Deepening Democracy Through Public Deliberation? Reflections from the Province Huanta in the Peruvian Andes

En los ultimos anos se han realizado intentos de encuadrar la participacion deliberativa dentro de la teoria de la democratizacion. Leonardo Avritzer, uno de los mas optimisticos teoricos de la democracia deliberativa sostiene que las nuevas practicas democraticas se desplazaran del nivel social a la sociedad politica en un ‘espacio publico deliberativo’, consolidando de este modo la democracia. Razonando en contra de la ola de idealizacion y de las concepciones normativas de la democracia deliberativa, este articulo trata los intentos politicos gubernamentales de lograr un proceso politico de negociacion y participacion en el seno de la sociedad civil en la provincia de Huanta, en los Andes peruanos. El estudio abarca el periodo autoritario del ex presidente Alberto Fujimori (1996–2001) y el subsiguiente regimen democratico (2001–2006). El articulo procura demostrar que los estratos marginados de la poblacion no son capaces de explorar plenamente los potenciales politicos y economicos de los procesos politicos deliberativos si no son respaldados por organizaciones externas e instituciones estatales. Al contrario, estos sectores corren el riesgo de ser explotados aun en mayor medida que antes, al ser abdicado el poder del Estado en beneficio de los tradicionales detentadores del poder, individuales y colectivos. English: In recent years attempts have been made to ground deliberatory participation into democratization theory. Leonardo Avritzer, one of the most optimistic proponents of deliberative democracy, states that new democratic practices will be transferred from the societal level to political society in ‘deliberative public space’, thus consolidating democracy. This article deals with governmental political attempts to achieve a political process of negotiation and participation within civil society in the province Huanta in the Peruvian Andes. In this respect, it goes against the wave of idealization and the normative assumptions of deliberative democracy. The study spans the authoritarian period of the former President Alberto Fujimori, 1996–2001, and the subsequent democratic regime of 2001–2006. The article aims to demonstrate that the marginalized strata of the population will not be able to fully develop the political and economical potentials of the deliberatory political processes unless they are backed by external organizations and state institutions. Rather, these strata will risk being even more exploited than before, when the state’s power is abdicated to the benefit of traditional individual and collective power-holders.


I. INTRODUCTION
The crisis of representative institutions in Peru after the fall of the authoritarian regime in 2001 has brought deliberatory democracy to the center of political discussion. In this optimistic discourse, popular participation in deliberatory processes (in contrast to representation through political parties), will not only democratize and strengthen the public spirit but also render the local political process more effective. Through a kind of local assemblies (mesas de concertación), a rational debate should replace power politics. Nevertheless, popular participation was also a central feature of the politics of the authoritarian regime, as a means to extend the control of the central regime. In this context, political scientist Martin Tanaka argues, that civil society participation in Peru has not contributed to democratic attitudes, due to the vulnerability of the local associations to clientelistic penetration. On the other hand, members of social organizations manifest a higher degree of tolerance to authoritarian values than those who do not participate (Tanaka 2001:23).
The difficulties with the government's political attempts to achieve a process of negotiation and deliberation within civil society are clearly illustrated in the Andean province of Huanta. As a strategic spot for the devastating conflict between the guerrilla, Shining Path (Sendero Luminoso), and the National Armed Forces, the province is also renowned as a stronghold for 'el fujimorismo'. Huanta is also one of the few examples in the Peruvian context, where the processes have persisted for over a decade. The studied period spans the authoritarian period of the former President Alberto Fujimori, 2 1996-2001, and the subsequent democratic regime 3 of 2001-2006. A comparison is made between La Mesa de Concertación 4 during the first period and the participatory budgetprocess 5 The article draws on an analytical framework of deliberatory democratic concepts of Leonardo Avritzer, Jürgen Habermas and John during the second period.
Dryzek contrasting these optimistic theoretical scholars with researchers having a more empirical and less idealistic approach, like for instance Günther Schönleitner. The author uses Schönleitner's broad definition of deliberation as a "discursive process in which equal and free individuals reach collective decisions through reasoning, argumentation and persuasion" (Schönleitner 2004:79). The article's analytical point of departure is the question of how the different purposes (authoritarian versus democratizising) of the deliberative arenas are perceived at the local level. How are the deliberative processes counteracting and/or reproducing the social stratification in the provincial society? Were the new urban elites during the two periods really prepared to incorporate the interests of the most excluded groups in the rural areas in the dynamics of the process? A number of analytical reflections of the structural relations between mainly urban and rural social actors will be presented with focus on the continuity during the two periods which is explained by the fragmented, subsistenceoriented civil society in the post-war society. As the case study suggests, effective deliberation depends on the existence of a tocquevillian, 6 The article is based on a fieldwork between July and August in 2006 in the province of Huanta in the Peruvian Andes. Fifty two semi-structured interviews have been conducted with a sample of leaders in social organizations, politicians and employees in the provincial authorities and NGOs. It is important to make clear that the article does not pretend to evaluate the concrete outcomes of the process, but rather present some analytical reflections of the perceptions of the actors. Even though the statements of all actors have not been included in the study, a careful selection of the testimonies has been made, in order to contrast the perceptions of different informants.
prodemocratic civil society and relative socio-economic equality. In the absence of these demanding conditions, the risk is that the deliberative arenas become yet another arena for the local elite to reproduce existing social stratification.
The article starts with a brief overview of some theoretical ideas concerning the consequences of the deliberative spaces in a context of clientelistic and vertical power structures; thereafter, a short social and historical background of Huanta is presented. Subsequently, the study analyzes the dialogue processes during the authoritarian and the democratic governments followed by a final discussion of the theoretical implications.

