Project R-7311

Determining the optimal forecasting method for the simultaneous equation system of criminal litigation. In search of accurate forecasts for the influx of prosecutor's offices and criminal courts in Belgium. (Research)

In the past decades, scholars started to develop crime forecasting models. Yet a comprehensive forecasting tool for the entire criminal justice system (i.e. prosecutor's office and criminal courts) is still lacking. This research proposal aims to forecast the influx of each prosecutor's office and criminal court in Belgium and thus goes well beyond the state of the art. Not only are prosecutor's offices and criminal courts in Belgium faced with backlogs, but major reforms drastically increase the need for these forecasts. However, solely accounting for prosecutors' and courts' influx could yield inaccurate forecasts because they are part of a simultaneous equations system. For example, the level of crime is determined by convictions and the police force, but the optimal size of the police force is dependent on, among other things, the prevailing level of criminal activity. Similarly, the number of court cases depends, among others, on crime rates. The necessity to incorporate these interdependencies of the criminal justice system makes this research groundbreaking in the field of law and economics and economics of crime. I use 4 (time-series) forecasting methods (ADL models, models with a Kalman filter, ARIMAX models and VAR models) to estimate the structural equations and 1 method (ARIMAX) to estimate the reduced form equations and test which approach is best suited to forecast the 41 prosecutors' and criminal courts' influx of cases.

01 October 2016 - 30 September 2019