An Unlikely Success : Peru ’ s Top-Down Participatory Budgeting Experience

This article focuses on the unlikely success of Peru’s top-down participatory budget experience. As part of democratization and decentralization efforts in the early 2000s, Peruvians mandated participatory budgeting in all subnational governments. The article suggests that, while success is constrained in many ways, Peruvians can point to two important accomplishments: 1) engaging a significant number of civil society organizations in debating public resources; and 2) an increased focus on “pro-poor” projects. The article concludes that the current challenge in Peru is to improve the process and engage an even more diverse array of participants. Only then will the process have real potential to improve local governance.


Introduction
To what extent can a top-down participatory budgeting initiative succeed?In some respects this question appears to be counter-intuitive: can national governments mandate participation at local levels of government?Will the process "stick" if there has not been a groundswell of support from below?Can a top-down initiative change the nature of local politics?As experiments with participatory governance spread around the globe, scholars and practitioners are documenting the various forms of institutions that engage citizens in subnational1 decision-making processes.Many of these experiences emerge from grass-roots initiatives to change the face of local politics.Others, however, are mandated by national governments; that is, they are "top-down" in nature.Top-down participatory projects have taken place in countries as diverse as India, Uganda, Bolivia, and the Philippines.They include citizen councils, townhall meetings, participatory development planning, and participatory budgeting (PB).
Participatory budgeting-the topic of this issue-often emerges from grass-roots activism at local levels of government.For example, the paradigmatic case of participatory budgeting in Porto Alegre, Brazil began as a local initiative.To date, only a few countries' national governments-including Peru, Bolivia, and the Dominican Republic-have mandated a participatory budget process to take place around the country in subnational governments every year.This article documents one of those experiences: Peru's mandated participatory budget process, which currently takes place in almost all of the 25 regions (akin to states in the United States), 195 provinces (akin to counties), and 1838 districts (also called municipalities) on an annual basis.Because it is obligatory at all subnational levels of government, it is one of the most ambitious participatory experiments in the world.
This article focuses on the unlikely success of Peru's top-down PB.I argue that while success is constrained in many ways, Peruvians can point to two important accomplishments: 1) engaging a significant number of civil society organizations in debating public resources; and 2) an increased focus on "propoor" projects.Data gathered during fieldwork in 2004, 2005, and 2010, as well as from evaluations conducted by development agencies, support this finding.The article concludes that the current challenge in Peru is to improve the process and engage an even more diverse array of participants.Only then will the process have real potential to improve local governance.

Peru's Top-Down PB Process
Peru's participatory budget process is closely linked to a democratization process that took place in 2000 after the fall of Alberto Fujimori's authoritarian regime.Fujimori had governed the country for ten years in a regime typified by the gradual concentration of power in the executive, extreme corruption, and the lack of transparency.After two national scandals came to light, involving high-level corruption and links to drug trafficking, Fujimori fled the country to live in exile in Japan. 2 He left a highly corrupt, centralized, and inefficient government in his wake.Immediately, reformers were faced with the difficult task of reconstructing a more democratic and transparent government.The mandated participatory budget process emerged in this context.
The legal framework of the PB is part of a larger decentralization reform, passed by Congress in 2002, that strengthened regional, provincial, and district levels of government. 3The reform is explicitly participatory in nature, as noted in the constitutional amendment designed to outline the very basic framework for decentralization.In addition to mentioning other participatory institutions,4 the constitution states that the regional government shall "formulate their budgets with the participation of the population and publicly disclose their execution annually according to the law" (Law 27680).
This mandate was clarified in a subsequent Participatory Budget Law (Law 28056), passed by Congress in 2003, which dictates that the capital investment costs5 of each regional, provincial, and local budget must be developed with civil society input.Following a series of steps-developed by the Ministry of Economy and Finance (MEF)-subnational governments must demonstrate that they have complied with this process in order to receive their annual budgets.
