44849 LOCAL GOVERNANCE & ACCOUNTABILITY SERIES Paper No. 111 / July 2008 Electoral Institutions and Local Government Accountability: A Literature Review Daniel Packel Social Development Papers LOCAL GOVERNANCE & ACCOUNTABILITY SERIES Paper No. 111 / July 2008 Electoral Institutions and Local Government Accountability Daniel Packel This Working Papers Series disseminates the findings of work in progress to encourage discussion and exchange of ideas social development issues. The papers carry the names of the authors and should be cited accordingly. The series is edited by the Community Driven Development team in the Social Development Department of the Sustainable Development Network of the World Bank. This paper has not undergone the review accorded to official World Bank publications. The findings, interpretations and conclusions herein are those of the author(s) and do not necessarily reflect the views of the International Bank for Reconstruction and Development/ World Bank and its affiliated organizations, or its Executive Directors, or the governments they represent. To request copies of the paper or for more information on the series, please contact the Social Development Department Social Development The World Bank 1818 H Street, NW Washington, DC 20433 Fax: 202-522-3247 E-mail: socialdevelopment@worldbank.org Printed on Recycled Paper Table of Contents Acknowledgments.....................................................................................................................................ii Acronyms...................................................................................................................................................iii Summary Findings....................................................................................................................................iv Introduction................................................................................................................................................1 Key Electoral Variables .............................................................................................................................3 Partisan vs. Non-Partisan Electoral Systems......................................................................................3 Electoral Systems....................................................................................................................................4 Nomination Rules...................................................................................................................................6 Competitiveness .....................................................................................................................................9 Re-election and Recall..........................................................................................................................10 Affirmative Action ...............................................................................................................................12 Conclusions...............................................................................................................................................14 Discussion..............................................................................................................................................14 Research Questions and Hypotheses.................................................................................................14 References .................................................................................................................................................17 i Acknowledgments This paper benefited from the comments and suggestions of Jesse Ribot, Varsha Venugopal, and Serdar Yilmaz. ii Acronyms BPL Below Poverty Line FPTP First-Past-the-Post PR Proportional Representation SC/ST Scheduled Caste/Scheduled Tribe iii Summary Findings In studies of decentralized governance in the developing world, electoral institutions have largely been overlooked as explanatory factors for the performance of local governments. Instead, much of the current research on accountability and performance in local governments has focused on macro-social factors as determinants of accountability in local governments. This paper does not discount the importance of these macro-social variables, but it proposes an institutional focus on local government success, examining the role of local electoral arrangements on accountability. The paper argues that even if electoral institutions cannot compensate for social forces that are hostile to the exercise of downward accountability; given favorable social conditions, certain electoral arrangements will likely improve accountability and provision of public goods in local communities. Through evaluating results from Latin America, Africa, and Asia, the paper argues that the mixed findings surrounding the relationship between degree of electoral competition and public goods provision in local governments illustrates the secondary importance of electoral institutions. The paper identifies rules surrounding partisanship in elections, nomination rules, term-limits, and recall provisions as key variables likely to influence accountability and performance in local governments. The paper also examines local electoral systems (proportional representation vs. first-past-the-post systems), the degree of competition in local elections, and rules surrounding affirmative action in local elections. iv Introduction Across the developing world, within the last several decades, decentralization has been touted as a strategy for countering corruption and inefficiency and improving the provision of public goods. However, empirical work in a variety of national contexts has shown that the formal establishment of local governments does not necessarily lead to the improvements outlined above. Consequently, current research has become more concerned with accountability in local governments and with the conditions under which decentralized governance succeeds in providing public goods, from services like sanitation, water, and other types of infrastructure development to maintenance and development of local natural resources. Much of this research has focused macro-social factors and their influence on accountability and efficacy in local governments. A vibrant civil society is one key factor that has been linked to the success of decentralized governance in meeting the expectations outline above; a prominent example of this relationship is the state of Kerala, in India (Heller 2000). Elsewhere, central and regional states have succeeded in taking steps to empower civil society at the local level, thus strengthening the capability of local governments, as illustrated by Tendler (1997) with regard to Northeast Brazil. Indeed, a concern with the substantive dimension of democracy rather than the process dimension of democracy has dominated recent research interrogating the conditions under which decentralization succeeds or fails. More attention has been focused on the social factors that determine whether individuals at the local level have the capabilities to engage in democratic decision-making and whether local institutions have substantive powers than has been focused on the actual procedures for exercising voice, specifically, rules surrounding elections for local councils.1 Scholars of decentralization do assert that fair, competitive and regular elections compel local politicians to exercise power in such a way that decentralized institutions provide efficient and fair outcomes (Echeverri-Gent 1993; Crook and Manor 1998; Blair 2000), but few provide any systematic insight into whether certain electoral arrangements produce better outcomes than other arrangements. While elections may be only one of several instruments of downward accountability (Ribot 2003), little attention has been devoted to how specific electoral mechanisms fare in delivering accountability, or even whether elections are used by voters to hold officials accountable for certain policy decisions (Rodden 2004). While this lacuna is indicative of a fuller understanding of democratic decentralization that recognizes the insufficiency of a sole focus on formal institutions, the general literature on elections does show that differing electoral arrangements shape how citizens exercise influence on policy makers (Powell 2000). This oversight regarding local elections is not just the case in the literature on democratic decentralization, but also in the even larger literature on electoral systems. These studies are almost entirely concerned with electoral rules at the national level, with a particular focus on the relationship between electoral systems and national party systems. Even where electoral procedures differ between national and local levels (for example, certain countries ban parties 1In a way, this tendency can be seen as a response to previous trends in the study of democratization in the developing world that over-emphasized formal institutions of democracy and failed to capture how social factors often eroded democratic authority. See O'Donnell (1993) and Heller (2000) for more details on this phenomenon. 