www.elsevier.

com/locate/worlddev

World Development Vol. 28, No. 1, pp. 2139, 2000 2000 Elsevier Science Ltd. All rights reserved Printed in Great Britain 0305-750X/00/$ - see front matter

PII: S0305-750X(99)00109-6

Participation and Accountability at the Periphery: Democratic Local Governance in Six Countries
HARRY BLAIR * Bucknell University, Lewisburg, Pennsylvania, USA
Summary. Democratic local governance (DLG), now a major subtheme within the overall context of democratic development, promises that government at the local level can become more responsive to citizen desires and more eective in service delivery. Based on a six-country study sponsored by USAID (Bolivia, Honduras, India, Mali, the Philippines and Ukraine), this paper analyzes the two topics of participation and accountability, nding that both show signicant potential for promoting DLG, though there seem to be important limitations on how much participation can actually deliver, and accountability covers a much wider range of activity and larger scope for DLG strategy than initially appears. 2000 Elsevier Science Ltd. All rights reserved. Key words democracy, decentralization, participation, accountability

1. INTRODUCTION As democratization generally has assumed a central role in the developing world over the past decade in both reality and international donor thinking, democratic decentralization has also taken on increased importance. Regimes have found themselves having to democratize at the local as well as at the national level, and donors have been attentive to supporting such initiatives. The major promise of democratic decentralization, or democratic local governance (DLG) as it will be called in this paper, is that by building popular participation and accountability into local governance, government at the local level will become more responsive to citizen desires and more eective in service delivery. In this paper, I will endeavor to analyze the two themes of participation and accountability in DLG in the context of a six-country study sponsored by the United States Agency for International Development (USAID) during 1996-97. The central ndings show both aspects exhibiting signicant potential for promoting DLG, though there seem to be important limitations on how much participation can actually deliver, and accountability covers a much wider range of activity and larger scope for DLG strategy than appears at rst glance. Finally, I will bring both aspects together to present a picture of their centrality to the general process of DLG.
21

(a) Democratic local governance as a donor initiative in the 1990s Democratic local governance as it is employed in this paper combines the devolutionary form of decentralization (in which real authority and responsibility are transferred to local bodies) with democracy at the local level. 1 Accordingly, it can be dened as meaningful authority devolved to local units of governance that are accessible and accountable to the local citizenry, who enjoy full political rights and liberty. It thus diers from the vast majority of earlier eorts at decentralization in developing areas, which go back to the 1950s, and which were largely initiatives in public administration without any serious democratic component. 2 This new mode of support for decentralization emerged in the later 1980s, in the wake of the democratization wave that swept so many countries toward the end of that decade and

* Research reported in this paper was sponsored by the

United States Agency for International Development, but interpretations or conclusions presented here do not necessarily reect USAID thinking or policy. They are the responsibility of the author, along with any errors or shortcoming. An earlier version of the paper was prepared for the annual meeting of the American Political Science Association, 36 September 1998, in Boston, MA. Final revision accepted: 25 May 1999.

22

WORLD DEVELOPMENT

that inspired both aid-recipient governments and donors to support democracy at local as well as at the national level. By the mid-1990s, USAID was supporting about 60 DLG activities around the world, and other donors were quite active in the eld as well, most notably the United Nations Development Programme, which during the course of the 1990s has assisted over 250 decentralization activities in various countries. Perhaps the most impressive testimonial to the perceived ecacy of DLG has been its endorsement by the Development Assistance Committee (DAC, which consists of all the principal bilateral donors) of the Organisation for Economic Co-operation and Development (OECD) in its 1997 ocial report on ``Participatory Development and Good Governance'' (OECD, 1997, espectully. I-11). At the decades end, DLG constitutes a major item in the assistance portfolio of most donors, and as such it deserves an analysis of its ecacy as a development strategy. (b) Participation and accountability as key themes in DLG DLG comprises a number of aspects in addition to participation and accountability performance in service delivery, resource allocation and mobilization, and degree of power devolution are among the most important ones. But what makes DLG dierent from earlier forms of decentralization is the inclusion of these two new themes. The central idea of participation is to give citizens a meaningful role in local government decisions that aect them, while accountability means that people will be able to hold local government responsible for how it is aecting them. Together, these two processes are what constitute the heart of the ``democratic'' component of democratic local governance. DLG has been advocated for what it is (or ought to be) and for what it does (or what it should do), i.e., as a process or end-in-itself and as a means to further ends, in this case the outputs of DLG. On the process side, through participation DLG promises to increase popular input into what local government does, and through accountability it bids to increase popular control over what local government has done or left undone. On the output side, DLG nds its justication largely in the ideas that it can improve local service delivery and that for a good number of donors it can contribute signicantly to poverty reduction as well. 3 In this

paper, I will concentrate on the two process themes of participation and accountability, while taking some issue with assertions about poverty alleviation. (c) Study, sample, methodology During 199697, USAIDs Center for Development Information and Evaluation (CDIE) undertook an assessment of DLG in six countries with ongoing programs. The primary aim was to distill from the experience of the last decade what USAID in particular and the international development community more generally had learned about DLG and how such knowledge might inform future donor initiatives supporting DLG. 4 The sample was chosen to include a variety of regions as well as a range of conditions in which DLG initiatives had been launched. In addition, given the rather unhappy track record of earlier administrative eorts at decentralization over previous decades, there was a strong incentive to nd current initiatives that showed some prospect of succeeding, so countries were chosen to include good cases rather than bad ones. The sample, in short, was purposeful and illustrative, not scientic or inclusive. It included two cases in Latin America (Bolivia and Honduras), two in Asia (India and the Philippines), and one each in Eastern Europe (Ukraine) and Africa (Mali). All six countries were essentially democratic at the national level at the time of the USAID assessment. Except for India, all the other countries had ongoing USAID-assisted DLG initiatives under way by the mid1990s (although Malis eort will be fully implemented only in 1999). None of these ve had begun before the 1990s, however, and it seemed worthwhile to include at least one system with a longer history in DLG. India had begun its eorts in this sector in 1959 with its Panchayati Raj program, which with occasional interruptions had continued into the 1990s. So although there had been no American assistance to Panchayati Raj for more than 25 years, the Indian state of Karnataka, which was reputed to have one of the most eective DLG programs in the country, was selected for inclusion in the study. 5 Assessment teams from USAID in Washington studied DLG in the ve independent countries, spending about three weeks in each one. Methodology consisted largely of key informant interviews, document review, and eld visits to a sample of local government units in

PARTICIPATION AND ACCOUNTABILITY

23

each country. The country studies were then written and published as separate reports. 6 All six studies were comparatively analyzed in a synthesis report, which emphasized several aspects of DLG; the present paper draws in signicant part on this report but attempts a deeper analysis of the participation and accountability themes. 7

higher that for national ones after DLG is introduced). Of course, political participation includes much more than voting itself, encompassing campaigning, demonstrating, lobbying decision-makers, and the like. The question, though, is whether the increase in participation leads to increases in the subsequent elements in the formula presented in Figure 1. (b) Representation

2. PARTICIPATION, REPRESENTATION, EMPOWERMENT, BENEFITS Much of DLGs attraction as a development strategy lies in its promise to include people from all walks of life in community decisionmaking. The hope is that as government comes closer to the people, more people will participate in politics. All sorts of constituencies women, minorities, small businessmen, artisans, parents of schoolchildren, marginal farmers, urban poorwill then get elected to oce (or have greater access to those in oce). That will give them representation, a key element in empowerment, which can be dened here as a signicant voice in public policy decisions that aect their futures. Local policy decisions reecting this empowerment will serve these newer constituencies, providing more appropriate infrastructure, better living conditions, and enhanced economic growth. These improvements will then reduce poverty and enhance equity among all groups. As a formula, the argument 8 could be written as in Figure 1 (with the D symbolizing ``an increase in''). (a) Participation The very act of launching a DLG initiative ensures a certain degree of participation, in that citizens will be voting for elected local council members. This tends to increase local interest in politics and so involves people, in the political process (often turnout in local elections is

Figure 1. DLG causal formula.

