The hegemonic processes of globalization are bringing about the intensification of social exclusion and marginalization of large bodies of population all over the world. Such processes are being met with resistances, grassroots initiatives, community innovations, and popular movements that try to counteract social exclusion, opening up spaces for democratic participation, for community building, for alternatives to dominant forms of development and knowledge-in sum, for social exclusion. They are, in general, very little known because they do not speak the language of hegemonic globalization and often times present themselves as promoting the case against globalization. They are of very different kinds or else their diversity has become more apparent after the collapse of the models of grand-scale social transformation that the old Communist Manifesto, celebrating in 1998 its 150th anniversary, portrayed so well (revolution, socialism, communism). Today's forms of counterhegemonic globalization occur in rural as well as urban settings, involve common citizens or especially vulnerable groups, and deal with issues as diverse as land rights, urban infrastructure, drinking water, labor rights, sexual equality, self-determination, biodiversity, environment, community justice, and so on. Finally, they may entertain a wide variety of relations with the state, from no relation at all to complementarity or confrontation.

In this paper I analyze one urban experiment aimed at redistributing city resources in favor of the more vulnerable social groups by means of participatory democracy: the participatory budget adopted in the city of Porto Alegre since 1989. In the first part, I briefly describe the recent history of Porto Alegre and its government in the context of the Brazilian political system and provide some basic information about the city. In the second part, I describe the main features of the institutions and processes of the participatory budget: institutions and processes of participation, as well as criteria and methodology for the distribution of resources. In the third part, I analyze the evolution of this institutional innovation since its creation until today. Finally, in the fourth part, I analyze the participatory budgeting process along the following vectors: redistributive efficiency, accountability and quality of representation in a participatory democracy, autonomy of the participatory budgeting vis-a-vis the executive government on the city, from technobureaucracy to technodemocracy, dual power and competing legitimacies, and the relations between the participatory budget and the legislative body vested with the formal legal prerogative of budget approval.

I. URBAN POLITICS: THE CASE OF PORTO ALEGRE

Brazil is a society with a long tradition of authoritarian politics. The predominance of an oligarchic, patrimonialist, and bureaucratic model of domination has resulted in a state formation, a political system, and a culture characterized by the political and social marginalization of the popular classes, or their integration by means of populism and clientelism; the restriction of the public sphere and its privatization by the patrimonialist elites; and the "artificiality" of the democratic game and liberal ideology, resulting in a huge discrepancy between the "legal country" and the "real country." Brazilian society and politics are, in sum, characterized by the predominance of the state over the civil society and by huge obstacles against the construction of citizenship, the exercise of rights, and popular autonomous participation. Brazil is also a society characterized by outrageous social inequalities, which have in fact increased tremendously in the past fifteen years because of the crisis of the developmental state, the deregulation of the economy, and the dismantling of the already utterly deficient welfare state. According to a recent World Bank report, Brazil is one of the most unjust societies in the world: "the level of poverty in Brazil is well above the norm for a middleincome country. On the other hand, it would be possible to eliminate poverty in Brazil (by giving every poor person enough money to bring them up to the poverty line) for a cost of less than one percent of the country's gross domestic product.... The reason why the recent record on poverty has been so dismal is deficient public policies."'1

The crisis of the developmental state coincided with the democratic transition in the late seventies. The political debate at the time put the democratization of Brazilian political life and the actual construction of citizenship at the very center of the national political agenda. The concerns in this regard surfaced in the emphasis on rights of citizenship, political decentralization, and strengthening of local power in the debates that led to the 1988 constitution. This new political context created the conditions for political forces on the left, whether out from clandestinity or meanwhile organized, to set up innovative experiments of popular participation in municipal government. This political opportunity was facilitated by the fact that the political forces in question were closely related to the popular movements that in the sixties and seventies had struggled locally, both in the cities and in the countryside-and in a doubly hostile context of technobureaucratic military dictatorship and clientelistic patrimonialism-for the establishment and recognition of collective subjects among the subaltern classes.

Among such political forces, the Partido dos Trabalhadores (Workers Party, henceforth PT) is to be singled out. The Workers Party was founded in the early eighties out of the labor movement, which was particularly strong in the state of Sao Paulo and one of the most important forces in the struggle against the military dictatorship. The electoral gains of PT have been dazzling. It has twice run for the presidency of the republic and with such results that its candidate, the charismatic Lula, became the major opposition leader. In the late eighties, the PT, in coalition with other leftist political forces, won the local elections in several important cities-such as Sao Paulo, Porto Alegre, Santos, Belo Horizonte, Campinas, Vit*ria, Goiania-and introduced in all of them institutional innovations encouraging popular participation in municipal government. Of all these experiments and innovations, those implemented in Porto Alegre have been by far the most successful and with wide recognition both inside and outside Brazil.

The Porto Alegre democratic experiment is one of the best known worldwide, acclaimed for both the efficient and the highly democratic management of urban resources it has made possible.2 The "popular administration" of Porto Alegre was selected by the United Nations as one of the forty urban innovations worldwide to be presented at the Second Conference on Human Settlements (Habitat II), which was held in Istanbul in June 1996.

In the past seven years, Porto Alegre has staged several international conferences on democratic urban management and, together with Montevideo (where a similar local government innovation has been implemented), is leading a movement toward the introduction of participatory budgeting institutions in the "Mercocities," the cities integrating the regional economic pact, Mercosul (Brazil, Argentina, Uruguay, and Paraguay).

In Brazil, there have been many manifestations of the success of Porto Alegre,3 the most significant being the electoral gains of the PT in the nineties and the public acceptance of its municipal government. In the first election, won in 1988 with a coalition of left parties-the Popular Front-the PT carried 34.3% of the vote. In the second election, in 1992, the PT and the Popular Front carried 40.8% and in the last election, in 1996, 56%. Another manifestation of the success of the PT government of Porto Alegre is the fact that Exame, an influential business journal, nominated Porto Alegre for the fourth consecutive time as the Brazilian city with the best quality of life, on the basis of the following indicators: literacy, enrollment in elementary and secondary education, quality of higher and postgraduate education, per capita consumption, employment, child mortality, life expectancy, number of hospital beds, housing, sewage, airports, highways, crime rate, restaurants, and climate. Lastly, in opinion surveys at the end of the second municipal term (1996), the municipal government was assessed as excellent or good by 65% to 70% of the inquired; if an evaluation of "medium plus" is considered positive, the government has actually earned approval by 85%. Thus, it is hardly surprising that the mayor of Porto Alegre for the last term, who had already been vice-mayor in the first term (1988-1992), was for a while insistently indicated as the PT candidate for the presidential election (October 1998).

What is the secret of such a success?

When, in January 1989, the PT took over the administration of Porto Alegre, a new modality of municipal government was installed, known as "popular administration." It was based on an institutional innovation aimed at guaranteeing popular participation in preparing and carrying out the municipal budget, hence in the distribution of resources and definition of investment priorities. This new measure, which became known as "participatory budgeting," is the key to the success of the PT municipal government. In this paper, I shall begin by describing how the participatory budgeting works, with special emphasis on its evolution since it was first set up until today. I shall then attempt an evaluation of its impact on the redistribution of municipal resources and on the political culture and system of the city, namely by analyzing both the tensions between representative and participatory democracy and the reach of participatory budgeting into other areas of urban government. Lastly, I shall try to define the contribution of participatory budgeting as institutional mediation for the reinvention of democratic theory, while questioning the potentialities and limits of its universalization as an organizing principle of democratic and redistributive municipal government.

The City of Porto Alegre"

With a population of 1.3 million inhabitants and 489 km^sup 2^, Porto Alegre is of major economic importance in Rio Grande do Sul with an estimated gross national product (GNP) of US$6.7 billion (1994). As an individual municipality, it is the largest industrial city, producing 12.4% of the state's industrial gross product and being responsible for nearly one-third of the income produced in the service sector. Its total population corresponds to 18% of Rio Grande do Sul (all figures from 1992). In national terms, its influence is mainly political, for many prestigious local politicians have served in the national government in this century. The most significant of them-"perhaps the most important Brazilian politician of all times"5-was Getulio Vargas, who served two mandates as president (1930-1945 and 1951-1954) and led in the first mandate the authoritariancorporatist regime that presided over the transformation of Brazil into a modern industrial country.

Like other Brazilian capitals, in the past decades Porto Alegre has experienced an accelerated process of urbanization. Its population doubled in twenty years (1960 to 1980), but in the past decade, the total population grew only 12% when new industrial centers in the state attracted migrants from the capital. Between 1970 and 1980, the participation of Porto Alegre industry in the total industrial production of the state of Rio Grande do Sul declined from 26% to 18%.6 It is a city that has been traditionally organized around service and government sectors. In 1949,73% of the city's income came from the service sector, and in 1980,78%. The relative deindustrialization in the early eighties did not affect the centrality and hegemony of Porto Alegre as a regional metropolis. Moreover, "the industrial sector of Rio Grande do Sul still persists in maintaining its nucleus and in developing its 'logic' in Porto Alegre, a tendency based on the importance of the local industry as a share of the state's industrial production; in the participation of the city in the state's internal income; in the production; in the regional links with the commercial and service sectors of the city; in the evidence that a reasonable part of the state's internal income is spent in the capital . . . without mentioning the weight of the state and local public administration established in its territory."7

The state of Rio Grande do Sul presents some of the best social indicators of the country. According to Navarro, citing official statistics,8 among the fifty best Brazilian cities in educational performance (eradication or lower levels of illiteracy), thirty-two are in the state. Other social indicators show that life expectancy in the state reaches seventy years for men and seventy-five years for women, the highest if compared to other Brazilian states, and infant mortality rates fell, in the past two decades, from 52.6 to 16.5 deaths per one thousand children of less than one year of age. In the city of Porto Alegre, the latter was reduced from 37.2 deaths, in 1980, to 13.8 deaths, the best performance among all Brazilian capitals. Nevertheless, there are also contrasting negative indicators, such as deep social inequalities (like the rest of Brazil), the housing problem, and unemployment. Onethird of its population lives in slums, and a recent report indicates that the total population in these areas more than doubled between 1981 and 1990. In contrast, according to former Mayor Tarso Genro, some fifteen families own the urban land available for development.9

Porto Alegre is a city of ample democratic traditions, a strong, highly organized civil society. The military dictatorship met with fierce political resistance in Rio Grande do Sul, especially in Porto Alegre. For example, because of the pressure exerted by the democratic opposition against the repressive institutions of the dictatorship, political prisoners could not be "safely imprisoned" in the city and were often sent outside the city, usually to Sao Paulo. The opposition was led by intellectuals, labor unions, and the only legalized opposition party, Movimento Democratico Brasileiro (Brazilian Democratic Movement, henceforth MDB). The MDB attracted all the clandestine organizations-whether socialist, communist, or revolutionary-Christian-opposed to the military dictatorship. Since the political situation rendered inviable almost all political struggle at the national (macropolitical) level, the above-mentioned organizations focused their activity on strengthening the unions and such community movements as neighborhood and street associations, soccer clubs, cooperations, mothers clubs, cultural groups, and so on. These movements and organizations were of a general nature now concerned with specific demands, such as the struggle for bus lines, the struggle for sewage or street pavement, the struggle for housing or health centers, and so on. A powerful, diversified popular movement thus emerged, which in the early eighties became deeply involved in local government.10

In the first half of the decade the grassroots movements, however, highly heterogenous both in political and organizational terms, gained a new political clout in local politics. In 1983, the UAMPA (Federation of Neighborhood Associations of Porto Alegre) was founded and in 1985 held its first congress. Besides "specific demands" on housing, education, health, nutrition, human rights, and unemployment, the congress called for the "effective democratization of political structures at the federal, state and city level."11 In the first democratic municipal elections, in 1985, the PDT (Democratic Labor Party), with a long tradition in the state, won easily the elections with 42.7% of the total votes. The PT, still struggling to expand its influence among the popular and labor movements, received 113%.12 Heir to a pro-labor populist tradition, the newly elected mayor decreed the establishment of "popular councils" in the city but in real terms exercised the municipal power in the old clientelistic, paternalistic way, frustrating the democratic expectations and failing most of the electoral promises.

In 1988, the PT started its amazing political success. Without precedent in the city, already in 1992 and again in 1996 the party in government could elect its successor: Tarso Genro, vice-mayor in the first PT mandate, became mayor in the second, and Raul Pont, vice-mayor in the second mandate, became mayor in the third.

