Participatory budgeting (PB) — a process in which residents deliberate and decide how part of a municipal budget is spent — has been shown, using panel data on Brazilian municipalities from 1990 to 2004, to shift spending toward health and sanitation and to reduce infant mortality. That evidence, however, rests on a binary adoption indicator: it cannot say which design features of PB drive the effect, how much operates through spending reallocation rather than through information transmission or monitoring and accountability, or how the effect varies with political context and institutional maturity. Later work has produced associational evidence linking specific PB rules to lower infant mortality and experimental evidence on PB in other countries and for other outcomes, but no study causally identifies the operative design features or decomposes the mechanisms within the original expenditure-to-health chain. The problem is to supply that causal and mechanistic accounting.

Background

Participatory budgeting (PB) is a democratic innovation, pioneered in the Brazilian city of Porto Alegre in 1989, in which residents meet in neighborhood and thematic assemblies to deliberate on local priorities and to decide how a portion of the municipal budget is allocated; councils of delegates then monitor implementation. After Brazil's 1988 constitution decentralized spending and service delivery to municipalities, PB spread to hundreds of Brazilian local governments, usually promoted by the Workers' Party (PT), a nationally organized left party.

The empirical landmark for PB's welfare effects is Gonçalves (2014). Using a panel of 3,650 minimum comparable areas — geographic units that keep municipal boundaries consistent over time despite splits and mergers — covering 1990–2004, she coded a binary indicator for whether each municipality operated PB in each year and estimated regressions with area and year fixed effects (that is, exploiting within-area changes in PB status over time). Because adopting municipalities were initially richer and more urban, she also constructed a control group by nearest-neighbor matching, a method that pairs each adopting unit with a non-adopting unit having similar 1991 income or education. The findings: PB adoption raises the share of the municipal budget spent on health and sanitation by roughly 2–3 percentage points without raising total per-capita spending, and reduces infant mortality (deaths of children under one year of age per 1,000 resident infants) by roughly 1–2 per 1,000, about 5–10% of the baseline rate. Crucially, the mortality-reducing effect of health and sanitation spending is significantly larger under PB (a negative interaction between the spending share and the PB indicator), which the author reads as a hint of an accountability or monitoring effect beyond pure reallocation of funds.

Three causal channels are posited in this literature. First, allocative reallocation: PB redirects funds toward services residents prioritize, such as sanitation, which lowers mortality through reduced waterborne disease. Second, information transmission: deliberation reveals residents' needs to officials and officials' constraints to residents, reducing information asymmetry (situations where one side of a transaction knows more than the other). Third, monitoring and accountability: residents who help set priorities watch execution more closely, disciplining waste and improving how effectively a given budget is spent. Gonçalves' data cannot distinguish these channels, nor the design features — rules, procedures, the share of the budget covered, how representatives are selected, how intense participation is — that plausibly drive them, because her PB measure is simply an adopt/do-not-adopt dummy.

The broader evidence base is thinner and more mixed than the landmark result suggests. A systematic scoping review of 37 evaluations of PB's health and social outcomes found the infant-mortality evidence coming from five quantitative observational Brazilian studies with mixed results, and concluded that evaluations identifying impacts 'and the processes by which they are realised' were scarce. Subsequent large-N studies, experimental and quasi-experimental designs in other countries, and descriptive work on which municipalities sustain PB have advanced pieces of the question — design rules, institutionalisation, prior administrative experience, information channels — but none has delivered a causal, mechanism-decomposed account of the original spending-and-mortality result. That explanatory gap is the problem stated here.

Problem Statement

The problem is to convert the established association between participatory budgeting and improved municipal outcomes into a causal and mechanistic explanation — primarily in the Brazilian setting of the original result, and, where identification requires, in comparable decentralized settings — along three jointly necessary fronts that together answer one question: how does participatory budgeting actually work?

  1. Causal design attribution. Identify which concrete, time-varying design features of participatory budgeting — breadth of citizen participation, scope of deliberation, share of the municipal budget subject to PB decisions, rules for selecting neighborhood representatives, intensity and duration of operation, and degree of institutionalisation — causally drive its effects on the composition of municipal expenditure and on health outcomes such as infant mortality, as opposed to being correlated markers of already-reforming municipalities.

  2. Mechanism decomposition. Within a single credibly identified research design, quantify how much of participatory budgeting's total effect operates through each of the posited channels: (a) allocative reallocation of spending toward health, sanitation, and other mortality-reducing services; (b) information transmission between residents and officials; and (c) monitoring and accountability pressure that raises the effectiveness of a given amount of spending — the last being the channel Gonçalves' spending-by-PB interaction suggests but cannot isolate.

