The Difference in Design: Participatory Budgeting in Brazil and the United States

Participatory Budgeting (PB) is conceptually powerful because it ties the normative values of non-elite participation and deliberation to specific policymaking processes. It is a democratic policymaking process that enables citizens to allocate public monies. PB has spread globally, coming to the United States in 2009. Our analysis shows that the types of institutional designs used in the United States are quite different from the original Brazilian programs. What explains the variation in PB institutional design between Brazil and the United States? Most PB cases in the US are district-level whereas in Brazil, PB cases are mainly municipal. We account for this variation by analyzing the electoral system; configuration of civil society; political moment of adoption; and available resources. We use case study analysis to account for this variation in institutional design. We then assess how the different rule design is likely to create a different set of institutional outcomes. Author Biography Hollie Russon Gilman is a Lecturer and Post-doctoral fellow at Columbia University's School of International and Public Affairs and a Fellow at New America and Georgetown's Beeck Center for Social Impact and Innovation. She is the author of Democracy Reinvented: Participatory Budgeting and Civic Innovation in America. Professor Brian Wampler focuses his research and teaching on Brazil and Latin America. Wampler has lived and conducted research in Brazil, Mexico, and Spain. Wampler earned his Ph.D. at the University of Texas, Austin, and his BA in Politics at the University of California, Santa Cruz. He lives in Boise, Idaho with his family.


Introduction
Participatory budgeting (PB) began in Porto Alegre, Brazil, in 1989, after two decades of military dictatorship as part of an effort to simultaneously build a democratic culture and deliver public goods to underserviced communities. By 2013 it had been adopted by over 2,500 local governments across Latin America, North America, Asia, Africa, and Europe (Sintomer et al. 2013). The World Bank and United Nations singled out PB as a "best practice" in democratic innovation and policymaking. In Brazil, several billion US dollars have been allocated through this public, participatory process. Recent research demonstrates that in the last 20 years PB in Brazil has enhanced governance, citizens' empowerment, and the quality of democracy (Avritzer 2002;Baiocchi et al. 2011;McNulty 2012;Wampler 2007;Russon Gilman 2016). PB in Brazil is associated with increases in civil society organizations, spending for health care, and decreases in infant mortality rates (Touchton and Wampler 2014).
PB came to the United States in 2009, when Chicago Alderman Joe Moore put $1 million of his discretionary funds into this participatory process. PB has grown to nearly 50 distinct programs across the United States. The US PB programs are within the larger family of PB because of there an emphasis on participation, deliberation, direct involvement in decision making, and social justice (Wampler 2012;Pateman 2012). In this article we provide an overview of the differences and similarities of PB in the major Brazilian cities of Porto Alegre and Belo Horizonte and the major US cities of New York City and Chicago. We refer to these four cities when we refer to the Brazilian and US cases.
Our analysis shows that the institutional designs being used in the United States are quite different from the original Brazilian programs. Most importantly, most early PB cases in the United States are adopted at the district (sub-municipal) level whereas in Brazil, most PB cases are at the municipal level. Although there are some US-based PB cases adopted at the city level (Vallejo, Boston, Greensboro 1 ), the two most prominent cases of PB are the district-based programs in Chicago and New York City. We focus on PB in Chicago and NYC, as they are the longest standing and most visible U.S. implementations to date. In the case of New York City, it represents the largest amount of dollars being allocated through a continuous PB cycle in the United States.
At the broadest level this article addresses the following question: What explains the variation in PB institutional design between Brazil municipalities and the district-based programs in the United States? More narrowly, we ask, why is PB adopted at the district level in the US? To answer these questions, we consider a number of possible explanations, including electoral system and districting rules; configuration of civil society; political moment of adoption; and available resources. We draw from case study analysis and secondary materials to account for this variation in institutional design and program configuration.
Given the differences in PB basic rules, we then turn to the question: How might the institutional design affect the outcomes generated? There are several subset questions to address, including: How do these rules alter how people/citizens engage the processes? In what ways do different types of rules alter the impacts? In other words, how should scholars and practitioners alter expectations for PBgenerated outcomes based on institutional design? In particular, the article examines the potential consequences of institutional design and process outcomes on (1) participants, (2) deliberation, (3) resources, (4) scale of money, (5) public learning, and (6) elections. By answering this question, we hope to provide preliminary answers to the "Why PB?" question that is often asked by politicians, citizens and civil servants.
This article should be of interest to policymakers and academics based in the US and Europe because it directly compares the first generation of reform (in Brazil) to the second generation (in the US). Given the vast literature on PB in Brazil, this will better enable these readers to understand the key differences in these programs. And, perhaps most importantly, the comparison of the potential impact will enable us to develop a clearer understanding of what we should expect from these programs.
The article unfolds in the following manner: First, the article explores the theoretical and normative values that guide PB. Second, the article discusses the potential impact of institutional design through exploring variation among PB in Brazil municipalities and sub-municipal programs in the US. Third, after outlining key institutional differences and similarities, the article analyzes five key variations: (1) political moment at adoption, (2) institutional context, (3) available resources, (4) civil society, and (5) internal PB rules that regulate how and when citizens participate. Fourth, the article uses a process-tracing method to analyze how the institutional design affects a range of outcomes. This co-authored article draws from the field research and work carried out by a US-specialist and a Brazilspecialist. It is our hope that our specific case study knowledge can be fused to contribute to provide insights into how variation in PB rules generates distinct outcomes. The article concludes by arguing that these rule differences shape future research and implementation questions. Given the likely differences in outcomes between Brazil's most successful cases and the growing number of cases in Brazil, we consider how this will shape practitioners' and citizens' attitudes.

