Participatory budgeting after 35 years: an evidence review

Participatory budgeting (PB) — letting residents decide directly how to spend part of a public budget — is the most widely copied democratic innovation of the past forty years. From its origin in Porto Alegre, Brazil, in 1989, it had spread to some 2,700 governments by 2013, according to the World Resources Institute, and has kept spreading since. This review asks what the evidence shows about its effects. The strongest findings come from Brazil, where studies of hundreds of municipalities associate PB with more spending on sanitation and health, lower infant mortality and higher local tax collection, especially where programmes lasted many years. Elsewhere the evidence is thinner and the programmes smaller: in Paris, New York and Seoul PB controls a few per cent of spending at most, and its main documented effects are on who participates and on public trust rather than on welfare outcomes. Porto Alegre itself shows how fragile the model is: after its founding party lost office in 2004 the process lost priority, project completion collapsed, and in 2017 a new mayor suspended it; it has since been revived, but its 2024 round was cancelled after the floods.

Key findings

• Brazilian evidence links long-running PB to better health and revenue outcomes. Gonçalves (2014) found that Brazilian municipalities using PB in 1990–2004 channelled more of their budgets to sanitation and health and saw lower infant mortality. Touchton and Wampler (2014) found PB associated with higher health spending, more civil-society organisations and lower infant mortality, with stronger effects the longer PB lasted.
• PB municipalities collect more local tax. A World Bank study of all 253 Brazilian municipalities with at least 100,000 residents found tax outcomes about 34% higher in municipalities with PB and about 39% higher where PB had run for over eight years. The paper does not know how the extra revenue was spent.
• Porto Alegre's decline is well documented. More than 80% of the 5,556 projects in the city's PB investment plans for 1994–2004 were completed by the end of 2004; of the 3,712 projects in the plans for 2005–16, only 42% were completed by early 2017. The share of the budget decided through PB fell from about 15% in 1991–94 to around 8–9% in 1995–2000. A new mayor suspended the process in March 2017; assemblies were being held again by 2023, the 2024 round was cancelled because of the floods, and a 2025 round of 23 assemblies took place.
• The poor participated — but not the poorest. In Porto Alegre's 1990s assemblies, participants were poorer and less often white than the city average, but analysts noted that the very poor took part less than their share.
• Outside Brazil, PB is small and its effects modest. Paris has allocated €768 million through PB since 2014, 34% of it in working-class neighbourhoods; 162,395 residents voted in 2025. In New York, PB engagement raised the probability of voting in elections by an estimated 8.4 percentage points, with larger effects for under-represented groups.
• Capture is a real risk. By the 2000s, studies and interviews in Porto Alegre pointed to rising patronage and to "community oligarchies" in alliance with councillors; the removal of term limits for budget councillors reduced renewal.
• Robust health evidence is scarce beyond Brazil. A 2018 scoping review found 37 evaluations, mostly case studies, and concluded that robust evaluations of health and wellbeing outcomes were rare outside Brazil.

Research question and scope

The question is: what does the evidence show about who takes part in participatory budgeting, what it changes in spending, services and wellbeing, and how it fails? The review covers the canonical Brazilian case and large-N Brazilian studies, the World Bank's work, Europe (with Paris as the largest programme), the United States (New York and Chicago), South Korea (Seoul) and India (Kerala's People's Plan Campaign, a cousin of PB that devolved planning and funds to elected local bodies). It excludes participatory planning without budget authority, and school or housing-association budgets.

"PB" covers very different things. Sintomer, Herzberg and Röcke (2008) identified six ideal types in Europe alone, from a Porto Alegre model adapted to Europe to community funds, public–private negotiating tables, consultation on public finances and "proximity participation". Wampler, McNulty and Touchton (2021) distinguish six PB types worldwide and find that changes in rules and design now produce significantly different outcomes. Any general statement about "the effects of PB" should be read with this variety in mind.