II. PARTICIPATORY BUDGETING IN PORTO ALEGRE

In the current, New Republic period of Brazilian political system, the municipal power lies on two separately elected bodies: the mayoralty (Prefeitura), the executive body, and the Chamber of Deputies (Camara de Vereadores), the legislative body. According to the 1988 constitution, the competence to approve the budget is vested on the Camara de Vereadores. Since 1989, the Workers Party and the Popular Front have controlled the Prefeitura but do not have the majority in the Camara de Vereadores.

In Brazil, the public budget includes three levels: federal, state, and municipal. Municipalities have relative autonomy in determining revenue and expenditure. Revenue is either local (taxes and tariffs of various kinds) or the result of federal or state transfers.13 Expenditure is classified in three large groups: (a) personnel, (b) public services, and (c) investment in works and equipment. The relative autonomy of municipalities occurs mainly in the third group of expenditure. Since the budget does not have to identify the works and services to be carried out-the establishment of expenditure ceiling sufficing-the executive has ample leeway for budgetary implementation. The budget must, however, be approved by the legislative body.

The participatory budget promoted by the Prefeitura of Porto Alegre is a form of public government that tries to break away from the authoritarian and patrimonialist tradition of public policies, resorting to the direct participation of the population in the different phases of budget preparation and implementation, with special concern for the definition of priorities for the distribution of investment resources. The participatory budget and its institutional framework have no formal legal recognition. Such legal recognition could only be provided by the Camara de Vereadores, albeit within the limits of federal and state legislation. As we will see below, the issue of the legalization of the participatory budget is a major topic in an ongoing conflict between the executive and the legislative in Porto Alegre politics. As things stand now, since the definition and approval of the budget are a legal prerogative of the Camara de Vereadores, the Prefeitura, in strict legal terms, limits itself to submitting to the Camara a budget proposal that the Camara is free to approve, change, or defeat. In political terms, however, because the executive's proposal is sanctioned by the participatory budgeting institutions and thus by the citizens and community organizations and associations that participate in them, the executive's proposal becomes a fait accompli for the legislative body in view of the political risks for the deputies in voting against the "will of the citizens and the communities." The majority of the chamber thus claim that by institutionalizing the participatory budgeting without involving the legislative body, the executive has in real terms emptied out the latter's jurisdiction over budgetary matters-hence the political conflict that will be dealt with in greater detail below.

Institutions of Participation 14

The participatory budgeting (henceforth PB) is a structure and a process of community participation based on three major principles and on a set of institutions that function as mechanisms or channels of sustained popular participation in the decision-making process of the municipal government. The three principles are

(a) all citizens are entitled to participate, community organizations having no special status or prerogative in this regard;

(b) participation is governed by a combination of direct and representative democracy rules and takes place through regularly functioning institutions whose internal rules are decided upon by the participants;

(c) investment resources are allocated according to an objective method based on a combination of "general criteria"-substantive criteria established by the participatory institutions to define priorities-and "technical criteria"-criteria of technical or economic viability as defined by the executive and federal, state, or city legal norms-that are up to the executive to implement.

The basic institutional setup of the PB consists of three kinds of institutions. The first kind of institutions consists of the administrative units of the municipal executive charged with managing the budgetary debate with the citizens: Gabinete de Planejamento (Planning Office, henceforth GAPLAN), Coordena*ao de Relacoes com as Comunidades (Coordination of Relations with the Communities, henceforth CRC), F6rum das Assessorias de Planejamento (Forum of Advisors for Planning, henceforth ASSEPLAS), Forum das Assessorias Comunitarias (Forum of Community Advisors, henceforth FASCOM), Coordenadores Regionais do Orqamento Participativo (Regional Coordinators of the Participatory Budgeting, henceforth CROPs), and Coordenadores tematicos (Thematic Coordinators, henceforth CTs). Of this set of institutions, the two most important ones are the CRC and the GAPLAN. The CRC, both directly and through its regional or thematic coordinators (CROPs and CTs), is a mediating agency linking the municipal government with the community leaders and their associations. It has also a central role in coordinating the assemblies and the meetings of the COP (see discussion below). The GAPLAN, which shares with the CRC the coordination functions, is in charge of translating the citizens' demands into technically and economically viable municipal action by submitting them both to the general and the technical criteria.

The second kind of institutions are the community organizations, with autonomy vis-a-vis the municipal government, constituted mainly by regionally based organizations, which mediate between citizen participation and choice of priorities for city regions. Since they are autonomous structures, hence depending on the organizing potential of each region, these popular organizations do not occur in every region concerning the PB. They bear different kinds of organization and participation according to the local traditions of the regions. They are the Conselhos Populares (Popular Councils), Unioes de Vilas (Township Unions), and Articulaq*es Regionais (Region Articulations).

The third kind of institutions is designed to establish a permanent mediation and interaction between the two first kinds. They are regularly functioning institutions of community participation: Conselho do Plano do Governo e Orgamento (Council of the Government Plan and Budget), also known as Conselho do Orgamento Participativo (Participatory Budgeting Council, henceforth COP), Assembleias Plenirias Regionais (Regional Plenary Assemblies), Forum Regional do Orgamento (Budgeting Regional Forum), Assembleias Plen,rias Tematicas (Thematic Plenary Assemblies), and F6rum Temitico do Orgamento (Budgeting Thematic Forum).

The Participatory Process

The main goal of the PB is to encourage a dynamics and establish a sustained mechanism of joint management of public resources through shared decisions on the allocation of budgetary funds and of government accountability concerning the effective implementation of such decisions.

In a brief summary, the PB centers on the regional and thematic plenary assemblies, the Fora of Delegates, and the Council of the PB (COP). There are two rounds of plenary assemblies in each of the sixteen regions and on each of the five thematic areas. Between the two rounds there are preparatory meetings in the microregions and on the thematic areas. The assemblies and the meetings have a triple goal: to define and rank regional or thematic demands and priorities, to elect the delegates to the Fora of Delegates and the councillors of the COP, and to evaluate the executive's performance. The delegates function as intermediaries between the COP and the citizens, individually, or as participants in community or thematic organizations. They also supervise the implementation of the budget. The councillors define the general criteria that preside over the ranking of demands and the allocation of funds and vote on the Investment Plan proposal presented by the executive. Next I will describe in greater detail how the PB works (see Figure 1 ).

The city is divided into sixteen regions and five areas for thematic discussion of the budget. The thematic areas (or themes) for the discussion of the municipal budgeting are more recent (1994). These are (1) Transportation and Circulation; (2) Education, Leisure, and Culture; (3) Health and Social Welfare; (4) Economic Development and Taxation; and (5) City Organization and Urban Development.

Two rounds (rodadas) of regional and thematic assemblies are held annually. They are coordinated by members of the municipal government (CRC, CROP, and GAPLAN) as well as by PB delegates and councillors.

Before the annual assemblies, there are preparatory meetings of the citizens, which ordinarily take place during the month of March, without the interference of the municipality. The aim of these preparatory meetings is to collect the demands and claims of individual citizens, grassroots movements, and community institutions concerning regional or thematic issues; they also initiate community mobilization to select regional delegates. These meetings are crucial to ventilate community demands and to discuss their relative priority. These meetings are convened and chaired by the popular councils or by the community leaderships and are at times very conflictual, since the different political orientations of the community organizations surface in the identification and formulation of demands and tend to impregnate the whole debate.

Figure 1.

Enlarge 200% Enlarge 400%

The first rodada of assemblies, held in March and April, has the following objectives: rendering of accounts, by the executive, of the Investment Plan of the previous year and presentation of the plan approved for the current budget; evaluation, by the citizens (by region or themes) and the executive, of the Investment Plan of the previous year; and first partial election of the delegates to the Fora of Delegates (regional and thematic). The remaining regional or thematic delegates will be elected during the next step of the process. The regional assemblies are open to the public, but only the registered inhabitants of the region have the right to vote.

The evolution of the criterion to determine the number of delegates to the regional and thematic fora bears witness to the increasing involvement of the citizens in the PB. Initially, the criterion was 1 delegate for every 5 people attending the assembly; in the early nineties, it changed to 1 delegate for every 10 people and later on to 1 delegate for 20 people in force until 1996. The most recent criterion, adopted in the COP meeting of April 4, 1997, which approved the new rules, is more complex, comprising different ratios according to the size of attendance: up to 100 people attending-1 delegate for every 10 people, from 101 to 250-1 for every 20, from 851 to 1,000-1 for every 70, and more than 1,000-1 for every 80.

The delegates elected in the first round of plenary assemblies and then in the "intermediary meetings" are in general selected by the leaders of the associations present at the meetings, and thus there is not much chance for a citizen not integrated in a collective structure to be elected delegate (more on this below).

Between the first and the second rodada of the assemblies, March through June, the so-called intermediate preparatory meetings take place. They are organized by the community or thematic organizations and associations though now "coached" by the regional or thematic CROP and other representatives of the executive. At such meetings, the demands approved by each association or organization (e.g., neighborhood associations, mothers clubs, sports or cultural centers, housing cooperatives, unions, nongovernmental organizations, etc.) are ranked by the participants according to priorities and general criteria. The better organized regions have an internal microregionalization for their choice of priorities. Later, the resulting priority lists will be fought for at meetings involving the whole region or at thematic plenaries. In the intermediate meetings, there is much discussion and voting, but the real negotiations leading to proposals to be voted for tend to take place behind the scenes at informal meetings of the community leaders. The levels of conflictuality depend on the level of community organization and on the level of political polarization among the leaderships.

At these intermediate meetings, each region or theme hierarchizes the sectorial priorities. Until 1997, the regions hierarchized four priorities among the following eight sectors or themes: sewage, housing, pavement, education, social assistance, health, transportation and circulation, and city organization. In 1997, the COP introduced some changes in this regard to be effective in the preparation of the 1998 budget. The regions now hierarchize five priorities among twelve sectors or themes. To the previous eight sectors or themes were added leisure areas, sports and leisure, economic development, and culture. As we will see later, these changes reflect the discussions in the COP in recent years in which the majority of councillors had urged the expansion of themes covered by the PB. Each sector or theme is divided up in subthemes. For instance, housing includes land legalization, relocation, urbanization, and housing construction.

The elected priorities are given grades according to their ranking: first priority gets grade 5; fifth priority gets grade 1. Likewise, the specific works proposed by the citizens in every theme or sector are hierarchized as well (e.g., pavement: first priority, street A; second priority, street B, etc.). Sectorial priorities and hierarchy of works in every sector are forwarded to the executive.'5 On the basis of these priorities and hierarchies, adding up the grades of the different priorities in all the regions, the executive establishes the three first priorities of the budget in preparation. For instance, for the 1997 budget, the three priorities were housing (44 points), pavement (42 points), and sewage (30 points). At this time, also, the remaining delegates of each region or theme to the Forum of Delegates are chosen.

The second rodada of regional and thematic assemblies held in June and July is coordinated and chaired by representatives of the executive in conjunction with the popular organizations of the region or theme. In these assemblies, two effective councillors and two substitutes in every region and theme are elected for the COP. The councillors are elected for a one-year mandate and can only be reelected once. Their mandate can be revoked by the regional or thematic Forum of Delegates in a meeting especially called for that purpose and announced with an advance notice of two weeks. Once the quorum has been established (50% + 1 of the delegates), the mandate can be revoked by a two-thirds majority vote.

The institutional organs of community participation are then constituted: the Fora of Delegates (sixteen regional and five thematic ones) and the COP. The Fora of Delegates are collegiate organs with consulting, controlling, and mobilizing functions. The fora meet once a month, and the two major tasks of the delegates are to supervise the works and to act as intermediaries between the COP and the regions or thematic areas. As we will see below, the information flows are not without problems.

The COP is the main participatory institution. There the elected citizens get acquainted with the municipal finances, discuss and establish the general criteria for resource allocation, and defend the priorities of regions and themes. At the council sessions, the institutional mediation between citizens and community organizations on one side and municipal government on the other concerning budgetary decisions is conducted at the most concrete and intense level. Once inaugurated in July/August, the council meets once a week on a set day, usually from 6 to 8 p.m. The tasks of the COP are carried out in two phases. In August and September, the COP discusses the revenue and expenditure items (works not specified) and the general criteria for resource allocation until the budget proposal is drafted and sent by the executive to the legislative on September 30. From September until December, the COP prepares the Investment Plan, which includes a detailed list of the works and activities prioritized by the council and thus the specific allocation of resources programmed for every region and thematic area. The debate concerning investments is limited by the estimated general revenue and expenditure with personnel and other expenses estimated by the executive, including fixed expenditures enforced by federal legislation, such as the percentages constitutionally ascribed to education and health.