1 from participating in local elections), electoral studies still focus almost entirely on the national level. As one example of how local elections are frequently overlooked in the existing literature, the chapter length overview of India's electoral system in the recent comprehensive volume The Politics of Electoral Systems (Gallagher and Mitchell 2005) does not even acknowledge the presence of an elected third tier of governance in this country even though local governments have existed since a 1993 constitutional amendment and the country has approximately three million elected representatives at the local level (D'Souza 2003). This review therefore aims to begin filling this void at the intersection of these two literatures, identifying existing relevant research findings and suggesting opportunities for expanding the inquiry in the relationship between electoral arrangements and accountability. It identifies key areas where variation in both electoral institutions and the de facto practice of elections (specifically, the degree of electoral competition) is likely to have an impact on the efficacy of local governments. For some of these areas, little information exists beyond snapshots of different practices in place in different countries. In other areas, some studies have taken advantage of either intra-country or inter-country variation to provide a more systematic understanding of these relationships. Thus, while a growing literature exists on the relationship between electoral competition and local government performance; less literature considers the influence of partisan systems as compared to non-partisan systems as well as the difference between majoritarian and proportional systems of representation, nomination rules and provisions for recall and reelection. Other areas that may have a notable impact on accountability and performance, including rules surrounding campaign finance, information provision regarding elections, and the extent of the franchise are completely absent from the existing literature. Thus, this review is to be read as an examination of the existing literature that also contains some suggestions on how to improve our understanding of areas that have already received minimal to moderate attention. 2 Key Electoral Variables PARTISANVS.NON-PARTISANELECTORALSYSTEMS One key institutional variation in local governments across the developing world is the issue of whether parties have a place in local elections. Advocates of non-partisanship in local elections maintain that local government pertains to "bread and butter" issues on which there can be no division along party lines (Olowu 2003). Consequently, inclusion of parties at the local level risks allowing policy-making to become contaminated by patronage and clientelism. Political pressures thus prevent policy-makers from retaining a focus purely on long-term benefits. This is because elected officials may be focused on securing re-election or delivering benefits to their narrow client base, rather than delivering policies that benefit the entire community in the long run (Lankina 2007). Additionally, officials may be more concerned with taking measures to ensure their promotion and advancement within the internal party structure, rather than in passing policies that benefit the community (Ahmad, Devarjan et al. 2005) cited in Lankina (2007). In both Ghana and Uganda, parties are outlawed in local elections, with the rhetoric that this rule will ensure that merit, not party affiliation, is the basis of representation (Crook 1999; Francis and James 2003). India's panchayats (local governments) also operate on a non-partisan basis by law. Proponents of non-partisanship also argue that the presence of parties at the local level impedes openness and competition. Where parties are present, proposals may be determined in committees, not in open government council. This arrangement can stifle debate and separate the public from decisions. The presence of competing parties may also make campaigns more savage, discouraging some qualified candidates. Additionally, the nomination procedures of partisan elections prevent members of the same party from running for election, further winnowing the pool of qualified candidates (Ayee 2004). Allowing parties to participate in local government, on the other hand, acknowledges the link between local government and national government. Examples of systems allowing partisanship in local elections are nearly all of the recently decentralizing Latin American countries, including Bolivia, and Mexico. In Africa, prominent examples of party-based systems are Nigeria, Kenya, South Africa, Cote d'Ivoire, and Senegal (Agrawal and Ribot 1999; Olowu 2003). Even in political systems that explicitly prohibit parties from contesting seats in local elections, national and regional parties often unofficially back candidates in local elections, even if this support is not denoted on the actual ballot. This is certainly the case in India, where party officials at the state level proffer statistics about the number of local governments controlled by the party and village political leaders will admit to being aligned with a certain party, even though local elections operate on a non-partisan basis by law (Packel 2007). Ghana also witnesses this practice, where parties unofficially sponsor individual candidates (Ayee 2004). Thus, permitting parties to compete in local elections allows the inescapable link between local government and higher levels of government to exist openly, allowing for greater transparency. Where a strong national party system exists, parties have significant incentives to mobilize at the local level in order to improve their electoral chances at higher levels (Blair 2000). 3 Thus, preventing parties from participating in local elections permits decentralized governance to retain the negative qualities of partisan politics, without gaining the benefits. These benefits include the scrutiny and public accountability that comes from an officially acknowledged party opposition on local councils, an opposition that has clear incentives to be vigilant and uncover instances of wrongdoing (Crook 1999; Blair 2000). Additionally, non-partisan candidates are often elected based on their individual popularity, rather than the strength of their ideas (Ayee 2004). Recognizing the presence of parties at the village or municipality level thus may allow more transparent representation of interests. Furthermore, political competition built around personalism may be just as likely, if not more likely to lead to relationships of clientelism and patronage than political competition built around partisanship (Garcia-Guadilla 2002). These observations, collected from a variety of national contexts, suggest that permitting partisanship in local elections may ultimately be conducive to promoting accountability. The empirical evidence shows that non-partisanship is either unenforceable or used as a mechanism to promote the hegemony of a specific party. Indeed, the opening of local politics to real competition among parties in a number of Latin American countries is a key component of the shift away from single-party rule at the national level. As a corollary to this point, the mere presence of partisan elections at the local