🤖 AI Summary
This paper addresses the challenge of post-hoc performance evaluation for losing projects in participatory budgeting (PB). We propose the first quantitative framework grounded in counterfactual variability: it measures a project’s relative competitiveness by computing the minimal cost reduction, additional voter support, or removal of competing projects required to make it win. The framework ensures both interpretability and polynomial-time computability, and we design efficient algorithms tailored to three major PB rules—Greedy Approval Voting, Phragmén’s rule, and Equal Shares. Extensive experiments on multiple real-world and synthetic datasets demonstrate that our metric effectively uncovers nuanced differences in the latent competitiveness of losing projects across rules. This enables principled ex-post attribution of PB outcomes, enhances decision transparency, and supports evidence-based policy refinement.
📝 Abstract
We study ways of evaluating the performance of losing projects in participatory budgeting (PB) elections by seeking actions that would make them win. We focus on lowering their costs, obtaining additional approvals, and removing approvals for competing projects: The larger a change is needed, the less successful is the given project. We seek efficient algorithms for computing our measures and we analyze them experimentally, focusing on GreedyAV, Phragmen, and Equal-Shares PB rules.