At the same time, the COP follows the debates on the budget proposal in the Camara de Vereadores and puts pressure on the legislators by meeting with individual members of the chamber, mobilizing the communities and thematic areas to attend the debates or to organize rallies outside the building.

During the whole process, the executive participates in the definition of investments through its Planning Office (GAPLAN), and also through the municipal secretariats attending the council meetings, by proposing works and projects of general interest and multiregional ambit or even works deemed necessary, upon technical evaluation, for a given region of the city. Thus, the Investment Plan includes works and activities suggested by the regions and thematic areas as well as works and activities involving several regions or even "the whole city." In the last phase of the procedure, the Investment Plan approved is published as a booklet and becomes the basic document to refer to by the community delegates in their supervising capacity and by the executive when rendering account before the organs of the participatory budget.

Throughout the participatory budgeting process, the executive plays a decisive role, and this is particularly evident in the COP meetings. Through the CRC and the GAPLAN, the executive coordinates the meetings and sets the agenda. The meetings start with information given both by the government representatives and by the councillors. The themes of the agenda are then introduced by the permanent government representatives or by the representatives of the different municipal secretariats in charge of the theme under discussion. In a disciplined way, the councillors, in three-minute interventions, raise issues and ask questions. After a number of interventions, the executive representatives answer the questions and give the information requested. A second set of interventions by the councillors is followed by the representatives' answers and so on. Both regional and thematic delegates may also intervene, but they cannot vote.'6

At times there are more direct and intense debates, but the most conflictual and even tumultuous debates occur in those rare "special" meetings of the COP that are not coordinated by the executive but by the councillors themselves. The political and even personal cleavages surface then more openly. In recent times, the executive's coordination of the meetings has been questioned by some councillors in the name of the autonomy of the COP (more on this below). According to the rules, the coordination belongs to the Parity Commission, so designated because it is composed of an equal number of councillors and government representatives, four each. But in reality the government representatives do most of the coordination, if for no other reason than their privileged access to relevant information. In any case, probably to accommodate increasing concerns about the limited autonomy of the COP expressed by the councillors, the most recent version of the rules determines that the government representatives and the councillors of the Parity Commission will alternate in chairing the meetings.

The Distribution of Investment Resources: Methodology and Criteria of Decision Making

The distribution of investment resources follows a method of participatory planning that begins when priorities are indicated at the regional and thematic plenaries and at the intermediate meetings and reaches its climax when the Participatory Budgeting Council approves the Investment Plan with detailed works and activities specified for each investment sector, region, and the whole city. The regions define and hierarchize the specific demands within each priority. Once the priorities of the different regions are established, the distribution of investments is carried out according to the general criteria defined by the COP and the technical criteria defined by the executive. The general criteria are lack of urban infrastructures and/or services, total population of the region, and priority given by the region to a specific sector or theme. To each criterion is ascribed a weight in a scale from 1 to 3 in direct proportion to the importance attributed to it by the COP. In the 1998 budget, both the lack of services and the priority of the region carry the weight 4 while the size of the population carries the weight 2. On the other hand, in its application to a given region, each criterion is given a grade. For instance, concerning the lack of services or infrastructure, the greater the need, the higher the grade. Let us take the example of pavement. A given region has a lack or need of 20% if 20% of its streets need pavement. A region with up to 20.99% need gets in this criterion grade 1, while a region with a need of 80% or more gets grade 5.

The grade received by each region in each criterion is then multiplied by the general criterion's weight. The sum of the partial points (grades x weight) gives the total grade of the region in that specific sectorial demand. This total grade determines the percentage of the investment resources that will be allocated to the region in that sector.

An example may illustrate how the general criteria are translated into a quantified allocation of resources. In 1997, the relative priority given by the sixteen regions to the street pavement determined the inclusion in the Investment Plan of a global expenditure item for street pavement corresponding to 20 kilometers (20,000 meters) of streets to be paved. The distribution of this amount by the different regions was the result of the application of the criteria, their weight, and the grade of the region in each one of them. Let us analyze the case of two contrasting regions: the Extremo Sul, a region with 80.21% need of pavement, and the Centro, with 0.14%. Concerning the need criterion, which carried a general weight of 3, the Extremo Sul had the highest grade (4)17 and accordingly got 12 points (3 x 4), while Centro, with the lowest grade (1), got 3 points (3 x 1). Concerning the criterion of total population, which carried the general weight of 2, the Extremo Sul, with a population of 20,647 inhabitants, had the lowest grade (1) and hence got 2 points (2 x 1), while the Centro, with a much bigger population (293,193 inhabitants), had the highest grade (4) and hence got 8 points (2 x 4). Finally, concerning the criterion of the priority given by the region, which carried a general weight of 3, the Extremo Sul gave the highest priority to pavement and, accordingly, had the highest grade (4) and thus got 12 points (3 x 4), while the Centro gave a very low priority to pavement and thus had the lowest grade (0) and consequently no points (3 x 0). As a result, the total sum of points of the Extremo Sul in the item of street pavement was 26 points (12 + 2 + 12), while the Centro's total sum was 11 points (3 + 8 + 0). Since the global number of points for all regions was 262 points, the Extremo Sul received 9.9% of the investment, that is, 1,985 meters of street pavement, while the Centro received only 4.2% of the investment or 840 meters of pavement.

III. THE EVOLUTION OF PARTICIPATORY BUDGETING: ON LEARNING PARTICIPATORY DEMOCRACY

The structure and development of the PB have undergone important transformations since it was first initiated in 1989. This evolution illustrates the internal dynamics of the PB and, above all, the institutional learning by both state and civil society.

When, in 1989, the PT took over the government of Porto Alegre, the party leadership was involved in an intense internal debate, which essentially may be summarized as follows: is the PT government a government for the workers, or is it a leftist government for the whole city though with a special commitment to the popular classes? Initially, the position that the PT should govern for the workers alone prevailed. Such a position was deeply rooted in the political culture of the PT, closely linked to the political theories of the popular urban movements in the seventies centered on the core concepts of dual power and popular councils derived from the Bolshevik Revolution. Under these premises, considering that the state is always particularist and exists only to fulfill the interests of the bourgeoisie, it should likewise be the task of the PT to carry out a particularist government only now favoring the interests of the workers. Since such government would be exercised in the institutional context of the bourgeois state, its major objective would be to provoke confrontation and bring about crisis so as to unveil the classist nature of the state.la

Such a political stance held for the first two years. The aim was to hand over power to the popular councils derived from community organizations so that they could be the ones to make decisions about municipal policy, especially the budget. According to Tarso Genro, who was then the vice-mayor and became mayor in the following term ( 1992-1996), by the end of the first year it was already obvious that such a political and administrative strategy reflected a "romantic conception" of popular participation and was destined to fail for three main reasons. First, neither the party leaders heading the executive nor the community leaders had any experience in promoting institutionalized participation. Both had been socialized in a political culture of confrontation and were not ready to go beyond protest and confrontation. 9 Such a context did not allow for the creation of spaces for negotiation capable of articulating and making compatible all the different claims and demands from different regions, let alone establish a political contract and take part in the institutional mediations necessary to make it effective. Second, it was soon quite evident that the community leaders were not only socialized in a political culture of confrontation but also in a political culture of clientelism, on the basis of which they channeled resources to the communities. Thus, careerism in community power went on reproducing careerism in traditional politics. Popular participation of a nonclientelist type was therefore upsetting both for traditional politics and community power structures. And finally, the municipality was bankrupt. During the previous decade, the municipal revenue had decreased and the former mayor had approved a dramatic salary raise for the municipal workers just before he left office. As a consequence, in the budget for 1989, expenditure with personnel carried around 96%, and only 3.2% of the revenue was left for investment. With such scarce resources, it was impossible to meet adequately the demands of the communities.

In this first year, then, the experience of the participatory budgeting could not but be frustrating.TM Very few of the works planned were carried out. For example, none of the 42 kilometers of pavement projected for the communities was completed. In the following year, the extent of the frustration was quite visible. Tarso Genro recalls that meeting attendance, which had been relatively large in 1989, dropped in 1990. At a particular meeting in one of the regions, there were more members of the executive (twenty-five) than people from the community (sixteen).

The years of 1990 and 1991 were devoted to recuperating the financial and investment capacity of the municipality. Expenditure control, combined with a municipal fiscal reform, and larger federal and state transfers allowed by the 1988 constitution were the policies that increased investment percentage of the budget to 10% in 1990, 16.3% in 1991, and 17% in 1992. As regards the municipal fiscal reform, progressivity was introduced in the tax on urban property (IPTU, Property Tax on Urban Land and Homeownership), and several tariffs concerning municipal services were updated (for instance, garbage collection) and indexed to inflation (then skyrocketing) at the same time that surveillance of tax and tariff payments was made more efficient. The most dramatic change concerned the IPTU. If in 1990 it amounted to 5.8% of municipal revenue, in 1992 it reached 13.8% and today varies between 17% and 18%.21

The tax reform, which was crucial to relaunch the popular administration, had to be approved by the Camara de Vereadores. Because the Popular Front did not have the majority in the Camara, the PT and the executive promoted a massive mobilization of the popular classes to pressure the legislators to approve the tax reform law. As Tarso Genro recalls, the rightist and centrist legislators, taken by surprise, could not understand why the people would pressure them to raise taxes.22

The executive's response to the initial failure of the participatory budget did not limit itself to overcoming the financial crisis. It also included deep politicaladministrative changes inside the executive itself and a significant swerve in the political-ideological debate inside the party leadership. Also introduced was a conception of strategic planning influenced by Salvador Allende's experience in Chile.23 The coordination of the PB was taken from the Secretariat for Planning, whose technical body was prey to clientelist policies, and centralized in two organs answering directly to the mayor's office: the GAPLAN and the CRC.

On the other hand, institutional mediation between the executive and the community organizations was started so as to combine effective participation with the preparation of an efficient, coherent, and realistic budget. This kind of mediation amounts to the structures and processes of the PB, which have not stopped being improved to this day. Thus, a new regional division was discussed with the community delegates. The previous five regions gave way to sixteen, a figure that still holds.

In the 1991 budgeting debate, a methodology was introduced for the first time for the distribution of resource investments among the city regions and choice of budgetary priorities. Always as a result of discussion with community delegates, it was decided to concentrate 70% of the resources for investment in five regions considered priority. The choice of regions was based on the following criteria: popular mobilization in the region, importance of the region for the organization of the city, lack of public services and/or infrastructures in the region, and number of people living in conditions of extremely deficient public services or infrastructures in the region.24

At the same time, there was evident progress concerning the creation of community-based representative institutions for the specific discussion of the budget. The Comissao dos Representantes Comunit*rios (Committee of Community Delegates), which had initiated popular participation in the preparation of the budget, was eventually replaced by two important institutions still existing today: the COP and the Forum of Delegates. This model of institutionalized participation and decision making, based on a strong binding link between the municipal administration and the communities, amounted in practice to putting aside the dual power thesis. This, however, did not mean the marginalization of the popular councils. On the contrary, they continued to be acknowledged as autonomous regional organizations vis-a-vis the state.

From 1992 onwards, popular participation increased significantly as a result of the recuperation of the credibility of the PB, which in turn was due to investment increase, particularly from 1991 on, as well as to the fact that investment was carried out in strict compliance with the decisions taken by the COP. New changes and improvements were then introduced as regards to the methodology used to distribute the resources. Thus, the discontent of the eleven regions considered nonpriority and, therefore, granted only 30% of the investment, led to the abandonment of the criterion that prescribed concentration of investment in priority areas and the adoption of a systematic selection of priorities within the different investment sectors (e.g., pavements, sewage, land legalization, etc.) throughout all the regions of the city.

On the other hand, while the notion of objective criteria for determining priorities and selecting investments was maintained, the criteria underwent many changes. Two criteria were abandoned-"popular mobilization in the region" and "importance of region for the organization of the city"-the former for being considered subjective and allowing for manipulation (for instance, artificial promotion of participation to suggest high levels of mobilization), the latter for being vague and allowing for technicist deviation, since the grades for each region in this criterion were given by the technical staff of the mayoralty. The remaining criteria were kept and two others added: "priority of investment chosen by the region" and "total population of region." The former reflected the demand of community delegates to have the priorities of the inhabitants of each region better contemplated in the allocation of municipal resources. The latter, claimed by the delegates of some regions and by the executive itself, aimed at making the allocation of resources a more universalizing process.25

From 1993 onwards, when the second PT term began, the pattern of participation and institutionalization of the PB entered a phase of consolidation, as witnessed by the significant increase of participation in assemblies and meetings, as well as the acknowledgment of the COP by both the executive and the community movements, as a crucial institutional mediation for the democratic distribution of budget resources. However, this pattern of participation and institutionalization left intact the conflictive nature of participation and indeed drew its strength from the permanent tension between conflict and negotiation (more on this below).