The original participatory budget law outlined eight phases that occurred over the course of the year, including a call for participation, registration of participants, a training period for participants-called "participatory agents" or PAs-the formation of a technical team, and several meetings during which participating agents prioritize and vote on investment projects.The final phase consisted of setting up an oversight committee, made up of representatives from civil society organizations who are also participating agents, to monitor spending and progress on prioritized projects.In 2009, the Peruvian government reformed the original Participatory Budget Law to reduce these eight steps to four. 6Law 29298, paraphrased below, outlines the phases for the regional level; these phases also take place at the district and provincial levels: 1) Preparation, or identifying, registering, and training participating agents.2) "Concertation" 7 : During this phase the participating agents meet to discuss the region's development plan 8 and prioritize the "themes" of projects that should be funded in the new budget.This discussion should be based on the development plan.The technical team then evaluates each proposed project and, based on the agreed upon priorities, recommends the projects that should be funded.3) Coordination among the different levels of government, which consists of meetings between the regional president and the local mayors to make sure that spending is coordinated, sustainable, and has regional impact.4) Formalization of investment projects.This takes place during a regional meeting where all participating agents are given a vote in the final project list.This final list is sent to two regional governmental bodies, the Regional Coordination Council and the Regional Council, for approval.Who is a participating agent?In other words, who can participate in this process at the subnational level?Participating agents are defined as representatives from civil society organizations, members of the Regional or Local Coordination Council, and government officials.This is an important distinction from many PB experiences, where only individuals or representatives from neighborhood organizations are invited.Civil Society Organizations (CSOs) should register for the process in advance, assuming that they meet certain criteria.The national law allows each regional, provincial, or district government to determine the registration process for CSOs and codify it in a local ordinance.As such, registration criteria vary around the country.Generally, it is common to mandate that an agent representing a CSO should have legal status9 and work in the entire region/province/district, not merely representing one neighborhood or city.Some governments are more flexible about the criteria in order to allow more informal groups to participate.For example, one government official in Lambayeque, a region in the north of Peru, told me in an interview that he tries to be flexible about the issue of legal status since it can be costly to register.In my analysis of registration lists, available online, I have also noticed that individuals occasionally register.However, neither of these examples is the norm in Peru.
While both participating agents and government officials can submit proposals for investment projects in any city, province, or region, one goal of this process is to encourage local CSOs to submit project proposals.Slowly, this is starting to happen.For example, in Junín, a region in the central highlands, women's groups organized to propose a job creation project.They proposed to produce yogurt to sell to the "glass of milk" program, a program that provides a nutritious beverage to children in public schools each morning.Other women's organizations have proposed shelters for victims of domestic violence.However, the majority of the projects tend to fund infrastructure like roads, bridges, dams, and irrigation systems.
The technical team-with members from the government and civil society who are invited to participate by the government officials-plays several important roles in the PB process, especially since the national government revised the process in 2010.First, the team receives the initial project proposals and reviews them to ensure that they meet the MEF's criteria (discussed in more detail below).Second, the technical team develops a preliminary list of projects to recommend for approval.This list of projects is then debated and approved in a larger PB meeting with all participating agents.Third, in some regions and municipalities, the technical team might also serve as a resource as participating agents debate the projects.This is not a formal role; however, when observing PB debates in one region, I witnessed the technical team helping participating agents understand project proposals when there were questions.
In practice, the role of the technical team is extremely varied.There are over 2,000 PB meetings taking place annually in Peru, which translates to over 2,000 technical teams.Some technical teams are very small (with four or five members) and some are large (with twenty-five members).Some are made up of mostly governmental officials and some have a mix of government and nongovernmental members.To date, there is no analysis about how reliable or legitimate these teams are, mostly because this kind of research is extremely difficult to undertake.
What information or criteria guides the technical team as the members make decisions about project viability?The MEF provides templates for project proposals to any official or CSO that wants to propose a project.The proposal should include information about how much the project will cost and who will benefit.In theory, the project must meet the MEF's National Public Investment System's (SNIP) requirements.The MEF installed the SNIP system in 2000 to verify and approve investment projects.To meet SNIP requirements a project profile has to include reliable statistics about its impact, have information about cost-share, and be ready for a feasibility study (if that had not already been done).One problem is that these criteria are hard to meet for many governments in poor and rural areas.