Background

Porto Alegre's Workers' Party (PT) administration introduced PB in 1989. Residents met each year in regional assemblies — and later thematic ones — to set investment priorities and elect delegates and councillors to a municipal budget council that negotiated the annual investment plan with the city. The PT held the mayoralty from 1989 to 2004. The process attracted international attention through the World Social Forum and through agencies such as the World Bank, whose 2007 volume on PB described it as a tool for inclusive and accountable governance while warning that participatory processes "run the risk of capture by interest groups".

Early outcomes in Porto Alegre were striking, though not all attributable to PB. Sewer connections rose from 46% of households in 1989 to 85% in 2002; the city built 900 km of sewage lines and storm drains between 1989 and 1996, compared with 1,100 km built by all previous governments; and per-capita spending between 1992 and 2000 clearly favoured the poorest regions. Part of this reflected an overall rise in infrastructure spending after 1989, so PB is better seen as the channel through which new resources were distributed than as their source.

Method and sources

The review draws on peer-reviewed large-N studies, systematic and scoping reviews, World Bank and World Resources Institute reports, city data and case studies. Where quantitative results are reported, the study design is noted, because most PB research is observational: municipalities choose whether to adopt PB, and those that do may differ systematically (for example, by having left-leaning or reform-minded mayors). Difference-in-differences and matching designs reduce but do not remove this concern. Figures from city governments (Paris, New York, Seoul) are the cities' own and have not been independently audited.

Finding 1: Brazil — the large-N evidence

Spending and infant mortality

Gonçalves (2014) studied Brazilian municipalities in 1990–2004 using difference-in-differences and matching methods. Municipalities that used PB favoured an allocation of spending that closely matched popular preferences and channelled a larger share of their budgets to sanitation and health; this shift was accompanied by a reduction in infant mortality. Touchton and Wampler (2014), using an original database of Brazil's largest cities over twenty years, found PB "strongly associated" with higher health-care spending, more civil-society organisations and lower infant mortality, and found that the association "strengthens dramatically" the longer PB remains in place. It was also stronger under mayors from the Workers' Party.

These studies are the backbone of the claim that PB improves wellbeing. Two cautions apply. Both are observational, so the correlation between PB and outcomes could partly reflect the kind of government that adopts PB. And the stronger effect under PT mayors suggests that political commitment, not the PB format alone, may be doing much of the work.

Tax collection

Touchton, Wampler and Peixoto (2019), in a World Bank working paper, examined all 253 Brazilian municipalities with at least 100,000 residents from 1989 to 2016. On average, municipalities with PB had locally generated tax outcomes (property, sales and fees) 34% higher than those without; where PB had run for more than eight years the difference was 39%, equivalent to about 3.3% of the mean budget and 40% of capital investment. The authors argue that participation builds trust and a sense of fair exchange that encourages compliance. They acknowledge that they do not know how the additional revenue was spent.

Implementation of proposals

Evidence from outside Brazil points the same way on one narrow question: whether participants' proposals are carried out. In Font and colleagues' study of 571 proposals from 39 participatory processes in Spanish municipalities, proposals from PB processes had a predicted 51% chance of full implementation, compared with 30% for proposals from other participatory formats, and were 17 percentage points less likely to be rejected. Budget authority, in other words, makes participation bite.

Finding 2: Porto Alegre — rise, drift and suspension

Who participated

The World Resources Institute's case study by Abers, King, Votto and Brandão (2018) summarises the evidence on inclusion. In 1998, 30.3% of regional and thematic assembly participants earned less than twice the minimum wage, against 11.4% of the city's population. A 1995 study found 26.6% of participants were non-white, compared with 15% of the population. Thematic assemblies drew a more affluent crowd: in 1995, 54% of thematic participants earned more than five times the minimum wage, similar to the city as a whole. And although most participants were poor, the very poor took part proportionally less, likely because they lacked the information, resources and above all time that participation requires. Ethnographic work such as Baiocchi's Militants and Citizens (2005) argued that the process also built new neighbourhood associations and civic skills.