In the following years, institutional learning and dynamics continued to be the main features of the PB. We might even say that from 1993 onwards, the structure and functioning of the PB gained increasing operational complexity, which did not prevent, quite the opposite, the number of participants in the official meetings from rising (from under 1,000 in 1990 to 3,700 in 1991 and 10,000 in 1993) and their social composition from getting more diversified. One of the criticisms addressed by the opposition and the media at the experiment of the PB was that, all in all, the PB was just a new version of the "bean 'n rice politics," that is to say, a formula for solving a few of the urgent problems affecting the popular classes, perhaps a less clientelist version than the traditional one, but no less imediatist and electoralist. Also, the fact that participation had a regional basis made impossible any discussion of the city as a whole, any definition of sectorial policies concerning every region, and, above all, any definition of a strategic plan for the city. On the other hand, focusing on the basic needs of the popular classes had resulted in neglecting issues of interest to other social sectors: the middle classes, business groups, and even the trade unions. For this reason, these social categories had been largely absent from the PB up to then.

Such criticisms coincided in part with the evaluation the PT government itself made of the first years of the PB. To respond to them, two important initiatives were taken right at the start of the second term (1993): the Congresso da Cidade (City Congress) began to be held regularly and the plenarias tematicas (thematic plenaries) were created. The first Congress, known as the First Congress of the Constituent City (echoing the large democratic mobilization at the root of the 1988 constitution), took place in December 1993, its main topic being urban development. Participants were all the PB delegates and councillors, civil society organizations of all kinds (community organizations, trade unions, cultural and business associations, etc.), the university, and organs of municipal state and federal government. The conclusions of the congress defined the "major lines of economic and urban development," which from then on became the guiding principles of the municipal government and the PB. Two years later, the Second Congress was held focusing on the strategic plan of the city, known as piano director.

The meetings, known as thematic plenaries, which from 1994 onwards became part of the PB cycle, were the way found to commit the PB to the principles approved by the City Congress. To further the directives of the first City Congress, five thematic plenaries were created: city organization and urban development, health and social assistance, economic development and tax system, circulation and transportation, and education, culture, and leisure.26 The thematic plenaries are organized in a similar way as the regional assemblies: they include two rounds of general assemblies (rodadas) and elect delegates according to the number of participants in the assemblies as well as two representatives for the Participatory Budgeting Council. However, while in the regional plenaries, even though anyone may participate, only the inhabitants of the region have the right to vote; in the thematic plenaries, any citizen, whatever his or her region, may participate and vote. Among the most important decisions of the first few thematic plenaries was the decision to depollute the river and beaches of Porto Alegre, an issue of general interest for the city as a whole and not just for the region where the beaches are located (Ipanema), and the decision to restore the public market, a public space of great architectural value and with great symbolic value in the social and cultural imaginary of the city.

The thematic plenaries were the means of expanding both the matters for discussion and participatory decision making and the social composition of the participants, thereby improving the quality and complexity of the participation. According to municipal data, of a total of 1,011 people attending the second rodada of thematic plenaries in 1994, 11.5% belonged to the trade union movement, 14.3% represented business interests, 20% belonged to community movements, 35% belonged to other institutions of the civil society and the state, 14.4% were individuals with no organizational affiliation, and 0.7% were representatives of political parties. Nevertheless, the fact that the participation of community or regionally based associations was still predominant may have resulted in a certain overlapping of spaces of participation and representation in the regional and thematic plenaries. Fedozzi mentions a survey in 1995, according to which, when asked about what distinguishes regional from thematic plenaries, 60% either answered that they did not know the difference or did not answer at all.'

The relationship between regional participation and thematic participation is not a mere question of an overlapping of spaces of participation. It is, above all, a question of urban politics, and it has over the years become ever more contentious. In the early nineties, the regions were mainly concerned with the lack of physical infrastructures, and this explains why leisure or culture was not even considered as possible topics for prioritization. But the regions had always a tradition of lively leisure and cultural activities. Many of them had cultural and sports clubs, theater groups, and so on. As the lack of physical infrastructures was attenuated as a result of the success of the PB, the demands for "postmaterialist" improvements increased and hence the relative overlap with the thematic plenaries. The overlap hides a conflict of conceptions about city culture that may be related to the different social composition of the regional and thematic plenaries. However, the conflict in this respect is mainly between the regions and the executive itself. The COP, dominated by the regions, has been fighting with increasing aggressiveness for the expansion of the topics of regional interest under PB decision making. Among such topics, culture is always mentioned.

Through the CRC and the GAPLAN, the executive has in several occasions resisted such an expansion with the argument that such topics, rather than being of regional interest, concern the city as a whole. This argument has not convinced the councillors, and in 1997, they were very critical of the cultural proposals presented by the executive that, in their opinion, were biased in favor of "high culture" activities. The truth of the matter is that many regions have their own cultural traditions, programs, and facilities and want above all to improve them. In the COP meeting of March 1, 1997, one thematic councillor, representing the theme Education, Culture, and Leisure, challenged the executive representatives: "our concern in the thematic is that in fact the activities in the area of culture end up being chosen by the Mayor's Secretariat for culture and regions, and the people are excluded. Notwithstanding the thematic's attempt it has not received any response from the Secretariat. Our interest is that the popular will is respected in the cultural programs of Porto Alegre. Because the thematic believes that culture is how we live."

During the second term, and also to respond to a demand by the PB councillors that the realm of the PB be expanded to other areas of municipal spending, the Three-Party Commission (Comissao Tripartite) was set up, composed of six PB councillors (three regular and three substitutes), representatives of the municipal workers union (SIMPA), and members of the executive. Its purpose is to participate in decisions concerning policies related to personnel and municipal administration.

Two other institutional changes introduced after 1993 must be mentioned. Since the beginning of the decade, the Forum do Orqamento (Budget Forum) was the means of gathering together all the delegates of all the regions. Though its existence was certainly justified because of the need for a transregional, citywide mediation, the truth is that the forum did not have clearly defined functions, and its dynamics were deficient. The decision was, therefore, taken to cancel the Budget Forum and create in its stead forums of delegates in every region and in each one of the thematic plenaries. On the other hand, with the objective of refining the methodology of participation and representation, the election of delegates stopped taking place only in the second rodada of the regional or thematic assemblies and began to occur in two moments: a part of the delegates was elected in the first rodada of the plenary assemblies and the other part in the intermediate meetings that took place between the first and second rodada of the regional or thematic assemblies.

The PB learning process in the course of its first ten years (1989-1998) reveals itself not only in formal institutional innovations but especially in the internal operation of the existing institutions. As we saw above, some substantial changes in the criteria for resource allocation have been introduced. Moreover, the conflicts of interests and the political cleavages have been surfacing ever more openly. In recent years, the COP has become more assertive, challenging what is sometimes considered tutelage or even manipulation by the executive. In sum, the PB has become more transparent regarding its core character: a democratic political struggle centered on different conceptions of fair distribution of scarce public resources in an extremely unequal society.

IV. PARTICIPATORY BUDGETING: TOWARD A REDISTRIBUTIVE DEMOCRACY

In this section, some of the most salient political features of the PB will be briefly analyzed. I will also identify the major challenges facing the PB as well as the problems and even dilemmas lying ahead. In order to structure the discussion, Figure 2 will be helpful.

Participation, Negotiation, and Redistribution

The PB is a process of decision making based upon general rules and criteria of distributive justice discussed and approved by regular, institutional organs of participation in which the popular classes have majority representation. The communities in which they live and organize are recognized as having urban collective rights that legitimate both their claims and demands and their participation in the decisions taken to meet them.

The selection of priorities and works included in the Investment Plan is reached by means of objective criteria, defined through a complex process of negotiation that takes place at the intermediate meetings, regional assemblies, and PB council. It is today generally recognized that the PB changed the political culture of community organizations, from a culture of protest and confrontation to a culture of conflict and negotiation. Indeed, conflict and negotiation do not occur among the regions alone but inside each region as well, and it is equally complex and tense at the intraregional as at the interregional level.

The objectivity and transparency of the criteria are expressed in the points earned by each region and the percentage of investment resources into which they are converted. The point system was the methodology conceived to hierarchize priorities and to turn them into resource and investment quantities determined by general criteria. For each one of the priorities, the weight of the criterion and the grade given to the region define the points, which in turn decide the percentage of resources to be invested. The point system aims at converting the political decisions reached through complex negotiations in the detailed distribution of resources included in the Investment Plan and makes sure that such conversion is as faithful and objective as possible.

The latter concern implied successive refinements of the distributive methodology that endowed the PB with great operational and functional complexity. The increase and diversity of participation, together with the increasing intensity and differentiation of demands, have also contributed decisively to making calculating methodologies even more complex and sophisticated. The complexity of the point system resides in the fact that it seeks to articulate measures of participation, on one hand, with measures of priority and recognized necessity, on the other. The participation measure guarantees the democratic legitimacy of political decisions, while the priority and necessity measure guarantees the fidelity, objectivity, and transparency of the conversion of political decisions into distributed resources.

Figure 2.

Enlarge 200% Enlarge 400%

Once the amount to be invested in the region according to the priorities defined by the region has been decided, that sum has to be distributed inside the region itself. The latter distribution is often extremely difficult, given the internal diversity of the regions and the political struggles concerning community leadership. These difficulties led to the creation of microregions, that is to say, social spaces with some identity inside the regions themselves. The aim was to reproduce inside the regions the same decision processes and criteria adopted for resource distribution among regions.

In 1995, the PB council approved nonbinding directives for resource distribution inside the regions, proposing the adoption of objective criteria for the hierarchization of priorities and choice of investment that were similar to the ones adopted for the interregional hierarchization and distribution: priority of microregion or neighborhood (Vila), lack of services or infrastructures, and population benefiting from the work demanded. According to Fedozzi, in most regions, especially those not divided into microregions, decisions did not obey the point system; there was, rather, political negotiation and the direct vote of the delegates of each neighborhood for the choice of priorities.28 Regardless of the difficulty in measuring, for example, the lack of services or infrastructures in each microregion or neighborhood, most community leaders chose the distribution of the resources according to the criterion of the participation of the neighbors in the meetings. Only four of the sixteen regions used some kind of system based on the calculation of the choice of priorities.

Because of its major concern with the democratic nature of the distribution, the PB may be considered the embryo of a redistributive democracy. As I have indicated, the democratic process of the distribution has been organized through an increasingly complex and sophisticated methodology. One could say that, when it does not evolve in a Weberian way, that is to say, together with an increase of bureaucracy, democracy evolves together with an increase of decisional complexity. The following working hypothesis could indeed be formulated: in internally differentiated societies, the stronger the bond between democracy and distributive justice, the more complex the methodology that guarantees such bond tends to be. The decrease of complexity that bureaucracy allows for cannot but bring about the loosening of the bond between democracy and distributive justice.

The redistributive efficiency of the PB has been fully confirmed. Suffice it to mention that in the PB, the poorest region of the city, Ilhas (nowadays a microregion of Humait*/Navegantes), with a population of 5,163 inhabitants almost entirely classified as needy people, has the same decisional weight as the wealthiest region, the Centro, with 293,193 inhabitants of whom only 7,586 are considered needy.29 When, during the 1992 election campaign, those opposed to the PT candidate argued that the PT government only provided "bean 'n rice works," this criticism was actually the great lever for popular mobilization in favor of the PT government. The communities then participating in the PB assumed as being in their favor the "bean 'n rice works," for the communities themselves had voted for them in the PB. The pejorative nature of the phrase "bean 'n rice works," pointing to the popularclasses as socially devalued subjects, was neutralized by the popular classes themselves as citizens of the democratic decisions that led to the works. Thus, the unequal and conflictive nature of power relations in the city became visible and a field for political strife itself.

By reversing the priorities that traditionally presided over the resource distribution by the municipal government, the PB reached striking material results. As regards basic sanitation (water and sewage), in 1989, only 49% of the population was covered. By the end of 1996,98% of the households had water and 85% were served by the sewage system. According to the influential paper, O Estado de Sao Paulo, while all the previous municipal governments of Porto Alegre had built some 1,100 kilometers of sewage, the two PT administrations alone built 900 kilometers between 1989 and 1996. As concerns street pavement, 215 kilometers were built at the rate of 30 kilometers per year. Even so, only one-half of the street pavement deficit (approximately 500 kilometers) was eliminated. The legalization of land ownership, which, as we have seen, is a high priority in many regions where the popular classes live, is an issue where the power relations of the city have eloquent expression, since most of the urban land available for development is owned by fifteen families or entities. Nonetheless, in the past seven years, it has been possible to urbanize the slums and build many houses for the marginal populations. As regards education, between 1989 and 1996, the number of students enrolled in elementary and secondary schools doubled.