level may not lead to increased downward accountability where there is little political competition at the national level. With a single party dominating at the national level, a local government led by a different party may receive very little in discretionary funds or earmarks from the national government. This may serve as a powerful disincentive for electing representatives from parties that have little power at the national level. This condition would then obviate the expected benefits of political competition, particularly when national parties exert great control over nominations for candidates at the local level (discussed in greater detail below). Thus, the proposed advantages of openly partisan elections appear to be linked to substantive political competition at the local and national level. ELECTORALSYSTEMS In the study of elections, from the local level to the national level, the choice between proportional representation (PR) and election through single-member districts through plurality votes, also known as first-past-the-post (FPTP), is a key institutional variation. In systems of proportional representation, voters generally vote for a party, rather than a specific candidate. Electoral districts contain multiple representatives, and the share of votes received by a particular party is then translated by a fixed formula into the number of seats to be held by that party. Thus, proportional representation works to ensure that a political party's degree of representation appropriately reflects the electoral support for the party. The alternative to systems of proportional representation is election based on majority or plurality voting within single- member districts. Here, voters in a given electoral district vote for a specific candidate. With only one representative per electoral district, voters have greater clarity of representation; they have no doubt who is directly charged with accounting for their interests. The chief drawback with this system is that there is no guarantee that minority interests receive any electoral representation. This arrangement becomes a particular concern where the minority interests are equally distributed across the polity; a party or group that has only a slight minority could easily lose elections in each electoral district, leaving it with no representation whatsoever (Farrell 2001; Gallagher and Mitchell 2005). 4 Ultimately, while electoral systems2 receive a great amount of attention in the larger political science literature, there are very few systematic studies evaluating the relationship between varying institutional arrangements and accountability in local governments. Electoral studies almost always examine the effect of varying electoral rules on political outcomes at the national level, and are frequently concerned with elucidating relationships with national party systems (Taagepera 2007). This lacuna with regard to local governments is the case for both advanced industrial democracies as well as states in the developing world.3 In the advanced industrial democracies, majority or plurality votes in single-member districts determine elections for most local councils. In the emerging systems of democratic decentralization in the developing world, there is more variation, ranging from pure PR systems, systems mixing PR with elections from single-member districts, first-past-the-post elections in single member districts, to arrangements where the winning party takes all the council seats allocated through the election. But one reason for the limited presence of PR systems in the developing world is that many polities do not permit partisanship in local elections, as is discussed above. Without official recognition of parties, candidates can only be elected on an individual basis, so elections based upon party list are impossible. Because of the lack of studies critically assessing the outcomes of different institutional variations in electoral systems, the following paragraphs simply overview systems in place in a variety of states, identifying their salient features. In the several nations that do employ PR at the local level, the size of the local government is often larger than one or several villages. In Latin America, Mexico and Bolivia are two prominent examples of local governments receiving increased autonomy within the last several decades. Bolivia, where the lowest tier of government is the municipality, divides its population of roughly 8 million among 314 municipalities, which include the nation's largest cites as well as rural areas. Mexico, similarly, is divided into approximately 2,400 municipalities, which vary dramatically in size, ranging from over one and a half million people to just over one hundred.4 Both these countries elect municipal councils through PR, where the mayoral candidate is situated at the top of the party list. In both of these systems, the mayor is the key figure in the government, setting the agenda for action making determinations about what projects to pursue (Hiskey and Seligson 2003; Grindle 2006). In the case of Bolivia, this combination of a strong mayor with PR appears to have negative consequences on accountability. When no party wins an absolute majority of seats on the municipal council, the mayor is decided by a vote of the council. This arrangement leads to the mayor being selected on the basis of "back-room bargaining" by the local elites and national party figures, rather than a mayor that is directly accountable to the electorate (Hiskey and Seligson 2003). 2Here, the concept of "electoral systems" is defined narrowly, as a subset of "electoral laws." The latter refers to the laws regulating all facets of the election process, including laws on nomination procedures, the characteristics of the franchise, and how campaigns are conducted, while the former refers only to the rules that determine "the means by which votes are translated into seats in the process of electing politicians into office." (Farrell 2001: 15). 3One study from Switzerland does suggest that PR at the local level, especially in smaller municipalities, increases the amount of partisan activity in local elections, which then stems the general trend of declining voter turnout (Milner and Lander 2006). 4Municipalities in Mexico are roughly equivalent to counties in the United States. They are composed of a county seat (cabacera) and the surrounding rural communities (Grindle 2006). 5 One example of a mixed system is South Africa. Here, the 284 municipalities across the country elect their councils with half of the seats allocated by PR and the other half allocated by first- past-the-post elections (Beall 2005). Many remaining systems rely on single member districts where council members are elected by a plurality or have committee members elected on an at-large basis from the entire village. Within India's panchayati raj system, the lowest tier of government, the gram panchayat, covers several villages or one village if it is large enough. These panchayats generally have a size range of between 4,000 and 10,000 people, although some panchayats are as small as 500 people, and in the southern state of Kerala, panchayats can be as large as 60,000 people. Panchayats are divided into wards and each ward has a single elected representative. China's system of village representation involves elections for village chairperson, vice chairperson as well as committee members.5 As party competition is not allowed at any level within China, candidates are elected on their personal merits. Likewise in Uganda, villages elect nine independent candidates on an at-large basis to local councils (Bazaara 2006). Among partisan systems that do not employ proportional representation, Senegal's local council elections sit at the opposite end of the spectrum from PR. Candidates for rural councils must be presented for election from nationally registered parties, and the winning party slate receives ¾ of the seats in a winner-take-all election (with the remaining members chosen by a council of state-organized producer and marketing cooperatives) (Agrawal and Ribot 1999). Agrawal and Ribot assert that accountability suffers in this particular institutional arrangement, as villagers have limited