Towards more Equitable Housing Policies via AI and Agent Computing
Building holistic computational models of the housing market.
1 October 2019
Grant: Grand Challenges Small Grants
Year awarded: 2019-20
Amount awarded: £7,500
- Dr Omar Guerrero, Department of Economics, Social & Historical Sciences
- Dr Stephen Law, Bartlett School of Architecture, Bartlett
This project examined the UK housing market, creating a holistic computational model of the London housing market to understand how its dynamics drive inequality. To do so, the project integrated actual housing data and spatial considerations, so as to be able to test experiments regarding London infrastructure and its effects on prices.
In order to address housing inequality, policymakers often debate policies such as rent controls, inheritance taxes and social housing. However, evidence about the effectiveness of policies is scant. Partly, this is due to a lack of adequate methods.
Thus, this project sought to fill a key evidence gap to develop a spatially aware agent-computing model, where individual agents make realistic economic decisions within a particular geography. By combining AI, spatial analysis and economic computational modelling, the project provided novel analytic tools for bespoke policy designs.
The policy recommendations that it produced are pertinent not only to urban planners and housing regulators, but to a broader community of policymakers to understand how housing ownership underpins wealth distribution in the UK and to design evidence-based policy instruments. Results about estimating the impact of infrastructure in the housing market were produced and have formed the basis for policy papers.
Outputs and Impact
- Journal article: Decentralised markets and the emergence of housing wealth inequality, Computers, Environments and Urban Systems (vol. 84, 2020)
- Created dataset of the London housing market including housing demand and housing attributes
- Informed policymakers - white paper forthcoming