Loss of priority after 2004

In 2004 the PT lost the mayoralty to José Fogaça. The WRI case study finds that the two non-PT administrations of 2005–16 kept PB but gave it lower priority, favouring a partnership-based programme in which community groups helped implement and fund projects rather than decide public spending. The budget data are stark:

The share of the municipal budget decided through PB had already fallen under the PT: from around 15% in 1991–94 to around 8–9% in 1995–2000, with one year at 5%. In the 2000s, interviews and studies suggested that patronage relations were rising again; the removal of term limits for budget councillors reduced turnover; and by 2009, for the first time, most assembly participants did not belong to any civic association. One scholar interviewed by the WRI team said the PB had been "captured by community oligarchies in direct alliance with members of the City Chamber and the municipal executive". In March 2017 the new mayor announced a two-year suspension. The process was later revived in reduced form: the city held regional assemblies in August 2023, cancelled the 2024 round because of the floods that hit the state that year, and in 2025 held 23 regional and thematic assemblies, which it says drew more than 18,000 participants, the second-highest attendance in the process's history. The city's figures have not been independently checked, and the share of the budget now decided through PB was not found. The WRI authors conclude that PB sustains transformative change only if it has well-structured arrangements, adequate resources, political commitment with enough flexibility to survive changes of government, and a government commitment to implement what it produces.

Finding 3: Europe — Paris and the European variants

Paris

Paris launched its participatory budget in 2014. According to the city, €768 million had been allocated through it by its tenth anniversary, of which €263 million (34%) went to working-class neighbourhoods (quartiers populaires); 1,345 projects had won and more than 3,400 works had been completed. The eleventh edition, in 2025, put 261 ideas to a vote; 162,395 residents voted and 104 projects were selected. Proposals must be investment spending. These figures make Paris one of the largest PB programmes in Europe in absolute terms, but they also show its limits: turnout is a small fraction of the city's roughly two million residents, and PB covers investment only, not the operating budget that funds most services. No independent evaluation of who votes, or of the distributive effect of the 34% earmark, was found for this review.

European variants

Sintomer, Herzberg and Röcke (2008) found that the transfer of PB to Europe was "highly differentiated", shaped by existing participatory traditions. They argued that "empowered participatory governance" was most likely to develop in the Porto Alegre-adapted and community-fund models, while consultative models mainly served administrative modernisation. Many European PB processes are small and consultative, and participation rates can be very low; this is one reason why European PB rarely shows the redistributive effects reported in Brazil.

Finding 4: the United States and South Korea

New York City

New York's PB began in 2011, when four City Council members launched it in their districts; it has since grown to include council members across the city, who ask residents how to spend part of their discretionary capital funding on projects costing at least $50,000. The FY2027 cycle's district-wide vote ran from 11 to 19 April 2026.

Two quantitative studies stand out. Johnson, Carlson and Reynolds (2023) estimated that engaging in PB raised the probability of voting in subsequent elections by an average of 8.4 percentage points, with larger effects for young people, the less educated and lower-income voters, Black voters and people who are a racial minority in their neighbourhood. Sihotang (2023), exploiting the staggered adoption of PB across council districts in a difference-in-differences design, found that PB reduced total complaint rates to the city's 311 service line — a proxy for service performance — but found no effect on specific complaint types such as school maintenance, potholes or new trees.

Chicago

Chicago's 49th Ward ran the first PB process in the United States. Stewart, Miller, Hildreth and Wright-Phillips (2014) found that residents chose very different projects from those produced by standard procedures, often ones the research literature classes as low priority, and that social-justice outcomes were mixed: there was evidence of geographic clustering, with some areas receiving gardens, dog parks and playgrounds while others received street resurfacing and sidewalk repairs.