Since 2010 the MEF has also developed criteria about the cost and impact of projects.For the most part, there is no stated rule about how much a project should cost, nor are there fixed cost ceilings for most projects.Each regional or local government has an estimated budget within which to work during the PB discussions.In the 2010 instructions, the MEF does state that regional projects should cost at least 3 million nuevos soles (more than 1 million USD) and benefit at least two provinces and 5% of the population.Provincial projects should cost at least 1.2 million nuevos soles and benefit at least two districts and 5% of the population.Further, investment projects should be linked to the overall development plans.These stipulations are meant to prevent participating agents from approving small projects with limited impact.It is possible, however, to get around these stipulations.
Over time, more and more subnational governments are paying attention to these suggestions.For example, 60.1% of the regional projects that were financed in 2008 cost more than 8 million nuevos soles (MCLCP 2011).In 2009, the percentage of funded projects at that same amount increased slightly to 63.2% (MCLCP 2011).Almost 13% of the projects cost between 3.5 and 8 million nuevos soles in 2008 and, again, that percentage increased to 15.1% in 2009 (MCLCP 2011).10Thus, regions seem to be approving larger projects with greater impact.In theory, it is up to the technical team to think about these issues.
Once a list of projects is approved in the final participatory budget meeting, the subnational executive (the regional president or local mayor) sends the list of projects to his or her Regional (or Local) Council, similar to a legislative body, to approve.The final list is then sent to the Ministry of Economy and Finance, which evaluates technical viability.The MEF sends a final budget to Congress to approve as part of the national budget process.11While civil society participation is an important part of the process, it is only one part of a long and complicated budget process that involves elected regional authorities and central government officials, all of whom can legally change the final list of projects that leaves the final participatory budget meeting.
Participants, both governmental and non-governmental, are aware of several budgetary limitations related to the PB process.First, because only capital investment costs are debated, the funds up for discussion are limited.For example, the World Bank (2010, 8) estimates that in 2007, 36% of subnational budgets (an estimated $393 million US dollars) had been debated in the participatory budget process.Second, participating agents are updated about ongoing projects each year, thus, many realize when the budgets are changed after meetings.While this update is important for the oversight process it can also contribute to "participation fatigue," which is discussed at more length later in this article.
For the most part this process has remained the same since the early 2000s.It is worth noting, however, that the national budget process did undergo a major overhaul in 2007, when Peru implemented a reform called "Result-based Budgeting" (Presupuesto por Resultados, or PpR).12According to the MEF, the PpR is "a public policy strategy that links resource allocation to measureable results" (MEF n.d.).In a mostly top-down process, the national government developed eight results that all parts of the budget should work toward addressing: improved nutrition, prenatal maternal health, access to education, access to basic social services and the market, access to formal identity papers, improved water and sanitation, access to telephone and Internet in rural areas, and access to energy in rural areas (MEF 2009).One current challenge in Peru is to define how the PB process, which is mandatory in all subnational governments and eventually feeds into the national budget, fits into the PpR framework.

Why? Explaining the Top-Down Nature of Peru's PB
What forces led to a top-down, mandated participatory budget process in Peru?In previous work, I explore this question more closely, and it is worth restating that argument briefly here. 13Three factors explain the emergence of this reform.First, wide levels of public disgust with the Fujimori regime meant that drastic measures were both called for and acceptable.After the scandals broke, Peruvians watched their televisions as leaked videos captured Fujimori's top security advisor, Vladimir Montesinos, paying off judges, congressmen and -women, and business leaders to support the regime.Citizens, demanding change, subsequently elected many officials who openly espoused the need for more direct, participatory channels for citizens to engage in and oversee politics.Thus, public opinion supported these leaders' efforts to produce radical means to create more democratic and accountable institutions.