The Promise of Participatory Democracy
The goal of PB is to provide a context for people to engage more deeply in their democracy. Participatory institutions aim to enhance governance, citizens' empowerment, and the quality of democracy (Avritzer 2002;Fung 2006;Pateman 2012). In addition to the broader effects on democracy, PB has an individual-level impact on participants. The institutional rules of PB are designed to improve the quality of participants' discourse, roles, responsibilities, and impact. The direct engagement of non-elite citizens is thought to provide a corrective to elite, technocratic policy decisions as well as to inculcate democratic values among participants.
PB processes contain a deliberative element, through which non-elite citizens are brought into discussion, dialogue, and negotiation with one another. The degree to which citizens engage in deliberation and dialogue is not consistent across different forms of PB and has a direct impact on the process and its outcomes. There are numerous opportunities for PB to expand the ability for everyday citizens to form preferences, effectively communicate, and improve democratic health. PB is conceptually powerful because it ties the normative values of non-elite participation and deliberation to specific policymaking and decision-making processes. Political theorists have been moving toward more unified and systemic approaches for analyzing deliberative democracy in context (for discussion see Mansbridge 2013;Smith 2009;Thompson 2008). Mansbridge (2013) identifies three criteria to govern a systemic approach for deliberative democracy. The first is epistemic democracy to "produce preferences, opinions, and discussions that are appropriately informed by logic and are the outcome of substantive and meaningful consideration of relevant reasons" (11). Democratic health enables citizens' considerations to be discussed, aired, and appropriately weighed (see also Habermas 1996). Second, is the ethical function to foster mutual respect towards effective communication (Mansbridge 2013, 11). Promoting mutual respect is an intrinsic part of the process that helps ensure that a deliberative process keeps running. Finally, the democratic function supports an inclusive political process with equality (12). It requires the inclusion of all types of people, which is critical to have an informed, contested environment. It also enables individuals in deliberative settings to conduct rational, good faith discussions to enhance democratic governance. In this capacity, citizens are agents who can make active choices to govern their society (Gutmann and Thompson 2004).
Although theorists have articulated norms for increased participation in a broad variety of political interactions, budgets are sometimes missed as opportunities for meaningful deliberative and participatory engagement. Fung and Wright (2001) articulate a concept of Empowered Deliberative Democracy (EDD) in which Brazilian PB is given as one compelling example among others, including: neighborhood governance in Chicago to check urban bureaucratic power over public schools and policing; Wisconsin Regional Training Partnership (WRTP/BIG STEP), which enables organized labor, firms, and government to assist workers in employment transitions; and Panchayat reforms in West Bengal and Kerala in India that have created both representative and direct channels to empower local villages (Fung and Wright 2001).
According to Fung and Wright (2001): "Conceptually, EDD presses the values of participation, deliberation, and empowerment to the apparent limits of prudence and feasibility" (7). EDD places PB in dialogue with diverse initiatives meant to give citizens an additional voice in decision making. Giving citizens a voice in decision making can lead to several types of improved democratic outcomes. Contemporary conceptions of participatory democracy include a deliberative element that interacts with governance (Fung 2007). Barber discusses participatory democracy in terms of the values of reasoned rule, self-government, and political equality (Barber 2003).
Participatory democracy offers an institutional design framework to empower citizens to have a more substantive role in governance beyond a simply consultative or advisory one. There is a diverse body of literature, which suggests that participatory programs might enhance state accountability, perceptions of efficacy, and efficacy (Avritzer 2002;Baiocchi 2005;Baiocchi et al. 2011;Fung and Wright 2003;Gaventa and Barrett 2012;Gibson and Woolcock 2008;Labonne and Chase 2009;McNulty 2012;Wampler 2007). Touchon and Wampler (2014), studying PB in Brazil, identify three aspects of PB programs that can uniquely strengthen democracy. First, governments adopting PB incorporate community-based organizations (CSOs) and citizens into the governance process. Second, the design of PB programs allocates greater levels of resources to underserviced, poor neighborhoods while also increasing spending on social services which the benefit the poor. Finally, directly empowering citizens to make public decisions provides hands-on civic education. The very process of engaging in participatory democracy provides a powerful "school of democracy" (Baoicchi 2005).
The participatory and deliberative aspects of PB can serve as citizenship training, providing a kind of learning whereby citizens leave with more knowledge, increased self-efficacy, and fewer antidemocratic attitudes (Almond and Verba 1963). "Individuals learn to participate by participating" (Pateman 2012, 10). Scholars have suggested that when people engage in participatory democracy they are better able to assess the performance of elected officials on both local and national levels (Santos 2005;Abers 2000). Furthermore, according to Amartya Sen, expanding human capabilities offers the greatest promise for producing broad social development (for discussions see Touchton andWampler 2014, 1446). Broadening these capabilities, in turn, can enable citizens to have more agency in their governance.
However, there are also concerns about the institutional design of PB and a fear that the concept has traveled too far and lost its original intent. First, there is a concern that participatory democracy has come to be seen as "ideological, oriented to personal transformation, and-no coincidence-as white" (Polleta 2005, 271). Organizational choice impacts the way people view a given process and the types of people who may be compelled to be a part of it. Polleta, used process-tracing to analyze the Student Nonviolent Coordinating Committee (SNCC) of the mid-1960s, demonstrates how participatory democracy became viewed as driven by principles instead of impact, focused on personal self-liberation rather than political change, and dominated by white, rather than black, participants (Poletta 2005, 272).
An additional critique of deliberative democracy, and participatory democracy in particular, is the lack of concern or relationships with wider society. According to Pateman (2012), "This means, for the most part, that 'democracy' in the wider society and political system is outside of their purview" (10). Instead of one-off deliberative experiments, Pateman calls for the creation of a "participatory society" which requires structural changes (10). Within this context, the institutional rules governing PB shape the degree to which the process can empower and enfranchise diverse, traditionally marginalized participants and the opportunity for PB to become more than a series of small-scale experiments. PB's ability to affect wider societal concerns is inextricably linked to its process configurations.