II. THE 'NEW POLITICS' IN A CONTEXT OF VERTICAL POWER-STRUCTURES
Peruvian sociologist, Julio Cotler, describes the organizational structure in the Peruvian rural areas before the land reform in 1968 as "a triangle without a base" (in Diez 1999:15). He referred to the absence of horizontal ties between the subordinated, landless peasants and to the patronage relations that were central to the understanding of social stratification between the landowners and the peasants. These concepts are however still relevant in order to explain the patterns of articulation of the citizens versus the local government. In a modern, semi-democratic context, patronage is usually understood as particularized exchange of votes and support for goods and favors between subordinated groups and the political elite. While political clientelism could be defined as "a more or less personalized, affective reciprocal relationship between actors, or sets of actors, commanding unequal resources and involving mutually beneficial transactions that have political ramifications beyond the immediate sphere of the dualistic relationship" (Lemarchand 1972:151).
One risk with political institutionalization for subordinated groups is getting co-opted. Co-optation was defined long ago by Selznick as "the process of absorbing new elements into the leadership or policydetermining structure of an organization as a means of averting threats to its stability or existence" (Selznick 1966:13). Thus, co-optation in this context means being absorbed without any real power-sharing occurring.
Certainly, the vulnerability varies widely among the organizations, depending on their capacity to negotiate with the state and struggle for social and political space. Political scientist Jonathan Fox argues that dispersed populations at a strictly local level have the greatest difficulty in defining their common interests, and are therefore vulnerable to the "divide and conquer efforts from above" (Fox 1996(Fox :1091, while organizations at a regional level can realize collective action by overcoming locally confined solidarities. John Harriss is another scholar who stresses the need of alliances in both horizontal and vertical directions. He refers to this as bonding, bridging and linking social capital. Bonding refers to the strong ties between people in a limited geographic area. Bridging refers to the weaker ties between people with a similar socio-economic background and political influence but with different geographical, ethnic and occupational backgrounds. Linking may often have similarities with the earlier mentioned patronage relations, as it refers to the ties between poor people and those in influential positions in formal organizations (Harriss 2002:86).
Since the mid-1990's the locally confined associations in Peru have increased dramatically. However, anthropologist Ludwig Huber considers this problematic since many of the associations (the women committees are the most discussed example) are dependent on external actors (the state, NGOs) for their existence (Huber 2005:40). While NGOs in some cases contributed to organizational development through capacity-building and institutional support, it is still important to emphasize that power inherent in empowerment. Barbara Cruikshank argues that citizens are particular kinds of subjects created through the "technologies of citizenship" (for instance, educational programs and social services). For powerful agents such as NGOs to take an active role in this process is not necessarily the most effective way, since it is not possible to mold citizens into selfsufficient, politically active, empowered democratic citizens in this way (Cruikshank 1999:86).
Despite the image the NGOs like to give of being autonomous from the 'corrupt' and bureaucratic Latin American states and international development agencies, they are in practice dependent on financial resources from these public and private organizations (ibid:238). After 2001, NGOs gained an increased role in capacity-building to strengthen popular participation in Peru. From their technocratic perspective they proclaim a firm ideological belief in their ability to render local politics more democratic, without taking into account the enduring mechanisms of clientelism and the power struggles among competing groups. Romeo Grompone, a political scientist at IEP (Instituto de Estudios Peruanos), argues that the idealization of popular participation in Peru has led to an apolitical system where participation is being put against representation (through labour unions and political parties) (Grompone 2005:24).
Another central issue for successful democratization is the role of the central state to counteract the power of the local elite. Judith Tendler claims that improvements in local government depend on a tripartite dynamism between the local government, civil society and an active central state (Tendler 1997:145-6). More common, however, is a network-like, regulating state structure, absent at the local level. 7

III. DELIBERATION AND INEQUALITY
John Sidel argues that this structure produces the conditions for local bosses to establish themselves in the local political arena. It is not the traditional oligarchic power elite with its base in society that constitutes this elite; it is rather the weak network state that gives rise to new local bosses in the decentralized politics (Sidel 2004:53).
In recent years, deliberative participation has been presented as the solution to failing elitist political systems imposed from above. Leonardo Avritzer (drawing on the ideas of Habermas), argues that a deliberative public space 8 will link and transfer new democratic practices from the societal level into the public decision-making, and in this way transform elite practices into full democratization (Avritzer 2002:5). He thereby assumes that there is a fundamental difference in the political attitudes and practices between civil society and political society. The first is seen as characterized by democratic renewal while the second is characterized by clientelism and authoritarianism (Schönleitner 2004:78). While Schönleitner emphasizes that bottom-up democratization via deliberative public spaces may not occur due to the congruent attitudes and practices in the civil and the political society, he argues that pressure for democratic renewal also could flow from the political society (e.g. committed governments), rather than from institutionalized public spaces (ibid:79).
Deliberation could be defined as a discursive process in which equal and free individuals reach collective decisions through reasoning, argumentation and persuasion. While in liberal democratic theory the preferences are unchangeable (they are prior to the political process), in deliberative theory the assumption is that the preferences are (and should be) transformed through political interaction (ibid:79). Grompone claims that the entire democratic value of deliberation is dependent on the actors' willingness to change their preferences during the process in the search of the 'common good' (in Remy, 2005:80). While these authors assume consensus-building as the goal of deliberation, replaces consensus with the more realistic aim of ´reasoned agreement´. Since deliberation itself "eliminates preference orderings which cannot be [publicly] defended" (Dryzek 2000:49) it is not necessary to restrict the preferences or the options of the participants. The deliberative process itself encourages the actors to argue in favor of public interest and thus they acquire a public spirit during the process (ibid:47). According to Pellizoni the public spirit could be acquired by replacing the 'myth of the best argument' with mutual recognition. Through the joint management of concrete problems, an understanding and mutual recognition can be reached. Moreover, as a consequence, a redefinition of the division of labor and competences with respect to these concrete problems can emerge (in Schönleitner 2004:86).