Second, the Ministry of Economy and Finance played a decisive role in this reform.The MEF designed the program-which began as a pilot project in 2002-as a means to control corruption and clientelistic politics.Several officials from the MEF's Office of the Public Budget then presented a national-level proposal that both the Budget and General Accounts Committee and the Congress' plenary passed quickly and with little resistance.
Why would the MEF, a conservative and technical institution, take on this project?According to MEF officials, the institution had become frustrated with the lack of transparency behind funding decisions.As one expert stated to me when interviewed in 2004, "Our goal was to improve public spending at the local level and make sure that public funds were being spent well.The typical mayor was corrupt and did not understand public policy."Further, officials believed that Congress made ad hoc decisions that were not always linked to local needs.Empowering civil society to participate in the process would offer more accountability and oversight.
Third, previous experiences, both domestic and abroad, also informed this reform.During the late 1980s and 1990s, several communities had interesting experiences with participatory planning that were taking place completely independent of the central government.These were emerging in Peru at the same time that participatory governance methods were becoming more prominent around Latin America.For example, in Ilo, a port town in the south of Peru, with the direction of CSOs and local leftist leaders, participatory planning and budgeting took off in the late 1990s. 14In Villa El Salvador, a shantytown outside of Lima, elected leaders from the left began to undertake participatory planning in the early 1980s and began participatory budgeting in 1999. 15Many of the officials who had been involved with these experiences had been elected or appointed to public offices in the post-Fujimori years.
When I interviewed the mayors and civil society actors who had instituted some of these experiments, they mentioned a variety of reasons for undertaking such innovative measures to engage citizens in local planning.Interestingly, several interviewees mentioned having gone to conferences, both in Peru and in other Latin American countries, where different participatory experiments were discussed.One, the participatory budget process in Porto Alegre, stood out, especially to the mayor of Ilo, as particularly worth attempting to replicate.Thus, the Porto Alegre process provided a template of sorts for local (mostly leftist) innovators in Peru to undertake similar processes in the late 1990s.
While informed by the Porto Alegre process, Peruvians designed a fundamentally different process. 16For example, Peruvians decided to invite civil society representatives to vote on subnational investment projects and not individual citizens.Generally, policy makers can design participatory institutions based on one of two models: 1) the "individual model," which opens up spaces for individual citizens to participate; and 2) the "corporate model," in which participants represent civil society organizations (World Bank 2008).In Porto Alegre, individuals are invited to a series of public assemblies and then elect delegates to represent them, mostly based on territorial criteria (e.g.neighborhoods).Peru, on the other hand, opted for the corporate model.
Why not invite ordinary citizens?The design decision to engage civil society, as opposed to an open call to all citizens, was not entirely new to Peruvians.During Peru's very brief experience with decentralization in the late 1980s (cut short by Fujimori's self-coup in 1992), 17 regional assemblies (akin to state legislatures) included elected members from civil society.The regional assembly design, in turn, had been influenced by the corporatist structures that had been implemented in Peru during its military regime (1973-1989) and had gained popularity in other Latin American countries during the 1970s. 18Thus, the decision to invite CSOs to participate was based on an organizational logic that can be traced back to the 1970s.
There is a clear consensus in the literature on PB experiences around the world that design decisions matter, and this particular experience is no exception.Having said that, there are few analyses of design decisions regarding who attends (individuals vs. representatives from organizations) and what kind of organizations attend (sectoral, territorial, or both).One clear outcome related to design, discussed below, is the number of participants attending meetings.We should expect to see fewer people attending meetings when the corporate model is chosen.However, beyond this distinction, we do not yet fully understand the implications of this particular design decision.
Another difference lies in the oversight process.In Porto Alegre, there is a built-in, longer-term oversight mechanism that allows participants to discuss the local budget in a more comprehensive way over the course of the year.Each June, a municipal budget council is elected to meet twice a week about "nearly all aspects of the municipal budget, including the annual budget law (Lei de Diretrizes Orçamentárias) and investment plan, the execution of the current year's budget, tax policy, the application of extrabudgetary funds, and the hiring of new personnel" (Goldfrank 2011, 194).Thus, government officials and participants sit down together regularly to discuss the status of all budget-related issues.
