Public Good Provision and Traditional Governance in Indigenous Communities in Oaxaca,Mexico

Abstract

Can ethnically distinct communities ruled through “traditional” assemblies provide public goods and services better, than when they are ruled by leaders elected through “modern” multiparty elections? We exploit a unique institutional feature in the state of Oaxaca, Mexico, where municipalities are ruled by traditional governance institutions, to explore the effect of these forms of governance on the provision of public goods. Using locality-level census data, we study the provision of local public goods through a geographic discontinuity approach. We demonstrate that communities ruled by traditional governance practices offer more effective provision of local public goods than equally poor communities ruled by political parties. Relying on qualitative fieldwork and household surveys, we argue that the significant differences in the provision of public goods according to governance regime derive from community practices that solve collective action problems, enhance citizen participation in public decisions, and restrain elite capture.

Keywords

subnational politics Latin American politics decentralization political economy politics of growth/development

Introduction

States in developing countries wield uneven authority over their territories. Tribes, clans, religious groups, and other traditional leaders often control parallel governance zones beyond the reach of the state. African chiefs continue to play important roles in local government (Baldwin, 2015), and in some countries, “traditional”¹ political authority mechanisms have been accorded greater status in local and regional administration (Holzinger, Kern, & Kromrey, 2016). An all too common view is that traditional authorities should be understood as historic relics, waiting to be replaced as nation-states gradually pave the road to institutional modernity.²

In recent years, however, there is increasing awareness of the pervasive role that traditional governance structures may play in development (Acemoglu, Reed, & Robinson, 2014; Díaz-Cayeros, Magaloni, & Ruiz-Euler, 2014). When the centralized nation-state is unable to improve people’s lives substantially, because of weak or dysfunctional “modern” institutions, societies often tap into cultural traditions usually preserved by some distinct ethnic communities to offer alternative forms of governance. This article explores whether traditional governance may actually be better able to shape an equitable provision of local public services and prevent the all too common propensity for political parties and privileged elites to seize the benefits of development projects. In the political science literature, communities with distinct ethnic identities have often been studied as prone to clientelism and even political violence (Chandra, 2007; Posner, 2005). Here, we explore, instead, the role of such distinctiveness as a source of alternative forms of governance that lead to greater accountability and well-being.

We explore this question through the case of Oaxaca, Mexico, where a constitutional reform in 1995 formally validated traditional governance structures, in what is known as usos y costumbres. In Mexican regions with a strong indigenous presence, traditional governance practices coexist in parallel to formal institutions. Holzinger et al. (2016) suggest that such coexistence is not unusual but that in fact more than half of the countries in the world have “dual polities” embedded in traditional governance arrangements. The Mexican “first polity” establishes democratic multiparty elections as the prime mechanism for selecting local leaders, and the court systems as the main form of adjudication of conflicts. However, in indigenous places like Oaxaca, a “second polity” exists, where assemblies, rather than mayors or local councils, make the most important budgetary decisions, citizens are compelled to carry out tasks or duties for the provision of public goods, and local leadership itself can take various forms including rotating appointments. Conflicts are solved through informal community mediation, rather than resorting to courts.

The greatest challenge in the research agenda of dual polities lies in the difficulty untangling the effects of those practices—which are usually informal—tolerated or even sanctioned by the state, from the general social, economic, and political conditions of the places where they are used. We exploit a constitutional reform in 1995 in the state of Oaxaca, Mexico, where municipalities were assigned to be either ruled by traditional usos y costumbres institutions or by mayors elected through multiparty elections.³

It is important to note that the choice of governance institution did not change the fiscal relations of local governments with the federal government: Municipalities kept identical systems of intergovernmental transfers, public debt, public finance accounting and auditing rules, and, in general, the administration of public expenditure and investment projects. Such unique setting provides an opportunity to tease out the effect of traditional governance institutions on public goods and services provision, controlling for the fiscal conditions and administrative processes that in this instance remain constant across governance types.

To explore how traditional governance in indigenous Mexico determines the provision of public goods, we use a multimethod research strategy based on qualitative ethnographic work,⁴ a household survey that sheds light on differing patterns of local political accountability and citizen participation in public affairs, and a statistical analysis that models the provision of basic local public goods like sewerage and drinking water through a geographic discontinuity (GD) approach, relying on census data. The statistical approach aims to address problems of causal identification by allowing us to study the provision of public goods in villages that are very similar in terms of their sociodemographic features and geographic location, except that they differ by governance type.

We demonstrate that the allocation of local public goods is more effective and equitable in usos y costumbres municipalities. Survey and ethnographic evidence suggests, moreover, that traditional institutions distribute local public goods in a relatively egalitarian fashion and tend to favor the poor. In communities governed by political parties through elections, in contrast, water and other public good provision are more inequitable in their distribution and their allocation is politically mediated, tending to favor households aligned to the mayor’s party.

The multiple method approach (Box-Steffensmeier, Brady, & Collier, 2008; Dunning, 2012; Humphreys & Jacobs, 2015; Seawright, 2016) allows us to better interpret our statistical findings and to understand mechanisms. By triangulating inferences emerging from qualitative fieldwork, secondary ethnographic materials, a carefully designed field survey, and statistical analysis mindful of selection bias problems, we provide insight into forms of governance that are often elusive to outside observers.

