Statistically Assisted Programme Selection (SAPS)
Active labour market programmes frequently generate disappointing or negative employment outcomes because public employment services fail to match unemployed jobseekers to measures that maximize individual employment probabilities. Statistically Assisted Programme Selection addresses this misallocation by providing caseworkers with econometric predictions of individual net programme impacts derived from comprehensive administrative historical datasets.
Picture this
Think of Statistically Assisted Programme Selection like a digital navigation app for job training. Instead of a driver guessing which route is fastest based on personal habit, the app analyzes millions of past trips taken by similar drivers under similar weather conditions to show the driver which specific turn gives them the highest probability of arriving on time.
What the evidence says
Providing econometric predictions to treatment caseworkers produced zero statistically significant change in programme assignment behaviour, with short-term compliance between recommended and assigned programmes remaining identical at 12% (Definition 1) and 29–30% (Definition 2) for both treatment and control groups.
- Who
- Historical estimation sample of N = 460,442 jobseekers aged 25 to 55 from Swiss administrative databases (2001–2003); field experiment evaluated N = 18,713 treatment jobseekers and N = 16,677 control jobseekers across 21 Swiss employment offices.
- How
- Randomized Control Trial (RCT) randomized at the caseworker level (142 treatment vs. 137 control caseworkers), utilizing biweekly econometric potential outcome predictions matched with Multiple Comparison with the Best routines.
What to do
Calculate individual potential employment outcomes using disaggregated administrative data, and integrate statistical prediction sets into caseworker workflow interfaces alongside explicit compliance monitoring protocols.
From the source
"The evaluation results show that caseworkers did not change their behaviour in any significant way due to having access to the additional information. Caseworkers either decided to ignore the system or were over-confident in that their own experience clearly dominates any information that a statistical system might provide."
Targeting_Labour_Market_Programmes_Results_from_a_Randomized_Experiment.pdf