Assuming that a central government has a genuine commitment to devolving power, the very fact that democratic decentralization is taking place means many new constituencies can gain representation through public oce. Businessmen, local notables, large farmers, professionals, and possibly some labor leaders will quickly nd a place on local councils. Even at this relatively elite level, expanding representation will increase the likelihood of pluralism and competition, for dierent constituencies will often disagree with each other and among themselves. Some local businesspeople may want to restrict outsiders from entering the retail trade in dry goods, for example, whereas others will see opportunities for themselves in opening local markets to external participants. But local elites are more likely to collude in their own material interest than to compete. For example, local business operators will likely agree to limit the number of licenses a municipality can issue for selling fertilizer or running taxis, thereby keeping prices higher for the public and adding prots for themselves. Thus if only the local elites gain representation on local government councils, the wider public is unlikely to be well served. In our six cases, two other constituencies have gained representation: minority ethnic groups and women. In large areas of Andean Bolivia, Quechua and Aymara community representatives now sit on municipal councils that did not exist before implementation of the Popular Participation Law. In the Philippines Cordillera region, indigenous groups such as the Kalingas and Gaddangs now enjoy membership on local government bodies. Similarly, councils in new rural communes in Mali are likely to show heavy representation from such ethnic groups as the Songhai and Dogon in areas where they are numerically strong. Women in signicant numbers sit on local councils in countries where their inclusion is mandated. Karnataka requires that one-third of elected members of all local bodies be

24

WORLD DEVELOPMENT

women, and that women hold one third of council presidencies and vice presidencies. Karnataka has also mandated that the Scheduled Castes (the former Untouchables or Dalits) be allotted memberships and executive positions equal to their proportion of the population in the area. 9 In Mali the major political parties are promising similar reservations for women, but no decisions have been made yet. Where female representation on local councils is not mandated, women have fared poorly in elections. Honduras and Bolivia boast few women on local councils. In Bolivia the percentage of local elected oces held by women is actually lower (8%) than it was under the earlier, more restrictive system (11%). And in the Philippines, women hold only about 10% of elected local oces. In Ukraine, where the overriding concern has been to keep local governments functioning at all in an era of vastly shrunken resources, gender representation has not surfaced as an issue. There are also other ways to ensure representation. In the Philippines, the Local Government Code requires that at least 25% of the voting members on local development councils be from nongovernmental organizations (NGOs), chosen by local NGO constituencies. Bolivia has taken a dierent approach, setting up parallel vigilance committees to monitor elected bodies. Vigilance committee members are selected from traditional local governance systems, such as peasant syndicates and neighborhood councils, and they tend to be nonelites. (c) Empowerment The third element of the formula argues that a groups increased representation will lead to empowerment, but DLG has delivered only partially on this promise. There is some good news: local governance has brought empowerment to ethnic groups that are minorities nationally but geographically concentrated in certain areas. Andean councils can now steer investments toward primary schooling in Bolivia, for example, and Cordillera municipalities in the Philippines can enforce legal restrictions on land purchases by outsiders. In some urban areas, poor neighborhoods can take charge of their own community destiny by winning electoral control of local councils, as in many sections of El Alto, the large lower class bedroom community next door to wealthier La Paz in Bolivia.

On empowerment for women, however, the news is not as good. There is no indication that women acting consciously as a group have had much eect on local public aairs in Bolivia or Honduras, possibly because they have not achieved signicant representation. Worse, in Karnataka, while women do indeed hold one-third of the seats, they have played little part in council aairs, tending to remain silent or participate only as their husbands direct. In the municipal corporation of Bangalore, Karnatakas capital city, which has a population of well over a million, the mayor and deputy mayor are both women, as are 34 of the 100 elected corporation members. Yet here as elsewhere in the state generally their husbands continue to manage public aairs through them. At the very least, we must conclude that even bold armative action on representation does not easily or rapidly empower women. Similarly, the Scheduled Castes on Karnatakas councils either do not participate or do so essentially at the direction of their local patrons. This is discouraging, because they have had mandated representation in elected local government systems since Panchayati Raj was introduced around 1960. After almost four decades, they still do not exercise a signicant voice in local aairs. For example, they are typically unable to demand more equitable siting of water supplies and electricity lines, two areas of public infrastructure investment that have been notoriously absent in areas where they live. If this has been the story for the Dalits after almost 40 years, is it reasonable to expect much more for women, whose representation has been mandated for less than a decade? (d) Distribution of benets As the formula set out at the beginning of this section suggests, the degree of empowerment has an impact on the distribution of DLG benets. In Karnataka, local elites still control elected councils and steer most benets to themselves, reecting the lack of empowerment of marginal groups. For example, they upgrade existing high schools (which their children attend) rather than spend public funds to expand primary and secondary education (to serve all children). In Ukraine, businessmen (or ``biznizmen'' as they are referred to pejoratively) and entrepreneurs who have sprung up with the collapse of the Soviet

PARTICIPATION AND ACCOUNTABILITY

25

system have directed many benets their own way by capturing privatized state enterprises through bribes and inuence peddling and by evading local taxes through intimidation and payos. In areas of Bolivia and the Philippines where ethnic minorities have taken over local councils, it is too early to tell whether benets will be broad based or go primarily to a few community leaders. In one Bolivian municipality visited by the USAID research team, the council decided to invest in school construction in remote hamlets, an initiative that should spread benets widely. But the same council also decided to build an automobile racetrack (prodromo'') and moted as the ``worlds highest auto a sports coliseum. Income generated by these ventures could go toward improving primary health care (for example, to promote treatment of scabies, a chronic disease in the area), or it could go to a few council members as kickbacks for construction contracts or skimming from gate receipts. As for the Cordillera, it is unclear whether community leaders will regulate local mineral extraction to minimize environmental damage and maximize royalties for local investment, or sell o long-term contracts for mining rights and pocket the proceeds themselves. Outside the regions where specic ethnic groups dominate, the evidence is also mixed. In some areas of the Philippines, for example, local bosses have simply spread their tentacles to envelop the new councils, rigging elections and appointing their own retainers to the seats reserved for NGOs (or ignoring the provisions for NGO representation altogether). The ow of benets remains unchanged. In other Philippine localities, the story appears to be dierent. 10 Latin America, where civil society arguably has developed more than elsewhere in the universe of aid-receiving countries, oers some further evidence of local ability to undertake activities beneting constituents across the socioeconomic spectrum. At the insistence of its vigilance committee, Cochabamba municipality in Bolivia is building primary schools in its outlying (and poorer) neighborhoods, moving away from its old pattern of building them mostly in the wealthier central area. Other initiatives in Cochabamba to improve the citys central hospital and reconstruct a municipal park will benet everyone. In many Honduran municipalities, considerable eort is going into providing sanitary drinking water for everyone. 11