Our approach stresses that the profound institutional differences among governance forms affect the provision of public goods through three mechanisms: (a) stronger social embeddedness of municipal presidents leading to greater local government accountability in usos y costumbres municipalities, (b) broader civic engagement in collective decision making, which produces more informed citizens and decisions that are more representative of the median voter preference, and (c) an effective system of community service enforced through credible sanctions, which allows for widespread cooperation with activities that promote the well-being of the community.⁵

Our findings are related to the literature on direct participatory democracy and public good provision. In a seminal study, Olken (2008) used an experimental setting to evaluate the impact of direct democracy on the provision of public goods in villages in Indonesia. Projects were decided through either representative-based meetings or direct election-based plebiscites. His results suggest that direct participation in political decision making substantially increased satisfaction and legitimacy, but it had little effects on the actual types of projects that were chosen. Although our study uses observational data, our results reveal more profound differences between direct and representative democracy that go beyond perceptions of government legitimacy to include patterns of distribution of public goods and citizen engagement.⁶

Our study also sheds light on the quality of Mexico’s local democracy in the context of fiscal decentralization. Diaz-Cayeros, Estevez, and Magaloni (2016) demonstrate that the decentralization of federal transfers for financing public goods in Mexico had positive effects on development, as measured by improvements in the provision of local public services and ultimately an improvement in the quality of life by reducing infant mortality rates. However, fiscal decentralization is no panacea, given the high levels of political corruption and lack of electoral accountability that have prevailed in Mexico that are in part the result of a constitutional prohibition of reelection at all levels of government.⁷

To explore if local government corruption could be restrained through increases in citizen information, Ana de la O and her coauthors randomly assigned information to voters on political corruption and the misuse of decentralized federal funds in various municipalities in Mexico (Chong, De la O, Karlan, & Wantchekon, 2014). In their study, exposing rampant corruption leads not only to incumbents’ vote erosion but also to a decrease in political participation as measured by electoral turnout. Thus, their study demonstrates that, under some circumstances, exogenously provided information about local government corruption disengages voters from the political process and can potentially be counterproductive.

Our study complements Chong et al. (2014), suggesting that we need to pay more attention to the way citizen information is endogenously produced. Our work in Oaxaca reveals that preexisting levels of citizen information about public affairs differ in important ways between party and usos y costumbres municipalities. Only in the later do citizens appear to be better informed about public affairs. Citizen information results from the system of government, based on active participation in community assemblies and deliberation involving the allocation of public funds. Better informed citizens, in turn, appear to be associated with more accountable local leaders.

The article is organized as follows. Drawing from extensive qualitative and quantitative fieldwork, in the following section, we discuss the institution of usos y costumbres in Oaxaca. The section provides insights into how to understand citizen participation, information, and perceptions of legitimacy, as well as the distribution of services and degree of elite capture in each of these governance types. The “Traditional Governance Institutions” section presents the geographic discontinuity design, contrasting the effect of governance regimes on community-level indicators of public goods coverage. The final section concludes.

The Origins of Usos y Costumbres in Oaxaca

Southern Mexico is primarily populated by indigenous peoples.⁸ Municipalities in that region are governed by a combination of informal indigenous forms of self-rule, dating back to the colonial times, and formal institutional rules, where candidates nominated by political parties compete for political authority in democratic elections. Indigenous governance varies across states, from one ethnic group to the other, and among communities within any given region.

Traditional governance institutions are not static but are always evolving according to the formal institutional environment indigenous communities interact with (Recondo, 2007; Stephen, 2013). However, some general patterns can be highlighted, including participatory democracy practices, mandatory unpaid community work which ranges from short-term to long-term commitments, strong systems of social control that put the community above the individual, and a parallel system of conflict resolution that is essentially conciliatory.

During the colonial period, indigenous towns were granted significant autonomy provided they converted to Catholicism and submitted to the powers of Crown and Church. After Mexico’s independence, the autonomy of these indigenous communities was increasingly challenged. Liberal politicians interested in expanding the powers of the central state imposed stronger military control over the territory. Hacienda landowners despoiled Indian towns from their ancestral lands. Indigenous communities often rebelled to keep their land, and sometimes succeeded in escaping the reach of a dominant nonindigenous State.

In the aftermath of a bloody civil war at the beginning of the 20th century, that included many elements of indigenous rebellion, the Mexican Constitution of 1917 paradoxically undermined the autonomy of indigenous communities. Although the Constitution incorporated reforms to enact land reform and assert peasant rights, it did not truly recognize indigenous communities. The new Constitution sought to assert a peasant identity, rather than an ethnic, linguistic, or cultural distinctiveness for rural communities. Although communal land and collective ownership became recognized, this was sanctioned by federal laws (within the ejido structure) rather than any significant form of local indigenous autonomy. Furthermore, municipal governments and the Comisariado Ejidal, a collective body overseeing communal landholding arrangements, were under the firm control of the Partido Revolucionario Institucional (PRI). Hence, indigenous traditional governance practices were only tolerated by the hegemonic party as long as they refrained from challenging the party’s central political monopoly and the corporatist intermediation of peasant organizations affiliated to the party.