The investment effort made by the executive was possible because the revenue increased dramatically, due to federal and state transfers and to the tax reform. In the period under analysis, a 48% increase of revenue was made possible. The former mayor, Tarso Genro, believes that the transparency in municipal spending brought about by the PB contributed to the motivation to pay taxes.

The PB and the People

The main issues concerning the relations between the PB and the citizens and popular organizations are the articulation of representation with participation and the quality of representation.

In 1996, 14,267 people participated in the two rodadas of regional and thematic assemblies. According to the estimate of the Office of the Mayor, if the hundreds of intermediate meetings, both regional and thematic, were to be taken into account, total participation would come close to 100 thousand people, that is to say, 8% of the city population. According to a survey conducted in August 1995, the majority of participants-a balanced group of men and women with an average age of forty-one-belong to the popular classes: 40% have a household income of one to three times the minimum wage3" and have elementary education. Around 60% of the participants have an income of up to five times the minimum wage. A significant number of the people surveyed have flexible jobs as regards time and schedule, for example, self-employed, retired, and at-home workers.3" In comparison with a similar survey done in 1993, an increase in income and education is detected.

Concerning the participation of women, though rather balanced at the base, it decreases as the decision scale rises. The gender factor is particularly noticeable in the following categories: board of directors of a neighborhood association (20% are women), PB council (20%), and Forum of Delegates (16.9%).32 On the other hand, the participation of women in community associations and basic structures of the PB (assemblies) is higher than is usually reported on similar participatory experiences in Brazil and other countries of Latin America.

The people participating in thematic assemblies have considerably higher levels of income and schooling. Fulfilling their original purpose, the thematic plenaries are evidently attracting a more varied set of entities and organizations than the regional assemblies. Nonetheless, most participants indicated that they represented neighborhood associations, whether in regional or thematic assemblies; 75.9% declared they belonged to some entity or association, and 50.5% declared they belonged to neighborhood associations. That is to say, of those belonging to associations, 66% belong to neighborhood associations. Although the thematic assemblies were conceived as a privileged space for the participation of the labor unions, the average participation of unions in thematic assemblies is the same as their average participation in regional assemblies. As regards "second-level" autonomous structures33-popular councils and regional articulations-surprisingly, they show more participation in the thematic than in the regional assemblies, even though their organization is based on the region.

As we have seen, the concern has always been to achieve a fine-tuning between participation and representation, that is, to improve the mechanisms of representation needed for participatory democracy to function adequately. Indeed, even in such a vibrant participatory environment, there is no guarantee that representation may not be thwarted, either because the principles of the mandate are not respected and the priorities decided by the assemblies are manipulated, or because representation becomes professionalized when a delegate holds the post for too long. In order to neutralize the possibility of such deviations, the term of the PB councillors may be revoked at any time by the assemblies that elected them, and no candidate can be elected more than twice to a given position. Moreover, members of any other municipal council, holders of elected public positions, and people with a contractual relationship with the municipal administration cannot be elected to the COP.

This same concern with binding participation to representation has led in the past few years to the addition of a few alterations in the electoral system. Thus, the increase in participation and the need to safeguard minority positions led to the adoption of the proportional method in the election of the PB delegates and councillors.

Finally, since 1994, the incentives to participation have stopped being merely materialist and have become cultural as well, although these changes were also seen as an attempt to fight the boredom of some meetings. Thus, in order to make the assemblies more attractive and lively, before they begin, plays and sketches prepared by cultural associations in the region are performed, and a video is shown displaying the works-in-progress or works already completed, which had been decided the previous year in the Investment Plan, along with tables and charts demonstrating their conformity to the decisions of the PB council.

The relations between popular participation and popular representation in the PB are not as smooth as they appear. The problems center on the following two questions: the ratio of represented/representatives and the quality of representation (autonomy, accountability, and retorno or transparency). From the very beginning, there was a tension between the executive and popular movements concerning the criteria used to determine the ratio between the number of people attending the meetings and the number of delegates elected. As the attendance increased, the executive proposed that the ratio of one delegate for every five people change to one delegate for every ten people. As I mentioned above, the ratio or proportion suffered subsequently two more changes: one for every twenty and then the current system. Some of the leaders of the popular organizations of the PT have always contended that the number of delegates should be as large as possible. Considering that the PB is a learning process and that important information circulates inside its institutions (mainly the COP and the Forum of Delegates), the exposure to such learning and information should be as widespread as possible. The executive has always responded with the argument of deliberative efficiency and the need to make compatible direct democracy and representative democracy. As Olivio Dutra, the first PT mayor, said, "We are not selling the illusion of the direct democracy in the Greek plaza which, let us bear in mind, was not the democracy of all but the democracy of the best."34

The question of the quality of representation involves a number of issues. The first concerns the autonomy of the popular representatives in the PB vis-a-vis the government. Leaders of the popular movement not affiliated with the PT claim that the government, through the PB, has co-opted the popular movement, distorting its priorities and subjecting it to the executive's political agenda. As a result, non-PT leaders argue, the popular movement has concentrated too much on local politics and neglected the critique of and confrontation with the state and federal government. Moreover, the tensions inside the community movement have been submerged and left unresolved, since the new political culture aimed at by the PB has not yet filtered down to the popular movement. The new agendas have silenced the old ones instead of incorporating them. The issue of the autonomy of the PB vis-A-vis the executive will be dealt with below in greater detail.

Related with this issue is the issue of how closely the views and positions of the delegates and councillors reflect those of the regions they represent. This issue was not very relevant as long as the institutions of the PB were concerned exclusively with the physical infrastructures of the communities. More recently, however, as the debates and demands have expanded into other areas (such as culture, etc.) and both the delegates and the councillors have started participating in numerous events on very different topics, the bond between the population of the region and its representatives has loosened up. The positions assumed by the delegates or councillors may reflect their personal preferences more than anything else. It is feared that this "autonomy" of the representatives vis-a-vis their constituencies may bring back in a new guise the old populist, clientelist system of resource allocation and vote exchange. CIDADE, a nongovernmental organization (NGO) that follows very closely the PB, reports on a certain uneasiness in the communities both because the councillors assume positions without previously consulting with them and because they fail to report back to the communities and inform them about the debates and decisions in the COP and other committees.35 Retorno-literally, return-has become a key word in this debate. It means the demand of transparency, of reporting back, of diffusion of information. Retorno has been demanded by the popular organizations vis-a-vis delegates and councillors, by delegates vis-a-vis councillors, by councillors vis-a-vis the councillors who are members of the Parity Commission, and the Three-Party Commission. Retorno, the flow of information, is crucial for the effective control over the representatives of popular participation and thus for their accountability.

The retorno raises another issue, related to the quality of representation: the issue of specialized knowledge and its impact on training and reelection. To have access to and master relevant information is probably the most basic condition for the effective operation of the PB. Such information is often technical and difficult to grasp by people without a high level of education. As we will see below, the popular administration has made a genuine effort to make accessible much of the information that was previously a monopoly of the technical staff (engineers, lawyers, public administrators, architects, etc.). Still, the councillors and delegates claim sometimes that they have been denied important information or that they have had access to it too late or even that it is too cumbersome to get it. In the COP meeting of April 6, 1997, the difficulty in having access to information was raised by one of the most influential councillors. The executive representative, from the GAPLAN, answered that the criticism was fair and that the electronic dataprocessing system of the municipality had suffered some delays. The councillor counterargued that the firm in charge of installing such a system should be summoned to the COP for a hearing. The GAPLAN representative commented later in the meeting: "I would like to clarify the following. There is information. It may be incomplete, but it is there and although the councillors keep talking about autonomy the fact of the matter is that they don't use information that is available. For instance, the record (cadastro) of the streets. We in the GAPLAN have insisted that the cadastro is at your disposal and that you should go through it to detect possible mistakes. How many councillors have conferred the cadastro? Three!"

After the new COP is inaugurated, some training seminars are organized to familiarize the new councillors with the complex operation of the PB. Moreover, in recent years, through protocols celebrated between the Prefeitura, the university, and the NGOs, the training of councillors and delegates has intensified. The need to get familiarized with the PB process and to master the relevant information has raised still another issue, the reelection of councillors. According to the rules, the councillors can only be reelected once. But both inside and outside the COP the question has been raised whether the reelection should be admitted without term limits, the justification being that one year is too short a period to get fully acquainted with the operation of the PB. The NGO CIDADE has even proposed that the councillors be elected for a two-year term or that, as an alternative, the effective councillors in a given year become substitute councillors in the following year, thus allowing for the transmission of their knowledge and experience to the incomers.

In the popular movement, this position is often looked at with suspicion, fearing that reelection might lead to a new breed of professionalized elected officials easily prey to the old populist, clientelist system. In the COP meeting of March 4, 1997, in which the internal rules were on the agenda, the issue of reelection was once again raised. Some councillors defended the reelection with two very different arguments. One argument was about the quality of representation: the knowledge and the experience acquired would improve the quality of representation. The other argument was about the autonomy of popular participation. One councillor said, "To limit the reelection amounts to saying to the communities and the delegates: `Look, you don't know how to vote and for that reason you are not allowed to vote on anyone more than twice.' If someone is a good councillor why should he not be allowed to stay for 4 or 5 years? The council denies the delegates the option to vote how they like and the assumption of responsibility for the way they voted." Other councillors counterargued. One said, "Look, in my region for a long time only X knew about PB. Only one person. The essence of the PB is to form leaders, in the plural, not one leader that knows everything while the rest knows nothing." Another said, "Why are new councillors here today? Precisely because of this provision of the rules. Fortunately it privileges the arrival of new councillors every year. Otherwise many of you would not be here today." The representative of the executive expressed the same opinion: "We had a similar discussion three years ago. In the first years of the PB various councillors stayed for three or four years. The idea was `let us support the one with greater experience: But then we came to the conclusion that we were perpetuating the same people in the same position. We were preventing the emergence of new leaderships. Today the PB is the vanguard. Neither the bourgeois parliament nor the trade unions or other entities do as we do. They allow for consecutive reelections and thus perpetuate the same leaders. They are lagging behind." After some more interventions, the issue of reelection was voted. There were two proposals: to maintain the current system and to allow reelection without limits. The first proposal won with seventeen votes in favor and eleven votes against.

Concerning the quality of representation, there is still another issue to be mentioned: the degree of participation of the councillors at the meetings and other activities of the COP. Throughout the years, the COP met regularly once a week, but in 1997, it met twice a week for several months. The meetings lasted two hours and sometimes longer, and since most councillors live in the peripheral regions, the time spent in long bus rides should also be added. It is thus a very intense type of voluntary work, and some councillors find it impossible to attend all meetings. Moreover, whenever the meetings take too long, some councillors must leave before the meeting ends in order to catch the last bus to his or her region. The more assiduous councillors have been very critical of absenteeism. In the meeting of March 6, 1997, one of them said, "The councillor does not represent one person. He/she represents thousands of people. They represent the city of Porto Alegre. If a councillor assumes the responsibility of discussing the problems of the city but when the time to vote comes he leaves because he has another engagement I think this behavior amounts to an insult to the COP and to the people."

At this meeting, the question of quorum to take a vote was raised and vividly debated. There were three proposals. According to the first proposal, the quorum should be a simple majority of the councillors (twenty-three, even though the government representatives do not vote) established at the beginning of the meeting. If the quorum was fulfilled at the beginning of the meeting, then decisions could be taken, even if in the course of the meeting the number of presences came down below the quorum. The second proposal defined the same quorum but required it for all the decisions taken. The third proposal, which had already been adopted in the previous year, proposed two formulas for the quorum: simple majority of the councillors or, in its absence, the simple majority of the regions and thematic areas, that is, nine and three, respectively. The first proposal was defended by some councillors who felt "penalized by the colleagues that abandon the meeting making it impossible to take decisions. These colleagues are the ones that should be penalized. We should be able to decide without them and without regard for their positions." Other councillors and the executive representatives were in favor of the third proposal. Said the GAPLAN representative, "Often the discussion becomes heated, takes more time and many councillors start leaving to catch the bus. It is frustrating because we discuss and discuss, and all of a sudden we don't have quorum. I am also concerned with the representativeness of the discussion and debate. The third formula is a compromise. At least the majority of the regions and of the thematic areas are present." The proposals were then voted: first proposal: seven votes; second proposal, four votes; third proposal, twelve votes.