(e) Poverty reduction Although poverty alleviation is a high priority for a number of donors, there is little evidence so far that DLG initiatives can do much directly to reduce poverty, at least in the short run. The main reason for this short-term pessimism is that when governance is decentralized, local elites get most of the power and steer benets to themselves, or at least maintain the existing distribution patterns (which largely benet them, anyway). Local elites may be even less likely than national elites to target government resources to the poor, as Manor reports, noting that he has ``yet to discover evidence of any case where local elites were more benevolent than those at higher levels'' (Manor, 1999, p. 91). (f) Implications It appears that as we proceed analytically through the above formula, the results registered to date become successively less substantial. DLG initiatives have encouraged participation and have increased representation, but they have provided little in the way of empowerment, and even less in making the distribution of benets more equitable or reducing poverty. This does not mean however, that promoting DLG is a futile endeavor. Increased representation oers signicant benets in itself. When women or Dalits hold public oce in Karnataka and sit on local councils, it demonstrates that males of hereditary privilege are not the only ones who can enjoy such positions. Increased representation on local councils also makes it clear to children of both genders and all ethnic communities that they can aspire to public service. For example, a Dalit girl who has an aunt or neighbor on the village council is more likely to have higher ambitions than one who perceives inevitable exclusion as her future. The presence of NGO representatives on local development councils in the Philippines shows people there are alternative ways to participate in local decisionmaking where traditional political bosses still have too much power. A second benet of increasing representation among marginal groups is its attendant leadership experience. Karnatakas new structure provides 85,000 elective positions, which means more than 28,000 women are panchayat members at any given time. Among them, some 1,900 women are presidents and an equal

26

WORLD DEVELOPMENT

number are vice-presidents of their panchayats. While many will fail at these tasks or enjoy only modest success, some will acquire skills that will equip them to be leaders in other endeavors, for example, in civil society or in state or national politics. Similarly, a substantial number of representatives of the lower castes have gained leadership skills in their reserved slots and then parlayed those abilities into career advancement. These promising benets of increased representation, however, will take some time to unfold, as suggested by the Panchayati Raj's nearly 40 years of experience with Scheduled Caste reservations. In Bolivia and the Philippines, it will be some time before previously marginal minorities can translate the empowerment they have gained in their geographic areas of strength into improved lives. People new to making public decisions must be expected to have a fairly long learning curve before they can actually shape local public policy in ways that benet their constituencies. Even when they do, benets may be long in coming. Building new schools, for example, does not mean that eective teachers will immediately be there to sta them. Moreover, it takes time to educate successive groups of children, and longer still for them to use that education in making their lives more productive. In sum, while we are most probably talking in terms of decades here, not ve-year donor projects, these kinds of improvements in public life can start with donor-supported DLG initiatives in localities where minorities predominate numerically. The central question stemming from the formula in Fig. 1, however, is whether if at all DLG can enable marginal groups to use local politics to press for policies of direct benet to them in areas where they are minorities (or in the case of women have traditionally been relegated to the political periphery). Is it feasible to expect marginal groups to pressure local governments into delivering services more

equitably by targeting eorts or redistributing ongoing activities? Or can such groups more likely garner benets if they participate in coalitions that push for eorts that benet the entire community? For example, can marginal groups persuade local governments to set up programs to educate lower-class girls and build sanitary wells in poor areas, or alternatively is there more promise in pushing for programs that oer universal primary education and provide potable water for everyone in the community? Perhaps in some places traditionally marginal groups can be mobilized to act in their own behalf locally, even where they are minorities. This appears to be true in some parts of the Philippines, but not in many other parts where local bosses still maintain control. In other places, such as Bolivia and Honduras, marginal groups do seem able to direct benets to their members through universal eorts that benet everyone locally. 12 Table 1 illustrates the issue here for three sectors of development activity: education; health; and agriculture. The left column oers examples of the kinds of eorts traditionally implemented in development programs, which favor local elites. They will send their children to quality secondary schools, take advantage of high-tech medical therapies, and place themselves rst in line to irrigate their landholdings. The middle column indicates what would ideally happen if weak and vulnerable groups were able to gain a dominant voiceor even a signicant onein local development planning bodies. But because it is so dicult for such groups to assemble the public support needed for such targeted initiatives, their best bet lies in the last column, which lists some ``universalistic'' initiatives that stand to benet all groups and thus have a much greater chance of appealing to a wider constituency. Pluralism, after all, is compromise, and this is the kind of compromise that seems most likely to succeed in DLG.

Table 1. Examples of three types of local initiative Type of initiative Education Health Agriculture Traditional High quality secondary schools Research hospital in urban center Subsidies for purchasing tubewells Poverty-focused Special education for the poor Clinics for poor neighborhoods Micro-credit for the landless ``Universalistic'' Primary education for all Campaign against malaria Improved pest management

PARTICIPATION AND ACCOUNTABILITY

27

3. PUBLIC ACCOUNTABILITY Democratic governance at the local or national level can succeed only if public servants are held accountablegovernment employees must be accountable to elected representatives, and representatives must be accountable to the public. 13 A wide range of mechanisms can serve as agents of accountability. Each has its strengths, but, as will become evident, none appears able to carry the whole load itself. (a) Bureaucratic accountability to elected ocials In all but Mali, signicant control over local civil servants was transferred to local elected bodies. When the health sector was devolved to the local level in Bolivia, for example, so was supervision of health ocers. But in every case this transfer of authority has been incomplete. For example, the central government of Bolivia still determines salary, posting, and tenure, so a mayor could order his health ocers to keep posted clinic hours, but could not discipline them if they failed to do so. The same is true, more or less, in Honduras, Karnataka, and the Philippines. Ukraine transferred far more authority than the others did, largely because the central government was unable to do more on its own. 14 Ukraine aside, the pattern in the other countries is that devolution, and hence accountability, has been incomplete. One reason for this has been the reluctance of central governments to truly decentralize. But another signicant reason has been the unwillingness of government employees to be decentralized. This is especially true of those in the professional elds, such as health workers, teachers, and engineers, who invariably want to keep their national career opportunities, enjoy urban amenities, and send their children to good schoolsall ambitions likely to be frustrated if they are posted indenitely to rural areas or small towns. Small wonder, then, that when central governments decree decentralization, bureaucrats in the lower echelons are quick to declare that such initiatives should not apply to them. Central governments in turn almost always back down. At a minimum, they arrange programs so that eld ocers maintain links with their original line ministry and thereby enjoy some insulation against local control. Such an arrangement sounds like a recipe for impasse between local bureaucrats and elected

ocials, but this need not be the case. In Karnataka over the several decades of Panchayati Raj, for the most part the two have achieved an accommodationelected ocials direct government civil servants in their jurisdictions, while the line ministries write annual evaluation reports (and determine promotions and postings). The arrangement is not perfect, but it can work. 15 (b) Elected ocials' accountability to the public Free, fair, regularly scheduled elections and universal surage are the most direct mechanism for ensuring that those who govern are accountable to the citizens. Without elections, local government is not democratic. But elections are crude instruments of popular control, since they occur at widely spaced intervals (two to ve years in the countries studied here), and address only the broadest issues. In a well-established system they allow voters to register general approval or disapproval for an incumbent's performance or to select new public managers when incumbents vacate their oces. DLG needs more than this kind of blunt direction from the citizenry, however. People must be able to indicate their likes and dislikes between elections, as well as their views on specic proposals. There must also be ways to publicize citizens' views and uncover wrongdoing in local government. For this more continuous and ne-tuned accountability, varied arrangements are available. Altogether seven such mechanisms have proven at least potentially viable in the six systems studied. (i) Elections Five of the case study countries have managed at least one free and fair local election. 16 Some have managed more. India has conducted elections periodically (if not always on schedule) since Panchayati Raj was introduced around 1960, and the Philippines has had elections since 1988, but in both countries there have been aws. People in the Philippines still speak of ``guns, goons, and gold'' as perennial factors in electionsand in some areas the charge rings true. Intimidation and vote buying also happen in India. In both places there are public ocials who have won and held onto their oces through fraudulent means. But on the whole, elections in these two countries as well as the others in the CDIE sample have reected the public will. India in particular has

