In many public institutions, productivity depends not only on workload, but also on whether tasks are assigned according to the capacities and specialization of public servants. This document presents evidence from a randomized controlled trial (RCT) conducted by researchers Alipio Ferreira and Sandra Cermeño León in partnership with IPA and the Organismo de Evaluación y Fiscalización Ambiental (OEFA). 

The study evaluated whether the automated assignment of case files on environmental infractions improves the performance of analysts responsible for administrative sanction procedures. The hypothesis was that, by considering variables such as the complexity of each case, the risk of expiration, estimated processing times, and the available capacity of analysts, the algorithm would optimize expected productivity by distributing the workload in a balanced manner and reducing the accumulation of case files close to expiration. The study evaluated two interventions: the first intervention replaced the discretionary assignment by supervisors with lists generated by an algorithm; the second intervention maintained the assignment of case files by supervisors but added suggested completion dates to help organize the work. 

The results show that automated assignment by algorithm reduced the probability of timely completion by 5 percentage points relative to a mean of 40 percent in the control group. In contrast, showing suggested dates increased the probability of timely completion by around 4 percentage points. No improvements were found in fines or appeals. In sum, replacing managerial discretion with automated assignment reduced productivity, while a light work organization tool did generate improvements.

Implementing Partners