However, it should also be noted that the inability to handle interest conflicts persists, which opens the door to bargaining, strategy and manipulation. Pierre Bourdieu claims that deliberation and participatory democracy tend to reproduce hierarchies. He criticizes what he calls a 'linguistic communism' on two points. Firstly, they tend to reproduce class hierarchies due to the uneven distribution of speech opportunities between distinct social groups. Secondly, they reproduce hierarchies of political competence of 'experts' over non-experts. Language is a medium of power through which relations of power between the speakers is articulated. The privileged class has a statutory ability, meaning that "not all linguistic utterances are equally acceptable and not all locators equal." (Bourdieu 1991:146). Jane Mansbridge draws attention to the problematic in a city council: These patterns imply that the psychic costs of participation are greater and the benefits fewer for lower status citizens. In contacting town officials, for instance, they feel more defensive beforehand and less likely to get results afterward. In speaking at meetings they feel more subject to ridicule and are less likely to convince anyone. Each act of participation not only cost them more but also usually produces less (Mansbridge 1980:103).
The main point here is that the consequences of inequality are not sufficiently taken into account in deliberative democratic theory. Therefore, the 'ideal speech situation' 9 Furthermore, as it is the state that invites organized civil society to participate in Peru, a central question is which groups are invited to participate. Since decisions are taken by consensus, plurality within the participating group is more important than sheer numbers (that is central in majority voting). It could, however, be difficult to reach an agreement when very distinct groups are invited. Theoretically, each is accountable to all, in a deliberative forum. A broader accountability that transcends geographic boundaries, classes and interest groups is required (Schönleitner 2004:81). However, the empirical material shows that in practice, the participants often defend the interests of their own constituency fearing that they might be questioned for not attending to the specific needs and interests of their own group. A related problem is that since only the organized interests are represented, this could give rise to a situation in which many social groups -perhaps even a majority -experience that they are left without representation. These arguments suggest that consensus politics demand the prior existence of profound communitarian social structures, rather than giving rise to them. It is questionable whether or not these exist in many of the Third World localities where participatory reforms are implemented with the encouragement of international development agencies that are attracted by the idea of empowered citizens constructing democracy from below. may not occur and could evolve into strategic calculation and manipulation.

IV. SOCIAL AND HISTORICAL BACKGROUND OF HUANTA
The province of Huanta is situated in the region of Ayacucho, in the central Peruvian Andes. It is divided into seven districts distributed over three areas: the valley, the highlands and the tropical rain forest. 10 The ethnic and class structures are heavily linked to geographical factors. The indigenous population in the rural highlands is the group most subjected to economic, political and cultural exclusion. Certainly, the 1969 land reform ended the hacienda system, but the clear social stratification between urban and rural areas has persisted. The land reform was a slow and conflictive process that left a feeling of discontent among the peasants. In the highlands, it led to a power vacuum (Coronel 1996:38) which facilitated the expansion of the Shining Path in the late 1970's. 11 As a result of the atrocities 12 Nevertheless, the consequences of the civil war were devastating for the rural peasant organizations since entire communities disappeared and many leaders were killed. Nowadays, the province is one of the poorest in the country with 92 per cent of the rural population without their basic needs satisfied (Ávila 2003:112). The social programs have therefore played an essential role in the economical re-construction of the region. However, the social programs caused the rural population to become linked and dependent on the authoritarian regime. While the paternalistic politics of Fujimori fragmented the challenging civilian and political representation, they also contributed to the creation of organizations directly dependent on the state, like for example the women's committees. Fujimori had a high level of support in the rural areas of the province and it was regarded as one of the strongholds of the regime. 42 per cent (mostly in rural areas) supported him in 2000 (ibid:144). committed by the guerilla and their lack of respect for the traditional organizational form, the peasants organized their resistance in Self Defense Committees supported by the National Armed Forces that had controlled the region since 1983. The Peruvian historian, José Coronel, stresses that even if the Self Defense Committees clearly were subordinated to the National Armed Forces, the peasants in Huanta succeeded in negotiating arenas of increasing independence (Coronel 1996).
The paternalistic practices are deeply reflected in the political and social structures of the Huanta province. The weak political and representative institutions permitted the emergence of strong caudillos, without mediation of collective organized interests and political parties. The paternalistic tendencies are clarified by Victor Ramos, a former mayor in Huanta: "To be mayor is to be the father of the society that has to take care of all" (in Degregorio et al. 1998:19). In general terms, the weakness of the political institutions in Huanta reflects the situation in Peru as a whole, with a fragmented party system and the emergence of 'outsiders' that appear as transitional electoral phenomena (see for example Lynch 1999;Tanaka 2005;Grompone 2005). The independent movements have no permanent organization but mobilize support close to the elections only to disappear shortly thereafter. Weak representative institutions also reflect a broader trend in Latin America that has prevented the construction of durable majorities, resulting in deadlocks and institutional crises (Levitsky and Cameron, quoted in Van Cott 2005:5). However, during the authoritarian regime of Fujimori the weakness of the representative institutions was reaching a culmination and political brokers came to play a key role in mediating state-society relations.
V. DELIBERATION UNDER THE AUTHORITARIAN REGIME (1996)(1997)(1998)(1999)(2000)(2001) The deliberative process, La Mesa de concertación, was introduced as a way to reconstruct civilian political power when the Shining Path and the military lost power in the province. Since Huanta was one of the most affected zones, the presence of public institutions and international development agencies was very high after the war. Through these organizations money poured into the province, though in an un-coordinated way. From 1993 to 2002 the annual budget increased from 6 to 43 million soles (Arieta 2006:29).
In 1995, the independent political movement Paz y Desarrollo, led by Milton Córdova, won the provincial elections with the promise to "reconstruct the province through technical efficiency and popular participation" (Ávila 2002:590). The movement consisted of a group of professional, young leaders without earlier political experience that were deeply influenced by the participatory discourse of the NGOs. Paz y Desarrollo succeeded in unifying a broad political alliance constituted by actors such as different human rights associations and local fujimoristas. Nevertheless, Córdova did not rely on a party apparatus that could bring institutional support to his leadership.