28

WORLD DEVELOPMENT

shown over many decades now that elections can be a powerful tool for accountability. (ii) Political parties In many systems, opposition political parties make for a powerful engine for enforcing accountability. The party in power all too often has strong incentives to evade accountability, but opposition parties have their own incentives to uncover wrongdoing at city hall and publicly hound incumbents for their misdeeds. They present a constant vision of a viable alternative for doing public businessa dierent and perhaps better way. Such a visible alternative helps keep the party in power on a path of rectitude. This scenario requires a strong party system at the local level, however, which is often not the case. 17 In Ukraine, after the political implosion of the Communist Party, the party system became feeble and fragmented. There is one regional party with some strength in the western part of the country, but for the most part there are only pieces of political parties at the local level. Although it has been more than a decade since democracy was restored in the Philippines, the party system there has yet to stabilize. Instead, it is weakened by personal loyalties, opportunism, and a tendency for parties to break up and re-form. At this early stage in its democratic history, Mali also has a highly fragmented political system, with nine relatively stable parties and dozens of splinter groups operating almost entirely at the national level. The other three countries have reasonably strong party systems. In Honduras a two-party system is vigorous and stable at all levels. Bolivia has been somewhat less stable, but ve parties currently enjoy serious national support and the party system has become the major organizing principle in municipal government. 18 Karnataka had a strong two-party system at state and district levels through the 1980s. In the 1990s, that expanded to a threeparty structure. Interestingly, however, and in contrast to what seems to be happening in Bolivia, the Karnataka parties seem to have been slower to establish themselves at the lowest level, though this process has begun. Such spreading is more or less inevitable, given the incentives parties have to mobilize activity at the local level in order to improve their electoral chances at higher levels. It follows that where there are strong national party systems, it is largely futile to impose ``non-partisan''

local elections as has been attempted in countries like Bangladesh. (iii) Civil society Civil society is usually dened as organizational activity between the individual (or the family) and the state. Its democratic role is to advocate for constituents, to act as watchdog over the state, and to support political competition generally (Hansen, 1996). Civil society is essential to DLG in USAIDs strategic vision and for other donors as well, 19 but in the six political systems studied it has not played a large role at the local level. Even in the Philippines, which is at the high end of the civil society spectrum among the six cases, the evidence is not all positive. At the spectrums low end, Ukraine had essentially no civil society at the local level at the time of the USAID study, nor was there much sign of social capitalthe trust that facilitates people working together toward a common purposewhich would help build civil society. The main reason, of course, is the countrys recent Soviet history, in which the state controlled almost all organized social activity. The whole web of organized life collapsed with the communist system, leaving families and individuals autonomous and without social moorings. The USAID study found people in the major cities just beginning to organize in their apartment buildings for such purposes as maintaining common areas, providing security, and dealing with outsiders encroaching on public space. A number of donors, including USAID, the European Union, and the Soros Foundation, are supporting eorts to build social capital and civil society, but the road ahead is clearly a long one. In a way, India is the most startling case, because DLG has been in place there the longest. One would assume that civil society had at least gotten o to a good start. It has at the national and state (and to some extent the district) levels, where organizations of women, environmentalists, professionals farmers, and others have been a prominent part of the political landscape for some time. 20 But at the village and mandal (township) level, such activity is hard to nd in Karnataka. Even civil societys precursor, social capital, appears not to have developed outside of caste groups. Social capital seems reasonably strong in Bolivia, probably reecting its tradition of peasant syndicates, indigenous communal institutions, and urban neighborhood organiza-

PARTICIPATION AND ACCOUNTABILITY

29

tions. 21 But it appears not to have carried over into civil society, for these bodies have yet to become actors or advocates in municipal political life. Honduras appears to be somewhat more advanced in terms of civil society activism, in particular with its patronatos (voluntary neighborhood groups) in the urban areas, which have pressed for better service delivery in health and sanitation among other sectors. Mali, if anything, has a stronger base of social capital, with its rich tradition of associational life and strong interpersonal networks at the village level. There are signs of a civil society emerging there, including the rapid growth of NGOs (there are now more than 600) and of registered village associations (more than 2,000) since 1992. But whether such groups can form a viable civil society will not be evident until Mali fully launches its DLG initiative after its 1999 commune elections. Among the six case study countries, only the Philippines shows signs of a civil society that is an eective instrument of public accountability at the local level. Even there the story is mixed at best. In a 1996 survey of USAIDs Governance and Local Democracy project, only 57% of the participating municipal governments demonstrated clear evidence of genuine NGO participation. 22 That participation is likely higher than in the country as a whole, since project sites were chosen in part because they were likely to succeed at DLG. Even with these qualications, however, it is clear that genuine civil society participation exists in the Philippines in some places, even as local bosses continue to dominate in others. For a country long plagued with elite control at all levels (Timberman, 1990), though, this is a promising start. Why civil society has made less progress in Karnataka and Bolivia is not clear. Perhaps the answer lies partly in the functional nature of traditional social structures, especially in rural areas. Caste councils and Andean farmer groups appear in many ways to have been mainly governing bodies, whose calling was to set rules for members and settle disputesnot to act as advocates with local government agencies or compete with other organizations in getting the attention and support of local government. There seems to be no obvious way to transform them quickly into players in new DLG systems, because their experience simply has not prepared them for such a role. Building social capital and civil society will probably take a long time, perhaps a very long

time, as Robert Putnam (1993) argues in his analysis of their growth over the centuries in Italy. Places such as Karnataka and Bolivia surely do not have as far to go as Ukraine, but there will have to be much social change before civil society can become a major engine of social accountability at the local level in all these settings. (iv) The media At the macro level, the media have two basic roles. The most important is to make political news public. Only if people know what is going on, good and bad, can they hold their government accountable. Without vigorous media to spread it, political news remains the property of the inside few. The medias second role is to help uncover government misdeeds. This investigative function is subordinate to the main public information function, however, for there are other institutionsopposition political parties, civil society, and the legal systemto uncover malfeasance. Moreover, at the local level the latter role becomes even more secondary, or perhaps unrealistic, for small newspapers and radio stations generally cannot aord to sponsor the work necessary to undertake good investigative journalism. Thus, it is the public information function that the local media must perform: to make political news from all sources (including government, opposition parties, and civil society organizations) available to the widest possible audience. What this means in most developing countries is radio. Newspapers tend to be published only in larger population centers, and are accessible only to those who can read them (although in many societies it is traditional for literate people to read aloud to nonliterate audiences). Television reception is conned to those who can aord a TV set and live within the eective radius of a transmitter. But radio, especially the AM band, is cheap to operate, does not require line-of-sight transmission like TV, and has great audience potential. 23 Local news, talk shows, and question-and-answer programs are all excellent ways to spread political news widely. This use of radio is probably most advanced in the Philippines, which has hundreds of local AM stations, many of which broadcast political feature programs. There is also considerable regional press in both English and vernacular languages, although it is mainly restricted to the provincial capitals. Bolivia also has many local radio stations and at least one very active

