Modelling demand on the prison population in England and Wales
Over the past four decades, the prison population in England and Wales has doubled in size and has shown no indication of decreasing. Prison shows little sign of working; reoffending rates remain high and there is no proven track record of rehabilitating offenders away from a life of crime, despite a £37,000 a year price tag per inmate. With the Ministry of Justice needing to find 15% additional savings by 2020, it will be crucial for the criminal justice system to find ways of managing demand on its services and exploring more effective, beneficial and economical methods of punishment than prison.
With this in mind, Crest Analytics conducted an analysis of what has driven the growth in the prison population, followed by the development of an illustrative model to enable policymakers to ‘apply’ alternative policy interventions to assess the impact on the prison population. We analysed a range of national data, from annual number of court proceedings to reoffending and recall rates, and conducted a desk-based review of prison policy reform over a twenty year period.
The model designed and produced by Crest allows users to identify the key drivers of the prison population and assess the impact of individual policy changes on the system as a whole, including the prison population.
Specifically, the model is able to:
- Predict future sentence numbers and prison population for one or more crime types
- Compare predicted trends with the historical trends
- Present numbers and data visualisations
- Create a scenario to understand the impact on the prison population of:
- Use of community sentences
- Sentence length
- Proportion of sentence served
- Reoffending rate
The advantages of the model we have designed are that it is:
- Granular: we can choose different crime types to model and therefore can answer policy questions readily
- Interactive: it can create scenarios so the user can ask ‘what if’ questions
- Rigorous: it is based on the simulation of thousands of results and therefore more reliable results are presented
- Expandable: we designed the tool so it can be expanded readily