During the first two years of the creation of the Development Plan, the process was successful in including the indigenous population in the political process 13 Nevertheless, the creation of the dialogue process was also a pragmatic way to play politics by restoring the power of the municipality. Since the deliberative arena had some influence (despite its lack of direct control) over the resources that were distributed, it was not only a strategic way to increase Paz y Desarollo's power in negotiating with the NGOs, but also a way of increasing legitimacy among the citizens as the municipality appeared to be promoting the investments. 14 Institutions of 'mesa de concertación' 1) General assembly. Composed of representatives from the social organizations or the registered public and private institutions. Its function is to define the general principles and approve the provincial Development Plan. 2) Thematic groups. Road network, economy and production, health and food, education, housing, environment and institutional development. Each group consists of a coordinator and a technical secretary. The purpose is to produce development plans and projects. 3) Technical secretariat. Composed of the technical secretaries of the thematic groups. Its function is to give technical support to the thematic groups and develop the provincial Development Plan. 4) Coordinating committee. Composed of the provincial and district mayors and the coordinators of the thematic groups. Its function is to coordinate and evaluate the process and to approve the projects developed by the technical secretary. 5) Presidency. The provincial mayor. His/her function is to represent 'la mesa de concertación' in public and private institutions. Source: Avila (2003). However, the position of Paz y Desarrollo in the loose network of NGOs and public institutions was too weak when it came to the implementation phase of the Development Plan. In fear of losing the coming elections to the traditional elite from the APRA-party, the mayor split his own provincial movement and joined the national movement of Fujimori that, as a result of the successful victory over the guerilla, already had great support among the poor and rural populations. 15 Apart from the need for a national ally, it was a strategic way of responding to the social demands by entering the clientelistic network of the authoritarian regime. 16

VI. PARTICIPATION WITHIN A WEAK REPRESENTATIVE SYSTEM
The deliberative arena has passed through different phases and the impact of local and national actors are clearly reflected in the participation patterns. In the table below, we can observe a significant change in the actors who participated in 1997 and in 2000. It is worthwhile emphasizing that in 1997 the deliberative process was still in the planning process, while in 2000 the implementation phase had started. The significant decrease of the social organizations and the districts/indigenous communities (many of them rural) could be seen as an evaluation of the concrete effects in terms of implementation of developmental projects. As the social organizations decreased, simultaneously the NGOs and provincial authorities strongly increased their participation. In a sense, the deliberative process lost some of its initial social mobilization force and became a more bureaucratic arena for private and public institutions. This has meant that the representation of the rural population has decreased significantly. In 1997, they constituted 38 per cent and in 2000 only 8 per cent (Development Plans 1997 and. Formally, a great number of organizations participated in the dialogue process. Ávila mentions sixteen social organizations that were registered (Ávila 2002:595-96). Indeed, not all of these participated on a regular basis (Development Plans 1997. In general terms, the committees for women, self defense and irrigation had the highest attendance rate at the meetings. With the exception of the Irrigation committee (that exclusively organized the order of priority for irrigation), these have their dependence on the state and NGOs in common.
The Self Defense committees 17 in Huanta were created in 1983 and later became tightly connected to Fujimori who provided them with weapons in the struggle against the guerilla. After the pacification of the province, former leaders who had fled during the war returned to their communities. With their return came internal divisions and conflicts concerning representation which obstructed the strong solidarity ties within the indigenous communities. 18 The internal divisions were expressed openly in meetings where different leaders claimed to be representatives of the same community. 19 The consequence of this fragmentation was that the highland population did not succeed in participating effectively in the dialogue process in the longer perspective through their community leaders or municipality authorities. After the initial massive mobilization of the rural district leaders, the representation of the rural areas was taken over by Llaqtanchikta Qatarichisun, an organization promoted by the evangelical World Vision. 20 However, their economic resources made them important actors with institutional influence. According to José Tavara: They were appreciated and their participation was important. When Llacctanchikta and World Vision participated it was a multitude of participation.
They had resources and were a powerful force in La Mesa de Concertación. They were maybe the only group that had resources and they were making projects. World Vision was almost like a parallel municipality. 21 However, according to other critical voices within the local society, the internal dynamic of this organization was led by the technical team of World Vision. They tended to mix their different roles (religious, political and social). According to Artemio Sanchez: The technical team is paid and responsible to World Vision and not to the leaders of Llaqtanchikta Qatarichisun. World Vision was governing the organization through the technical team that has the office here in the urban area. They participated in the Mesa de Concertación. Why? Because they have money, the peasants were just accepting their decisions. 22 The dependence on World Vision (or the vertical linking social capital) undermined the possibilities of the members of the organization to claim responsibility, which also undermined their legitimacy as collective decision makers. A consequence of the weak organization in the rural areas was that the poorest population could not put forward their demands. This is related to the earlier discussion of the necessary cross-boundary representation in deliberative arenas. In order to be democratic, the actors must be willing to change their preferences during the process in the search for a communitarian, common good that crosses geographical, social and ethnic boundaries. The consensus-oriented, participatory discourse of the NGOs and the political elite was, however, not shared by most citizens in the rural areas who instead had a more instrumental perspective. They viewed the process as an opportunity to apply for support from the mayor or create contacts with NGOs, which illustrates the continuity of the paternalistic structures. According to a poll about what the principal role of the mayor was, 83.6 per cent responded that it was to provide infrastructure, 8.2 per cent considered that it was to promote democracy and only 0.6 per cent answered that the mayor's main function was to promote participation (Ávila 2002:597).