  3. Conditional effectiveness. Estimate how these effects and channels vary across municipal contexts — political commitment (mayoral party and program survival), fiscal capacity, region — and across stages of institutional maturity from adoption through consolidation, distinguishing genuine maturity effects from selection and survivorship.

Any credible causal accounting must address the selection problems that a binary adoption indicator cannot, and must confront confounding from concurrent national programs that rolled out across the same municipalities and years — notably the Family Health Program (Brazil's national community-based primary health-care program) and Bolsa Família (Brazil's national conditional cash-transfer program) — both of which independently reduced infant mortality.

References

  1. Gonçalves, S. (2014). The Effects of Participatory Budgeting on Municipal Expenditures and Infant Mortality in Brazil. World Development 53, 94–110. DOI 10.1016/j.worlddev.2013.01.009. https://doi.org/10.1016/j.worlddev.2013.01.009
  2. Touchton, M., & Wampler, B. (2014). Improving Social Well-Being Through New Democratic Institutions. Comparative Political Studies 47(10), 1442–1469. DOI 10.1177/0010414013512601. https://doi.org/10.1177/0010414013512601
  3. Wampler, B., & Touchton, M. (2019). Designing institutions to improve well-being: Participation, deliberation and institutionalisation. European Journal of Political Research 58(3), 915–937. DOI 10.1111/1475-6765.12313. https://doi.org/10.1111/1475-6765.12313
  4. Touchton, M., & Wampler, B. (2020). Public engagement for public health: participatory budgeting, targeted social programmes, and infant mortality in Brazil. Development in Practice 30(5), 681–686. DOI 10.1080/09614524.2020.1742662. https://doi.org/10.1080/09614524.2020.1742662
  5. Wampler, B., Sugiyama, N. B., & Touchton, M. (2019). Democracy at Work: Pathways to Well-Being in Brazil. Cambridge University Press. DOI 10.1017/9781108675949. https://doi.org/10.1017/9781108675949
  6. Campbell, M., Escobar, O., Fenton, C., & Craig, P. (2018). The impact of participatory budgeting on health and wellbeing: a scoping review of evaluations. BMC Public Health 18, Article 822. DOI 10.1186/s12889-018-5735-8. https://doi.org/10.1186/s12889-018-5735-8
  7. Beuermann, D. W., & Amelina, M. (2018). Does participatory budgeting improve decentralized public service delivery? Experimental evidence from rural Russia. Economics of Governance 19(4), 339–379. DOI 10.1007/s10101-018-0214-3. https://doi.org/10.1007/s10101-018-0214-3
  8. Abbott, J., McKiernan, K., & McNulty, S. (2024). Technocracy for the People? The Impact of Government-Imposed Democratic Innovations on Governance and Citizen Well-Being. Comparative Political Studies 57(2), 187–220. DOI 10.1177/00104140231178725. https://doi.org/10.1177/00104140231178725
  9. Zheng, F., Lin, H., Liu, B., & Fei, R. (2026). Scrutiny and Spending Shifts: How Participatory Budgeting Reduces Local Government Debt. Sustainability 18(1), 399. DOI 10.3390/su18010399. https://doi.org/10.3390/su18010399 (published online December 31, 2025)
  10. Castro, V., & Rocha, F. (2026). Determinants of the Survival of Participatory Budgeting: Evidence from Brazilian Municipalities. International Journal of Public Administration 49(12), 763–782. DOI 10.1080/01900692.2025.2559371. https://doi.org/10.1080/01900692.2025.2559371 (published online September 29, 2025)
  11. Saguin, K. (2018). Why the poor do not benefit from community-driven development: Lessons from participatory budgeting. World Development 112, 220–232. DOI 10.1016/j.worlddev.2018.08.009. https://doi.org/10.1016/j.worlddev.2018.08.009
  12. Johnson, C., Carlson, H. J., & Reynolds, S. (2023). Testing the Participation Hypothesis: Evidence from Participatory Budgeting. Political Behavior 45(1), 3–32. DOI 10.1007/s11109-021-09679-w. https://doi.org/10.1007/s11109-021-09679-w
  13. Sihotang, A. D. (2024). Does participatory budgeting improve public service performance? Evidence from New York City. Public Management Review 26(11), 3176–3200. DOI 10.1080/14719037.2023.2212259. https://doi.org/10.1080/14719037.2023.2212259
  14. Calabrese, T., Williams, D., & Gupta, A. (2020). Does Participatory Budgeting Alter Public Spending? Evidence From New York City. Administration & Society 52(9), 1382–1409. DOI 10.1177/0095399720912548. https://doi.org/10.1177/0095399720912548