, including whether it results in greater social and income integration, and which policies could mitigate its harmful effects. In this thesis, we propose an agent-based model (ABM) to study the impact of gentrification on urban segregation and to explore how public housing policies could affect its dynamics. The model simulates a stylized city populated by two classes of agents: low- and high-income residents. Agents move across the city according to a Schelling-like rule: a utility function defines their satisfaction level, which determines whether an agent is happy or unhappy. At each step, happy agents remain in place, while unhappy ones relocate. A developer agent drives the gentrification process by renovating properties in response to market signals. The model is initialized using real-world data from the United States. Although it is designed to represent a generic U.S. city rather than a specific one, it could be calibrated in future work to match a particular urban context and used for predictive purposes. The model reproduces gentrification patterns that are consistent with observed trends in income and rent increases in real U.S. neighborhoods. According to our simulations, gentrification does not necessarily improve integration between income groups; from certain perspectives, it may even worsen it. The model also shows that even a small fraction of public housing units can be very effective in slowing down gentrification dynamics. While using ABMs to model gentrification and segregation is not new, introducing an external agent as the explicit driver of the process, as well as incorporating housing policy interventions into the simulation, represent an innovative and underexplored approach.