Institutional Design: PB in Brazil and the United States
We know from Ostrom's 1990 Nobel prize-winning work that institutional configurations greatly affect collective action. Ostrom worked within a rational choice institutional framework to provide a better explanation of why there is far more cooperation than rational choice theory would predict. The core of Ostrom's insight is that individuals are not only involved in one-time, single-shot interactions, but also in long-term interactions. When individuals have a past history and when they know that they will continue to work together in the future, there is a very different decision-making calculus than the classic rational choice models would suggest. Individuals are more likely to work together to find mutually agreeable solutions because democratic deliberation is not a zero-sum game but can produce positive-sum outcomes. It is vital to include an institutional analysis of PB, not just to supplement the broad number of sociological-based work, but because PB is a co-governance institution that involves intensive involvement of state officials.
Ostrom's insights are applicable to participatory democracy because citizens are encouraged to deliberate with each other regarding how they will spend scarce resources; citizens and government officials negotiating within PB have a past history and they are likely to work together again in the future, so there is a strong emphasis on cooperation. The deliberative characteristics of PB induce participants to listen to one another, present their arguments, and then vote.
Ostrom's insights are specifically applicable to PB in a couple of ways. First, PB encourages deliberation, which involves speaking and listening. As we argue in this article, PB in Chicago and NYC much more strongly emphasis small-group deliberation than is the case in Brazil. Brazilian PB programs sought to mobilize greater numbers of participants, which limited deliberation among participants. Second, PB encourages the formation of voting alliances among citizens, many of whom may not initially know each other. This encourages cooperation in subsequent years (Avritzer 2002;Baiochhi 2005;Wampler 2007). In a case from Ipatinga Brazil, Wampler uses the example of a rural community organization that agreed to delay its more expensive project until the fourth year of PB; during the intervening years, the community members voted for other groups' projects (Wampler 2007: 105). Third, PB encourages an ongoing commitment from its participants to hold government to account and keep the process running year-toyear. Unlike voting every two or four years, PB asks participants for a more sustained engagement. As a result, government officials know that they must treat participants with more respect because they will likely encounter each other in the following year.
We analyze the institutional design of PB in Brazil and the United States using the framework of Fung's (2006) "democracy cube," which offers an institutional approach to better conceptualize deliberation. The democracy cube outlines three distinct dimensions: (1) who participates, (2) how participants communicate, and (3) the degree to which discussions are linked with policy or public action. Within Fung's schema, the participants in PB are lay stakeholders as opposed to professional stakeholders or randomly selected participants. The mode of communication is deliberation and the degree of authority is on the higher end of the spectrum, with PB's voting mechanism leading to binding policy results that government officials implement in turn. Fung's Cube is particularly relevant to the study of PB because PB programs generally include participation of non-elite actors, public deliberation, and the delegation of authority to citizens.
Tables One and Two presents similarities and key differences in PB in large Brazilian and United States cities. The first important difference is the political and administrative context of PB. In Brazil, these programs are almost exclusively adopted at the municipal level, although there was a PB experiment at the state level (Goldfrank and Schneider 2006). Brazilian mayors are politically powerful, thus allowing them to implement very different policymaking processes. It was this concentration of authority in the mayor's office that created the necessary political space to create a new democratic process (Wampler 2007). In the US, city council members seeking to expand their connections to citizens initiated PB.
A second major difference that we identify in our comparative analysis is that Brazilian programs more strongly emphasize "within-PB representation" while US programs more strongly emphasize "within-PB deliberation." By "within-PB representation," we refer to internal selection processes in which elected or selfappointed community leaders are extended additional responsibilities not held by the average participants. For example, community leaders are called upon to publicly debate or to organize planning documents to propose documents. By "within-PB deliberation," we refer to a process in which all participants are induced to deliberate over policy proposals and community needs. For example, participants might be randomly assigned to a group where all participants have the opportunity to actively deliberate over their needs.  In Brazil, the emphasis on representation is notable in three ways. In the large regional meetings, where attendance may be between 500 and 1,000 individuals, it is only a relatively small number of participants who speak and are actively engaged. Most participants listen and are there to show their support. Second, Brazilian PB programs elect, from a pool of citizen-participants, "PB Delegates" who are then tasked with negotiation, deliberation and oversight. Finally, many PB programs use a municipality-wide PB council (Conselho do Orçamento Participativo or COP), which is an elected body of 20 to 30 PB delegates who are involved in oversight and making the program work. Thus, the Brazilian cases draws upon principles associated with participatory democracy as well as representative democracy.
In contrast, the New York and Chicago PB programs focus on the direct engagement of citizens at the local level. The district-level focus was led by city council members and the community partners; there were a strong emphasis on encouraging deliberation at the local level. There are several opportunities for deliberation, which first occur at the hyper-local neighborhood assembly meetings where residents identify neighborhood priorities for spending. Neighborhood residents learn about their city's budget process and the PB process, and then break up into groups to brainstorm. The organizers compile the results of these deliberations for the budget delegate phase of the process. At these neighborhood assemblies residents sign up to serve as budget delegates. The budget delegate phase follows, which offers opportunities for deliberation and dialogue for a selfselected smaller group of people than the idea assemblies. The US has a long history of using local deliberative processes; from the voluntary membership associations Tocqueville lauded to the idealized vision of New England Town Halls (Bryan 2003). PB fits into a longer history of a strong emphasis on public dialog. Thus, the US cases draw more heavily than the Brazilian cases on deliberative and participatory principles.