This accommodation of the indigenous communities to PRI hegemonic rule changed dramatically in 1994, when the Ejército Zapatista de Liberación Nacional (EZLN) erupted in the southernmost state, Chiapas (Trejo, 2009, 2012). Less well know than the Zapatista rebellion, the neighboring state of Oaxaca was discussing a constitutional reform that allowed its municipalities to adopt an alternative governance structure: usos y costumbres, an umbrella that encompasses various governance forms, but in which political parties are not allowed to formally endorse candidates to run for municipal office (Eisenstadt, 2011). The 1995 reform of the Constitution of that state recognized usos y costumbres, a complex and unique mixture of traditional practices where local authorities are selected in assembly elections according to indigenous practices instead of regular multiparty elections. Of the 570 municipalities in that state, 418 adopted usos y costumbres.⁹ Although there is some controversy over this point, we believe municipalities were allowed to select themselves into the traditional form of governance. No other state in Mexico has officially recognized such an overarching right to indigenous self-determination.¹⁰

Mayors are in charge of making decisions over investments in local public good provision: provision of drinking water, public lighting, sewerage, street pavement, markets, slaughter houses, parks, graveyards, and local public safety, among other. Both party and usos y costumbres municipalities have to abide by the federal Constitution that forbid, up to 2018, immediate reelection. It was not uncommon, however, for mayors in usos y costumbres to serve for more than one term, although normally not in consecutive terms. Citizens ruled for municipal affairs under usos y costumbres still hold the constitutional right to vote for political parties and their candidates in all state and federal level elections but do not allow political parties to take over local government decision making, including the allocation of public goods and services.

In contrast to the struggles in the state of Chiapas after the Zapatista rebellion, Oaxaca effectively preempted explosive social and political conflicts by granting an important degree of autonomy to indigenous communities through the validation of practices they were already following in the selection of municipal authorities. Several authors, however, believe that the institutional reform in fact enhanced conflicts, particularly, in the electoral arena (see Eisenstadt, 2007, 2011; Ríos, 2006; Eisenstadt & Ríos, 2014).¹¹ Allison Benton (2012, 2016, 2017) has argued that the selection into these system of governance is mostly a reflection of the authoritarian entrenchment of the PRI at the local level. In contrast, Díaz-Cayeros et al. (2014) and Recondo (2007; Chapter 8) find that electoral competition at the federal level is more prevalent in places ruled by usos y costumbres.¹²

David Recondo (2007) discusses the origins of those reforms (p. 213-215) as an impetus spearheaded by a group of anthropologists that became the main advisors of the incoming governor at the time, Diódoro Carrasco. Recondo’s nuanced view of the dynamics of political accommodation between the imperatives of the PRI (the party to which the governor belonged) and the capacity of indigenous communities to assert their rights is often lost in some superficial reading of his book. Eisenstadt (2007) suggests that the reform was motivated by the Zapatista uprising in Chiapas. But it is important to note that the state of Oaxaca was on a very different political trajectory compared with Chiapas before the rebellion. Oaxaca was not governed by White or Ladino elites or landowners but rather by an ascendant indigenous political class, well represented by Heladio Ramirez, governor from 1986 to 1992. Land conflicts were important, but they were mostly related to disputes between indigenous communities that already had communal land rights, rather than involving landless peasants demanding reform. We believe the state Constitution was reformed in 1995 allowing for traditional governance mostly due to internal processes in the state, rather than a demonstration effect from events in Chiapas.

We do not believe this system of governance is an authoritarian imposition. Survey data collected in rural and semirural municipalities in the state of Oaxaca show that in places ruled by usos y costumbres, a vast majority prefers these traditional forms of governance over the alternative of being ruled by political parties.¹³ An overwhelming 90.8% of respondents told us they prefer to be ruled by usos y costumbres than political parties. In contrast, 65.5% of respondents under party governance believe political parties divide their community and provoke conflict, whereas 54.7% feel that public benefits are captured by partisanship. Granting that there is considerable debate among scholars regarding the political motivations behind the reform of 1995, we should note that there is a general consensus suggesting that the shift had momentous implications for local governance in the state.

Traditional Governance Institutions

The key institutional difference between usos y costumbres governance municipalities and those ruled by political party competition are (a) the form of selection of municipal presidents, (b) participatory democracy practices, (c) existing sanctioning mechanisms to solve coordination problems, and (d) alternative forms of conflict resolution. We discuss each element in detail. At the end of this section, we also discuss intergovernmental fiscal relations, highlighting that municipalities, regardless of governance type, share the same administrative framework and equivalent financial resources for to pay for the provision of public goods. Thus, variations in public good provision cannot emerge from differences in funding or financial flows but are determined by institutional variation in governance forms.

Leader Selection

Usos y costumbres municipalities choose leaders in community assemblies. This is the most distinguishing trait that determined, when the reform was approved in 1995, which municipalities would be assigned to the traditional governance regime. Our ethnographic and survey work show that for the most part municipal presidents in usos y costumbres places tend to be elected after climbing the ladder of community service or having a known record of performing volunteer services (cargos) in the community. Moreover, these leaders often tend to remain in their communities, performing other cargos after their terms are over (e.g., they might join the Elders Council or become Regidores).