opportunity to select individual candidates and only parties organized at the national level can participate. Additionally, the rules of the election completely exclude minority interests from representation. Ultimately, the issue of electoral systems is in part dependent on the place of parties in local governments. Where parties are prevented from competing, officials can only be elected from single-member districts or on an at-large basis. Opening local governments to partisan competition enables the possibility of proportional representation. However, there is very little evidence regarding whether proportional representation provides more or less accountability than alternative institutional arrangements. Additionally, the electoral institutions in local governments in the various countries noted above vary significantly along multiple variables, not just electoral systems, and these variables are not necessarily independent of each other (like electoral system and partisanship.) This will provide a challenge to proving any relationship, especially when the existing observations provide little guide to even crafting hypotheses. NOMINATIONRULES The example of Senegal noted above, in which candidates for local elections can only be presented by national parties, suggests that rules governing nomination of candidates for local 5China's laws on local governments mandate that village committees are to consist of three to seven members. In addition to the chairperson and vice chairperson, villages generally elect accountants. Other possible positions include a public security officer and a representative to the state-organized women's federation (Tsai 2007). As such, committee members are elected on a village-wide basis, not according to specific wards. 6 councils are also likely have consequences on the accountability of elected officials. As with electoral systems, there is little literature that compares different arrangements for nominating candidates. However, a review of relevant case studies shows that one major distinction is the role that national parties have with regard to nominations. Where local elections occur on a partisan basis, nomination rules that favor national parties can serve as impediments to downward accountability. This is the case with Senegal, mentioned above, where only nationally registered parties can field candidates for local elections. In a study comparing six Latin American countries that have taken formal measures to decentralize forest management, Larson (2003) notes with concern that in five of these cases (Bolivia, Nicaraugua, Costa Rica, Honduras, and Brazil), local candidates are selected by national parties. Employing the logic used by Agrawal and Ribot, these arrangements are also likely to diminish the capacity of local actors to exercise accountability on local elected officials, as their choices for selecting representatives are likely to be limited. Heller (2001) finds a similar tendency on South Africa, where he cites the control of the "hegemonic" ANC (African National Congress) party over local nominations as both a cause and a symptom of weakened local governments. Thus, both de jure and de facto national party control over nominations for local councils appears to have negative consequences on accountability at the local level. The case of village elections in China provides another angle in evaluating the issue of nominations and barriers to entry. China, as a one-party state, does not feature multiple parties fielding candidates for local elections, so there are no laws restricting nominations to certain national parties. At the same time, the particular history of local elections in China has led to a wide degree of institutional variation.6 As a result, rules for nomination of candidates for village elections vary across China, and power over these rules rests with provincial authorities. Several researchers have taken advantage of this internal variation, using it to produce statistical studies about the effect of institutions for nominations on government performance. The conclusions of two of these statistical studies suggest that at least within China, nomination procedures that are more open to popular participation at the village level, do not lead to improved outcomes in accountability. Wang and Yao (2007) identify three different categories of institutions for nominating Village Committee chairpersons: nomination by villager representatives, popular nomination, or mixed nomination (combined government appointment and popular nomination.) In a statistical analysis that uses the type of nomination as a proxy for the competitiveness of elections, Wang and Yao compare villages employing these three procedures with the baseline category of government appointed candidates for chairperson (a procedure existing prior to the electoral law of 1998). They measure accountability of village committees by examining the share of public expenditures and share of administrative costs in the village's budget, positing that where village committees are more accountable, the former should increase while the former should decline. Their findings show that villages with popular nominations (the most competitive form of nomination), when compared to elections with government appointed candidates, result in a lower share of public expenditures. Manion (2006) also takes advantage of the uneven progress of grassroots democratization in China to test the 6Local elections in China began following the 1987 Village Committee law, which mandated direct elections every three years at the village level, but did not provide guidance on what kind of elections were to be held and how they were to be conducted. Consequently, even after the new 1998 law on village elections, which aimed at greater standardization of practice, there is still significant variation in procedures across the country (Tan 2004). 7 influence of the openness of the nomination process on how villagers perceive their local leaders, especially with regard to probity and corruption. Her statistical study, coding nomination as a three-level ordinal variable, with higher levels indicating greater openness to the public, finds that high openness in nomination has no discernable influence on how villagers view the trustworthiness of local leaders. Still, both the studies cited above only consider China, which is a rather unique case, as democratic institutions only exist at the local level, while the authoritarian Communist Party controls politics at the regional and national level. As a result, there are good reasons to be skeptical about the applicability of these findings to polities that feature some degree of democratic competition at all levels. At the very least, these studies provide a template for future research that could test differing nomination procedures on a cross-national basis. One additional variable, largely ignored in the existing literature, is whether or not independent candidates are allowed to contest elections in partisan systems. This provision, which exists in Mali, presents the possibility of checking the dominance of national parties over candidates, providing an additional mechanism for downward accountability. Still, the efficacy of this provision is unexplored, as it is uncertain that independent candidates can receive sufficient backing and support to challenge candidates supported by institutionalized parties. 