30

WORLD DEVELOPMENT

network that provides political features. Bolivia's print media, however, originate only in the largest cities. Honduras has some local radio, although not as vigorous a system as in Bolivia, while its print media have similarly limited coverage. Mali has 77 private rural radio stations promising to become strong instruments of accountability as the new DLG system takes shape. 24 But the low distribution of radio receivers (four per 100 people) is a serious constraint. In India, as in so many former British colonies, the government retains a monopoly over radio. There is some local press, and newspapers published in district towns do feature political news, but districts are large (a district in Karnataka averages about two million people), so coverage of subdistrict levels is bound to be thin. In Ukraine, both print and broadcast media are still emerging from the heavy control imposed during the Soviet period. Newspapers, television, and radio are all nding their way in the uncertain new world of press freedom. There is some regional news coverage, and there are local talk shows. At least a few mayors have gone on television to publicize their budgets and invite public debate about them. Thus, there is some promise that the media can become an eective instrument of accountability, but for the most part that promise has yet to be realized. (v) Public meetings Several countries have instituted public meetings to insert civic opinion into local governance. A number of Ukrainian mayors have launched public budget hearings, and public hearings have become common for many local government bodies in the Philippines. In both countries, ocial council meetings are open to the public. In two countries, law mandates open meetings at the lowest level of local governance: ve cabildos abiertos a year in Honduras; and two gram sabhas in Karnataka. In Mali the central government organized an ongoing series of regional and local meetings to involve people directly in reorganizing the local government units into the new communes. The results have been uneven. Some mayors in Ukraine and the Philippines use council meetings and hearings to solicit citizen views and mobilize support for their programs, but others ignore them. In Karnataka, elected ofcials were evidently so burdened with embarrassing questions at village meetings that they quickly abandoned their gram sabhas, or held

them at odd times or in obscure locations. Some Honduran ocials have made their mandated meetings largely ceremonial functions, but others have used them to involve citizens in local government decision-making. In Mali, the meetings have been quite successful, informing people about decentralization and helping them become stakeholders in its implementation, in particular by having them participate in the redistricting process that established the new communes. (vi) Formal grievance procedures Two of the case study systems have instituted formal procedures for citizens to redress grievances against elected ocials. The procedures in Bolivia are especially ambitious. The vigilance committees, whose main duties are to make plans for local infrastructure investments and to monitor municipal budgets, are also empowered to wield a legal instrument called a denuncia against local councils. If convinced that its municipal council has acted wrongly, a vigilance committee can lodge a complaint with the national executive branch, which can then pass it on to a special committee of the Senate, which in turn can suspend central funds to the erring council. A recent law also allows municipal councils in certain circumstances to formally dismiss a mayor. In the Philippines, a recall procedure can be used against governors, mayors, and municipal council members, somewhat similar to those introduced in many US systems in the early part of this century. (vii) Opinion surveys The USAID Governance and Local Democracy project in the Philippines has pioneered the public opinion survey as an instrument of accountability by building local capacity to design and conduct them. These polls measure citizen satisfaction with government services. Results to date are promising (Van Sant, Blair, Razon-Abad & Amani, 1998, pp. 314 to 316), but the polls are still experiments. Thus, it is not clear whether they can be replicated on a larger scale or be self-sustaining once project support is withdrawn. Unlike the national level, where the media sponsor public opinion polling as in the United States and then make public the results, there is little local scope for such backing at the local level, for the media are too embryonic as yet. One hope, though, lies in creating a local commercial polling capacity that can sell its wares to the

PARTICIPATION AND ACCOUNTABILITY

31

private sector and realize sucient income thereby to nance political polling that can then be done pro bono. So far, local surveys have been restricted to the Philippines among our six countries, but the approach oers considerable promise in places with a fairly sophisticated survey capacity at the national level, such as Bolivia, India, and Ukraine. (c) Patterns of accountability The mechanisms used to promote accountability in the six systems studied range from reasonably eective to virtually useless. Table 2 gives an overall picture of how they have been working for accountability of government ocials to elected leaders and of the latter to the citizenry. To be sure, such an exercise has to be impressionistic; it was not possible to develop and apply strict standards or indices of ``eectiveness'' in the study. Moreover, it is somewhat misleading to give the dierent mechanisms seemingly equal weight. Public meetings, for example, are clearly less vital than elections as instruments for popular control. The table lays out the seven mechanisms for elected leaders-to-citizens accountability in a very rough order of what appeared to be their salience in the six case study countries. It might be more accurate to assert a simple dichotomy, in which the rst four (i.e., elections, parties, civil society and the media) can be labeled ``more important,'' while the remaining three (meetings, formal procedures and surveys) become ``less important.'' These distinctions are indicated by boldface and normal typeface respectively in Table 2. Another qualier is that some instruments (parties, civil society, the

media and surveys) exist independently of the local governance structure, while others (local elections, public meetings, and formal procedures) exist only as part of them. Finally, there are obvious measurement problems: how does one tell if the media or civil society is eective in assuring accountability? Despite all these problems, however, considering the mechanisms in this comparative way does yield some interesting insights. The table distinguishes between ``viable sources of accountability,'' mechanisms that appeared eective in ensuring accountability, and ``potential sources,'' mechanisms with some promise of making elected ocials accountable, but not enough track record yet to be considered viable. Mechanisms that seem to be in place and viable (that is, eectively ensuring accountability) are indicated with two Xs; those showing promise of being a source of accountability have one X. While these estimations are approximate, they give a rough idea of the accountability the seven mechanisms provide in the six settings. At least six observations emerge from the table. First, the case-study DLG systems have used varying combinations of accountability mechanisms, though none has instituted them all. Nor have any two countries tried the same combination of mechanisms. Second, there is no country in which one mechanism by itself has proven sucient to realize genuine accountability in DLG. Some kind of mix of several instruments would appear necessary even to have a chance of realizing it. Third, so far no one mechanism has proven viable in all settings, or even most of them. Elections exhibit the most potential for working at present, but

Table 2. Public accountability and DLG: Viable and potential mechanismsa Philippines Bureaucrats to elected ocials X Bolivia X Honduras X Karnataka XX Ukraine X Mali

Elected ocials to the citizenry X Electionsb Political partiesb b XX Civil society X Mediab Public meetings X Formal procedures X Opinion surveys X
a b

X XX XX

X XX XX

XX X

X X

X X X

XX

XX Viable source of accountability, X Potential source of accountability. More important mechanisms.

32

WORLD DEVELOPMENT

they have only shown much real success to date in Karnataka. In the other cases, serious constraints remain. Fourth, the systems have achieved quite different rates of progress with the mechanisms of accountability they have used. Bolivia has three that appear viable, Honduras and Karnataka two each, the Philippines one, and Ukraine none. Potential sources of accountability show another pattern, with the Philippines far out in front (six mechanisms), followed by Honduras, Mali, and Ukraine (three each) Bolivia (two), and Karnataka (one). Fifth, success does not appear to be a function of time. The country with the most recent DLG program in place (Bolivia) has the most accountability instruments up and running, while the system with the longest experience (Karnataka) has done less well. Finally, it must be noted that each of these mechanisms can serve for ill as well as good: Elections can be fraudulent, parties can foment hostility and conict, civil society can advocate the destruction of the body politic, the media can become captive of an authoritarian central government or self-seeking elite elements, public meetings can turn into controlled puppet shows, formal redress procedures can be manipulated by demagogues, and opinion surveys can be doctored to show false results. Just because these mechanisms are in place, in other words, does not mean that they will inevitably conduce to the public good. (d) Implications This analysis presents several implications for DLG strategy. First, a number of approaches can workthere is no ``one best way'' to promote accountability. DLG programs, accordingly, can be crafted to many dierent settings. A second implication is that for accountability to be eective, several instruments must take root and succeed. One, or maybe even two mechanisms, even if they work, will probably not suce to establish DLG on a selfsustaining basis. A third and more sobering implication is that many of these approaches will take a long time before they function properly, and indeed may never do so fully. As long as local bosses retain their power in various parts of the Philippines, for example, they will likely be able to bend elections in their favor. Moreover, if civil society has not emerged at the local level in Karnataka over the past several decades, it would