VII. THE ROLE OF THE NGOS
The most salient NGOs in Huanta during the late 1990's were World Vision, CARE 23 and SER. 24 The latter two were closely connected to the provincial authorities (while World Vision focused their intervention in the rural areas). The close relationship between the provincial authorities and the NGOs can, in the first years, be described as an alliance (Degregori et al. 1998:39). They were deeply immersed in the provincial institutions and did not question or counteract the leadership around the mayor. 25 However, when Mayor Milton Córdova joined the political movement of Fujimori, SER distanced itself from the provincial level, focusing its work on strengthening the process from the rural districts (mainly Santillana). On the other hand, the NGOs were also dependent on the international development agencies. 26 According to Javier Torres at SER, they always take a very independent position in relation to the development agencies, but at same time they are well aware of the trends in the development world. He manifests that in the 1990's the NGOs in Peru were deeply influenced by a general participatory discourse: Habermas and public space were widely discussed, and deliberation was seen as the solution to the political exclusion of the rural population. 27 Despite being a relatively small NGO, SER came to have influence over the institutional design of the process. CARE and World Vision had more resources, which later became clear since they could employ technicians who worked within the deliberative institutions permanently. 28 One remarkable example is Augustin Soza Chamba, who was vice Mayor in 2001 while in the employ of CARE. 29 According to various critical voices, the will of the NGOs to empower the population also led them to push the processes and in some cases take control over setting the agenda. 30 La Federación de Clubes de Madres de Ayacucho (FEDECMA) emerged in the late 1980's. The federation consisted of eleven provincial affiliations, one of them the Provincial Federation of Huanta. Two leaders from the Federation are incorporated in the municipality to work with the implementation of the social programs assigned to the organization. In some cases, this provided the women a platform of negotiation with the provincial authorities and increased female political participation.
Hence, this is problematic since it entails power relations that restrict the possibility for social groups to define their identities and political alternatives. Nevertheless, the dependence on external actors was also in relation to the provincial authorities as with the example of the women committees. 31 However, other informants witnessed the weak presence of the Federation on the local political scene, which could be explained by the dependence on paternalistic social programs that inclusively result in an external influence in the internal decision-making of the Federation. 32 We have meetings every week, but just a few participate. But when we are going to distribute the milk everybody arrives and starts to discuss who are going to receive the milk first. They are not sufficiently prepared for the participation.
The president of the Federation reflects upon the difficulties with the subsistence-oriented perception of the members: 33 According to other informants, dependence on the goodwill of the mayor gives the committee a rather passive, legitimizing role 34 without the capacity to develop policy proposals and its own agenda. The proposals are often developed in the municipality and the women are instructed on how and at what moment they should go forward and hand in the applications. 35 The leaders who emerged in the of context of weak representative civil society seem to play the role of brokers with the ability to respond to the social demands by establishing linkages to the central state, development agencies and citizens, thereby taking over the conventional role of political parties as the primary link between state and society. Since the capability to create a network and knowledge has been prioritized, a technocratic leadership has been consolidated through the deliberative process. This is an inescapable dilemma inherent to deliberation -that rhetorical capacity brings power.

VIII. THE POWER OF KNOWLEDGE -FROM LANDOWNING TO EDUCATION
Paradoxically, the political violence has also meant significant democratization of the local politics, since the traditional political elite migrated to Lima. For the first time, indigenous peasants could hold political positions as mayors or councillors. Many leaders have studied outside the community. Thus education plays an important role in reestablishing social structures of inclusion and exclusion. The Mayor, Milton Córdova, himself came from a humble family, but has, through academic studies, achieved a career within the state administration. A majority of the actors in his political movement have the same social background. 36 Nevertheless, in rural areas the majority is quechua-speaking and illiteracy is widespread whereby the differences in the level of education between urban and rural areas still persist. José Coronel states that the word 'ignorante' 37 Among other things, the deliberative process was intended to be a form of democratic control of the local government. The model and the rules that regulated the access to information and control over the agenda were elaborated by the employees of the provincial authorities in collaboration with NGOs. They were also responsible for inviting the participants, which could be somehow problematic since they were both elaborating the rules and chosing the actors that would control them. The highest authority was the mayor. According to Elida Ramirez, promoter of is often used to describe the poor, which illustrates the importance of education both as a posible democratizing force but also as a way of reproducing social exclusion. The following analysis traces how this social stratification was incorporated in the institutional design of the deliberative process. The focus will be on how the poorest, largely indigenous rural regions were articulated through these institutions.
the NGO Sisay, 38 there are certain advantages with having the mayor as president since the financial resources of the province could be used to pay for the transportation of rural participants. However, it could also be negative since the mayor, as she considered what had happened in the case of Milton Córdova, could use the deliberative institutions for his own political purposes. 39 The dialogue process consisted of six committees that should have a technical, supportive role and a general assembly with decision-making power over the general guide-lines and the ratification of the Development Plan. The committees consisted almost exclusively of the population from the urban areas. The majority of the presidents were provincial councillors representing Paz y Desarrollo (Development Plan, 1998: 65). Employees of NGOs (especially CARE) played important roles within the committees not only in maintaining the institutional structures but also, according to some critical voices, by trying to integrate their own projects in the plan, presenting them as popular demands. 40  Then what perception did the rural population have of their participation? According to Feliciano Medina, (Mayor in Santillana 1996: admits that NGOs came up with proposals, but claims that this was due to the incapacity of the rural population to come up with their own independent proposals. "La Mesa de Concertación was an arena with institutions with great financial resources that supported us […]. They were specialists, anthropologists and economists and they came from other countries to teach us about politics and how to work with dialogue processes". 42 Marisol Ovalle Roca, a poor urban street-vendor, did not dare to put forward her proposal about making waste containers available on the streets, since she felt that she could not express her ideas in a proper way. 43 Since the roles were so divided, the citizens were never integrated in a close co-operative form. Instead the definition of the problems and the development of a plan were performed by the technical teams while the final plan was written by two specialists from Lima who did not have a deep knowledge of the local context (Àvila 2002: 596). The rural population informed the technicians about their material needs which later were not necessarily integrated into the plan. According to José Tavara, an NGO-employed in Huanta, the committees exhibited unwillingness to integrate the opinions of the rural population in the plan (Tavara 1999:32).
Similar experiences that illustrate a vertical relationship where knowledge and economic resources constitute barriers preventing an equal dialogue were expressed by various informants in rural areas.
