This paper aims to examine the effectiveness of different social housing allocation mechanisms, comparing direct-offering (DO), choice-based letting (CBL), and the Gale-Shapley Matching Scheme (GSMS). While CBL offers applicants greater choices compared to the traditional DO system, it remains unclear whether alternative matching mechanisms could improve allocation efficiency and overall welfare. The findings aim to inform housing policy decisions, offering insights into how more effective and equitable allocation strategies could be implemented in social housing systems.
This study uses a simulation-based approach to evaluate different social housing allocation mechanisms: DO, CBL and the GSMS. The model is constructed using simulated data, which is based on empirical tenant profiles and probabilistic distributions that match real-world housing demand in the London Borough of Southwark.
The results suggest that the GSMS provides better results in terms of matching rate, aggregate welfare for applicants and aggregate welfare for landlords. On the other hand, CBL is an improvement of DO, in terms of the above indicators.
The limitations of this research are threefold. First, the GSMS algorithm created in this research is conceptual, which omits some factors that might have real-life relevance for analysis purposes. Second, there is no access to real-life data that shows the profiles of social housing applicants and properties, and preferences of social landlords. Therefore, the author used simulated data based on empirical data. Third, because of the limitation of computational power, the author could not conduct simulation of GSMS for a large data set, which also limits the sample size of the research.
There are two main practical implications of this research. First, it extends the discussion of GSMS into social housing matching and allocation. In addition, it uses simulated data set as a method of policy evaluation. Both enable policymakers to explore alternative and supplementary allocation schemes. Second, using algorithms to allocation social housing can also improve efficiency in public administration. In particular, GSMS also creates higher matching rates than DO and CBL.
The main social implication of the research is to discuss and improve social housing allocation mechanism to better incorporate social tenants’ and landlords’ welfare. It also aims to help create long-term and sustainable tenancies.
This research makes an original contribution by integrating theoretical insights from matching theory with computational modelling to evaluate social housing allocation mechanisms. The findings offer a policy-relevant framework for improving housing allocation efficiency, fairness and tenant welfare, providing data-driven insights for policymakers seeking to reform social housing systems.