Assessing Variation
What accounts for the variation between the municipal-level implementation of PB in Brazil and district-level implementations in the United States? We analyze five categories that help to explain the variation between the Brazilian and US processes: (1) political moment at adoption, (2) institutional context, (3) available resources, (4) civil society, and (5) internal PB rules that regulate how and when citizens participate. For example, an internal rule is the degree of deliberation and representation within a given process. These criteria help inform the potential consequences of institutional design and process outcomes.

Political Moment at Adoption
PB developed in Brazil in a political moment marked by re-establishment of democracy and, importantly, with the widespread interest in creating new democratic institutions. PB was one among many new democratic institutions created in Brazil during the 1980s and 1990s (others include policy councils and conferences). Brazilian PB's roots lie in the post-authoritarian left in Brazil, in a political situation that grew out of 21 years of a military dictatorship. New thinking revolved around the concept of "radical democracy," also known as "direct democracy," "deepening democracy," and "democratizing democracy" (Goldfrank 2007). It was influenced by Marxist ideology and initiated by a specific political party, the Workers' Party (Santos 2005;Abers 2000;Avritzer 2002Avritzer , 2009Baiocchi 2001Baiocchi , 2005Goldfrank 2011;Wampler 2007).
The Workers' Party was a reformist, outsider political party that had deep ties to social movements and community organizations. The Workers' Party sought to accomplish a number of goals through PB: Engender greater participation, reform the policymaking process, build a base of political support, and attend to the policy and political demands of its base. There was an incipient party system in which party leaders were seeking to establish their base of support and establish a party "brand." The Workers' Party sought to brand themselves as participatory, democratic, and social justice oriented (Wampler 2008).
In the original campaign for PB, the Workers' Party (PT) outlined four basic principles for PB: (1) direct citizen participation in government decision-making processes and oversight; (2) administrative and fiscal transparency as a deterrent for corruption; (3) improvements in urban infrastructure and services, especially in aiding the indigent; and (4) change in political culture so that citizens could be democratic agents (see Goldfrank 2012;Ganuza and Baiocchi 2012;Peck and Theodore 2015).
In contrast, the United States has a more rigid party system that is dominated by two political parties. In most urban cities in the (e.g., Chicago, New York, Boston, San Francisco, Seattle, Portland), a single party-the Democratic Party-is the dominant political actor. In these urban environments, the most important elections are often, but not always, the primary elections. Democratic candidates must distinguish themselves vis-à-vis their Democratic rivals. Local-level politicians, such as aldermen (Chicago) and city council members (New York), work to organize and establish their own political bases. The PB process serves as one instrument to reach their constituents. Chicago aldermen often have personal relationships with their constituents and effectively serve as "mayors" of their wards, with significant decision-making power. For example, wards in Chicago are much smaller than council districts in New York City. There is roughly 1 aldermen for every 57,000 residents. In contrast, New York City has 1 city council members for every 165,000 residents. 2

Institutional Context
PB in Brazil has been adopted at the municipal-level, which allows their PB process to access greater resources. By contrast, in the Chicago and NYC, elected officials representing single member districts implemented PB. These officials bypassed mayors and other political leaders, and instead have used their discretionary funds to begin the process. Because Chicago and NYC have single member districts, council members have greater flexibility to implement PB within their districts. In Brazil, city council members are elected from the city as a whole and those don't have a specific incentive to support a district-level PB.
The relative strength of Brazilian mayors over local legislators plays a role in the structure of PB. Through powerful mayors in Brazil, PB often enters in conflict with the city council, whereas in Chicago and New York City, it has been the councilors or aldermen themselves who have instituted PB. In recent decades in the US, local-level elected officials have been concentrating more power. This is in part a reflection of the devolution of services from the federal to subnational governments for implementation but it is also reflected in the growing power of city council officials (Hall and O'Toole 2004;Milward and Provan 2000). For example, since reforms in 1989, the New York City Council has expanded in terms of size and its roles and responsibilities (for earlier history pre-reforms see Sayre and Kaufman 1960).
PB first implementation in the United States, at the behest of one Chicago alderman, was instituted with neither a strong partisan nor ideological framework. The New York City process is bipartisan, in distinct contrast with the partisan framing of PB in Brazil. Citywide processes are emerging-as seen in Boston, Massachusetts, or Greensboro, North Carolina, but to date these have been limited to small amounts of public monies or pilot processes. These cities are Democratic strongholds; in one-party dominant districts such as these, the process can be viewed as partisan and a key issue becomes, "What kind of Democrat are you?" PB is one example of how elected officials within these one-party dominant urban areas may work to distinguish themselves.