8 COMPETITIVENESS Tied to the issue of procedures for nomination is the question of whether the competitiveness of elections influences accountability in local governments. Indeed, nomination rules can be seen as simply one particular barrier or enabler of competition. The intuitive hypothesis is that greater competition will fuel accountability, as local officials will have a greater incentive to provide visible services and engineer effective programs if there are more individuals competing for election to their seats. This hypothesis is rooted in certain strands of democratic theory: following Joseph Schumpeter (1943, 1976), the act of electoral competition for leadership positions allegedly underpins the very concept of democracy. But as with the other areas of inquiry addressed thus far, there are a limited number of systematic studies that evaluate the connection between competition in elections and the performance of local governments. With regard to competitiveness, valuable evidence comes from China, again, as well as Latin America, where a surprising number of studies examine the case of Mexico. While Wang and Yao (2007) above used the nomination process in local elections in China as a proxy for the degree of competition, finding that popular nomination is linked with lower levels of public expenditure, Manion (2006) separates nominations from competition, identifying the ratio of candidates to positions on the village committee as the measure of electoral contestation. With this ratio as an explanatory variable in her statistical model, she finds a clear positive relationship between contestation and trust in elected officials. In Mexico, similarly to China, even if for differing reasons, the democratization of local politics has moved forward, if haltingly and unevenly, over the last several decades. As a result, the dissolution of one-party rule has left a landscape in which some municipalities have higher levels of electoral competition for control of local governments than do others. Ward (1998) hypothesized that in Mexico, the shift from one-party rule to increased competition at the local level would lead to more technical competence in public administration. Cleary (2007) exploits the presence of variation in electoral competition, examining the relationship between the margin of victory in municipal elections and the performance of elected governments. He measures the latter by focusing on public utility provision, specifically sanitary sewers and potable water. Ultimately, Cleary's statistical analysis finds no clear relationship between the margin of victory in a specific municipality (averaged over a ten year period) and either sewer or water coverage. A similar study by Grindle (2007), using a smaller sample of 30 medium-sized municipalities, likewise finds no discernible relationship between competition and performance, the latter measured by an index of indicators in five different categories (efficiency, effectiveness, responsiveness, change initiatives, and development orientation). Moreno (2005) also failed to find a discernable relationship between electoral competition (again, measured by the margin of victory) and basic service provision in Mexico.7 However, in one contradictory study from Mexico, Hiskey (2003) treats electoral competition as a categorical, rather than a continuous, variable and finds that municipalities where the dominant PRI (Institutional Revolutionary 7Moreno does find that voter turnout rate has a statistically significant positive influence on service provision. This finding, along with a positive relationship between literacy and service provision, allows him to conclude that "demand" factors (participation rates, literacy rates) have a greater impact on service provision than "supply" factors (closeness of electoral competition, number of parties competing.) 9 Party) won all elections in a five-year period with a vote share of over 65% had poorer records of public service provision than did municipalities in which the PRI lost at least one election during the same period and had an average vote share of less than 60%. The results from Mexico, excepting Hiskey's findings, appear to be typical for Latin America. Kauneckis and Andersson (2006) test the relationship between competitiveness and local government performance in a cross-national statistical study of nearly four hundred municipalities across four Latin American countries: Mexico, Brazil, Chile, and Peru. They find that that competition, measured by the difference in vote share between first and second place parties in the last election for mayor, had no discernable influence on provision of natural resource management services. These results, when viewed together, may not be an indictment of competitiveness as a completely inconsequential factor, but rather a reminder that electoral variables are only one potential determinant of accountability in local governance. In the absence of a proper enabling environment, or where other institutional hurdles exist8, the proposed benefits of competition may not come to fruition. RE-ELECTIONANDRECALL Re-election procedures are another potential determining factor of accountability on local councils, although, inductively, two opposing relationships are possible. Term limits could foster accountability by preventing local politicians from becoming entrenched in their positions and locked into relationships of patronage. Representatives that were very recently private citizens would thus be more attuned to the concerns and interests of the community than career politicians. Yet if term limits are too restrictive, then council members may not have sufficient opportunity to master the responsibilities of their position before their terms expire. Even where other institutional variables may favor accountability, if elected officials never gain familiarity with their responsibilities, then their ability to govern will be compromised. Furthermore, the theory of retrospective control of politicians asserts that, given perfect information, voters use elections to reward and punish politicians--without the possibility of re-election, elections lose their power as an instrument of control, and politicians may engage in greater degrees of rent- seeking (Maravall 2007). Even when the assumptions of perfect information and the utility- maximizing behavior of politicians are relaxed, reducing the explanatory power of the retrospective theory, eliminating the possibility of re-election may at the very least reduce the opportunities for accountability. Most of the research concerning the relationship between re-election and accountability comes from national elections. The existing literature on local governments largely ignores the question of term limits, such that many case studies do not even reveal whether they exist or not for the country in question. Overviews of local councils in Africa provide no information on whether term limits exist, even for prominent cases like Senegal and Ghana (Agrawal and Ribot 1999; Crook 1999; Olowu and Wunsch 2004). The lack of discussion of this topic suggests that term limits do not exist, but this is still unconfirmed. India's panchayats have no term limits, 8Cleary (2007) cites term-limits as one such hurdle, which will be addressed below. 10 although reservation of seats for women and certain caste groups can serve as a de facto term limit, for these reservations rotate between different panchayats.9 One country that does impose term-limits on local elected officials is Mexico, where mayors and local council members are elected for three years and are legally barred from holding the same position again for one term (Grindle 2006). Both Cleary and Grindle argue that these term limits served as an impediment to performance. The limited amount of time that any councilor is in office makes it difficult to gain policy expertise and familiarity with bureaucrats and institutions (Cleary 2007). Additionally, the countervailing belief is that the longer programs are in place, the more likely they could be sustained beyond the three-year incumbency (Grindle 2006). This attitude, generated by the institutional constraint of term limits, reduces the window for policy implementation, therefore limiting the efficacy of local government. Comparative studies between Mexico and Latin American nations that share other institutional features might be able to provide more robust evidence for the detrimental effects of term limits. Other countries have different institutions that can limit the life of local elected governments, particularly procedures for recall. Because they generally occur only every 4 to 5 years, elections are a rather blunt instrument of accountability. To compensate, some countries have implemented procedures for recall, a process allowing councils or popular bodies the ability to dismiss elected leaders because of wrong-doing. Unlike elections, in which voters' decision- making is likely to incorporate expectations for the future as well as evaluation of actions taking in the