Skip to main content


Volume 461: debated on Monday 11 June 2007

To ask the Minister of State, Ministry of Justice how prison population projections are calculated; and if she will place in the Library a copy of the models used to do so. (140913)

The models used to calculate the current prison population projections are described in Annex 2 of the Home Office Statistical Bulletin 11/06, ‘Prison population projections 2006-2013, England and Wales’, published on the Home Office website in July 2006. There are five elements to the modelling. In the very short term (the first three months) the XI2-ARIMA method developed by the USA Census Bureau is used. This method is publicly available at:

In the short term (between four months and two years) the population is modelled by a stock-and-flow model. This is not publicly available as it links to Prison Service databases that cannot be published. In the longer term (between two and seven years) the Grove-Macleod model is used. This has been published in OR Insight Vol. 11 Issue 1, January-March 1998, pp. 3-9, ‘Forecasting the prison population’. More detail is also available in Occasional Paper 80, ‘Modelling crime and offending: recent developments in England and Wales’ published on the Home Office website in 2003. The impacts of some changes in legislation and operational procedures are estimated using the Criminal Justice System Model and, if necessary, one-off bespoke calculations. The Criminal Justice System Model is owned by the Office for Criminal Justice Reform and is available to CJS practitioners but is not publicly available. Any one-off calculations are described in the aforementioned mentioned Home Office Statistical Bulletin.