In my view, the way the different issues involving the quality of representation have been debated inside and outside the PB institutions bears witness to the engagement of the popular sectors of Porto Alegre in preventing the PB from falling into the trappings of the old clientelist, authoritarian system. Indeed, the specter of the continuities between the old and new system surfaces occasionally in the debates.

Autonomy and Co-Government

For its founders and activists, the PB is the manifestation of an emergent, nonstate public sphere where citizens and community organizations, on one hand, and the municipal government, on the other, converge with mutual autonomy. Such convergence occurs by means of a political contract through which this mutual autonomy becomes mutually relative autonomy. The experience of PB configures, therefore, a model of co-government, that is to say, a model for sharing political power by means of a network of democratic institutions geared to reaching decisions by deliberation, consensus, and compromise.

The problems facing a system of power sharing are well expressed in the relationship between the COP and the executive. In the beginning, while the community leaders wanted the COP to have unconditional deliberative power, the executive searched for a formula capable of reconciling the decisions of the COP and the political representativity of the mayor inscribed in the constitution of the republic. The formula is as follows: the deliberations of the COP are taken by simple majority; the decisions are forwarded to the executive; in the case of veto by the mayor, they return to the council for a new evaluation; rejecting the mayor's veto requires a qualified majority of two-thirds of the vote; if rejection occurs, the matter goes back to the mayor for evaluation and final decision. Since, according to the constitution, the power to approve the budget is vested on the legislative body, this formula accommodates the constitutional requisite: formally, the budget proposal submitted to the Camara de Vereadores is the mayor's proposal.

The mayor's veto must be substantiated and can only be exercised for technical reasons and financial evaluation. To this date, however, the veto was never exercised, since whenever the executive had reservations concerning a work, its position was explained to the community by its technical staff, and the community ended up agreeing.

The consensus-building process is complex because the problems under discus.sion as well as the decisions taken often have, besides the political dimension, a strong technical dimension. Moreover, "technical criteria" constitute one of the limits of participation and are sometimes the object of debate and conflict themselves. The internal rules of the PB include the technical criteria established by the various departments of the executive and justify them as follows: the presentation and clarification of the technical and legal criteria, utilized by the secretariats and departments, will make the PB procedure more transparent for, having been previously enlightened, the community, when discussing its priorities, will avoid the selection of works that cannot be implemented by the municipal mayoralty. The totality of technical criteria was submitted to the Parity Commission (Comissao Paritaria) for evaluation, debate and deliberation.

As mentioned above, the Parity Commission is another of the institutional creations of the second term of the popular administration. It is now composed of eight councillors of the COP (four effective and four substitute), two representatives of the CRC, and two representatives of the GAPLAN. The CRC and GAPLAN are the two main institutions of the executive that guarantee institutional mediation between the executive and community organizations and associations. The main function of the Parity Commission has been, up until now, to legitimate the definition of the technical criteria by submitting it to some kind of participatory decision making. In real terms, given the technical complexity and knowledge behind the criteria, the commission has always rubber-stamped the executive's proposals. Since 1997, the Parity Commission has been endowed with broader tasks of coordination of the COP's activities and meetings but, as I suggested, the real coordination belongs still to the government representatives.

Here are some of the technical criteria currently in force: all community claims and demands found technically nonviable by the municipality are cancelled; preference is given to works-in-progress; the pluvial network will not be installed in streets without pavement because the network, being open to allow for the captation of rainwater, might be blocked by sand and rubbish; and in streets with heavy traffic, a minimum of 10 meters width is required: 7 meters for the lanes and 3 meters for the sidewalks.

In such a system of co-government, the executive does have a very active role, if for nothing else because it controls technical knowledge and also because it either generates the relevant information or has a privileged access to it. Its presence in the PB is quite strong, by reason of its coordinating functions both in the COP through its two representatives (one from GAPLAN, another from CRC), even though they do not have the right to vote, and in the regional assemblies through the CRC delegate (the CROP) in the region. Furthermore, the executive itself forwards autonomous investment proposals to the COP, the so-called institutional demands that have their origin in departments of the executive and usually concern maintenance or improvement of urban infrastructures of the city.

Besides technical limits, there are financial limits not always duly considered by the assemblies. Suffice it to mention that, for financial reasons, only 30% of the demands originally formulated by the community can be taken care of. Sometimes, the way the demands and priorities are formulated does not take into account certain technical conditionalities that increase the cost beyond what the communities themselves consider reasonable. For example, the fact that street pavement must include street lighting increases its costs considerably. Nowadays, the percentage of investments included in the budget varies between 15% and 20%, a percentage that, according to Brazilian standards, must be considered high. On the other hand, community councillors in the COP have always questioned the amount of expenses with the personnel and services of the executive and argued that the PB process should contemplate such expenses. To meet this demand to some extent, a representative and a substitute of the union of municipal workers (SIMPA) now take part in the COP; on the other hand, as mentioned above, the Three-Party Commission was created in the COP-composed of councillors, representatives of SIMPA, and representatives of the government-"whose objective is to debate and deliberate on the admission of personnel to the Municipal Government."

The decision model resulting from power sharing in the PB is quite complex, a complexity that emerges in the way the PB is seen by the participants in regional assemblies and thematic plenaries. According to the 1995 survey already mentioned, to the question on the decision power of the communities, 33% replied that the population decides "always," 27.3% replied that the population decides "almost always," and 23.8% replied that it decides "sometimes." Significantly enough, 15.3% did not or did not know how to reply.

The credibility of the political contract that constitutes the PB resides in the effectiveness of the decisions and in the accountability both of the executive and the representatives of the civil society in the COP. The fact that only 30% of the demands may be considered is less important than the effective satisfaction of the demands selected for inclusion in the Investment Plan. Several mechanisms guarantee effectiveness and accountability. First, the political will of the executive must be mentioned. The basic principle of the municipal government is to fulfill as rigorously as possible the Investment Plan and justify what is left unfulfilled. Second, there are committees-created within the Fora of Delegates-whose function is to supervise the works. In the case of delays or alterations, the delegates have direct access to the mayor's office to ask for explanations. Third, the very structures of the PB strongly encourage accountability themselves. The two institutions of regular functioning-the council of the PB and the Forum of Delegates-are bound to the grassroots institutions: the regional assemblies and the thematic plenaries. The two latter organs, because they are open to the individual and collective participation of all citizens, exercise a double popular control: upon the performance of the executive and upon community representation itself. As we saw above, in practice, the exercise of control is problematic, as witness the debates about the quality of representation and about transparency (retorno).

A possible demonstration of the effectiveness of the decisions can be observed in the 1995 survey I have been referring to: 56.5% of participants in regional and thematic assemblies declared to have benefited from works and services of the PB. It is significant that this percentage increases with the number of years of participation in the PB. Thus, among those who had participated in the PB for six years, 72.7% claimed to have benefited. The percentage is also higher among leaders of neighborhood associations (67.9%) and those who have already been elected (either as delegates or councillors) to the organs of the PB (743%).36

As I have indicated, this linkage of participation to resource distribution and effectiveness of decisions is one of the nuclear features of the PB. This alone explains why for five months, the PB councillors meet at least once a week, often twice or thrice a week, with no remuneration, even without fare expenses (fare expenses are actually a demand that has not been met to this day). This linkage of participation to distribution is, no doubt, one of the virtues of the PB, but perhaps also its limit. According to the former Mayor Tarso Genro, it is common for a region or microregion to stop participating in the meetings and assemblies after their demands have been met. Later, they usually come back, once having realized that in the year in which they did not participate, there was no investment in their region or microregion.

As far as accountability goes, rendering account and providing information are crucial for the intelligibility and transparency of the whole process. As early as 1990, the executive declared the Municipal Day of Accountability, on which, at a public meeting downtown, the executive was to render an account of the works decided upon in the PB. Nowadays, accountability is performed in many ways, often by means of flyers widely distributed throughout the city and at the beginning of the assemblies and plenaries of the first round (rodada). On the other hand, the GAPLAN publishes a booklet with the Investment Plan, a list of all the approved works described in detail, as well as a list of names and addresses of every councillor, as well as the telephone number of the PB coordinator (CROP) in every region. This document circulates widely and reaches all the regions, giving citizens the opportunity to see if their decisions are being executed. In the 1995 survey, when asked about the degree of satisfaction concerning the accountability of the executive, 47.6% replied it was satisfactory, while 23.6% said it was "in part" satisfactory.

The close binding articulation of participation, distribution, and effectiveness of decisions may eventually provoke some additional tension in the already tense field of co-government that constitutes the political contract between the executive and the organized communities. Two reasons are especially important: the limits to investment and the problem of pursuing major works. In the past seven years, the municipality of Porto Alegre has managed to increase its investment resources more than any other Brazilian city. According to Navarro, budget resources available for investments accounted for US$54 million in 1992, US$31.5 million in 1993, US$82 million in 1994, US$65.7 million in 1995, and US$70 million in 1996.37 The global investment figures indicate that municipal investment has probably reached its maximum limit. Since increases of participation generate an increase of demands, it is to be expected that the struggle for resource sharing will become fiercer in the near future. If, as a consequence, the percentage of demands actually met decreases significantly, we may well witness a lack of interest in participation emerge, as indeed happened in the early years of the PB. This problem becomes more serious in light of the budgeting crisis of 1998. In fact, the level of investment has decreased in the execution of the 1998 budget due to the sudden reduction of federal transfers at the end of 1997. Such reductions were the result of the measures taken by the federal government to carry the structural adjustment made necessary by the Asian crisis. The reduction of federal transfers derived mainly from the Kandir Law (after the name of the then minister of planning), which granted generous tax exemptions to the export industrial sector. The sustained growth of municipal revenues could not compensate for the reduction of federal transfers.

By its very genesis, the PB has been the privileged mechanism to decide the works that are directly relevant to the communities. It has been, in a word, closer to "beans 'n rice works" than to "major works." The 1993 and 1995 congresses of the constituent city, as well as the creation, from 1994 onwards, of the thematic plenaries, were an attempt at expanding the reach of the decisions. However, given the predominance of the regions in the COP, it is not easy to keep the political contract when the material results turn out to be less anchored in the immediate needs of the regions. By way of example, let us consider the case of the loans that the executive was forced to take from international banks, in view of the ceiling of the investment with locally obtained revenue, in order to carry out works deemed important for the city as a whole.

Having obtained a loan from the World Bank to build various infrastructures, the executive proposed to the COP the construction of five avenues. There was great resistance on the part the community councillors, who wanted the money to be invested in street pavement in the regions. Tarso Genro, former mayor, recounts: "I, myself, and the Executive's staff engaged in a dispute right in the middle of the Council and I threatened: `if you want to shred, OK, we'll shred and build a tiny street in every region. But you will be held responsible and shall answer before the city and will give arguments to the Right for whom you have no vision of how the city should be developed. The five avenues are crucial for all the city population, especially for those living in the periphery.' "3 After a long debate, the council approved the construction of the five avenues, with one vote against.

Actually, resorting to international loans to promote urban development immediately poses problems to the PB. Such loans require the previous, detailed indication of the investment to be made, which may collide with the decisionmaking process of the PB. It so happens, however, that meanwhile, the PB gained some international recognition as a transparent and efficient means of resource distribution. So, recently, the Interamerican Development Bank agreed to grant a loan for the construction of a turnpike (the Ill Perimetra), approving at the same time a loan for street pavement in areas around the turnpike without requiring the usual specifications. In other words, the money will be released to be applied in street pavement, but the specific streets and extension (up to 100 kilometers) will be decided later on by the PB instances.

Is the political contract of co-government that sustains the PB a contract among equal partners? This question raises the issue of the autonomy of the participatory budgeting institutions and processes. I said above that this political contract is based on the premise that the autonomy of both the elected mayor and the popular movement becomes a mutual relative autonomy. The question is, whose autonomy is more relativized by entering the contract? The PB is an initiative of the PT popular administration of Porto Alegre, and its basic institutional outline has been designed over the years by the executive. It is part of a political program of the redistribution of public resources and democratization of the state. This political program is also the meeting ground for a demand with a similar political orientation advanced by the popular movement and sustained over the years by much struggle. The issue is, then, how this convergence of political will has been carried out, in whose terms and timings, and with which outcomes.