In on the other hand, participants in the final PB meeting elect civil society representatives to form an oversight committee (described above).This committee is tasked with monitoring the implementation of the budget that is passed during that year.This committee has a limited scope and it is the responsibility of civil society actors to convene and attend the meetings.As will be discussed below, the Peruvian oversight committee design has not proved as effective as the Porto Alegre model of accountability and oversight.

Surprising Success
The PB process in Peru has been relatively successful in spite of the fact that it was mandated by national politicians, involves a relatively complicated process, and must take place annually in over two thousand political units around the country.Many watchdog groups and experts have described the PB as quite successful.In my own interviews, experts on decentralization often referred to the PB as "institutionalized" and one that they fully expect to continue.The World Bank (2010, 38-39) writes: "There is a general recognition that the participatory budget is the most valid mechanism in the decentralization framework."Another national watchdog group, Grupo Propuesta Ciudadana19  (2009, 5), writes that "the participatory budget is the most important participatory mechanism that has been established since the return of democracy and decentralization in our country."Although these same documents and interviewees note several problems with the PB (discussed below), these are most often posed as challenges that need to be addressed; few would argue that this new institution should be eradicated.
Beyond expert opinion, what indicators of success exist?In any PB process, success should be evaluated in light of process and outcome.In other words, we should explore the actual meetings and fora that channel this participation as well as the kinds of projects that are approved in the meetings and the impact that these projects have on the region, province, or district in which they are approved.In Peru, due to limitations in data collection, it is much easier to capture process than outcome. 20To the extent possible, however, this section discusses both.

Process
Two process indicators illustrate the success of the PB in Peru: 1) Do regions comply with the law?Do they hold meetings to debate investment projects in a participatory way? 2) Do people attend these meetings?Do civil society actors find that these fora are worth spending their limited time and resources?

Compliance
In a top-down process like Peru's, it is perhaps useful to begin with the most basic indicator of success: Do the different subnational governments undertake the process?Because the PB in Peru did not emerge from regional or local actors, this is an important question.It is easiest to track this indicator at the regional level.
When the process began, not all regions called meetings.According to the MEF's database on participants, four of twenty-five regions did not hold meetings in 2006 and three did not in 2007.However, since 2008 all regions have reported on PB meetings.While important, it is not terribly surprising that all regions are now holding PB meetings in Peru.By law, a region cannot get its annual budget if it does not hold these meetings.

Attendance
A second way to evaluate the success of an experiment like this is to assess if CSO representatives actually attend meetings.Do they consider these fora worthy of their limited time and resources?It is difficult to know exactly how many representatives from civil society have attended meetings around the country since 2004.When I travelled to Peru to research this question in 2010, my interviewees consistently responded, when asked, that "hundreds of thousands" CSOs have sent representatives to PB processes around the country since 2004.
The World Bank, which has done the most recent and most reputable study of the process at all levels of government, estimates that 150,000 participating agents attend PB meetings on an annual basis (2010: 6).They find that from 2005 to 2009 the average number of participants in a given municipal, provincial, or regional process had increased.The average number of participants in a given meeting in 2005 was 41; this grew to 82 in 2009 (World Bank 2010, 42-43).