Seoul and South Korea

South Korea is the main example of PB imposed from above. A 2005 amendment to the Local Finance Act allowed local governments to involve citizens in budgeting, and later national legislation made PB a requirement for local governments. Seoul enacted its ordinance in May 2012, with a citizen budget of 50 billion won and a 250-member committee combining nominated members with residents chosen from applicants, by lot when oversubscribed. In 2014, 1,533 proposals produced 352 approved projects for 2015. Participedia describes the 50 billion won as 2.2% of a city budget of about 23,000 billion won (23 trillion); the arithmetic gives about 0.2%, so the PB share was a fraction of one per cent. Touchton, McNulty and Wampler (2023) count South Korea, with Brazil and Peru, among the contexts where there is evidence that PB improves wellbeing, while stressing how few rigorous comparative studies exist.

Finding 5: Kerala's People's Plan Campaign

Kerala's campaign, launched in 1996 by the state's Left Democratic Front government, went further than most PB: the state decided to earmark 35–40% of its plan funds for elected local governments, which then prepared plans through village assemblies (grama sabhas), development seminars and expert task forces. T. M. Thomas Isaac, a State Planning Board member and one of the campaign's architects, reported that devolved resources reached Rs 1,025 crore in 1997–98 and Rs 1,178 crore in 1998–99, compared with untied funds of around Rs 20 crore before 1996–97. In the first campaign grama sabhas in August–September 1996, around 2.5 million people took part — about 180 per assembly and a fourfold increase on the previous round, but only around 10% of voters. Participation declined in later rounds; women were only 25% of participants in the first round; the educated middle class increasingly stayed away; and there were allegations of party manipulation. Heller, Harilal and Chaudhuri (2007) provide a later, independent evaluation of the decentralisation's effects on local democracy; their work is the standard reference for readers who want evidence beyond the architects' own account.

A comparison of the main programmes

What the evidence does not show

• It does not show that PB causes better health everywhere. The infant-mortality results come from Brazil, where PB often coincided with reformist governments and with growing municipal resources. The 2018 scoping review by Campbell, Escobar, Fenton and Craig found 37 evaluations, most of them case studies focused on political outcomes; all but one of the quantitative population-level studies were in South America.
• It does not show that small PB programmes redistribute much. Where PB covers a few per cent of investment spending, its fiscal weight is limited, whatever its civic value.
• It does not show that PB reaches the poorest. Even in Porto Alegre's best years the very poor were under-represented; time and information costs fall hardest on them.
• It does not show how robust gains are to political change. Porto Alegre's experience suggests that much of PB's effect depends on a committed executive.
• It does not settle whether digital voting broadens or narrows participation. Online voting can make taking part easier, but it bypasses the deliberative assemblies through which Porto Alegre achieved its inclusion effects; the evidence reviewed here does not resolve the trade-off.

Implications for institution design

1. Give PB real money and a protected share. Effects in Brazil grow with duration and scale. A fixed, legally protected share of the investment budget is harder to erode than a discretionary allocation.
2. Protect implementation. Porto Alegre's collapse in project completion after 2005 shows that the decisive moment is not the vote but delivery. Publish a project tracker with deadlines and reasons for delay.
3. Design against capture. Term limits for delegates, rotation, open meetings and published criteria for allocation between districts guard against "community oligarchies" and patronage.
4. Target inclusion deliberately. Equity weights (as in Paris's earmark for working-class neighbourhoods or Porto Alegre's distributive criteria), outreach through community organisations, childcare and paper ballots help counter the time and information costs that exclude the poorest.
5. Link PB to the wider budget. Isolated PB on small capital items rarely changes overall priorities. Kerala's devolution and Porto Alegre's budget council show what tighter links look like — and the risks when local capacity is weak.
6. Evaluate with comparison groups. Staggered roll-outs across districts, as in New York, offer natural experiments; cities should plan for them from the start.