seem unrealistic to expect it to do so in the near future. In some cases, however, it may be possible to devise a mix of instruments to make up for those that are unlikely to work. For instance, to hold local governments accountable in the Philippines a good approach might be to strengthen the party system so that it could function in tandem with civil society and the media, but if this seems altogether unfeasible (which, given the continually changing nature of parties there in recent years, it might well be), a better approach could be to concentrate on making elections more transparently honest and developing opinion surveys as additional instruments of accountability. Finally, a fourth implication is that there is no apparent sequencing of appropriate mechanisms. It would certainly be convenient if some kind of formula could be constructed, for example that the rst step for donors supporting DLG should be to ensure that an operating civil society is in place at the local level, then move to the media, formal procedures and parties in that order. But if the Philippines is relatively strong in civil society but weak on parties (as per Table 2), while in Bolivia the picture is exactly reversed, then there would seem little reason to think that a sequencing approach could be devised. The answer to the question of when should the various mechanisms be used, then, is that each case depends on what is already there and can be used to build up or perhaps substitute for weak or absent mechanisms.

4. BRINGING TOGETHER PARTICIPATION AND ACCOUNTABILITY It is now possible to bring together the two principal themes of the paper, which is attempted graphically Figures 25. Their main purpose is to oer a plausible set of linkages tting together the various components of DLG. As such, the gures are intended to be heuristic rather than denitive, and one could argue that other items could have been included, that some of those shown are superuous, that the arrows could have been drawn dierently, etc. Certainly greater parsimony could be exercised, although the complexity of the full presentation (Figure 5) is ameliorated somewhat by oering its components in stages as Figures 24.

PARTICIPATION AND ACCOUNTABILITY

33

Figure 2. Accountability mechanisms in democratic local governance.

Figure 2 presents the seven accountability mechanisms discussed above, showing some of the linkages between them (the arrows are intended to indicate causality). The four more important ones are indicated in boldface. Civil society and political parties have the most arrows, indicating their primacy as the most active mechanisms, a point to which we shall return.

Figure 3 shows where the impact of all this activity is directed: to elected local governments and thence (as per the analysis earlier) to the local bureaucracy and DLG performance. As the arrows indicate, civil society and parties have both a direct and an indirect impact here, while the medias impact is only indirect (operating through these other mechanisms), and the others have only a direct aect on local government. Figure 4 presents the outcomes in terms of what DLG performance actually gives the citizenry and what they were looking for. The degree of overlap between the two ovals at the lower right of the gure can be interpreted as showing how eective local government is in meeting citizen needs: the greater the overlap, the more DLGs eectiveness. Finally Figure 5 lls out the chart by showing the principal citizen actors who participate in the DLG tableauthe forces that drive the whole processand the underlying social capital and civic culture that gives rise to them. It makes sense to distinguish several kinds of

Figure 3. Accountability mechanisms and impacts in democratic local governance.

Figure 4. Accountability mechanisms, impacts, and outcomes in democratic loca governance.

34

WORLD DEVELOPMENT

Figure 5. Causal linkages for participation and accountability in democratic local governance.

groups; four types are shown here, but the number could be enlarged or decreased. At the outset, we can say that initiating a DLG structure will surely encourage traditional elites (``old elites'' in Figure 5) to participatelocal bosses, substantial property owners, wholesale traders and the like. It can also be counted upon to lead new elites to join the political arena; the most obvious candidates here are the professionals that generally get transferred to local government units in the course of a DLG initiative (e.g., health workers, teachers, engineers, agricultural specialists) and the contractors that spring up to build the physical infrastructure that newly installed local governments inevitably tend to make their rst priority. This is all to be expected. The critical question is to what extent can other elements beyond these elites get involved in local governance, indicated in Figure 5 as ``middle groups'' and ``minorities and vulnerable groups.'' The designation ``middle groups'' would of course depend on country context, but is intended here generally to denote shopkeepers, small and medium landholders (``middle peasants'' in a political economy sense), oce workers, and so on. Finally, ``minorities and vulnerable groups'' would comprise ethnic, religious and caste strata with low social status, sharecroppers and landless agricultural workers, urban daily wage laborers, and in most settings women. As we have seen earlier, it is possible, given a DLG setup that incorporates democratically elected local governance, for many if not most

of these latter two categories to enjoy some real participation and even representation (particularly when seats on local bodies are mandated for various groups in the population), but actual empowerment appears signicantly harder to attain. ``Harder,'' however, should not be taken to mean necessarily ``hopeless.'' Even if middle groups and especially minorities and vulnerable groups nd it dicult and probably impossible to gain a dominant position in local elected councils, there are other paths to inuence. Erstwhile marginal groups, for example, can combine to use their numbers in seeking power. This is potentially a winning strategy, given that, when taken all together, such groups often constitute a majority of the population. But to expect these ``groups in themselves'' to suddenly become ``groups for themselves'' and collectively one large ``group for itself'' (to take some liberties with political economy analysis) is surely more than a bit utopian in most Third World settings where elites have so long dominated the local scene. The civic culture indicated in Figure 5 is, after all, the product of many decades and even centuries of local relationships, and it has an inertial momentum that is not easily redirected, as is indicated by the hexagon labeled ``socio-political history'' in Figure 5 and the heavy arrow linking it to the civic culture oval. Thus the construction of a new civic culture that will facilitate cooperation across middle and lower groups will be at best a long process. Even the accumulation of sucient social capital to begin such a process may

PARTICIPATION AND ACCOUNTABILITY

35

take a good while in places like Ukraine, although in others like the Philippines it appears to be more advanced. Far more likely is the prospect that new elites may nd themselves quarreling with old elites and in need of allies who are willing to give support for a political price. Similarly, older elites may fall out among themselves as the changes brought about by DLG unsettle old balances and modes of operation, thus giving factions among elites incentive to seek help from other, previously unrecognized quarters. This is an approach that has enjoyed some success in India, particularly at the state level, as for instance Dalit communities have been able to press for some benets. At the local level, however, as the Karnataka evidence illustrates, the path is likely to be a long one. This discussion hints that civil society and political parties have to be the key institutions in bringing marginal groups into the DLG political arena, an idea supported in Figure 5, where these two mechanisms are connected to more elements in the chart than any others. Such a centrality ts in with our earlier analysis, where it was observed that elections, while critical, form crude instruments used only at widespread intervals, while civil society and parties are continuously operating institutions. 25 One possibility raised by Figure 5 is that, to the extent it is an accurate depiction, there may be signicant substitutability between civil society and political parties at the local level. In the Philippines, for example, the political party system continues to have little structure or continuity and is largely personality-based, but civil society shows considerable promise at local as well as at higher levels. In Bolivia, on the other hand, local civil society appears quite underdeveloped, while the party system has become relatively well institutionalized and may well be able to carry some of the pluralistic load that civil society cannot as yet. Countries such as Honduras (see Table 2) that have possibilities for both mechanisms are at an advantage, while those without much going in either sector (e.g., Ukraine, where fragmented parties are feebly attempting to coalesce into

viable units and civil society continues in a state of chaos after the collapse of communism) face a double disadvantage.