Civil Society
Brazilian PB is associated with a broader mobilization of civil society. Adopting PB in Brazil is associated with an increase in the number CSOs. PB's rules create specific incentives for citizens to participate in ongoing policymaking processes and to mobilize themselves into associations. This difference helps, in part, to explain the type of people who participate and their preferences. Research on PB in Brazil has demonstrated that broad majorities of participants and elected PB delegates have low income, low levels of education, and are often women (Abers 2000;Avritzer 2009;Baiocchi 2005;Nylen 2003;Wampler 2007), thus confirming that PB rules have successfully expanded public venues to include poor and traditionally excluded sectors.
Brazilian PB has been a tool to empower traditionally marginalized community members and enact policies that reflect the priorities of the poor. Many PB programs now adopt a "quality of life index," which allocates greater resources on a per capita basis to poorer neighborhoods (Wampler 2007). This creates a preferential bias in favor of the low-income residents, thereby encouraging them to participate. The process is also designed to allocate public dollars to the types of policy problems that most strongly affect poor neighborhoods (e.g., access to public health care and public housing, building basic infrastructure). In Brazil, there is an emphasis on infrastructure projects.
In the Chicago and NYC PB programs, there are not specific social justice rules that determine the allocation of public resources to low-income communities. However, social justice considerations are part of the broader deliberation about how resources are distributed (Lerner and Secondo 2012). 3 In both Chicago and NYC, considerable time and efforts was spent on incorporating poor and politically marginalized groups (Lerner 2014). The active involvement of individuals from poor and politically marginalized groups greatly increases the likelihood that policy concerns of central importance in these communities will be raised by participants. Of course, it doesn't guarantee that these citizens' demands will be meant. One task for researchers will be to determine if and how the inclusion of social justice values within the debates translates into projects that reflect these interests and if there is an accountability mechanism in place.
In Chicago and NYC, a strong emphasis on social inclusion is bringing new civic voices into the process, but there is wide variation among PB programs regarding who participates. Some communities have been able to encourage more diversity than others along various socio-economic indicators such as race, income, and education. In practice, New York and Chicago have been successful overall in mobilizing a wide cross-section of residents to engage as participants. While higher-income residents are still overrepresented in some districts, city officials have engaged in pro-active efforts to engage low-income and minority households, for example by targeting distinct communities, such as public housing residents, youth, and seniors. Roughly a third of assemblies included multilingual interpretation and translation support (CDP and PBNYC, 2015).
In New York City, 51,000 residents voted in the 2014-15 PB cycle. The majority of these voters, 57 percent, identified as people of color, in comparison to 47 percent for local election voters (CDP and PBNYC, 2015). Initial data from PB in the United States, including in New York City, demonstrates that PB is an effective gateway for getting people to vote. Comparing data from two districts in New York City, District 39 and District 23, researchers compared a PB voter with another similar person in the voter file who could not participate in PB because their district was not participating. The researchers' matched people from neighborhoods with similar racial composition, income, education, and voting patterns. Through this process, researches demonstrates that people who vote in PB on average are 7% more likely to vote in subsequent other elections (Lerner, 2018).
Research on PB in the US demonstrates the black residents and white residents are generally overrepresented or represented proportionally to their community's general share of the local census track (Hagelskamp et al 2016). But Hispanic residents are often systematically underrepresented. With regard to education and income, there is a bit of bifurcation. Lower-income households were overrepresented or represented proportionally in most PB programs, but there was also overrepresented among those with the highest education levels (graduate and undergraduate degrees) (Hagelskamp et al2016). This bifurcation indicates the PB programs are partially successful in attracting new political actors into the political system, most notably among young, low-income residents. The programs still difficulties attracting Hispanics and those with lower education levels.
Finally, we lack information to know if the inclusion of these new voices is being translated into new community organizations or the strengthening of existing organizations. Citizens may organize themselves as they seek to have a greater voice in the process, but it is unknown at this time if new civic voices are being translated into new community organizations.

Available Resources
There is a greater concentration of resources in the hands of mayors in Brazil. Mayors prepare the budget and, after it is approved by the city council, they often have the right to reallocate up to 20 percent of the budget without additional approval. Although city councils have the right to reject a mayor's proposed budget, the outcome of a rejected budget is that the previous year's budget is adopted. This means that the mayor has considerable leeway to move resources from department to department and from project to project. At the height of Porto Alegre's PB experience, in the late 1990s and early 2000s, upwards of 15 percent of the entire municipal budget was being allocated through PB. This often represented 100 percent of new capital spending. However, this high mark soon dissipated in Porto Alegre and in other large Brazilian cities. It became much more common to allocate 1 to 3 percent of the entire municipal budget to PB. Nevertheless, this percentage is much larger than in the US PB programs.
In Chicago and NYC, resources come from legislators who have a decentralized set of discretionary funds to spend in their individual districts. As a result, the process in the NYC and Chicago is restricted to a set percentage of councilors' discretionary funds, whereas PB in Brazil often has no clearly defined amount of resources at its disposal. Currently a small level of discretionary funds are being allocated to the process in the Chicago and NYC, potentially constraining PB projects to smallerstake political issues with redistributive social impact. To date, Chicago and NYC's PB programs started with a pre-determined amount of funds and has not used a needs-based tool, such as Brazil's quality of life index. Without a bounded set of dollars, US PB may use different structures to determine need, which may have longer-term policy implications.
Since 1994, aldermen in Chicago have been receiving "menu money" in the amount of roughly $1.3 million per ward per annum for infrastructure projects (Russon Gilman 2016). This "menu money" is disbursed equally to all fifty wards in Chicago in a need-blind allocation. The PB process in New York City was able to leverage roughly $1 million per council member in discretionary funds. Starting in 2009 with a bipartisan group of four City Council members in 2009, the process in 2016 involved 28 (of the 51) council members and has a centralized support structure in the Speaker of the City Council's office.