Due to the sheer number of provinces and districts, as well as the fact that data are often missing for these levels of government in the database, the regional level provides the most reliable data on the number of participants over time.Taken from the government's database, Table One provides disaggregated data on the number of PAs that participated in making twenty-four21 regional budgets from 2008 to 2012 and the percent of those PAs who represented civil society organizations and not government officials.In regions, participation increased from 2592 in 200822 to 3596 in the 2009 process.Participation then declined to 3129 in 2010 and 2818 during the 2011 process.Most recently, it increased again to 3213.Because PAs also include government officials, it is also useful to document how many CSOs participate.Interestingly, while the actual number of agents has fluctuated over the years, the percentage that represents civil society organizations has remained in a similar range (57-59%) since 2009.I argue that these levels of CSO participation are significant, especially in light of the fact that Peruvian civil society has historically been relatively weak, especially outside of Lima, the capital (see López Jiménez 1997).Further, there are several costs of participating that an attendee must take on.These costs include (but are not limited to) transportation, lodging, and meal costs as well as time lost in engaging in other activities.An average meeting or workshop can take several hours or even a day, which takes time away from work and other duties, such as caring for children and/or other family members.CSOs face these costs disproportionately in the regional budget process because the government neither pays for travel nor honoraria for attending meetings (unlike the governmental employees).CSOs from cities outside of the capital, where most meetings are held, face steeper costs due to travel, which can take days.The World Bank has quantified the actual costs for a Peruvian rural dweller who participates in the entire budget process during a given year as approximately $195, or 95% of a monthly salary at minimum wage (World Bank 2010, 47).Finally, individual representatives of an organization may choose not to attend, as Goldfrank (2011,  30) notes, due to "fear of public speaking" or "doubts about the usefulness of their input."In Peru, CSOs are still going to meetings in large numbers in spite of these numerous challenges.

Outcome
What do we know about the outcome of the process?Is the PB improving public expenditures and increasing transparency?In Peru, two indicators allow us to evaluate the outcome of the PB process: 1) The nature of the actual projects that are being approved in the participatory meetings.Are they directed to meeting the basic needs of the poorest people in Peru? 2) Levels of support for the process.Do different actors view this as a worthwhile institution?

Projects
As noted above, data regarding outputs of the PB process, that is, the kind of projects that are prioritized, approved, and actually executed, are very limited.However, the data that do exist show that PAs are prioritizing "social" projects that are "pro-poor" in nature.The World Bank undertook one of the most rigorous attempts to study the impact of the PB process, commissioned by the MEF.
Noting that measuring long-term impact in this case was not yet possible,24 they did find that at least in the short term, participants are approving projects that are directed toward meeting citizens' most basic needs.They determined this by documenting the projects that were directed to improved access to water, electricity, and sanitation.Thus, the World Bank concludes that "participating agents in the participatory budget prioritize investment projects in social infrastructure and those that are oriented to serve the needs of the poorest" (World Bank 2010, 8).
However, there is an important caveat to this finding.While the projects that are approved are directed towards basic needs, the actual projects that are implemented are not.The World Bank documents a disconnect between the approved list and the final projects that can be implemented.This problem is discussed in more detail below.

Levels of support
When thinking about support it is important to note that the intended "audience" for this process is the civil society sector in each city, province, and town.Because individual citizens are not invited, it is not a process that is widely understood among average Peruvians.Many interviewees and experts have commented to me that this is still an elite process in Peru.
How, then, can we measure levels of support?It would be ideal to have access to survey data on a random sample for all Peruvian CSOs or even citizens to determine what percentage of civil society and individual citizens are aware that the PB exists and their levels of support for the process.Unfortunately, no surveys of this nature exist.In lieu of this data, we have to rely on two measures of support: 1) perceptions of PB participants and 2) national politicians' support for the process.
In terms of the first indicator, the Mesa de Concertación para la Lucha Contra la Pobreza (Roundtable for the Fight Against Poverty, or MCLCP) undertook a comprehensive monitoring of the 2007 regional budgets, which included surveying participating agents who attended meetings in 2006.When asked to "grade the process," 50.5 percent of the respondents reported that the process was "regular" (normal), 35.1 percent "good," and 5.4 percent "excellent" (MCLCP 2007, 51). 25 Eighty-two percent of the PAs responded that "civil society's predisposition, or inclination, to participate" was "regular" or "good" (MCLCP 2007, 61).When asked to rank the role of civil society in general in the participatory process, 61.9 percent responded that it was "regular" and 28.2 percent called it "good" (MCLCP 2007, 69).In addition, the majority (69.4 percent) of the respondents said that the process is improving with time (MCLCP 2007, 53).The MCLCP concludes that, although not perfect, the process is becoming more and more "settled" and participants are pleased with the quality of their participation (MCLCP 2007).