Open questions

• Do PB's health effects in Brazil depend on the format or on the political coalitions that adopt it?
• Can PB survive changes of government without legal entrenchment — and does entrenchment drain it of energy?
• Does a citywide digital vote, as in Paris, produce the same inclusion effects as neighbourhood assemblies?
• How large must the PB share of a budget be before it changes overall spending priorities?
• Does mandatory PB, as in South Korea, produce genuine participation or compliance rituals?
• What happens to PB under fiscal stress, when there is little new investment to allocate?

Sources

1. Abers, R., King, R., Votto, D. & Brandão, I. (2018). Porto Alegre: Participatory Budgeting and the Challenge of Sustaining Transformative Change. World Resources Report case study. Washington, DC: World Resources Institute.
2. Gonçalves, S. (2014). The effects of participatory budgeting on municipal expenditures and infant mortality in Brazil. World Development 53, 94–110.
3. Touchton, M. & Wampler, B. (2014). Improving social well-being through new democratic institutions. Comparative Political Studies 47(10), 1442–1469.
4. Touchton, M., Wampler, B. & Peixoto, T. (2019). Of Governance and Revenue: Participatory Institutions and Tax Compliance in Brazil. World Bank Policy Research Working Paper 8797.
5. Shah, A. (ed.) (2007). Participatory Budgeting. Washington, DC: World Bank.
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, 822.
7. Touchton, M., McNulty, S. L. & Wampler, B. (2023). Participatory budgeting and well-being: governance and sustainability in comparative perspective. Journal of Public Budgeting, Accounting & Financial Management 36(1), 105–123.
8. Wampler, B., McNulty, S. L. & Touchton, M. (2021). Participatory Budgeting in Global Perspective. Oxford: Oxford University Press.
9. Sintomer, Y., Herzberg, C. & Röcke, A. (2008). Participatory budgeting in Europe: potentials and challenges. International Journal of Urban and Regional Research 32(1), 164–178.
10. Schugurensky, D. & Mook, L. (2024). Participatory budgeting and local development: impacts, challenges, and prospects. Local Development & Society 5(3), 433–445.
11. Baiocchi, G. (2005). Militants and Citizens: The Politics of Participatory Democracy in Porto Alegre. Stanford: Stanford University Press.
12. Font, J., Smith, G., Galais, C. & Alarcón, P. (2018). Cherry-picking participation: explaining the fate of proposals from participatory processes. European Journal of Political Research 57(3), 615–636.
13. Ville de Paris. 10 ans de Budget Participatif and Budget Participatif 2025: les 104 projets lauréats.
14. New York City Council. Participatory Budgeting.
15. Johnson, C., Carlson, H. J. & Reynolds, S. (2023). Testing the participation hypothesis: evidence from participatory budgeting. Political Behavior 45(1), 3–32.
16. Sihotang, A. D. (2023). Does participatory budgeting improve public service performance? Evidence from New York City. Public Management Review 26(11), 3176–3200.
17. Stewart, L. M., Miller, S. A., Hildreth, R. W. & Wright-Phillips, M. V. (2014). Participatory budgeting in the United States: a preliminary analysis of Chicago's 49th Ward experiment. New Political Science 36(2), 193–218.
18. Participedia. Participatory Budgeting in Seoul, South Korea.
19. Isaac, T. M. T. (2000). Campaign for Democratic Decentralisation in Kerala: An Assessment from the Perspective of Empowered Deliberative Democracy. Working paper, Centre for Development Studies and Kerala State Planning Board.
20. Prefeitura de Porto Alegre. Orçamento Participativo (structure, 2023 internal rules, 2025 round) and Prefeitura empossa novos conselheiros do Orçamento Participativo para a gestão 2025/2026 (22 November 2025).
21. Jornal do Comércio (29 July 2025). Assembleias do Orçamento Participativo já receberam mais de 9 mil participantes (2024 assemblies cancelled because of the floods).
22. Heller, P., Harilal, K. N. & Chaudhuri, S. (2007). Building local democracy: evaluating the impact of decentralization in Kerala, India. World Development 35(4), 626–648.