Is there a prison size dilemma? An empirical analysis of output-specific economies of scale

We advocate a nonparametric multi-output framework to estimate outputspecific economies of scale and we apply this model to male prisons in England and Wales over the sample period 2009-2012. To estimate output-specific returns to scale in prisons, we consider not only the cost-per-place, but also qualitative outputs such as purposeful out-of-cell activity and successful reintegration. Furthermore, we introduce environmental heterogeneity using the characteristics of the prison(ers). England and Wales offers a unique example to study economies of scale in prisons as the UK has started to build new super-size prisons in order to replace the most outdated prisons.