Another important sign is that to date no national politician has attempted to reverse or shut down the PB.Since the 2004 process began, three presidents have led the country, each representing slightly different positions on the political spectrum: Alejandro Toledo (2001-2006), Alan Garcia (2006-2011), and the current president Ollanta Humala (2011-present).Although the three presidents disagree on many matters, no president has expressed interest in changing the process.Toledo went on record as supporting the PB.Garcia's support may have been unintentional.He was considered an "anti-participatory" president by most, yet he basically ignored the PB during his administration.The current president has stated his desire to mix representative and participatory democracy as well as to continue decentralization efforts, thus, we should expect national commitment to these institutions to continue under his administration.
While these indicators suggest that Peru's PB has demonstrated unlikely successes, there are still many aspects that need to be improved.The next section discusses some of the most glaring shortcomings, or blockages, that exist.These are preventing Peru's PB process from becoming an institution with real potential for improving democratic governance in this country.

Constraints to Success
Although the PB process has surprised many with its sustained levels of engagement, there are many shortcomings that exist. 26These shortcomings combine to threaten the sustainability of this project.Of the many problems facing the process, this section discusses four.
First, the process itself is often poorly organized and timed, making it difficult for actors to coordinate between levels of government as well as with the central government's budget-making process.As Maria Remy (2011, 43), a social scientist in Peru, points out, to be "truly participatory," the process "takes time."Each level of government-district, province, and region-has to coordinate the process so participants can ideally move from the smaller political unit to the larger.All decisions need to be made between July and September so the MEF can analyze the projects and get them into the final national budget for Congress' approval.This demands a level of political will-getting mayors and regional presidents to coordinate-and a level of information that does not always exist.In 2008, for example, the MEF did not approve the final instructions (02-2008-EF/76.01)until April.Further, the initial budget projections can change mid-year; for example, the MEF may realize that the actual national budget is more or less than originally planned.This challenge has led many to call for multi-year projections and a biannual PB process (Grupo Propuesta Cuidadana 2009; Remy  2011).
Second, also partly due to technical reasons, many of the projects that have been approved in participatory fora are not being executed, or implemented by the local governments.The World Bank's evaluation notes that in 2007, 50% of the approved projects had not been executed.This can be due to political factors, such as when a regional president or city council decides to fund different projects.However, it is most often due to technical limitations of the proposed projects themselves, which often do not meet the strict criteria of the MEF's National Investment Project System.This has led many analysts to point to the danger of "participation fatigue."Grupo Propuesta Cuidadana's (GPC) 2009 report argues that the accumulation of unexecuted projects had made some PAs think that the subnational governments were ignoring the approved lists.This led many participants to "lose confidence in the process" (Grupo Propuesta  Cuidadana 2009, 20).
Third, while most of the process is taking place and running smoothly, there is very little community or participatory oversight of the budget once it is approved.As noted above, the actual committee that should be overseeing the budget is the oversight committee, made up of CSO representatives who are elected at the final PB forum.This committee should meet throughout the year to monitor the budget process and spending.In fact, these committees do not always exist and when they do, they rarely meet.For example, the MCLCP (2007) documents that 7.5% of the districts that they monitored did not have oversight committees in 2007.Further, in the nine regions that Grupo Propuesta Cuidadana  (2009, 17) monitored in 2009, only one oversight committee had met as of October.Meetings do not take place because "the committee has problems convening meetings, members do not have resources to cover the operational costs of oversight, there is a lack of understanding of the functions of the committee, and there are difficulties in getting information and analyzing data about the approval and implementation cycle of investment projects" (Grupo  Propuesta Cuidadana 2009, 21).For this reason, some have called the PB process in Peru "pro-forma." The final and most problematic challenge for the continued success of this process, mentioned briefly earlier, is the fact that it is still mostly an elite process.Not only do few average citizens know about or understand the process, the actors who participate in the budget decisions do not represent the diverse kinds of regional and local CSOs.For example, according to Grupo Propuesta Cuidadana  (2009, 14), "[w]ithin civil society, accredited agents tend to be mostly from social organizations, NGOs, and professional associations."Further, the majority of these representatives live in urban areas, giving rise to an urban bias in the process (Grupo Propuesta Cuidadana 2009).Several kinds of organizations do not tend to participate, such as business organizations and universities (Remy 2011,  44).In my interviews with regional experts, many suggested that these organizations have more direct means of negotiating local resources, therefore they do not find the meetings necessary or useful.