Internal Rules
In the United States process, deliberation occurs early and often, and the emphasis is on within-PB deliberation rather than within-PB representation (for discussions see Russon Gilman 2016). Budget delegates volunteer to serve, contrasted with the Brazilian system of electing representatives to serve on administrative committees or the PB council). The United States process involves a greater degree of deliberation in the brainstorming phase than its Brazilian counterpart. In the New York City and Chicago programs, the focus is on educating citizens, providing learning opportunities, and facilitating small-group discussions. Informed by these practices, speech and deliberation have been a vital component of PB's starting formula in the United States. 4 The current PB process in Chicago and NYC is based on four phases. The first phase of the process, idea assemblies, offer both an educational and deliberative component. Neighborhood residents learn about their city's budget process and the PB process, then break up into groups to brainstorm. The organizers compile the results of these deliberations for the budget delegate phase of the process. Attendees at these neighborhood assemblies vary, with roughly forty people per community across the country during 2014-2015. At these events, people sign up to serve as budget delegates.
The budget delegate phase follows, which offers opportunities for deliberation and dialogue for a smaller group of people than the idea assemblies. Residents learn about city guidelines, hear from government experts, and work to reach consensus about which projects should appear on the ballot. The only requirement for budget delegates is that they are residents within the community they represent. The budget delegate phase often starts with an orientation to identify common themes among the ideas submitted; residents then break up into thematic committees, e.g., Parks and Recreation, Streets and Sidewalks, Public Housing, etc.
A facilitator leads these specific thematic subcommittees. Facilitators receive varying degrees of training across districts and have unique backgrounds. Some facilitators have previous experience moderating or are community leaders while others are relatively new to facilitation. The budget delegate phase is time-intensive, often requiring a significant time commitment to attend in-person meetings over several weeks or months. There is often attrition during this phase. What stands out is that ordinary citizen participants have multiple opportunities to deliberate over project selection and implementation.
By contrast, Brazilian PB often has a structured representative system-as seen in the municipal-wide PB council (COP)-as well as in the large-group assemblies. Deliberation takes places in Brazil, but it is spearheaded, especially in the case of the large assemblies, by a small number of CSO leaders. Citizen-participants often politely listen and may learn from the deliberation, but most people in the large meetings never speak. In the COP, there is also extensive deliberation but among a limited number of representatives (30 to 50 in large cities). Deliberation in Brazil is more likely to occur at the neighborhood level, when citizens and CSO leaders prepare for the regional meetings. There are often lively discussions regarding priorities and needs. Many of these meetings take place outside of the formal PB process. The ideas from the citizens are thus funneled into the main venues.
In sum, the PB program in Chicago and New York City programs provide greater opportunities for citizens to engage in public deliberation whereas the municipalwide cases in PB more strongly emphasized the role of leaders to carry out key functions such as public speaking and working on oversight committees.

Institutional Design Consequences
As demonstrated above, Brazilian and US PB programs differ in their institutional design in a number of ways: (1) participants, (2) deliberation, (3) resources, (4) scale of money, (5) public learning, and (6) elections. A key issue for this study is the extent to which PB generates different outcomes.

Who Participates?
The participants in the Chicago and NYC PB processes represent a wider crosssection of people than in Brazil. While higher-income residents are overrepresented in some US cases (e.g., Park Slope or the Upper West Side in NYC), there has been a dedicated strategy to engage low-income households. For example, the largest US PB process to date has been in New York City, where the process has made a deliberative strategy to target distinct communities, such as public housing residents, youth, and senior to participate (Russon Gilman 2016). At the idea collection phase, neighborhood assemblies, roughly a third of assemblies had language support including interpretation and translation (CDP and PBNYC, 2015).
NYC's PB has been effective at ensuring that PB voters represent a larger percentage of previously marginalized residents than in traditional elections. In the period from 2014 to 2015, 51,000 residents voted. The majority of NYCPB voters, 57 percent identified as people of color, in comparison to 47 percent of local election voters and 66% of the total population of the participating twenty-four districts (CDP and PBNYC, 2015). As one black public housing resident told an author, "I thought all the affluent white people would look down upon me because I live in public housing-in reality, they were all understanding and wanted to help." This comment illustrates how PB programs can generate new connections and forms of understanding among disparate communities. The emphasis on social inclusion helps bring individuals from politically marginalized groups directly into formal policymaking spaces.
The New York City process is successful at galvanizing typically marginalized communities, in part, because of strong community anchors with rich ties to community members. The lead technical non-profit organizing PB in North America, the Participatory Budgeting Project (PBP) is located in Brooklyn and Community Voices Heard (CVH) is the grassroots organizing partner for the city's process. CVH is a membership multiracial organization that organizes low-income populations to influence policy change. PBP and CVH have been advocated for PBP adoption and leveraging their networks to participate. A large organizing effort helped bring PB to New York City, as Baez and Hernandez (2012, 324) note: "The CMs [Council members] had never heard of PB before being approached by community-based organizations." In places in the United States where CSOs play a stronger role, there is more diverse representation in the PB vote (Russon Gilman 2016). In communities that conducted CSO outreach, there was an association with increased representation of traditionally marginalized communities at the vote. However, the role of civil society is unevenly dispersed throughout the process. Some communities have a strong support network of civil society that, in turn, can lift and amplify PB. Other communities have less well-established civic infrastructure. To date, the evidence suggests that PB programs in Chicago and NYC are able to incorporate individuals from poor and politically marginalized communities. However, we do not yet have evidence that their programs have stimulated the creation of new civil society organizations. Rather, in some communities has benefited from already wellestablished civic institutions.
In Brazil, early recruit efforts typically focused on poor neighborhoods with minimal attention to middle class communities. Those who participate are more likely to be women, over 40 years old, and with less than a high school education (Wampler 2007;Goldfrank 2011). In addition, evidence suggests that many come from large communities (favelas or low-income communities) (Abers 2000;Baiocchi 2005). The groups that are less likely to participate include those in very precarious economic situations (homeless, extreme poverty) as well as those who live in small communities (since they cannot generate enough votes to secure their policy preferences). In addition, middle-and upper-middle-class residents do not often participate because they know that they will be vastly outnumbered by poor residents from large favelas as well as because most middle-class residents do not depend on the types of social services provided by the municipal government. It is important to note that third generation PB reforms in Belo Horizonte and Porto Alegre sought to incorporate middle class and youth sectors into PB processes.