5. CONCLUSION Viability for the democratic component of DLG (and for much of its local governance component as well) depends in the nal analysis on participation and accountabilitybringing as many citizens as possible into the political arena and assuring that local governors are responsible to the governed for their actions. Fortunately, there is considerable scope for enhancing both participation and accountability at the local level. On the participation side, DLG can bring new elements, particularly women and minorities, into local politics in meaningful ways. They can gain representation in local decision-making and, where some degree of geographical concentration exists, empowerment as well. And although DLG oers only limited scope and for poverty alleviation through specially targeted eorts, it can be helpful in promoting more universalistic local development activities that will benet the weak and vulnerable along with everyone else. As for accountability, there is available a wide range of instruments or mechanisms through which citizens can exercise control over their ocials. Some like elections are more encompassing and nal, while others could be thought of as being more in the nature of netuning, like public meetings or opinion surveys. Any one or two of these mechanisms, even if eectively operating, seems unlikely to promise sustainable accountability. But they can be mixed and matched in a variety of combinations, and to some extent the more important mechanisms like political parties and civil society may be substitutable, so that if one fails another could be strengthened in its place. It seems feasible, then, that given sucient political will on the part of the central government to keep a decentralization initiative in place over time, eective democratic local governance can be achieved in a number of ways.

NOTES
1. The most widely used classicatory and denitional treatment of decentralization continues to be that pioneered by Dennis Rondinelli (see Rondinelli, Nellis & Cheema, 1984; Rondinelli, McCullough & Johnson, 1989). Following his taxonomy, devolution as a form of decentralization is the most ambitious, to be contrasted

36

WORLD DEVELOPMENT 1998), the Philippines (Lippman & Jutkowitz, 1997), and Ukraine (Lippman & Blair, 1997). All are available on the Web at http://www.dec.org/usaid_eval/. The Karnataka study was undertaken by an in-country team of social scientists (Inbanathan, Bhagyalakshmi & Doraiswamy, 1997); it was not published as one of the country series. Aside from the Philippines (which was the rst eld study conducted and needed some subsequent updating) and a few other exceptions as noted, all country-level information in the present essay is based on these reports and thus reects the situation as of 1996 or early 1997. 7. Blair (1998). This study is also available at the website noted just above. Eight themes are dealt with in the synthesis report: the country context; donor and host-country roles; resistance and political will; representation, empowerment and benets; scal autonomy and regional equity; public accountability; performance and accountability; and national advocacy. 8. The most committed proponent of this view is probably David Korten, who has made the case for it often (e.g., Korten, 1990). 9. In its latest (1993) version of DLG, Karnatakalike many other Indian statesrequires the so-called Other Backward Castes of Shudra communities (traditionally next to the Scheduled Castes at the bottom of the Hindu hierarchy) also be given membership in local councils proportionate to their share of the population. (So far there is no similar quota system for their becoming council presidents.) The mechanism for implementing these quotas is to declare specic seats reserved for women, Scheduled Castes or Other Backward Castes, allowing only people of that group to run. 10. Surveys by the Social Weather Stations' polling organization found that citizens believe they can deal more eectively with local government than national government and are also more satised with local government. Perhaps more signicant, there are smaller class dierences in this feeling of citizen eectiveness. Lower class people tend to see themselves operating with virtually the same degree of eectiveness as do the higher classes at the lowest (barangay) level. On national issues, class dierences are much wider (Rood, 1997). 11. Of course, it is possible for these eorts to produce less desirable results. The schools in poor neighborhoods may be so underfunded as to provide no real education, the central hospital may use its resources mainly on high-tech procedures serving few patients, the park may charge an admission fee that will restrict its use to the wealthy, and the pipes for drinking water

with the more commonly employed delegatory and deconcentrational forms. For more discussion in the DLG context, see Blair (1998, pp 13). 2. At times popular participation received strong support, as with USAIDs Title IX and New Directions initiatives of the late 1960s and 1970s (see e.g., Esman & Upho, 1984), but these eorts did not put much if any emphasis on elected local governments as engines for development. Participation, in other words, did not for the most part mean democracy. The major exception in this regard is India, which was included as part of the study reported on here. 3. Among the donors, USAID is arguably the most emphatic voice urging DLG as a democratic end in itself (see e.g., USAID, 1998, p. 42), while the multilaterals lead in seeing DLG as a means to promote development and particularly poverty alleviation (e.g., Binswanger & Shah, 1994, p. 4 for the World Bank; UNCDF, 1995, p. 5; UNDP, 1997a, p. 19; UNDP, 1997b, pp. 89; also OECD, 1997, p. I, p. 11). It should be noted that these positions are not exclusive; USAID does see DLG as a means to achieve developmental as well as democratic ends (e.g., USAID, 1997), though this appears to be a less central theme, and in the same sense the multilaterals support democracy as an end for DLG, albeit not the primary end. In contrast to the USAID study reported on in the present essay, some other recent assessments have focused more on the output side; for example Crook and Manor (1994) have analyzed local government performance, while Cohen and Peterson (1995) studied administrative decentralization. 4. For more on the conceptualization of the study, see Blair (1995). 5. Karnataka also had the advantage of having been the focus of two excellent recent studies of DLG (Crook & Manor, 1994; also Inbanathan, 1992). Inasmuch as India is a federal system and local governance is a ``state subject'' in its constitutional setup, there is considerable autonomy in managing decentralization, so that Karnataka can be considered as at least somewhat akin to being a separate country in the DLG sense. Moreover, its large population (48 million in the 1991 census) made it larger than all but the Philippines and Ukraine in the sample. Finally, the national government, which provides large grants to the states, can be thought of as being very roughly analogous to a ``donor'' with the Karnataka state ministry being similar to the ``host country government'' in development parlance. 6. The reports covered Bolivia (Blair, 1997), Honduras (Lippman & Pranke, 1998), Mali (Lippman & Lewis,

PARTICIPATION AND ACCOUNTABILITY may somehow never stretch beyond the more well-todo sections. 12. This idea has some resonance in industrial countries, too. Wilson (1987, 1996) in his well-known works, The Truly Disadvantaged (1987) and When Work Disappears: The World of the New Urban Poor (1996) argues that US social programs to help minorities and the poor succeed in the long run only if they benet all groups, because only then is it possible to build the constituency needed to launch and maintain such eorts. Hard quantitative data are hard to come by on this point, but a recent World Bank study analyzing Bangladesh data from the mid-1990s found that in a sample of 239 villages, less than half the villages had a majority of men favoring income-generating activities for women (cf. the middle column of Table 1), while three-quarters of the villages had a majority of women in favor of such programs. On the other hand, a majority of men favored female education programs in 85% of the villages (corresponding more to Table 1s right column, in that both sexes would receive the same education), while the analogous datum for women was 90% of the villages. (See World Bank, 1998, pp. 1112.) In a purdah-oriented rural society like that of Bangladesh, work for women generally means work for poor women, whereas female education would include females of all classes. Given that men have the dominant voice in rural Bangladesh, their views can be considered as essentially determining what is to be done in public aairs. 13. It could be argued that local bureaucrats are also accountable to the citizenry, but it is more practical to view the lines going from bureaucrat to elected representative to citizenry, because it is the elected ocial who must in the end sanction the bureaucrat (through his or her supervisory role) and it is the populace who must ultimately exert control over the representative at elections. To be sure, bureaucrats should be responsive to citizens (this responsiveness is one of the measures of good DLG), but they should be accountable to elected ocials. 14. The sole exception was the Finance Ministry, which largely retained control over budget ocials at all levels, to ensure the integrity of the governance structure. The result was a somewhat uneasy arrangement in which local ocers of the Finance Ministry in eect worked for both the mayor (or provincial governor) and the ministry in Kyiv. 15. That such an accommodation has been reached in Karnataka does not mean that this is the case everywhere in India. Mishra (1997), for example, in a district study conducted in Rajasthan state, found bureaucrats