As said above, the executive has a prominent role in the PB. The cycle, the agenda, and the timings are set by the municipal government according to legal requirements but certainly also according to a political strategy. But the initiative of the executive only becomes effective if the communities and popular movements participate actively in the process. Without any doubt, the popular participation in the PB is very active. Is it also autonomous? What does it mean to be autonomous? Should the issue of autonomy be discussed solely in the context of the relations between the popular movement and the government or, rather, also in the context of the relations of the popular movement vis-a-vis the other parties and political forces integrating the political field of Porto Alegre?

There is a long tradition of party involvement in the popular movements. The PT won the elections in part by creating a political base among the community organizations. Other parties tried to do the same. The PDT, for instance, has had for a long time a presence and an influence in the neighborhood associations movement and is still very strong in the UAMPA (the Federation of Neighborhood Associations of Porto Alegre). Autonomy cannot, therefore, be conceived as popular spontaneity, as a native capacity to organize poor people in degraded communities without the support or influence of external, organized political forces. Autonomy must rather be conceived as the popular capacity to channel external support and put it at the service of objectives, demands, and agendas generated in the communities. In the Brazilian context, autonomy is measured by the capacity to develop organizational strength and effectivity by maneuvering among competing external political influences, using such competition to impose demands that, however important for the community, do not represent a priority for none of the political forces in competition.

Since the PB is not a popular movement but an institutional constellation designed to function as a sustained, regularly functioning meeting place for the popular movement and the municipal government, the question of the autonomy of the PB must be formulated as the real capacity of the popular representatives in such institutions to shape agendas, timings, debates, and decisions. In this sense, autonomy, rather than a stable characteristic of a given political process, is the ever provisional outcome of an ongoing struggle.

Thus conceived, the autonomy of the PB must be discussed at two levels: the operational functioning of the PB institutions, including coordination, agendas, and timings, and the impact on the PB of changes in political orientation of the executive.

Concerning the first level, I have mentioned that the coordination of the PB institutions is in the hands of the executive's representatives and that the agenda and timing are proposed by them. But I have also stressed that the executive's role in this regard has been increasingly questioned and challenged by the councillors and delegates. The debates were particularly heated in 1997 in the process of preparing the 1998 budget. The discussion of the internal rules was very intense since many councillors opposed the control of the executive over the coordination of the meetings, whereupon some changes were introduced aimed at giving a greater role of coordination to the Parity Commission, composed of an equal number of councillors and representatives of the executive. The observation of the COP meetings in particular shows that the councillors have become more assertive and aggressive, and the procedural rules of the meetings have often been disrupted by heated debates. One of the widely violated rules is the prohibition of direct dialogue among the councillors. Such rule states that the interventions have to be previously registered by the coordination and take place by the order they have been registered.

Concerning the agenda, the conflict between some councillors and the executive is often quite open. The councillors have been consistently fighting for the expansion of the municipal activities to be submitted to the PB, and they have in general been met with the resistance by both the CRC and GAPLAN representatives. The basic argument of the government is that there are topics that engage the city as a whole and, for that reason, cannot be submitted to a debate that tends to promote particularistic solutions, be they relative to the regions or to the themes. The councillors counterargue that they represent the whole city and that the real issue is a different one: the opposition of the executive to the further decentralization of municipal services (culture, health, sports, leisure, etc.). In the preparation of the 1998 budget, the tension was particularly high when the councillors challenged the figures presented by the executive concerning the current spending of the administrative services and the increase in salaries of the municipal workers. According to CIDADE, the discussion of the budget proposal was the most intense in the history of the COP.39

The councillors have been more and more openly critical of the executive coordination and agenda setting. In an interview, one councillor, a very active woman, in the popular movement told me, "Sometimes I feel that I am being manipulated, that I am here to legitimate the popular administration and nothing else. The PB is the best thing that could happen in this city but it has to operate according to our way." Probably as a response to the greater assertiveness of the councillors, there are indications that the setting of the agenda is now more shared, and whenever there are overarching constraints, they are better explained. Besides, the proposals made by the executive representatives are sometimes voted down.

Concerning timings, deadlines, and times for debate, the discussions in the COP have also become more conflictual. On one side, the councillors claim that they need more time to process information, to ask for clarifications, and to consult with their constituencies. On the other side, the executive representatives claim that the deadlines are not an invention or whim of the executive-rather, they are established in laws promulgated by the Camara dos Vereadores. They also claim that to debate is fine, but it is very frustrating to verify that after heated, long discussions there is no quorum to vote because in the meantime, some less interested or more pressed councillors have left the meeting. One good instance of this conflict occurred in the COP meeting of August 8, 1996, when the municipal secretariat for housing submitted a vast housing program (PRO-MORADIA) to the COP and asked for a decision in two days in order to be able to comply with legal deadlines. Some councillors considered that request outrageous in light of the extension and complexity of the document. The same happened again subsequently, for instance, in the COP meeting of March 14, 1997, and as a result, the housing question has become a very contentious issue in the PB. In interviews, the councillors have recurrently voiced the concern that the discussions are rushed, there being not enough time to clarify doubts and vote with full knowledge of what is at stake. Sometimes voting is decided on the basis of a trust relationship with another, more knowledgeable councillor or with the executive representative.

The other dimension of the relative autonomy of the PB concerns the impact of changes in political orientation of the executive upon the PB institutions. Between the leadership of the municipal government and the leadership of the popular movement, there have always been some elective political affinities. Leftist political orientations have dominated both leaderships, and the conflicts among them, sometimes very sectarian, cannot be understood without contextualizing them in the historical conflicts within the Left. In light of these affinities, it was quite natural for the GAPLAN representative to start the COP meeting of April 29, 1997, by an exhortation inviting the councillors to participate in the May 1st (Labor Day) rallies and events and to mobilize their communities to participate.

The now three terms of popular administration have been dominated by different political tendencies inside the PT. These differences have expressed themselves both as different political languages and different political initiatives. In the third term, under a mayor of the Socialist Democracy tendency with a Trotskyite leaning, the plenary assemblies were renamed "popular general assemblies" and the housing cooperatives are now called "self-managed housing cooperatives." It remains to be seen whether these name changes correspond to real changes in the operations of these institutions. In the case of the rodadas, no real change has been detected so far.

But changes in political orientation have repercussions above all in policy changes. One such change concerns the Pluriannual Investment Plan, the plan for the whole term, in this case, 1998-2000. While Tarso Genro discussed and formulated the Pluriannual Plan inside the executive, he did not submit it in any meaningful way to the COP. In contrast, Raul Pont (the mayor elected in 1996), in a highly politicized statement, decided to submit the Pluriannual Plan to the COP in very much the same way as the annual budget. The objective was precisely to expand the ambit of the COP attributes, a demand frequently voiced by the councillors themselves, as we saw above. Another objective was probably to call the attention of the popular leadership to the changing macropolitical context: the cuts in social policies at the federal level and the consequent impact on the financial transfers to the cities and on employment and standard of living, as well as the struggles by the city governments around the country, some of them bankrupt, others forced to embark in aggressive and unpopular measures to attract foreign investment to the city.' Contextualized by this macropolitical environment, the political decisions of the executive would stand out in a brighter light and the difficulties ahead could be better understood.

The initial expectations about the debate on the Pluriannual Investment Plan have been frustrated. The debate on the pluriannual plan was added up to the debate on the annual budget, thus forcing the COP to an extra effort, as the need for two meetings per week well illustrates. Moreover, this effort was not well understood by all councillors. More familiarized with the annual budget, some of them could not see a clear distinction between the Pluriannual Investment Plan and the annual budget. Accordingly, they made suggestions and demands that could fit the annual budget but not the Pluriannual Investment Plan. This forced the government representative to endless explanations about the differences between the two documents and the criteria for their respective items. In the COP meeting of April 22, 1997, the GAPLAN representative stated, "We cannot include a public toilette in the region X in the Pluriannual Plan. The Pluriannual is a plan, it is not a budget. It is a reference planning to guide us in the elaboration of the annual budget." The frustration of the councillors increased when they became aware that their demands, no matter how just or justified, would be met only if the necessary funds would become available in the next coming years.

Irrespective of changes in political leadership, a gradual but consistent movement toward a greater autonomy of the COP and the PB in general vis-A-vis the executive can be detected. The autonomy of the PB has become an ever more cherished value for the councillors and delegates. At the end of 1996, when the Investment Plan for 1997 was presented in a public ceremony attended by the mayor and the councillors of the COP, there were some derogative comments against the councillors in the local press, which has been, in general, hostile to the PB. The councillors interpreted those comments as an insult against their autonomy and at the meeting of January 7, 1997, discussed ways of responding to the insults. A committee was nominated, and in the following month of April the COP decided to create a Media Commission in charge of following the media reporting on the COP and PB and to respond whenever necessary.4

From Technobureaucracy to Technodemocracy

Conflict and mediation between technical and political issues, as well as between knowledge and power, is one of the main features of the PB. If it is true that technical criteria limit the field of participation and deliberation, it is likewise true that the PB process has radically changed the professional culture of the technical staff of the executive. The technical staff has been increasingly submitted to a profound learning process concerning communication and argumentation with lay populations. Their technical recommendations must be conveyed in accessible language to people who do not master technical knowledge; their reasonability must be demonstrated in a persuasive way, rather than imposed in an authoritarian fashion; no alternative hypothesis or solution may be excluded without showing its inviability. Where earlier a technobureaucratic culture prevailed, gradually a technodemocratic culture has emerged.

This transformation has not been easy. According to Tarso Genro, there has been more progress in changing the language and discourse of the engineers when addressing the people in the communities than in changing their dismissive attitudes vis-a-vis what people have to say. In other words, the capacity to make himself or herself understood has improved more than the capacity to listen. More recently, it has been found out that the structure and process of the PB are very little known among the municipal workers and staff. In view of this, the new mayor launched a new program in 1997, called the "Program of Internalization of the Participatory Budgeting," targeted at the municipal personnel. This program was announced as part of a much broader program of an overall internal democratization of the state. In an interview, the official in charge of coordinating this program told me that "in order to be fully consolidated the PB must be part of the everyday work of a municipal worker." A working group has been set up to organize workshops with the workers and staff about the cycle, rules, criteria, and methodology of the PB. The targets of the workshops will be sequentially the following: the personnel who deal directly with the PB, the personnel who intermediate between the executive and the community (such as the community advisors of the FASCOM), and the supervisors and directors.

Once we analyze in detail the functioning of the PB, it will not be difficult to detect, among the multiple interactions between the participants of the PB and the personnel, situations that, no matter how apparently trivial, may be a source of tension even when the personnel support the PB. As an example of such a situation, I may mention the accreditation procedure-the process by which the people, the delegates, and the councillors identify themselves as they enter the room where the meeting is to be held. They must show their ID card and fill out a form. The accreditation is entrusted to a group of municipal personnel designated by the mayor. Even if we only take into account the regional and thematic plenaries, the staff must verify the credentials of hundreds of people in twenty-one meetings (sixteen regional and five thematic) per month. Because it resulted from a personal nomination by the mayor, the verification of credentials was understood for a while as a political job to be performed as militant work. As time went by, however, some of them refused to go on performing the job, invoking the many evenings they could not spend with their families. As a result, the coordinator of the Internalization Program told me that the executive was considering paying extra hours to the credentials personnel and that she was in favor of this.

The road from technobureaucracy to technodemocracy is a bumpy one. In the course of time, as the delegates and councillors have become more assertive, disputing more openly the technical criteria and solutions presented by the professional staff, the latter have become more defensive, but the conflict between competing knowledges has all but faded away. In my field observations, I witnessed many lively debates between residents and engineers about pavement, location of sewage pipes, and so on and was impressed by the argumentative capacity of the community leaders.

Competing Legitimacies: The PB and the Camara de Vereadores

In theoretical debates on the relationship between representative and participatory democracy, it is often forgotten that one does not exist without the other. Participatory democracy, in particular, in complex political processes always presupposes the creation of instances of delegation and representation. The PB experience is eloquent in this respect. As we saw, the basic structures of the PB aim at an institutional articulation not only with the institutions of representative democracy at the urban level (the mayor and his executive) but also with the representative institutions derived from participatory democracy at the community level. This articulation between participation/representation at the community level calls for careful reflection that cannot be undertaken here.