Deliberation in US vs. Representation in Brazil
The Brazilian PB programs include basic elements of deliberation and representation in their processes, whereas the US programs have a greater focus on deliberation. There are three key reasons for these differences. First, the US program was led by the Participatory Budgeting Project (PBP), a CSO that advocated for increased participation and better deliberative processes (Lerner 2014). As a CSO bringing PB to elected officials and galvanizing support, it has been focused on community engagement and empowering diverse stakeholders to have a say in the process. In the original PB process in Chicago, there was a set number of elected community representatives in the guidelines; the guidelines were later revised to enhance community participation. In Brazil, a political party was the principal proponent of PB; they promoted greater participation and less deliberation.
The second reason, intertwined with the first, is that the Brazilian CSOs sought to increase the number of participants to demonstrate the worthiness of their claims to a broader audience. This interest in broad participation was shared by the Workers' Party, since in the new democratic environment, political parties had to worry about turning out the vote.
Third, in the US, there is a longer, more well-established political tradition of deliberation. From New England town halls to public hearings, public deliberation was within the larger "political repertoire" of democratic politics in the US (Tarrow 1992;Bryan 2003). In the context of the US, there was a greater emphasis on the quality of the dialogue and debate, leveraging the North American tradition of participatory planning, community engagement, and small grants. This contrasts sharply with Brazil's new democracy, in which they had to experiment with new forms of deliberation and representation, and find a way to use different democratic tools.
There are several unexamined research questions about the quality and effectiveness of these deliberations. Currently, there is wide variation in the quality and training of moderators. There is also wide variation in the amount of resources expended by public administrators. Public bureaucrats are often volunteering on their off hours to help provide resources to budget delegate committees, which greatly influences the process of dialogue and deliberation. When and how can deliberations help participants? What is the quality of information delivered to participants and how does that impact the nature of deliberations? Better-informed deliberation may increase the quality of democratic debates and improve signaling back to elected officials about citizens' preferences.
A more diffuse, broad-based engagement from civil society could also potentially engage traditionally marginalized participants. Currently, the US PB process has had neither the resources nor people power to bring new civil society organizations into the formal governance process. PB has been able to tap into and leverage existing strong civil infrastructure. While the US process has aimed to engage several traditionally marginalized populations including youth and non-citizens, there further resources are needed to enhance the public learning impacts.

Resources, Scale of Money, and Type of Projects
We would expect Brazilian PB programs to have a larger impact on policy and social well-being outcomes because of the greater amounts of resources dedicated to these programs. In Brazil, the start of PB program in the 1990s coincided with an expansion of public monies spent by municipal governments. Mayors interested in PB thus had greater flexibility and more resources than their predecessors. PB administrators in Brazil had far greater resources than their US counterparts. In the larger Brazilian cities of Porto Alegre, Belo Horizonte, and São Paulo, it was common for citizens to propose medium-sized infrastructure projects, such as building health clinics or schools. Larger projects, such as housing projects, were less likely to be included because of the costs associated with implementing them. The policy selection of PB participants reflects the demand for basic infrastructure projects.
With an average of $1 million dollars of capital funds in each PB district in Chicago and NYC, what can be achieved? The small-scale investment limits the PB process's ability to have a large social policy or redistributive affect. Based on the current PB funding, further research is needed to see when and how PB addresses areas of greatest need-and not simply the preferences of those residents who selfselect to serve as budget delegates. It is currently unclear if PB in Chicago or NYC is having any type of redistributive impact or reaching communities with the greatest need.
PB in Chicago and NYC often revolves around community organizing and allowing local politicians to connect with new constituencies. It is not about building broadbased power or radically disrupting the status quo budget system in a visible way. The decentralized process, with limited funding, results in less tangible big budget allocation for the public at large. Smaller allocations result in less publicity drawn to specific projects which may lead to a feedback loop where constituents are less aware of PB projects and therefore do not place a lot of political pressure on elected officials to expand the resources dedicated to PB.
The amount of funding at stake also influences the salience of representation in PB. Currently, each US implementing locality places a very small portion of overall funds into PB. While across the United States nearly $50 million was allocated, this is distributed in amounts averaging $1 million per PB process, typically of capital funds. Within each district, there are price restrictions to ensure projects do not cost more than $500,000; given the costliness of implementing government services, this curtails the scale at which PB influences public policy. Across 45 PB projects in North America 5 during 2014-15, the average winning project cost $195,506 (Hagelskamp et al. 2016). Across the United States, an average of five projects were funded.
In the 2014-15 United States PB process, parks and recreation projects were the most common ballot items overall, followed by school projects (Hagelskamp et al. 2016). Overall, schools received the largest share of PB-allocated funds (33 percent). The least common types of projects on the ballot were public housing and public safety projects. Public housing projects rarely appear on ballots and also have a low chance of winning funding when they do appear because of their high cost.

