Agent-Based Modelling of Financialized Housing Markets: The Case of Vacancy Tax in Beirut

Abstract

Financialization of housing is a well-studied phenomenon, which has many effects, chief among them is decreased affordability of housing, and high vacancy rates in urban centres. Many policy proposals have been put forward to mitigate those effects, among which is the adoption of a vacancy tax, which aims to increase the cost of holding vacant housing units and encourages owners to release the housing units to the sale or rental markets. Decision makers often look to make use of decision support systems, or computer models, to test the complex and emergent effects of their policy proposals in a “virtual world”, before implementing them in the real world. Among the many possible modelling paradigms, agent-based modelling (ABM) fits the intellectual tradition of urban planning, insofar it acknowledges the complex interactions between agents and allows for the emergence of bottom-up structures, as well as its adaptability to participatory planning processes. I use ABMs to develop a model that describes the effect of a vacancy tax on a financialized housing market, integrating patterns on multiple scales, including pricing spatial distribution, vacancy rates, and neighbourhood definition. I apply the resulting model to the housing market in Beirut, using empirical data to calibrate it. I then run it for four scenarios, to examine the effect of a vacancy tax, the investment environment, changes in demand, and the search cost for households searching for housing. The results show vacancy rates decreased as vacancy taxes increased, demand increased, and with increased search time, but increased with higher interest rate offered from alternative investments. The most significant effect was from increased demand. Moreover, the effects were spatially heterogenous, with some neighbourhoods being more affected than others, depending on their class profile and proximity to important centres across the city.