Homelessness Prediction Tool
In 2022, during the COVID pandemic, Dr Maryna Ramcharan developed a Homelessness Prediction Tool to anticipate how the pandemic would influence homelessness risk across the West Midlands.
During the first national lockdown, one of the UK government’s key emergency responses was to introduce a temporary ban on evictions. This measure aimed to protect individuals who had lost their jobs due to the pandemic and were unable to meet their rental commitments. The policy had significant implications across the West Midlands, where many tenants and landlords were affected, placing additional strain on housing-related support services that were already under pressure.
The challenge was to anticipate how the pandemic would influence homelessness risk across the West Midlands, particularly under conditions where standard forecasting assumptions no longer held. The pandemic disrupted typical behavioural and economic patterns, violating the core assumption in statistical modelling that historical trends can be extrapolated into the future. A new approach was needed to predict which areas and populations would be most affected.
To address this, we developed a predictive tool grounded in pre-pandemic evidence on the risk factors associated with housing insecurity, which we then combined with emerging data on the economic and social impacts of COVID-19. We sourced data from multiple data sources and developed a new metric to estimate the likelihood of a rise in homelessness in each Parliamentary constituency in the West Midlands. This metric incorporated variables such as household income, tenure type, presence of dependent children, household composition, and others.
The first version of the Homelessness Prediction Tool was launched in 2020 and shared with local authorities across the region. We then undertook an iterative design process, gathering feedback from users to enhance the tool’s utility and geographic resolution. Following gathered feedback, we developed a second version that included breakdowns at lower geographical level, using additional proxy indicators to reveal pockets of vulnerability that might be masked within broader, more affluent areas.
Impact
The tool was used by local authorities across the West Midlands during the pandemic years of 2020 and 2021 as a radar toolkit, providing insights into areas at higher risk of rising homelessness. By identifying both the geographic areas and the population groups most likely to experience homelessness, the tool supported homelessness prevention teams across the region in implementing targeted actions to reduce and mitigate its impact.
Find out more
Homelessness in the West Midlands Prediction Tool - Policy briefing for the Midlands Engine Observatory, November 2022
Which Parliamentary Constituencies Will Most Likely Face a Homelessness Rise? - Blog, Jan 2022