In other cases, the demands were included, but then disappeared when the plan was updated which illustrates the difficulties in creating substantial change through the deliberative processes. The majority of the rural population defines their strategic interest as being agricultural development (SER 1997:23-25) During the period 1996-98 only two per cent of the provincial budget was invested in this area (SER 2006:19). According to German Analla, former leader of the Provincial Federation of Self Defense Committees: The rural population got tired. They do not have confidence any more, more words than practice. They went to represent their village and the villagers had contributed to the transportation. They had to bring back some concrete benefits for the village, if not it will be difficult to collect money at the next meeting. 44 The cost of attending the assembly was clearly weighed against the potential material benefits, thus the process was seen rather as a form of resource allocation from above than an arena for political advocacy from below. This sort of linking social capital can also be understood as a reproduction of personalistic leadership, or the caudillismo, which is a feature that is deeply entrenched in Latin American politics.
Another relevant aspect is the limitation of the public spaces in the presence of military bases in Huanta. The presence of the military in the meetings restrained the dialogue. Javier Torres, at SER, claims that it was not until 2000, when the military bases closed, that the population started to speak openly. At the same time, the civil war had left deep divisions between the groups that were for and against the guerilla, resulting in implicit conflicts not mentioned in the dialogue process. 45 To sum up, as the elites were unwilling to give up their power, the questions that threatened the status quo were never dealt with in the relevant decision-making arenas. In the absence of representative civilian and political organizations, a technocratic leadership emerged. This new local elite used the arena for networking with the central state and the NGOs, while the majority of the population was integrated in the political process through clientelistic mechanisms.
The deliberation became a way of covering all the micro-conflicts that threatened the consensus and the reconstruction process.

IX. DELIBERATION AND THE NATIONAL DEMOCRATIZATION PROCESS
With the fall of Fujimori in 2001, the transitory regime paradoxically institutionalized popular participation through a far-reaching decentralization reform that was aimed at strengthening local politics through the consensus model. The principal purpose was to counteract the socioeconomic cleavages and the clientelism of the authoritarian regime through strategic planning and public deliberatory processes. 46 Thus, in contrast to the apolitical focus of Fujimori, the aim was to create a more political process. In 2003, the related participatory budget 47 The question is how viable the implementation of deliberative institutions was in light of the weak civil and political societies in Peru. Was it really possible to democratize a state by using the same participatory reforms of the previous authoritarian regime to extend its control over society? The sociologist Irma del Águila Peralta argues that the dialogue and the budget processes tended to strengthen apolitical identities without connections to the general interests in society. A mosaic of special interests contributes to the fragmentation of the political system, rather than strengthening broad political identities based on class or ethnicity (Águila Perala 2005:40). The question is whether the focus on popular participation may be used as a pretext to avoid those institutional reforms which are essential for strengthening the representative system and the state. process was legislated. A criticism that has been raised against these institutional reforms is that they are part of the dismantling of the developmental state, since they bring about a market logic of supply and demand where the state has adapted itself to the demands of the citizens.
From the provincial perspective, the well-aimed reforms of the democratic regime do not have the expected impact and the process has been given increased autonomy by the central state. Still, the national processes are reflected in La Mesa de Concertación in Huanta. From the perspective of the rural populations, the fall of Fujimori is perceived as a withdrawal of the state from the local arena, since the dialogue process was strengthened through Fujimori's economic investment in the rural areas. The democratic government, on the other hand, has primarily a regulatory function that was perceived as imposed from above without respecting the earlier experiences in Huanta.
According to the district Mayor, Amador Barbossa: I do not deny that I have been a supporter of Fujimori. As I mentioned, in the time of Fujimori there was a lot of support. The [present] government works more with the macro economy, but not with the micro-level. Maybe the difference is that the policy of Fujimori was to develop, strengthen the micro economy and in that way get to the macro level. Nowadays, Toledo does the opposite, he does not care about these processes, he passes laws, but does not give any support. We can not see any big infrastructural projects in Ayacucho, even less in Huanta during the period of Toledo. You do not feel Toledo's presence here. 48 This argument reminds of Judith Tendler's tripartite dynamic, in which the central state plays an essential role in supporting the local processes. She, however, emphasizes the central state as counterbalancing the power of the local elite, rather than the resource allocating aspect (Tendler 1997). The local actors, however, seem to value the presence of the state, above all in relation to its transaction of economic resources. Viewed from this perspective, the second period offers fewer possibilities to network with the state. This state structure has more similarities with the network state which gives more autonomy to the local leaders.
At a provincial level, the death of Milton Córdova in 2001 led to the demobilization of La Mesa de Concertación, since there was no one left to promote the dialogue process. Clearly, the lack of efficient division of power has led to a process that was dependent on one dynamic actor. However, already before the death of the mayor, the process had lost its mobilizing aspect and was maintained by the mayor, his family and their allies in the civil society and the NGOs.