The PB decision model tries to reconcile the principle of democratic representativeness of the mayor and his executive with the principle of participatory democracy of the citizens organized in grassroots associations and assemblies. However, the government is not limited to the executive; it also includes the Camara dos Vereadores (Chamber of Deputies), which is the traditional elected municipal legislative body. The PT, which, together with the other leftist parties that compose the Popular Front, has been able to elect the mayor since 1988 but does not yet hold the majority in the Chamber of Deputies. Nonetheless, its electoral gains in the Camara over the past ten years are quite remarkable and have been generally attributed to the success of the PB. In the mid-1980s, the PT had a very small electoral base in Porto Alegre, and in the 1986 elections, the Popular Front did not elect more than two deputies (one from PT and one from the Socialist Popular Party) of a total of thirty-three seats. In the following years, the electoral gains speak for themselves: 1988 elections, ten seats (nine from PT, one from PPS); 1992 elections, eleven seats (ten from PT, one from PPS); and 1996 elections, fourteen seats (thirteen from PT, one from PPS). If this pace continues, the Popular Front may well have the majority of the seats of the Camara de Vereadores in the 2000 elections. It must, however, be pointed out that the non-Popular Front parties have been trying to resist the PT's electoral growth by nominating for the Camara very strong and high-profile candidates-former mayors or otherwise well-known personalities-who in other circumstances would be nominated for state or federal positions.

The political contract that exists between the executive and the communities has so far not been extended to the legislature. On the contrary, the relation between the PB and the legislature has been one of constant conflict (sometimes involving physical confrontation).42 The reason for the conflict is quite obvious. According to the constitution, it behooves the legislature to approve the municipal budget. Now, the PB has totally preempted this responsibility. To be sure, as we have seen, according to the PB cycle, the proposal of the budget law, after having been prepared in the COP, is forwarded to the legislature for debate and approval. Theoretically, the legislature could reject the proposal, but the fact that it has already been legitimated by the large participation of citizens mobilized by the PB compels the legislature to approve always the budget presented. It ends up, therefore, being a formality.

According to some, however, given the budgeting technique traditionally adopted in Brazil, the legislature has never actually deliberated substantially on the budget. The truth is that, given the fact it is not required that the budget indicate the concrete works to be carried out, the executive has always had ample leeway in budget execution. But the fact of the matter is that such a system also created the opportunity for the legislature to influence the execution by the traditional populist and clientelist methods. The legislators had their electoral folds in the different regions, and the votes they gathered from them were directly entailed to the works they managed to include in the budget. Now, this was precisely the clientelist system that the PB intended to put an end to, herein largely residing the hostility or distance with which the legislators not linked to the PT regard the participatory budgeting.

While the duality of power between the PB and the executive-notwithstanding the problems and tensions identified here-have been dominated by a logic of complementarity and cooperation, the duality of power between the PB and the legislative chamber has been dominated by a logic of open or latent conflict. It is both a duality of power and a duality of legitimacy. As one legislator told me in an interview, "The PT has co-opted and demoralized the popular movement. The PB is a diabolic invention of the PT to stay in power forever. Look how many people participated in the PB last year? A little more than 10,000 people. Well I was elected by a larger number. Why am I less representative than the councillors of the COP?" Another legislator less hostile to the PB said, "I think the PB is an excellent idea but I don't see-except for political reasons of the PT-why the Camara de Vereadores is not involved. We don't want to absorb the PB. We would like to have a part in it. For instance, a percentage of the investment fund should be left to the Chamber to allocate." As another legislator put it, "The budget arrives at the Chamber in a cast (engessado). We're tied up. It is not fair because after all we are the legislators."

The duality of legitimacy between the PB institutions and the Camara de Vereadores raises the issue of the relationships and hierarchies among two different kinds of representativeness: direct democracy councils and fora and territorially based representative democracy. The majority of deputies of the Camara defend the higher quality of the representativeness of the Camara if for no other reason because, as mentioned by one of my interviewees above, no matter how impressive the popular mobilization around the PB, the deputies are in general elected by a larger number of voters. Against this, the PB defenders use a twofold counterargument. On one hand, they argue, the vote for the Camara (as for other elected political bodies) is mandatory in Brazil, the unjustified nonvoting behavior being met with serious sanctions. Therefore, this compulsory democratic participation cannot be compared with the voluntary participation in the PB institutions and processes. On the other hand, the outcomes of representation in the PB are of higher quality than those obtained by means of representation in the Camara since the former derive more from political actors (the community leaderships) who are much more closely linked to and finely tuned with the needs and aspirations of the popular classes, and thus of the vast majority of the population, than the deputies.

One of the angles of the tension between representative and participatory democracy has been the debate in the course of the past few years on the official legal institutionalization of the PB. As it stands today, the PB is based on a political contract with the executive and is ruled only by its internal rules and the organic law of the Prefeitura. The crucial question is whether the future safeguard of the PB should not include its juridical consolidation. The positions diverge, even inside the PT and the executive themselves. While some believe that the legal consignation of the PB will help to defend its existence if, in the future, an executive hostile to citizen participation is elected, others argue that such legalization would be a submission of participatory democracy to representative democracy, do away with the political autonomy of the PB, and subject it in the future to legislative manipulation according to the majorities obtainable in the legislative chamber. Said one of the PT legislators in an interview: "I participate in the plenaries of the PB and I even have a vote as legislator. The legislators should integrate themselves in the PB and not seek a separate and privileged participation and decision making." Other PT legislators and leaders think that the tension between the PB and the chamber is not a "healthy one" and may be risky in the future. According to them, it is not in the interest of the PT to demoralize the legislative body and contribute to emptying out its prerogatives. Some of them have even presented law proposals concerning the legalization of the PB. One of them said in an interview: "I am in favor of a type of legalization that does not plaster the PB, and that contributes to consolidate it as an official component of our political system, a mark of our specificity." Up until now, the anti-legalization position has dominated, both in the executive and in the community movement. Whether this position prevails if the Popular Front obtains the majority of the seats in the chamber in the next elections is an open question.

The issue of legalization is one among many dimensions of the conflict between the executive and the Camara de Vereadores. The COP councillors have a clear understanding of this, and the divisions inside the COP on this issue reflect broader cleavages in the city politics and community movement. In spite of the political restrictions on the budgetary prerrogatives of the Camara de Vereadores, the latter makes many amendments every year not so much to the budget proposal, submitted by the executive until September 30, but to the proposal of budgetary directives submitted until July 15 of every year. These amendments are discussed in the COP. In the meeting of August 7, 1997, the GAPLAN representative read the most important amendments. One of them, coming from a rightist legislator and former mayor, consisted of restricting the mayor's expenses with publicity-a major weapon of the mayor to reach out to the communities and maximize the flow of information about the PB between the executive and the communities. The GAPLAN representative intervened then to draw attention to the negative impact of this and other amendments, emphasizing that through them, the legislators are trying to limit the autonomy of the PB. He exhorted the councillors and delegates to mobilize fast and strong to try to defeat these amendments in the Cam-ara. He concluded, "They want embarrass the PB. This is war and when you are at war you don't stop the war to prepare and debate." Some councillors were displeased by this comment and asked for more time to analyze the amendments because, after all, it is in the interest of the PB to cut some of the executive expenditures. One of them said, "I don't agree with X [the GAPLAN representative]. This is not a war. We are democratically debating and discussing with the Camara de Vereadores.... I don't disagree with the proposal of the coordinators, but the proposal is also a way of appropriating the issue. If we are going to discuss what is the autonomy of the COP then there is much more to be discussed." Implied in this comment is, of course, that the issue of the autonomy of the COP must be raised not only vis-a-vis the Camara but also vis-A-vis the executive.

On and off, the issue of legalization of the PB has been discussed in the COP. Some councillors have favored some kind of legalization. Others have opposed it in the name of the autonomy of the PB. In this debate, the international recognition and praise of the PB are often mentioned. One of the councillors commented in the COP meeting of March 3, 1997, "The way the PB has been operating in the last eight years without any regulation by the government is what makes it possible for us to go ahead and be internationally recognized."

The issue of legalization will probably remain as an unresolved tension in the PB of Porto Alegre for some time. The PB has, indeed, destabilized the old ways of doing politics in Porto Alegre, and the Camara de Vereadores is trying to reconstitute its political space in the new political conditions created by the PB. This reconstitution may, nonetheless, reveal some unexpected continuities with the "old ways." Indeed, the PB is contributing to expand both the political class and the circulation within it: two former PB councillors are now deputies of the Camara, and other former PB councillors hold positions in the executive.

CONCLUSION

The PB is a very dynamic social and political process, and it is difficult to make it yield many conclusions or to make a lot of projections on it. Up until now, the PB has been a remarkable way of promoting the participation of the citizens in decisions concerning distributive justice, effectiveness of decisions, and accountability of the municipal executive and of the delegates elected by the communities to the COP and the Fora of Delegates. Its success has been widely acknowledged not only in the city and in Brazil but also internationally. Over eighty Brazilian cities, 90% of which are run by the PT,"3 are adopting the PB system in some form or other, while international organizations have become increasingly sympathetic to it, although more interested in its technical virtues (efficiency and effectiveness in resource distribution and utilization) than in its democratic virtues (the sustainability of a complex system of democratic participation and distributive justice). The attractiveness of this system is such that recently, a small city contiguous to Porto Alegre (Viamao) had to be dissuaded by the executive of Porto Alegre from its desire to abandon its own existence as a municipality and turn itself into a region of Porto Alegre.

Over the years, the attractiveness of the PB has also become more transclassist in nature. While in the beginning, the middle classes, fearful of the "demagogic" democracy and of the PT's "fiscal rage," stayed away from the PB, they gradually became more and more involved in the PB. Middle-class participation has increased, first, because the popular administration was more effective, was virtually free of corruption in the use of municipal resources, and improved the kind of general services particularly cherished by the middle classes (such as garbage collection, public spaces, gardens and parks, and cultural activities) and, second, because the whole mobilization and public discourse around urban issues enhanced the self-esteem of the city as a whole, an elusive but nonetheless crucial symbolic urban value. A transclassist pride in the city developed in a way that is quite similar to what happened in Barcelona.'

Given the dependence of the municipal budget on federal transfers, an exogenously generated fiscal crisis may endanger the sustainability of the PB by undermining its capacity to deliver in a context of growing popular demands and expectations. The most recent budgeting crisis illustrates well this danger. The situation poses deep dilemmas because, to my mind, the future of the PB greatly depends on how its principles and practices of democratic participation are strenghtened and extended to areas or issues that are up until now not included in the PB45 and on how its autonomy is improved and consolidated so that the rupture with the old clientelist politics becomes irreversible.

The evaluation of the PB in the last part of this paper shows that these conditions are very demanding, involving difficult dilemmas. For instance, the consolidation of the PB only makes political sense as a rupture vis-a-vis the old clientelist-patrimonalist system. But is such consolidation possible without some form of continuity with the old system? As an emergent political reality, the PB tends to be destabilizing both in political and in ideological and cultural terms. However, a destabilizing idea that succeeds in converting itself into a sustainable practice always runs the risk of losing its destabilizing potential as it succeeds. The routine of mobilization calls for the mobilization of routine. Participation will remain high, but common citizens will gradually be replaced by specialized participatory citizens. The dilemma here is that while further radicalization of the experiment is the only weapon against routinization, there is an undeterminable threshold beyond which radicalization will irreversibly compromise the success of the experiment. There is no way out of this dilemma. However, the tension it creates may be itself sustainable-therefore contributing to the continuing, if always problematic, success of the experiment-provided that the participants engage in reflective self-subversion: a constant radicalization of political consciousness focused on the limits of the radicalization of political practice.

Another dilemma haunts the PB that, however, has less to do with the experiment in itself than with the interpretations and evaluations of it by both scientific and political observers and analysts. In a historical period of structural pessimism-as is ours, both the end of the century and the end of the second millennium-there is a tendency to be too complacent toward what exists and is familiar and too suspicious vis-a-vis what is merely emergent and thus unfamiliar. Suspicion consists of interpreting all the characteristics and developments of an emergent destabilizing reality as steps or movements toward a final and inevitable failure. This interpretation of the announced death, which can come from the left as well as from the right, operates as an intellectual trap. It sets the trap, and the rare bird of realistic utopia will sooner or later fall into it.

Related to this interpretation of suspicion is the interpretation of abbreviation. It consists of interpreting the institutional innovation in isolation from its historical and sociological embeddedness and specificity, thereby reducing it to a few abstract traits composing a model to be applied elsewhere by expert knowledge. Trapped between the interpretation of suspicion and the interpretation of abbreviation, popular initiatives such as the PB are subjected to a cruel dilemma: they are either conceived of as not working and bound to fail and therefore must be discarded as foolish utopias, or they are conceived of as working and bound to succeed and are adopted by the World Bank, where they are ground, pasteurized, and converted into new appendages of conditionality.