There is also a gender imbalance in the process.Very few women's organizations are represented in regional meetings.According to my analysis of the Peruvian governmental database, only 2% to 3% of the PAs that came to meetings officially represented women's organizations in the meetings. 27urthermore, female participants attend in lower rates than men.When documenting the participation by women in the 2009 regional, provincial, and district levels, the World Bank (2010, 47) finds that "women represent only 32% of the participating agents."Data from the Peruvian government's database on participating agents show that 27% of participants were female in the meetings to make the 2008 budget, 28% for 2009, 29% for 2010, and 30% for 2011.This number declined to 22% for the 2012 budget.Thus, it is clear that the process needs to be more inclusive.New actors are being engaged, which is especially important when we take note that many social organizations are making decisions to improve basic services like water and sanitation.However, it is imperative to improve the diversity of the participants so the many kinds of people who live and work in communities are engaged in the decision-making process through their local CSOs.
Some very small steps can be taken to improve diversity and inclusiveness.Local governments can make registration possible for more informal organizations, such as those that have members but lack the funds to register.Flexible registration criteria will allow more diverse kinds of organizations to participate.Additional steps can be taken to make attending meetings possible for more participating agents.For example, offering childcare would increase the number of women's organizations that are represented.Making sure that meetings are held in handicapped-accessible buildings would increase the number of organizations representing people with disabilities.Scheduling meetings after school hours would allow more youth to participate.Some of the best regional experiences in Peru have made it possible for less formal organizations to attend meetings by developing less stringent participation criteria and taking care to poll potential participants about ideal times before scheduling meetings.

Conclusions
To conclude, this article contributes to our understanding of participatory budgets in two ways.First, it documents Peru's top-down participatory budget process, instituted by the national government as part of a decentralization reform meant to improve democratic governance after years of authoritarian rule.This case has proven to be an unlikely success in many respects.Generally, a process that is designed and set up by national governmental officials faces many challenges in terms of implementation.There is no immediate and obvious group of stakeholders around the country to take ownership of the process.The fact that over two thousand governments around the country are undertaking this process on an annual basis, and that a significant number of civil society actors are attending the meetings, is a sign of success.
At the same time, Peruvians have a long way to go if they want to truly change the nature of subnational politics through this kind of reform.The process itself, as outlined in official instructions emitted by the MEF, is cumbersome and problematic.Due to this weakness, many projects that were approved by participants are never actually executed.Furthermore, many more kinds of organizations, especially those that represent the most disadvantaged groups (e.g., indigenous, women, and other minorities) need to become engaged in the process.
Second, the Peru case provides insight into how we might measure success in this particular kind of project.Especially in the case of Peru, where the design targets organizations and not individual citizens, we cannot expect the massive amount of participation that other experiences, especially in Brazil, have achieved.It may be more prudent to document the number and kinds of organizations that are engaged in the process.Second, several extraneous factors affect the final implementation of projects-for example, political interests of regional and local officials, technical viability of the actual projects, and the amount of the final national budget, which may change mid-year.It is therefore hard to measure the actual outcomes of the participatory process on citizen wellbeing.It is only possible to compare the projects that have been approved in participatory fora and the actual projects that have been executed by governments.
In sum, it is important to continue to document new forms of PB experiences that are emerging around the world in order to best evaluate the possibilities and limitations of these experiments.The many forms and designs suggest that there are numerous ways to craft efforts to improve transparency, accountability, and governance more generally.Our continued understanding of these forms will help improve their implementation as well as helping us decide if they are, in fact, the best tools to attack the myriad problems facing countries at the local level around the world.