This means that leaders in usos y costumbres municipalities not only tend to have a proven record of serving their community but also are more constrained to abuse their powers because they will most likely remain in their communities afterward. There are, of course, exceptions. In interviews, we identified cases wherein local leaders could not go back to live in their communities after they served their terms because they had abused their powers or stolen from the public coffers. But when a local leader is to continue living in the community after serving his or her term, the threat of expulsion is probably a strong enough incentive to induce greater accountability.

In party municipalities, the nomination of candidates for the mayor’s office and their promotion to future political positions respond to the political game at the state level. Given that reelection was not allowed, party elites and governors decided whether mayors got promoted to higher office after their term was over. There are little incentives for municipal presidents that are embedded in party politics to serve their constituencies, especially, because mayors often do not stay in their localities after they finish their term. Our field research revealed that municipal presidents in party municipalities more often than not climb the ladder of progressive ambition by becoming state officials (e.g., local deputies, state bureaucrats, or even federal senators). Political parties appear to be a vehicle for upward mobility that breaks the political relationship of mayors to their communities and enhance agency control problems for the community.

Our survey asked perceptions of how accountable municipal president are perceived to be. Table 1 compares these responses across governance types. Respondents in usos y costumbres are more likely to believe that the mayor cares about people’s problems, works to serve the community, handles money honestly, and consults citizens to make decisions. These differences are all statistically significant.

Table 1.

Perceived Accountability of Municipal President.

Municipal president	Usos (%)	Party (%)
Does not care for people’s problems	32	43
Works to serve the community	68	50
Handles money honestly	63	49
Consults citizens to make decisions	67	52

Participatory Democracy

A second important difference between party and usos y costumbres municipalities relates to participatory democratic practices. In usos y costumbres, municipal presidents actively consult community assemblies for making decisions. Many collective decisions concerning the everyday actions of government are taken item by item, mostly in assemblies where there is a process of collective deliberation. In most cases, attendance to community assemblies is mandatory. Most of the time, a long deliberation takes place—some assemblies can even last up to 10 hours—and citizens often raise their hands to vote. Traditionally, women were excluded from participating in many assemblies, although our field research revealed that this practice has changed.¹⁴

In party municipalities, executives usually make political decisions without much public scrutiny once in office. Formally, party municipalities are composed of a president (the executive) and a separate cabildo (assembly), although by design, the largest party has a majority and the government is always unified.¹⁵ In theory, the municipal president should consult the cabildo for making decisions and open assembly meetings could be held, where citizens may be called to participate. In reality, we found little of such participatory practices taking place. This means that in party municipalities, executives usually make decisions unchecked which stands in sharp contrast to usos y costumbres municipalities, where community assembles take place regularly.

Regarding public goods projects, our fieldwork showed that usos y costumbres municipalities tend to hold either biannual or yearly assembly meetings to set priorities as to how to spend these resources. In usos y costumbres municipalities, the leaders of the agencias, who are also elected by community assemblies in the surrounding villages, are often called to form part of what is sometimes called a Committee of Development (Comité de Desarrollo), where they discuss with the municipal president and with some of the traditional authorities how infrastructure budgets (primarily financed by a federal transfer, the Fondo de Aportaciones para la Infraestructura Social Municipal [FAISM]) should be spent.¹⁶ Moreover, there are often egalitarian rules regarding how projects should be distributed among villages (an example of such rule is that each village receives a large project every other year and a small project every year).

The visibility and publicness of assembly deliberations about what to do with public money is a fundamental difference between usos y costumbres and parties, which sets off control mechanisms that produce variation in terms of responsiveness and accountability toward the community. In party municipalities, decisions about project type, location, and benefited population are done within the municipal president’s inner circle, with little to no consultation among the population—except for the nominally public cabildo (town meeting) sessions, which citizens rarely attend.

Furthermore, usos y costumbres municipalities have mechanisms to broadcast information about public affairs that are absent in party representative systems. The process of direct democracy and deliberation that takes place in the assemblies reinforces a social network where individuals share information and deliberate. Preferences over public spending are deliberated openly both inside and outside of the assemblies, and the allocation of projects (type and population target) is done visibly for the community. Of course that participation varies across places. In the municipality of Diaz Ordaz, for example, we were told that “the majority does not attend the assembly, if for example we are supposed to be 500, only 100 show up.” In other places, however, almost everyone participates and citizens—especially the young—even complain that when meetings last for hours, the obligation can be taxing.

A crucial problem with these traditional institutions is that they require a substantial commitment from citizens to participate and volunteer. Such public mindedness is less prevalent among the youth. An elderly man in San Juan Guelavia told us, “Our children want another way of life,” referring to the gradual erosion in the practices of usos y costumbres. In his view, there were more leftover “customs” rather than the original governance by usos y costumbres: “People no longer cooperate in tequio and servicio, so they were often fined 200 pesos [20 USD at the time] for not participating, or, rarely, they lock you up. People fail to go to the assemblies, they participate less than before.”