Elections
We would expect PB to have an important impact on the election of city council members in cities like Chicago and NYC, but very little impact in Brazil. For mayors, we expect the opposite. In Chicago and NYC, the effect may be most important in the Democratic primaries for city races. But here, too, we observe that the political value of PB for US mayors is low. While New York Mayor de Blasio and Chicago Mayor Emanuel both ran on a platform of support for PB expansion, to date, neither mayor has made PB a priority issue or devoted significant time or resources to the process. In 2018, Mayor de Blasio indicated that his administration would support the expansion of PB into public high schools, allocating $2,000 for every public high school, over 400, to decide through the PB process. 6 After learning about the power of PB in Brazil, in part thanks to outreach from the non-profit Participatory Budgeting Process (PBP) 7 and other CSOs such as Cities for Progress, Chicago's Alderman Moore decided in 2009 to cede a portion of his discretionary funds for infrastructure to the PB process (Russon Gilman 2016; Lerner 2014; Baez and Hernandez 2012, 320). Having won the Democratic primary by a narrow margin in 2007, Alderman Moore wanted to try something new to galvanize supporters. After losing touch with his constituents, Moore enlisted PBP to engage his constituents to direct $1.3 million in public funds (see Lerner 2014).
Since implementing the process, he has been able to elevate his national profile and garner support, and voters have kept him in office. In 2011, Moore won with 72 percent of the vote. "According to Moore, PB was the most common reason people gave for re-electing him" (Lerner 2014). The alderman has spoken about PB across the country and has even been honored by the White House. 8 Following re-election after the initial New York City PB pilot, several of the four council members who first implemented PB have taken on new roles and leadership responsibilities within the New York City Council structure. One of the first four, Melissa Mark-Viverito, was elected to serve as the powerful position of Speaker of the City Council. Other early adopters have in since taken on leadership roles within the council's progressive caucus. The speaker and the City Council helped usher in a series of good governance reforms to the council's discretionary spending regime. 9 The reforms offer a formulaic, needs-based model of disbursement. Through PB, several City Council members have been able to elevate their profile and champion a series of governance reforms to increase transparency and accountability.
The process in the US depends heavily individual elected officials' desire and staffing capacity. In places where there is limited staffing capacity, such as Cambridge, MA, it is more difficult for participants to engage in the process because there is insufficient administrative support. Although there is no political or ideological mandate, the majority of the proponents are Democrats, who already dominate US cities. In Chicago, it took several years for the process to gain traction and for additional aldermen to support it. Even now, there are political concerns to maintain Aldermanic menu money discretionary dollars, which would severely limit PB in its current manifestation. 10 As mentioned above, aldermen effectively serve as "mayors" of their wards, so there is little top-down party pressure to implement PB. In fact, one Chicago ward decided after a year's trial that the process was too time-intensive and that the turnout was too low to justify continuation (Bishku-Aykul 2014). Although enthusiasm for the process was cited, the ward decided to implement a type of "PB lite" titled an "infrastructure improvement program" (Greenfield 2014).
Although PB in Brazil was a citizen-based participatory process, it was also embedded in a new representative democracy. Candidates running for office sought to build a base of support by differentiating themselves from other candidates and parties. In the most successful cases of PB, such as Belo Horizonte, Porto Alegre, Recife, Ipatinga, and Garulhus, the presence of PB was the central plank of a governing party that was re-elected multiple times. Although it would be nonsensical to credit PB as the primary reason behind these governments' successful re-election, the heavy emphasis on PB was one reason that they won reelection (Wampler 2007).

Conclusion
Under the broad brand of "Participatory Budgeting," Brazilian municipalities and US city governments are working to directly incorporate citizens into policymaking. In both countries, and across multiple programs, governments designed and are using participatory policymaking to give citizens a voice and a vote, as well as oversight opportunities. This article demonstrates that the adoption of PB in the US has several distinct differences from the original Brazilian cases, most notably the district-level adoption in US cities like Chicago and NYC and the municipal-level adoption in Brazil.
The institutional differences produced PB-processes in Brazil rely heavily on "within-PB representation" whereas the US-based processes promote greater "within-PB deliberation." The Brazilian cases drew more heavily from representative democracies, which allowed these programs to incorporate greater numbers of citizens. The US cases more strongly emphasize deliberation, which creates greater opportunities to engage in community building, but there is a greater difficulty in incorporating larger numbers of participants. This emphasis permits greater public learning and more opportunities to generate consensus around policy selection. However, this focus makes it harder to scale up these programs, highlighting a key tension faced by PB advocates in the US: A clear strength of their programs is the greater degree of deliberation, but this limits the number of people who will be involved and puts constraints on public bureaucrats to lend their time and expertise for small group dialogue and deliberation. A spillover effect of the small number of participants is that it is less likely that mayors, congressional officials, etc., will pledge extensive support to the US programs. The current programs link citizens to city council members; these council members seek to generate new opportunities to participate at the neighborhood level.
In addition, the institutional differences in program design (city-wide vs. district) means that the Brazilian cases had greater access to resources and more mayoral involvement, which created the possibility of a greater impact on well-being (Marquetti 2003;Touchton and Wampler 2014). PB in Brazil was created as part of a political project that sought to transform the lives of citizens and cities, whereas the US-based experiences have been much more focused on expanding citizens' voice in policymaking. The shifts in the US cities like Chicago and NYC may be more around attitudinal and behavioral shifts.
Finally, the Brazilian PB cases were created at a moment of democratic renewal, which created to explicitly creating "social justice" rules that ensured that poorer communities would receive greater resources than wealthier communities. This rule helps to account for the higher mobilization among poor citizens as well as the most positive impacts on well-being. In contrast, in the US, the emphasis on social inclusion and social justice are organizing principles but they are directly made into rules. PB administrators seek to recruit a broader range of participation and there are social justice issues discussed during deliberation. There is preliminary evidence that some PB programs are successfully incorporating a broader range of citizens into the process but we don't yet have enough evidence to more strongly demonstrate how and if these programs are producing social justice-related change.
In sum, PB programs in the US are likely to have different types of impacts than have been associated in the Brazilian cases due to differences in institutional design, local context and available resources. Given the stronger emphasis on deliberation as well as informal efforts to promote social inclusion and social justice, we encourage researchers and government officials interested in assessing impact to draw more heavily from academic and policy works on deliberation. As PB continues to expand across the US, we should expect that the outcomes generated will differ from the Brazilian cases, which means that researchers should cast a wide net to assess how and if social and political change is being generated.