← All signals

What the Evidence Says Emerging pattern Signal of the Week

Two towns, same budgets, far more projects funded: Aarau 17 instead of 7, Wieliczka 30 instead of 2

5 Oct. 2026

Aarau had CHF 50,000 for its first city-wide citizens' budget, 161 proposals submitted and 33 on the ballot. Under plurality voting, where you fund the most-voted project and then the next until the money runs out, an estimated 7 would have been financed: the costlier ones, with the largest voting blocs behind them. Aarau used the Method of Equal Shares instead. Every voter gets an equal slice of the budget, and each funded project is charged to the voters who asked for it. The same CHF 50,000 and the same 1,700 to 1,800 ballots funded 17 projects, across every district. That comparison is not the town's own account: it comes from a peer-reviewed AAAI/AIES evaluation. Nine of the 17 were realised by the end of 2024, five ran into 2025, and three could not be completed.

Wieliczka, in southern Poland, ran the same rule over PLN 1,000,000, with 6,586 voters and 64 projects on the ballot. Thirty won funding, worth PLN 995,079, where the town's previous standard-voting rule would have financed two. The share of voters who got none of their preferred projects fell from 28% to 18%. The furthest any voter lived from a funded project dropped from 6.1 km to 3.4 km.

A third Swiss case shows the method travelling beyond city-wide votes. In Winterthur, 38 residents drawn by lot allocated CHF 378,901 to 43 of 121 cultural applications in March 2025, using a co-designed variant that splits each member's virtual budget between their own vote and collective impact criteria. Of those committee members, 31% said the algorithm changed their decisions and 47% said it changed nothing.

In Aarau and Wieliczka the budget, the turnout and the proposals were unchanged. Only the counting rule was different. Plurality concentrates the money on the projects with the largest coalitions behind them; equal-shares voting spreads the same sum across more, smaller projects in more neighbourhoods. Two towns are not a body of evidence, and both comparisons are against a modelled counterfactual rather than a parallel vote. But both were measured, and both pointed the same way.

Supporting practices

Evidence package

Sources

3 of 3 practices evidence-assessed; 2 quasi-experimental with a published comparison against the previous voting rule

Audience

Municipal participatory-budgeting teams and councillors

The complete evidence package — limitations, confidence and per-practice sources — comes with Pro. See Pro