past, the decision to recall is more confined to evaluation of past actions. In Africa, recall provisions are present in Nigeria and Ethiopia (Olowu 2003). India's panchayati raj law varies across different states, but some of the most progressive legislation for recall exists in Madhya Pradesh, where the gram sabha (village assembly- open to all adult residents) has the right to dismiss the panchayat chairperson in the event of wrongdoing (Johnson, Deshingkar et al. 2005). In Latin America, Bolivia's law allowing council members to recall mayors has attracted attention, because of its widespread use. This provision, known as the voto constructivo de sensura, allows the municipal town council to remove the mayor with a three-fifths majority vote in cases of misconduct. However, the high rates of exercise of this provision (for example, in 1997, one year after the first mayors entered office following implementation of decentralizing reforms, thirty percent of mayors were replaced) indicated to observers that the voto constructivo was being used as a political maneuver, rather than a response to corruption (Hiskey and Seligson 2003). Indeed, this high use of the provision demonstrated that the procedure of recall actually served as a hindrance to accountability, as mayors wound up being selected by council members, rather than being chosen by the electorate. Hiskey and Seligson used survey data to compare municipalities in which the voto constructivo was employed to municipalities in which mayors served out their full terms, and they found that where mayors were recalled, citizens voiced a lower level of support for the political system than in municipalities where the recall 9India's laws for local governments mandate that 1/3 of all seats as well as 1/3 of all council heads (pradhans) are reserved for women. Additionally, Scheduled Castes and Scheduled Tribes receive reservations for these positions that reflect their numerical presence within a given state. To provide these reservations, particular seats rotate from general quota seats (in which anyone can stand for election) to seats reserved for these specific groups. For example, a pradhan might not be able to stand for re-election because her seat is being reserved for a Scheduled Caste individual in the next electoral cycle. 11 was not exercised. They use this finding to support the assertion that the recall provision was not used to stem corruption, but rather as a partisan weapon, used by local officials to obtain power for themselves and their parties. This finding ultimately suggests that if the provision for recall does exist for local officeholders, the specific design of this institution matters, if it is to be used as an instrument of accountability. Recall alone may not engender downward accountability, if the only actors capable of exercising this power are beholden to national political parties. In Bolivia, the particular combination of overt partisanship in local elections combined with a recall procedure restricted to the members of the municipal council rendered recall ineffectual with regard to the aim of downward accountability. The contrasting case, with regard to downward accountability, would appear to be Madhya Pradesh, India, which gives the power of recall to the entire village assembly, or gram sabha, an institution that is by law non-partisan. However, as of yet, no comparable studies of the "right to recall" in Madhya Pradesh exist. Ultimately, case studies evaluating the consequences of the use of recall in Madhya Pradesh or polities with similarly "populist" provisions could provide greater insight regarding the efficacy of this instrument. AFFIRMATIVEACTION One final variable worth considering is the subject of affirmative action, or reservations, for council seats. In systems that employ reservations, a certain number of seats in any given district are restricted such that only women and or members of certain ethnic groups can compete for election. Such reservations are generally adopted to compensate for inequalities in representation, where women and/or subordinated ethnic groups suffer from lack of a political voice and often a disproportionately low level of public service provision. As such, reservations are of a slightly different category from the variables considered above, for their presence does not aim to promote accountability of leadership or public goods provision, which benefit the community as a whole. Instead, reservations are aimed at improving the fortunes of certain groups within the community, and are thus often evaluated with regard to their impact on specific groups, rather than the entire community. Indeed, any community-wide effects are likely to be considered secondary to the effects observed on targeted groups. Prominent examples of affirmative action include India, which maintains reservations for both women and ethnic groups (Scheduled Castes and Scheduled Tribes), and South Africa. In South Africa, which features a mixed system for allocating seats to local councils, electoral laws mandate that for the seats allocated by PR, half of the names on the party list must be women, and these names must be distributed across the list, rather than concentrated at the bottom. Legislation also aims at ensuring equal numbers of men and women for seats that are allocated ward-wise, although evidently this policy does not entail a system of reserved seats, and is thus not as effective in ensuring equal representation at the ward level than the Indian system.10 Both the reservations for subordinate castes as well as women in India have been the subject of several 10Beall (2005) finds that in South Africa, these reforms aimed at increasing women's representation on local bodies have had a limited impact on the responsiveness of government to women's concerns, primarily because they have been accompanied by policies that have institutionalized the influence of male-dominated traditional authorities in local governments. 12 econometric studies, assessing whether reservations lead to improved public goods provision for members of the subordinate groups. The effects of reservations for women in India are largely mixed. In a study of two Indian states, Rajasthan and West Bengal, Chattopadhyay and Duflo (2004) exploit the institutional feature that seats reserved for women pradhans (mayors) are randomly distributed and find that in panchayats with women elected through reservations, investments in public goods that are associated with women's needs (drinking water in both states, roads in Rajasthan) are higher than in panchayats without reservations for women. However, in a study of four states in South India, Ban and Rao (2005) find that reservation of seats for women alone has a very limited impact on the activism of local governments, when the latter was considered as an index of eight different activity areas. Only when reservations are considered in combination with other variables do conclusions become less ambiguous: women in reserved seats with higher levels of education or political experience generate higher levels of activity in public goods provision. In a related study of the four South Indian states, Besley, Pande et al. (2005) found that reservations for women for the position of pradhan had a limited impact on the distribution of Below Poverty Line (BPL) cards, entitling holders to subsidized food and other benefits. The one noteworthy consequence of these reservations was that they increased the likelihood that politicians held BPL cards, suggesting that either reserved women pradhans actually hold less monitoring power or they are more prone to personal aggrandizement than non-reserved pradhans (Besley, Pande et al. 2005: 24). Related studies of reservations for Scheduled Caste and Scheduled Tribe (SC/ST) pradhans in India reveal the anticipated results that reservations are likely to influence positively the distribution of benefits to select groups but have limited influence on the entire community. In the study of four South Indian states, SC/ST reservations increase the likelihood that SC/ST households and reserved politicians have BPL cards (Besley, Pande et al. 2005). An additional related study examines the impact of SC/ST reservations on distribution of public goods in villages. In this study, Besley, Pande et al. (2004) found that reservations had little consequence on high-spillover public goods (for example, infrastructure improvements like roads, street lights and drains) that benefit the entire community, but reservations do increase the provision of low- spillover public goods (for example, programs that subsidize improvements to housing, private electricity, and water supply to beneficiaries). Specifically, SC/ST households were more likely to receive these benefits when SC/ST reservations were in place. 