Our survey asked questions about whether citizens participate in collective decision making and how informed they are about public matters. Our survey found that 44% of the respondents find out about the government actions in town meetings, whereas only 8% of the inhabitants ruled by political parties report that they learn about how government is being run by going to these meetings. Moreover, 62% of the respondents ruled by traditional governments report being informed about public decisions, including investments in public goods. This number is significantly lower, 38%, among citizens in party municipalities.

The typical voting method in community assemblies, according to our survey work, is 78% of the time through lifting hands in public. There is an additional 10% that our respondents report make decisions through secret ballot, and another 10% that count votes on a blackboard.

Mechanisms to Solve Collective Action

A third important distinctive trait of usos y costumbres municipalities is their system of cargos or tequio that allows these municipalities to solve coordination problems and to sanction those who refuse to collaborate with activities that promote the well-being of the community.

In usos y costumbres municipalities, members of the community are expected to perform public services which are usually mandatory. Traditionally, none of the cargos are for pay, although field research reveals that gradually many usos y costumbres municipalities are paying some cargos, most notably that of municipal president. Because of the system of cargos or tequio, citizens become deeply involved in the life of the community: organizing religious festivities; coordinating public works; cleaning the roads; policing the streets, rural roads, and fields; repairing the school; organizing the public market; decorating the church; and so on. According to our survey data, 60% of respondents had held a cargo of various types in the past. An almost universal 92% of respondents carried out a form of volunteer service (tequio) in the prior year.

Cargos can be exceedingly taxing. For example, Mayoles organize regular religious festivities and keep the local church. The water committee monitors water use and the wells’ reserves, sanctions wasteful practices, and fixes water pipes. Members of the education committee organize school-related activities, repair the classrooms, fix the school toilets, and make sure that school teachers show up to teach the children. In one remote municipality, for instance, we found that parents organized a committee that was rotated weekly to oversee the assistance and behavior of the school teacher. A mother of four told us that parents had problems with the previous teacher who gave classes drunk. This committee took the case to the assembly, and they decided to expel the teacher from the community. After some bureaucratic hurdles with the state government, they were assigned a new teacher of better quality.

Policing is another cargo done by the topiles (community police), in charge of patrolling the streets and rural roads, often without vehicles. In Teotitlán del Valle, a critical perspective on cargos came from this observation:

for me it is a little bit of a problem the matter of having to sometimes fulfill a duty, it is true, very true, that as part of our community it is our duty, right? Because we live here and we rule ourselves by usos y costumbres, but on the other hand I can see that in most cases for the young men, particularly when they decide to marry, they are affected.

This man went on to tell us that “when he got married the community started to have its eye on me and told me I had to perform my duty, starting from the lowest one.” This was discussed in the context of him being chosen to serve as topil.

Tequio is used for a wide range of small and short-term public works projects. In out fieldwork, we have reports of tequio being used for fixing the local school, church, or health clinic, digging a well, filling potholes, expanding the market, garbage collection, cleaning the town and the graveyard, fixing the gardens, planting trees, caring for the elderly, or unclogging the sewers. Larger public works projects—including street pavement, road construction, building a health clinic, or laying water or sewerage systems—are constructed through a mix of resources: FAISM funds for the bulk of the material and assorted construction equipment, whereas local men and women donate their labor to the community.

Cargos and tequio are a precondition to acquire community rights. Those who refuse to perform their cargos or tequio are sanctioned. Traditionally, sanctions include monetary punishment, imprisonment, disruption in the supply of water, physical punishment, or in extreme cases, banishment from the community. In our ethnographic work, we found evidence of a gradual abandonment of the latter sanctions. Some of our informants attributed the gradual abandonment of physical punishment to the presence of the Human Rights Commissions, which receive complains of human rights violations from members of the communities. Our qualitative field research revealed that the mandatory character of these public services is so binding that men living undocumented in the United States are also summoned to participate in a cargo, and they (or most often someone in their families) usually comply to avoid sanctions.

Women are also responsible for many cargos and increasingly seem to play roles of greater responsibility. In our field interviews, we found women engaged in caring for the public library, being in charge of the cooperativa store, providing community kitchens, taking care of the community phone, participating in committees for feasts (both civic and religious); all the way to serving as policewomen (topiles), treasurers, members of the municipal council, regidores in charge of various public services, and even mayors.

Compulsory, nonremunerated labor in these communities can be thought of as a form of taxation where the community decides tax structures, which most prominently include these nonpecuniary contributions, endogenous to the level of public goods provision. In party municipalities, by contrast, taxation is mostly exogenous to the provision of public goods: Political leaders at the national level decide tax bases and rates, and investment decisions at the local level are often made independent of these.

Dispute Resolution

Problems often arise within the communities when some families find the burden too high or unacceptable for some reason, including religion. Lynn Stephen (2005) provides a vivid account of a protestant family in a Zapoteco community Oaxaca who refused to perform cargos due to religious beliefs. The community sanctioned them by denying their right to burial within the community cemetery when one family member died. After days of negotiations and an obvious health hazard, they arrived to a middle-ground agreement by which partial participation was promised.