37

continuing to resist political control at the local level, even after decades of Panchayati Raj. 16. Mali, the sixth country, had local elections in 1992 for 19 urban areas. Local elections for the entire country are taking place during 1999. 17. As a corollary, it also demands that all parties adhere to some consensus on rules of the political game, for instance, that party competition be framed in peaceful terms and that for transferring power elections are the appropriate means (as opposed, for example, to violent confrontations and disruption of public life). 18. In the initial phases of the Popular Participation Law, municipal councils were elected on party tickets, but the parties did not appear to dominate the selection of vigilance committees in most places. This is likely to change as the vigilance committees become more integrated into the political system, with its strong party orientation. 19. For example, in its New Partnerships Initiative, USAID emphasizes interactive support for three sectorsbusiness, DLG, and civil society (USAID, 1997, passim). See also UNDP (1997c) and UNCDF (1995). 20. As for instance with women organizing against alcohol consumption, pressing states to change their liquor laws (Moore, 1993; Bearak, 1996; Mackinnon, 1997). 21. Many of these bodies became the building blocks for the structure of the vigilance committees that have taken shape under the country's Popular Participation Law. 22. Rood (1997, p. 18). ``Clear evidence'' meant some combination of representatives chosen by the NGO community actively sitting on local development councils, local government investment plans developed with real citizen input, and environmental plans developed in similar fashion. 23. There are at least twice as many radios per hundred households in the six countries as there are TV sets or daily newspapers received. The range in the early 1990s was from 81 radios per 100 households in Ukraine to four per 100 in Mali (UNDP, HDR, Various years). 24. This number, reported to the USAID team, was almost double the number of stations in place about a year previously (see Buckley, 1996). There are also a good many newspapers (Buckley cites 60 in operation),

38

WORLD DEVELOPMENT this continuous process of getting ready for an election is exactly what provides the major scope for many dierent groups in society to get involved.

but low adult literacy levels (29%, according to UNDP, HDR, Various years) are a problem here. 25. Parties, of course, are in many ways primarily focused on elections (especially opposition parties), but

REFERENCES
Bearak, B. (1996). The battle against the bottle. Los Angeles Times, December 12. Binswanger, H., & Shah, A. (1994). Decentralization, scal systems, and rural development. Revised proposal. Washington, DC: The World Bank, June 24. Blair, H. (1995). Assessing democratic decentralization: A CDIE concept paper. Washington, DC: USAID. Blair, H. (1997). Democratic local governance in Bolivia. CDIE Impact Evaluation No. 3, Washington, DC: USAID. Blair, H. (1998). Spreading power to the periphery: An assessment of democratic local governance,.USAID Program and Operations Assessment Report No. 21, Washington, DC: USAID, CDIE, October. Buckley, S. (1996). Mali: Tuned in and democratic. Washington Post, March 26. Cohen, J. M., & Peterson, S. B. (1995). Administrative decentralization strategies and for the 1990s and Beyond. United Nations Secretariat, Governance and Public Administration Branch, Division and for Public Administration and Management Development, Department and for Development Support and Management Services, New York. Crook, R., & Manor, J. (1994). Democracy and decentralization in South Asia and West Africa: participation, accountability and performance. Cambridge: Cambridge University Press. Esman, M. J., & Upho, N. T. (1984). Local organizations: intermediaries in rural development. Ithaca, NY: Cornell University Press. Hansen, G. (1996). Constituencies and for reform: Strategic approaches and for donor-supported civic advocacy programs. USAID Programs and Operations Assessment Report, Washington, DC, USAID/ CDIE. Inbanathan, A. (1992). The new panchayati raj in Karnataka: An evaluation. New Delhi: Institute of Social Sciences. Inbanathan, A., Bhagyalakshmi, T. A., & Doraiswamy, R. (1997). Democratic local governance in Karnataka State, India. Paper commissioned by CDIE, USAID, Washington, DC. Korten, D. C. (1990). Getting to the 21st century: voluntary action and the global agenda. West Hartford, CT: Kumarian Press. Lippman, H., & Blair, H. (1997). Democratic local governance in Ukraine. CDIE Impact Evaluation No. 2, USAID, Washington, DC. Lippman, H., & Jutkowitz, J. (1997). Democratic local governance in the Philippines. CDIE Impact Evaluation No. 1, USAID, Washington, DC. Lippman, H., & Lewis, B. (1998). Democratic decentralization in Mali: A work in progress. CDIE Impact Evaluation Report No. 5, USAID, Washington, DC. Lippman, H., & Pranke, P. (1998). Democratic local governance in Honduras. CDIE Impact Evaluation Report No. 4, USAID, Washington, DC. Mackinnon, I. (1997). Men sobered as Indias women seek new order. Scotsman, August 13. Manor, J. (1999). The political economy of democratic decentralization. Washington, DC: The World Bank. Mishra, S. N. (1997). Poverty alleviation programmes and gram panchayats. New Delhi: Mittal Publications. Moore, M. (1993), Indian village women ght state, husbands to ban liquor. Washington Post, December 19. OECD (Organization and for Economic Cooperation and Development) (1997). Final Report of the DAC Ad Hoc Working Group on Participatory Development and Good Governance, Parts I and II. Paris: OECD. Putnam, R. D. (1993). Making democracy work: civic traditions in modern Italy. Princeton NJ: Princeton University Press. Rondinelli, D., Nellis, J. R., & Cheema, G. S. (1984). Decentralization in developing countries: A review of recent experience. World Bank Sta Working Papers No. 581, Washington, DC: The World Bank. Rondinelli, D., McCullough, J. S., & Johnson, R. W. (1989). Analysing decentralization policies in developing countries: a political-economy framework. Development and Change, 20 (1), 5787. Rood, S. (1997). Democratic decentralization in the Philippines. Chicago: Paper for Annual Meeting of the Association and for Asian Studies, March 1316. Timberman, D. G. (1990). A changeless land: continuity and change in Philippine politics. Singapore: Institute of Southeast Asian Studies. UNCDF (United Nations Capital Development Fund), (1995). Poverty Reduction, Participation and Local Governance: The Role for UNCDF. New York: UNCDF. UNDP (United Nations Development Program), (Various years). Human Development Report. New York: Oxford University Press and UNDP. UNDP (1997a). Governance for sustainable human development: A UNDP policy document. New York: UNDP. UNDP (1997b). Draft. Management Development and Governance Division. New York: Bureau for Policy and Programme Support. UNDP (1997c). Participatory local governance. Technical Advisory Paper 1, Management Development and Governance Division, Local Initiative Facility for Urban Environment. New York: UNDP. USAID (United States Agency for International Development) (1997). NPI [New Partnerships Initiative]

PARTICIPATION AND ACCOUNTABILITY Resource guide: A strategic approach to development Partnering. Washington, DC: Report of the NPI Learning Team. USAID (1998). 1997 agency performance report. Washington, DC: Center and for Development Information and Evaluation. Van Sant, J., Blair, H., Razon-Abad, H., & Amani, T. (1998). Impact assessment: Governance and local democracy project. Research Triangle Park, NC: Research Triangle Institute. Report prepared for USAID/Manila.

39

Wilson, W. J. (1987). The truly disadvantaged: the inner city, the underclass, and public policy. Chicago: University of Chicago Press. Wilson, W. J. (1996). When work disappears: the world of the new urban poor. New York: Alfred Knopf. World Bank (1998). Bangladesh: From counting the poor to making the poor count. Report No. 17534BD, The World Bank, South Asia Region, Poverty Reduction and Economic Management Network, Washington, April 29.