13 Conclusions DISCUSSION Considering the amount of literature available on both decentralization as well as electoral politics, the limited attention paid to elections for local councils is rather surprising. Most likely, some degree of this oversight can be attributed to a welcome recognition that electoral arrangements, and formal democratic institutions in general, are only one determinant of local government performance. For example, where central governments hold weak commitments to fostering autonomy in local governments, and where local communities show high levels of social fragmentation and economic inequality, even the most "optimal" electoral arrangements could be expected to only have a minimal influence on the performance of local governments in providing public goods. This is not to say that electoral arrangements are without consequence, but rather to assert that they may be of secondary importance when compared with macro-social factors. Indeed, the multiplicity of variables hypothesized to have an impact on the success of local governments would make it very difficult to isolate the specific influence of electoral arrangements, even if a certain "optimal" set of electoral arrangements could be identified. The research on the impact of competition in local elections highlights the above points. Some of the most sophisticated research undertaken thus far on elections and local government performance has focused on competitiveness in local elections. While research on other electoral factors has involved little more than identification of varying institutional arrangements, the research on competitiveness proceeds from the expression of clear hypotheses to the testing of these hypotheses using systematically collected data. Even though the hypothesized relationship between competition and local government performance is that more competitiveness will lead to both increased provision of public goods and greater accountability in elected officials, studies from China, Mexico, and elsewhere in Latin America found ambiguous and somewhat conflicting results with regard to this relationship. One potential explanation for these ambiguous results is the presence of other influences that were not included in the statistical models. Perhaps the introduction of an interaction term measuring the favorability of macro- social factors might lead to a more robust relationship between competition and local government performance, although this would depend on the presence of some variation in this new variable. Furthermore, measuring macro-social factors or enabling environments with one variable in a statistical model presents another set of methodological concerns with regard to measurement of the variable. RESEARCHQUESTIONSANDHYPOTHESES Even if the consequences of electoral competition remain ambiguous, this review has identified several electoral institutions likely to have an influence on accountability and performance in local governments. Recognizing the underdeveloped state of findings of elections for local governments, these observations can be used to generate hypotheses for further empirical research. Up to the present, electoral variables have largely been examined as minor variables in studies of decentralization, or highlighted as descriptive features of how decentralization works. 14 To test the following hypotheses, future research should aim to treat electoral variables as primary variables and build empirical models around them for both intra-country and cross- country analyses. Variables that likely influence local government accountability and performance include presence of partisanship in elections, nomination rules, term limits, recall provisions, and campaign finance laws. · Partisan systems of local representation will lead to better outcomes in accountability at the local level. Existing empirical observations suggest that where an established party system exists at the national level (and at the state level, in the case of some federal systems), non-partisanship is an unenforceable arrangement. Formally recognizing the inevitable presence of some degree of party organization at the local level will make the interest groups at play in local politics more transparent to citizens. In some countries, the rhetoric of non-partisanship has also been used to prevent opposition to a single, dominant party; thus the presence party-based competition at the local level illustrates a greater range of competition and choice than non-partisanship that masks barriers to competition. Furthermore, some of the other alleged benefits of non-partisanship, including decreased patronage and an increased concern in governing for the entire community, rather than certain groups, likely do not stand up to increased scrutiny. In addition, recognition of parties expands the choice of electoral systems available to local governments. While little research exists on the influence that proportional representation, as opposed to majoritarian representation, holds on local governance; where local communities are divided along ethnic or religious lines, proportional representation might allow more thorough representation for minority groups. However, this is only possible if social groups are allowed to organize on a partisan basis. · Where higher-level officials in national parties control the process of nominating candidates for local elections, this exercise of authority is likely to shift accountability away from the community level. Even though the existing empirical tests of this proposition show no evidence that more open nomination practices lead to greater public good provision, these findings do come from the unique case of China, in which local quasi-democratic institutions exist in a larger environment of authoritarianism. Thus, we cannot yet discount the hypothesis linking local nomination of candidates to increased accountability without evaluating it within other national contexts. Instead, it is necessary to test this hypothesis cross-nationally in a sample of countries that have multi-party democracies. · Within partisan electoral systems, lower barriers to the entry of independent candidates will likely increase accountability, by providing alternatives to hierarchical political parties. While the phenomenon of independent candidacies is largely unrecognized in the existing literature, it is likely that under certain circumstances, openness to independents could provide a counterweight to the influence of national parties. This research will depend on preliminary efforts to generate basic data, identifying countries that by law permit independent candidates as 15 well as the subset of countries in which independents frequently contest local elections in practice. · Term limits may prevent the exercise of accountability though retrospective voting. They also may prevent elected officials from gaining competency in exercising their responsibilities, if terms are of too short a length. Term-limits are proposed as a strategy for preventing authorities from becoming entrenched in their positions and building networks of patronage. However, at the local level, where candidates may have limited formal education or technical expertise, term-limits may prevent the acquisition of experience and competence among elected officials, limiting their ability to serve the interests of the community and negotiate with a more experienced bureaucracy. Furthermore, term-limits may deprive voters of an important mechanism to exercise accountability via retrospective voting. · The efficacy of recall provisions, which allow the electorate or local elites to abbreviate the