This exemplifies how cargos and tequio are bundled to other political and community rights. Our field interviews give some insight into the forms of social control implicit in this system. In Villa Diaz Ordaz, a man told us, shrugging, that if one does not provide tequio, then he or she will not get paid work in the construction of public works or be allowed to use the communal lands. Discussing the temptation to shirk from those responsibilities, he told us that “lies do not work here, because if you continue lying or not collaborating, you will lose the right to your land.”

A fourth institutional difference between party municipalities and usos y costumbres municipalities is that the party municipalities have a professional armed police force and deal with conflicts through the judiciary. Conflict resolution and adjudication takes the regular judicial channels, which are often located outside of the communities and are administered by judges that have little knowledge of the locality. Usos y costumbres municipalities have a parallel system of conflict resolution and adjudication and a parallel security system. The community charges topiles with the task of protecting the security of its people, although they are seldom armed. The assembly, a counsel of elders, an indigenous tribunal, and/or the municipal president can hear conflicts and impose sanctions. The mechanisms for dispute resolution lie within the community and are exercised by its members who enjoy a high level of legitimacy. By contrast, in party municipalities, law enforcement and conflict resolution is done through the armed police and the judiciary, delegitimized almost throughout Mexico.

Fiscal Flows for Public Good Provision

Regardless of whether they are ruled by parties or usos y costumbres, all municipalities in Oaxaca share the same fiscal relationship with the federal government. The most important federal transfer used by municipalities to fund public works, the FAISM is granted directly to the municipal government, regardless of form of governance. The distribution of these resources is determined by a federal formula, based on poverty indicators, and its allocation cannot be modified by governors. Although municipalities sometimes fail to use all the available funds, there is no evidence of political manipulation in their allocation (Diaz-Cayeros et al., 2016).¹⁷

Once resources arrive to the municipality, the municipal president has ample discretion to select projects and distribute transfers among villages. Hence, mayors in both systems possess great leeway in deciding how these considerable financial resources get distributed within the municipality. But in usos y costumbres places, they are more likely to consult citizens through the assemblies for those decisions. As federal transfers for social infrastructure projects within the FAISM have significantly increased since 1997, municipal governments have played an increasingly larger role in the provision and distribution of local public goods.¹⁸ Thus, the budget for the provision of public goods such as sewerage or drinking water does not really depend on local taxation, and the possible effects of local governance on that revenue, but on predetermined federal transfers.

Municipalities in Mexico are administrative territories composed of both a central cabecera, the main town where the government seats, and a periphery of smaller villages. In Oaxaca, these outlying villages are referred to as “agencias.” Regardless of governance type, agencias often retain traditional governance forms, including assemblies and other direct participatory methods. It is not uncommon for FAISM funds to be primarily allocated to the cabecera towns, even though the most serious poverty and public service needs are concentrated in agencias (Fox & Aranda, 1996).

The administrative structure of Mexico’s fiscal federalism provides a unique research design opportunity because financial transfers are clearly exogenous: They are identically distributed to places with similar poverty levels. But there is significant institutional variation in the way formal local authority is selected and governance exercised. Parties and electoral competition governance prevails in some municipalities, whereas traditional assembly-based participatory rules of governance are found in other jurisdictions.

In the remainder of the article, we exploit this institutional variation to generate inferences about the effects of governance on the provision of public goods. To understand how the budgetary process works in usos y costumbres municipalities, we quote an interviewee who explaining this most lucidly:

Before we had no control of budgeting, as I explained, but this time we were stricter on how we elected the mayor, because we wanted him to work hard. Before construction companies would come and do the public works and take over most of the money, at most two or three projects were done, and the money would vanish. Since it was impossible to please the whole town that way, now the public works are given priority through the assemblies that say, the people say: How many works for the first section? Two or three, then, how much did the first section get? Now, on to the second section; then the third; and so on. Then priorities are being set. People approved that ranking, and then there were thirteen projects that were approved in that year, and the thirteen of them are being carried out. But as I was explaining, with the budget that is limited.

The quote suggests that assembly meetings play an important role in how projects are decided in usos y costumbres municipalities. There is little research on the dynamics of governance between agencias and the core towns in the cabecera. An exception is Eisenstadt and Ríos (2014) who suggest an important mechanism explaining conflicts in places ruled by usos y costumbres as a consequence of the disempowerment of agencias in the informal structure of decision making. We agree with them that this is a critical component of the way public resources are allocated in municipalities in Mexico but differ in that our fieldwork suggests that agencias in municipalities ruled by political parties are even more disenfranchised, than those ruled by usos y costumbres, through the lack of territorial representation in the rules for election of municipal councils.

A final difference highlighted by our qualitative interviews in party localities is that citizens believe resources inevitably get allocated to party supporters. In some localities, partisan divisions are so entrenched that our informants reported that water could be cut off from areas of towns where opponents lived. A man told us,

here you can’t get any service if you do not support the party of the municipal president. All this section of the town –pointing to a hill with many small houses—now does not get any water because the mayor is from the PRI and this section of the town is PRDista. Can you imagine that they do this?

One woman told us in a village from a party municipality: “Here even water is political. Those in that barrio who went with the PRD got removed from the service and now they have to get water by hiring pipas (private water trucks).”