INEQUALITY
After 2001, the mesa de concertación in Huanta is in practice replaced by the participatory budget as the principal deliberative arena. The budget process is however a much more limited process that should ideally be connected to a wider dialogue process. The participatory budget is organized once a year. Huanta is divided into five different zones. In each district, the citizens participate in meetings to decide what projects they should apply for. 49 According to the Ministry of Economy's general guidelines, the objective is to: reduce poverty, develop the productive activities and to contribute to the goals of the local Development Plan. However, the participants define the more specific criteria (Ministry of Economy 2005:10). Owing to the lack of sufficient resources, a decisive selection criterion is to what extent the community can participate with their own resources (financial or manpower). 50 Since the poorest groups have fewer possibilities to contribute with their own resources, they are not favored by this system. 51 However, the results have been varied. In some districts, the process, despite difficulties, has been quite successful in increasing the transparency of the budget process (for example the urban areas of Huanta). According to José Coronel: In the urban areas it [the participatory budget] has been a mechanism of participation. To learn to prioritize, not like in earlier times with Fujimori who was imposing the projects. Now the people have to choose, in that way it is democratic. But it is a very reduced deliberative process, to just meet a couple of times every year and decide over a limited part of the budget. 52 Nevertheless, in the rural areas the districts often lack the organizing ability and the capacity to network with external actors to mobilize forces for a common project. 53 Establishing a relationship with an NGO for a specific project requires organizational capacity and information about the project cycle and the institutions which not all rural districts have access to. In Iquicha, the mayor who belongs to APRA, the political opposition, searched for an NGO to support the development of a technical report for a road construction. The project was approved in the budget process but the construction was delayed six months according the locals due to lack of political will. 54 The NGO-promoter, Javier Torres, considers that the local capacity to create substantial change was overestimated, while the weaknesses and external threats were underestimated. The NGOs did not take this aspect into account when they introduced the strategic planning processes. 55 Compared to successful examples of participatory democracy as in Porto Alegre (Brazil) and Kerala (India), these were supported by transregional political movements (in contrast to the micro-local civil society in Huanta). Jonathan Fox claims that scaling up the organization to a regional level is especially important for representing the interests of isolated and oppressed populations (Fox 1996(Fox :1091. In the rural areas of Huanta, this kind of horizontal solidarity ties could transcend the micro-local perspective that the actors were lacking. Instead, the weakness of the representative institutions led to a situation where the leaders were picked by the mayor, taking their verbal capacities as a criterion rather than their representativeness. 56 These participants lack a solid mandate from the group they formally represent. At the same time, their rhetorical capacity brings material benefits to the group. 57 However, there is an important difference in the discursive equality. While in the earlier period the process was centralized to the province, the budget process in the second period was decentralized to five different zones. As the participants were more homogeneous within the zones, the Quechua population in the highlands was no longer forced to compete verbally with the educated mestizo population from the urban areas. Consequently, the effect of the discursive inequality was reduced. Nonetheless, unequal resource distribution and dependence on external actors remain decisive factors that prevent the reforms from generating the expected results. By not taking these complex local power constructions into account, the reforms lead to the opposite result, tending to cement socio-economic inequality.
As during the authoritarian regime of Fujimori, the participatory budget process has continued to reproduce local 'brokers', rather than having strengthened the organized civil society.

XI. THE CONSENSUS-BUILDING AND THE ATOMIZATION OF THE BUDGET
The deliberative process is based on a communitarian idea that the local society can be united in a common project. Since consensus-building is a long-term process, it is difficult to achieve even in relatively homogenous societies. The inherent dilemma with consensus-politics is that different groups cannot express negotiate their interests in a way that conflicts with the interests of another group, and channel these demands through the democratic institutions. The consequence is the marginalization of issues on which it is difficult to reach a consensus from these institutions.
In the post-war society of Huanta, there were still deep antagonisms to be dealt with. Isabel Limasna suggests that people who participated in the war on opposite sides did not feel sufficiently confident to be able to collaborate. In the budget process, these conflicts were expressed. 58 The process is conflictive as all the representatives strive for benefits for their own interest group. They question why their project has not been prioritized and it creates verbal confrontations and conflicts between different groups.
The budget process was therefore in some cases a conflictive process since the participants were suspicious about the selection and implementation of the projects. According to Eloy Robles, the leader of Frente Defensa-Huanta: 59 The deliberative process became an arena for political struggle between different interests groups instead of creating consensus-based broad political identities. In the absence of a development plan or a comprehensive strategy, the demand of the citizens has been fragmented into short-term, local projects. 60 The central issue in the current situation is whether the inclusion of marginalized groups in deliberative processes really has a democratizing effect. Eloy Robles claims that it is impossible to reach an agreement in questions of substantial change as the elite groups are not willing to give up their power. 61 The question therefore, is whether participation in deliberative processes could not have the opposite effect since it takes time and resources away from other activities such as mobilization, protests and campaigns. In Peru in general, it has become clear that there are other non-institutional, conflictive forms of popular participation, such as for example demonstrations and protests, which are more effective for creating immediate change. Nonetheless, this more confrontative participation has not strengthened the relationship between the state and the society, constructing a participatory democratic culture from below.

XII. CONCLUDING REMARKS
This article set out to shed some further light on the debate about deliberative democracy through empirical study of the impact of deliberatory reforms in the province of Huanta during the period 1996-2006. It has been argued that deliberative democratic theory is based on essentially normative assumptions of the democratic practices in civil society, without taking into account the power-structures embedded in civil society that could prevent the emergence of an ideal speak situation. Consensus-building depends on the elite being prepared to give up their power. Since this was not the case, the urban elite in Huanta, in alliance with the state and the NGOs, revealed a notable capacity to handle potential challenges from below through the exclusionary institutional design of the dialogue process, particularly during the first period. Consequently, the questions and actors that threatened the consensus were prevented from entering the relevant decision-making arenas. However, it is important to point out that there are differences between the two periods; while in the first period the control over the agenda was stricter with local and national fujimoristas deeply entrenched in provincial politics, in the second period the participatory budget opened up for underlying conflicts at the microlevel. Yet, the opening of a most pluralistic political space was not challenging the traditional power-holders, but has created conflicts between different micro organizations that competed for the same scarce resources. The result has been the emergence of a new technocratic leadership around the political movement of Mayor Milton Córdova, in the urban areas of the province, without a clear connection to the social constituencies which they were supposed to represent. Overall, the process has produced conditions for 'local bosses' to establish themselves in the local politics. Since the vertical power structures and the local elite have not broken down, public deliberation has not led to any profound changes in the relations between the urban and rural sectors.