terms of elected officials in the case of malfeasance, is likely to depend on the specific features of the provisions. Preliminary evidence from the Bolivian case suggests the hypothesis that where the power of the recall rests in the hands of the local political elite, downward accountability in officeholders is actually weakened, with corresponding negative effects on how citizens perceive their local governments. It thus follows that where community-wide referendums can be used to censure or remove underperforming officials from office, accountability in local governments will be strengthened. · More stringent regulations of the flow of money into local elections may lead to greater accountability. The question of financing of local elections is currently overlooked in the existing literature. Intuition and evidence from national-level elections across the world suggests where no limits exist to the financing of elections, policy outcomes are likely to be skewed to reflect the interests of those groups or interests with greater wealth. However, within the developing world, and particularly with regard to local elections, enforcement of these regulations may be limited. As such, studies of campaign financing must complement consideration of regulations with nuanced observation of how regulations are followed or skirted in practice in order to determine the impact of this variable. 16 References Agrawal, A. and J. C. Ribot (1999). "Accountability in Decentralization: a Framework with South Asian and West African Cases." Journal of Developing Areas 33(4): 473-502. Ahmad, J., S. Devarjan, et al. (2005). Decentralization and Service Delivery. Washington DC, World Bank. Ayee, J. (2004). Ghana: A Top-Down Initiative. Local Governance in Africa: The Challenges of Democratic Decentralization. D. Olowu and J. S. Wunsch. Boulder, CO, Lynne Rienner. Ban, R. and V. Rao (2005). Tokenism or Agency? The Impact of Women's Reservations in South India. Washington, DC, World Bank. Bazaara, N. (2006). "Subjecting Nature to Central Authority: The Struggle over Public Goods in the Formation of Citizenship." Africa Development 31(2): 21-37. Beall, J. (2005). "Decentralizing Government and Decentering Gender: Lessons from Local Government Reform in South Africa." Politics and Society 33(2): 253-76. Besley, T., R. Pande, et al. (2004). "The Politics of Public Goods Provision: Evidence from Indian Local Governments." Journal of the European Economics Association 2(2-3): 416- 26. Besley, T., R. Pande, et al. (2005). Political Selection and the Quality of Government: Evidence from South India. Washington, DC, World Bank. Blair, H. (2000). "Participation and Accountability at the Periphery." World Development 28(1): 21-39. Chattopadhyay, R. and E. Duflo (2004). "Women as Policy Makers: Evidence From a Randomized Policy Experiment in India." Econometrica 72(5): 1409-43. Cleary, M. (2007). "Electoral Competition, Participation, and Government Responsiveness in Mexico." American Journal of Political Science 51(2): 283-99. Crook, R. (1999). ""No Party Politics" and Local Democracy in Africa: Rawlings' Ghana in the 1990s and the Ugandan Model." Democratization 6(4): 114-38. Crook, R. and J. Manor (1998). Democracy and Decentralization in South Asia and West Africa. Cambridge, Cambridge University Press. D'Souza, P. (2003). "The Struggle for Local Government: Indian Democracy's New Phase." Publius: The Journal of Federalism 33(4): 99-118. 17 Echeverri-Gent, J. (1993). The state and the poor : public policy and political development in India and the United States. Berkeley, University of California Press. Farrell, D. (2001). Electoral Systems: A Comparative Introduction. London, Palgrave. Francis, P. and R. James (2003). "Balancing rural poverty reduction and citizen participation: The contradictions of Uganda's decentralization program." World Development 31(2): 325-37. Gallagher, M. and P. Mitchell, Eds. (2005). The Politics of Electoral Systems. Oxford, Oxford University Press. Garcia-Guadilla, M. P. (2002). "Democracy, Decentralization, and Clientelism: New Relationships and Old Practices." Latin American Perspectives 29(5): 90-109. Grindle, M. (2006). "Modernising Town Hall: Capacity Building with a Political Twist." Public Administration and Development 26(February): 55-69. Grindle, M. (2007). Going Local: Decentralization, Democratization, and the Promise of Good Governance. Princeton, Princeton University Press. Heller, P. (2000). "Degrees of democracy: Some comparative lessons from India." World Politics 52(4): 484-519. Heller, P. (2001). "Moving the state: The politics of democratic decentralization in Kerala, South Africa, and Porto Alegre." Politics and Society 29(1): 131-63. Hiskey, J. T. (2003). "Demand-Based Development and Local Electoral Environments in Mexico." Comparative Politics 36(1): 41-59. Hiskey, J. T. and M. A. Seligson (2003). "Pitfalls of Power to the People: Decentralization, Local Government Performance and System Support in Bolivia." Studies in Comparative International Development 37(4): 64-88. Johnson, C., P. Deshingkar, et al. (2005). "Grounding the state: Devolution and development in India's panchayats." Journal of Development Studies 41(6): 937-70. Kauneckis, D. and K. Andersson (2006). Making Decentralization Work: A Cross-National Examination of Local Governments and Natural Resource Governance in Latin America. Available at http://sobek.colorado.edu/~anderssk/06_Making%20Decentralization%20Work_Kauneck is_Andersson_v5.pdf, accessed 20 November 2007. Lankina, T. (2007). Cross-Cutting Literature Review on the Drivers of Local Council Accountability and Performance, World Bank. 18 Larson, A. M. (2003). "Decentralisation and forest management in Latin America: Towards a working model." Public Administration and Development 23(3): 211-26. Manion, M. (2006). "Democracy, Community, Trust: The Impact of Elections in Rural China." Comparative Political Studies 39(3): 301-24. Maravall, J. M. (2007). Accountability and the Survival of Governments. The Oxford Handbook of Comparative Politics. C. Boix and S. Stokes. Oxford, Oxford University Press. Milner, H. and A. Lander (2006). "Can PR Voting Serve as a Shelter Against Declining Turnout? Evidence from Swiss Municipal Elections " International Political Science Review 27(1): 29-45. Moreno, C. (2005). Decentralization, Electoral Competition, and Local Government Performance in Mexico, University of Texas, Austin. Ph.D. O'Donnell, G. (1993). "On the State, Democratization and Some Conceptual Problems: A Latin American View with Glances at Some Postcommunist Countries." World Development 21(8): 1355-69. Olowu, D. (2003). "Local Institutional and Political Structures and Processes: Recent Experience in Africa." Public Administration and Development 23(1): 41-52. Olowu, D. and J. S. Wunsch, Eds. (2004). Local Governance in Africa: The Challenges of Democratic Decentralization. Boulder, CO, Lynne Rienner. Packel, D. (2007). Unpublished Fieldnotes. Powell, G. B. (2000). Elections as Instruments of Democracy. New Haven, Yale University Press. Ribot, J. C. (2003). "Democratic decentralisation of natural resources: Institutional choice and discretionary power transfers in sub-Saharan Africa." Public Administration and Development 23(1): 53-65. Rodden, J. (2004). "Comparative Federalism and Decentralization: On Meaning and Measurement." Comparative Politics 36(4): 481-500. Schumpeter, J. (1943, 1976). Capitalism, Socialism, Democracy. London, Allen and Unwin. Taagepera, R. (2007). Electoral Systems. The Oxford Handbook of Comparative Politics. C. Boix and S. Stokes. Oxford, Oxford University Press. Tan, Q. (2004). "Building Institutional Rules and Procedures: Village Election in China." Policy Sciences 37(1): 1-22. 19 Tendler, J. (1997). Good Government in the Tropics. Baltimore, Johns Hopkins University Press. Tsai, L. (2007). Accountability without Democracy. Cambridge, Cambridge University Press. Wang, S. and Y. Yao (2007). "Grassroots Democracy and Local Governance: Evidence from Rural China." World Development 35(10): 1635-49. Ward, P. (1998). "From Machine Politics to the Politics of Technocracy: Charting Changes in Governance in the Mexican Municipality." Bulletin of Latin American Research 17(3): 341-65. 20