Traditional Governance and the Provision of Public Goods and Services

We hypothesize that traditional systems of usos y costumbres impact the provision of public goods because in this form of governance, there is greater participation in decision making, more accountability in the use of public funds, and community resources are leveraged through various forms of volunteering. Mayors have incentives to be more responsive to citizens than in municipalities ruled by political parties. However, this form of governance may not impact different types of public goods in the same way. In particular, education and health services in Mexico are provided primarily by the federal (and sometimes state) governments with little input from municipalities. And even electricity provision is primarily determined by the expansion of the public grid by the Federal Electoral Commission.

Hence, we will concentrate on two public goods that are crucial for development and happen to be under the direct jurisdiction of municipal governments. Drinking water and sewerage are crucial local public goods that do not just enhance household well-being but impact public health decisively. In our fieldwork, we witnessed how municipal governments under both usos y costumbres and multiparty competition emphasize these goods as local priorities and a good metric to evaluate their own performance. In almost every municipality we visited there was a water committee or a member of the cabildo (municipal assembly) in charge of supervising and overseeing the delivery of drinking water. And mayors proudly inaugurate sewerage systems (or expansions) as the hallmark of their good performance in office.

These public goods may be delivered with more or less coverage but may also be biased in favor of some citizens over others. In particular, a critical question regarding the impact of usos y costumbres on governance is to ask whether the poor are more or less likely to benefit from public goods under this form of governance. There is an additional question as to whether indigenous ethnic households are more likely to receive benefits vis-a-vis others. And a final concern is whether the core town of the so-called cabecera is more likely to receive benefits than the peripheral villages called, in the case of Oaxaca, agencias.¹⁹

Our ethnographic accounts suggest that traditional governance may have some advantages in the provision of local public goods. However, statistically evaluating the impact of usos y costumbres institutions is challenging because governance is not produced by an accident of nature. The complex historical, economic, social, and political processes that gave rise to this institution also confound efforts to assess a causal impact. And the institutional forms themselves have important differences in the way each particular town or village has adopted them. Notwithstanding those differences, Appendix A shows that there is enough commonality among these forms of governance to make a statistical analysis, complementary to the ethnographic evidence, a valid exercise.

A naive statistical model of the effect of the variation in these forms of governance on public good provision is not a reliable indicator from which one may infer causal effects. A frequency table can be built comparing public good provision according to governance types, using household level data from the sample of the 2010 Mexican census (Table 2). The census sample descriptive statistics would lead us to believe that usos y costumbres is unambiguously worse in the provision of public goods than political parties. The descriptive data are particularly striking for sewerage, where four fifths of the people living under usos y costumbres do not have access to sewerage, almost double the figure corresponding to parties.

Table 2.

Comparison of Usos Y Costumbres and Party Service Delivery in 2010.

	Parties	Usos y Costumbres
No water	31.6	42.4
No sewerage	43.1	79.6
No electricity	8.9	20.5
No basic schooling	40.8	56.1
No health insurance	70.1	85.9
Dirt floors	30.9	60.1
n	2,156,759	1,217,585

This purely descriptive data would suggest that traditional governance does not seem to perform well in public service provision: less drinking water, less electrification, larger gaps in basic schooling and health insurance, and a greater share of dwellings with dirt floors. However, from an inferential point of view, the descriptive table does not provide insight into the role of governance institutions. Sociodemographic conditions are widely different across these two groups: Inhabitants who live under political parties also tend to dwell in cities, for example, with much lower poverty rates and higher incomes.

Survey Data Evidence

Our survey was designed to allows us to explore some of the differences in how public services are distributed to citizens highlighting smaller municipalities and controlling for differences between the cabecera and agencias, under different governance types, and between households with different characteristics. The survey was stratified in its design to ensure a better coverage of places with traditional governance institutions. We are also better able to tease out household level differences that may be correlated with the provision of goods and services (such as income and ethnicity), allowing us to assess the distributional allocations of public services within the municipality.²⁰

We asked respondents to report whether their household had access to water inside the house, from a faucet in the yard or from other sources, including a nearby river, a well, or a private water truck. We code as 1 those who report having water at home or in their yard, and 0 otherwise. We also asked respondents whether they had access to sewage in their homes. We code as 1 those who report that their sewer is connected to the public sewerage system and as 0 those who report that they do not have connection. 48% of our sample has water, and 31% has sewerage.

To begin understanding how these public services are distributed, we perform a simple analysis in Table 3, which shows coefficients of two logit models, one for access to water and the other to the sewerage system. The table report coefficients for three relevant correlates: (a) whether the locality is a cabecera, (b) household extreme poverty, and (c) whether respondent self-ascribes as indigenous.²¹

Table 3.

Household Level Provision of Public Goods.

Water	Party	Usos	Party	Usos
Cabecera	1.60***	−0.25	3.97***	1.42***
Cabecera	(0.27)	(0.24)	(0.49)	(0.36)
Poverty	−0.69***	−0.15	−0.68*	0.43
Poverty	(0.30)	(0.23)	(0.39)	(0.27)
Indigenous	0.38	0.45*	0.69*	0.15
Indigenous	(0.29)	(0.24)	(0.36)	(0.29)
Constant	−1.36***	−0.03	−3.52***	−2.29***
Constant	(0.32)	(0.27)	(0.55)	(0.40)
n	278	335	278	335

***

Statistically significant at the 99% level.