The case study of Huanta suggests that the most marginalized indigenous population in the rural highlands did not have the capacity to challenge the social structures since they were represented by local, subsistence-oriented organizations dependent on NGOs and the state. In the absence of regional organizations, external support for the fragmented civil society, necessary for establishing a counter-hegemonic force, was lacking. Jonathan Fox offers a powerful alternative explanation to the lack of success, emphasizing the need for scaling-up the organization of dispersed population in order to be able to challenge concentrated elite-power. Instead of establishing broad political identities and creating effective links between the districts and the communities, the deliberative process has reproduced the fragmentation of the peasantry, including them in the semiclientelistic structures of provincial politics. The fundamental issue of the state structure is also raised. A comparison between the two periods indicates that the central state plays an important role in supporting and institutionalizing the deliberatory reforms. This suggests that the decentralization reforms should be preceded by institutional reforms strengthening the state and the representative system, in order to be able to counterbalance the role of the local elite.
The question of how to achieve democratization remains. As argued in this article, the conditions for altering the way that politics are conducted through deliberatory, democratic practices are highly demanding. Understanding the organizational and structural relationships between different social actors is decisive to the outcome of the reforms. Further research could focus more explicitly on contexts in which it is appropriate to introduce deliberative processes in order to create institutional designs that could deepen democracy. Gustafsson, Birgitta Genberg and John Sudmeier and anonymous reviewers for valuable commentaries. 2 Alberto Fujimori (without a political party) defeated the centre-right coalition in the presidential elections in 1990. He implemented a radical programme of free-market economic reforms and reduced the role of the state in almost all spheres of the economy. In 1992, he dissolved the Peruvian congress with the support of the military. However, his authoritarian methods were widely accepted as necessary in the struggle against the Shining Path guerrilla. Fujimori was forced to flee after the revelation of one of the greatest corruption scandals in Latin American history. 3 Alejandro Toledo won the elections in 2001. 4 Mesas de Concertación is a deliberative process promoted by NGOs and local governments to integrate the civil society, public and private institutions in the elaboration, negotiation and implementation of strategic development plans. 5 The participatory budget is a more limited (compared to Mesas de Concertación) method to integrate citizens in the budget assignment process and in that way increasing the efficiency and transparency as well as strengthening the state-civil society relations. 6 The tocquevillian model emphasizes the democratic functions of associational life within a liberal and democratic state. According to this model, the associational sphere is autonomous and occupies a central space between the individual and the state. The essence of the tocquevillian approach is that apolitical associations will help to produce a compromising public spirit that is necessary for efficient democratic governance (Goodhart 2005). 7 Chalmers Johnson makes a distinction between the regulating state (focused on regulating procedures) and the developmental state (focused on results) (Weiss 1999:81). 8 Public deliberation will lead to participatory publics. This concept involves four elements:

NOTES
The first is the formation of mechanisms of face-to-face deliberation, free expression and association. The second is that social movements will change the political culture by introducing alternative practices. The third concerns the transformation of informal public opinions into an arena for public deliberation and decision-making. The fourth is that they connect deliberation with the search for institutional format with the ability at the institutional level to address the issues made contentious at the public level (Avritzer 2002:7). 9 Central to Jürgen Habermas' ideas is the concept of the 'ideal speech situation' which could be interpreted as 'fair play' in dialogue. All participants must have equal opportunity to participate. 10 The districts are Ayahuanco, Huamanguilla, Huanta, Iguaín, Luricocha, Santillana and Sivia.
It has a diverse ecological environment with valleys, highlands and jungles. 11 For an in-depth study of the emergence of the Shining Path, see Degregori, 1990. 12 The rapid expansion of the Shining Path during the first years is explained by the strong alliance between the rural teachers and the central level of the guerrilla-organization. This organization came to have an explosive effect in the context of poverty and ethnic exclusion in the middle of the frustrating struggle for citizenship rights. However, while the ethnic administration during the hacienda system was based on reciprocity and a certain integration of the traditional Andean rituals, the Shining Path demanded that the peasants (women and men) worked hard for them without giving them anything but ideological training in exchange. They prohibited the traditional religious calendar of the Andean communities and imposed cruel, almost medieval punishment as for example decapitations and stoning, in the main square. The victims were not permitted the traditional funeral rituals (Degregori et al. 1998:24). 13 Interview with José Coronel (anthropologist at Instituto de Estudios Peruanos), the 22nd of July, 2006, in Huanta. He emphasizes the importance of distinguishing between the two first years and the period thereafter. He argues that in the light of the domination of the traditional elite from the APRA-party, the fact that Paz y Desarrollo came to power meant a significant democratization of provincial politics. Since the leaders of the movement came from a simple background this signified a considerable breaking-down of traditional power structures. But according to Coronel, after two years it became clear that only the actors had changed, since the way of doing politics remained the same. 14 Milton Córdova used the Development Plan in his political campaign in 1998, despite the fact that it had been created by all sectors, across the political party boundaries. 15 Córdova was also supported by provincial 'fujimoristas' in the election of 1995, as the political movement of Fujimori did not participate in the provincial elections. As a consequence of internal divisions between the 'fujimoristas' and the more pro-democratic groups, the movements split up. Interview with Artemio Sanchez Portocarrero the 3rd of July, 2006, in Huanta. 16 Interview with José Coronel, the 22nd of July, 2006, in Huanta. 18 In 1996 the process of returning to the communities started. The process was difficult as people fighting on opposite sides had to meet and collaborate in the reconstruction process. 19 Interview with Javier Torres, the 8th of august, 2006, in Lima. 20 World Vision is a Christian relief, development and advocacy organisation dedicated to working with children, families and communities in overcoming poverty and injustice (www.wv.org). The activities of World Vision were focused on humanitarian aid in the highland region of Huanta. 21 Interview with José Tavara (employee at the NGO Servicios Educativos Rurales, SER), the 2nd of August, 2006, in Lima. 22 Interview with Artemio Sanchez (Provincial Councillor 1996, the 3rd of July, 2006, in Huanta. 23 CARE is a humanitarian organization focused on poverty reduction and emergency aid to survivors of war (www.care.org).
24 SER (Servicios Educativos Rurales) is a NGO focused on popular participation and local democratic processes. 25 Interview with José Tavara, employee at SER, the 2nd of August, 2006, in Huanta.