Coefficients in Table 3 suggest that in party municipalities, provision of both water and sewerage is strongly skewed in favor of the cabecera in detriment of the agencias. The table also suggest that in party municipalities, provision of water is highly unequal, not reaching the poor. Moroever, those who self-identify as indigenous tend to be strongly disfavored in that they do not receive sewerage where parties rule. In usos y costumbres municipalities, by contrast, water is not skewed in favor of the cabecera. Importantly, provision of water does not disfavor the poor. Respondents who self-identify as indigenous appear to be better served with water in usos y costumbres municipalities.

This preliminary exercise suggests that there are differing patterns in the way public services are distributed in party and usos y costumbres municipalities. Distribution of public services in the later appear to be less skewed in favor of the ruling cabecera and to be more egalitarian, favoring provision to the poor and the indigenous. Suggestive as these findings are, it is important to recognize that there is a fundamental problem of inference related to the fact that, in spite of the careful sampling strategy in our survey, we cannot be sure that these differences are attributable to the form of governance, due to potential selection bias problems. In an attempt to offer a causal identification strategy, the next section uses census data from villages, providing a more robust empirical estimation of the significant difference between governance institutions on the provision of public goods.

A Geographic Discontinuity Approach

Whether traditional governance impacts public good provision is about showing evidence of a causal story, not simply an empirical regularity arising from correlations. The econometric identification challenge is to find a way to separate the effect of the governance type on public goods provision from modernization variables that are highly correlated with governance. In terms of the potential outcomes framework, the problem is that the performance in public good provision of a place ruled by usos y costumbres cannot be compared with itself under political parties as both states are not observed simultaneously. We need a counterfactual scenario to usos y costumbres that is credible, in which almost identical communities are compared, only varying governance institutions.

The academic literature proposes several strategies to deal with these issues, including instrumental variables and matching (Angrist & Pischke, 2008, 2014). Díaz-Cayeros et al. (2014) used propensity score matching to study public good provision in Oaxaca. Matching on observables allowed them to explore the effect of usos y costumbres at the municipal level, analyzing the coverage of public goods and the degree of political competition. We delve into the impact of governance type on the provision of public goods by using an empirical strategy that is more robust to selection on unobservables and studies conditions at the village and town level, namely, a geographic regression discontinuity design. This is a form of regression discontinuity design (Lee & Lemieuxa, 2010) in which a geographic boundary or threshold is used as a strategy for identification. The assumption in such design is that the boundary provides something akin to a “natural experiment” in which observations close to the municipal boundary behave “as if” randomly assigned.

Figure 1 shows boundaries between municipalities ruled by usos y costumbres and those governed through political parties. The figure shows 5 and 10 km buffers along the boundary, and the 7,721 localities with more than two households that existed in the state of Oaxaca at the time of collection of the 2010 census. Oaxaca municipalities are relatively small, so the threshold for comparison on the geographic discontinuity should be small. There is a large concentration of towns in the Central valley around Oaxaca city, which is the densely settled region at the center of the state. There are many areas in the mountains which are sparsely populated, with only a few villages close to each other along the discontinuity.

Figure 1.

Geographic discontinuity between usos and partidos municipalities.

The figure suggests that most of the peripheral areas of the state, by the coast and neighboring states, are ruled by political parties. The core of the Mixtec and Zapotec speaking regions are predominantly ruled by usos y costumbres. The geographic regression discontinuity design compares villages and towns (localities) within a buffer zone around the border that divides usos y costumbres from party governance. Neighboring villages would be treated as randomly assigned: From a statistical standpoint, their basic socioeconomic makeup should be undistinguishable, but some will fall in the party governance side while others would be governed by usos y costumbres.

Keele and Titiunik (2015) provide a careful critique of some of the reasons why geographic regression discontinuity designs should be used with some care. In particular, they note that potential threats to identification may come from sorting behavior along political borders as households might actually make efforts precisely to move onto either side of the discontinuity, to take advantage of the “treatment.” In the context of the indigenous towns in Oaxaca we study, however, this sorting behavior is unlikely. For our identification strategy to be credible, the choice of usos y costumbres among localities close to the boundary between the regions characterized by each governance regime should be exogenous. This is not just a matter of finding a statistically balanced set of variables characterizing the “treated” and “control” groups of villages but also about knowing something regarding the way in which settlement patterns in Oaxaca have historically come about.

Rules of communal land-tenure prevailing in most of Southern Mexico impose restrictions to movement across municipal jurisdictions. There are stringent requirements to buy or sell land within communities. These requirements arise both from the legal system of communal land-tenure developed historically, as well as from informal community rules that tie land ownership rights to broader patterns of civic engagement. Our ethnographic work suggests that communities are highly stable in their membership and the inhabitants sense of belonging. It is not easy for someone to simply move to another municipality to own land.²² Furthermore, linguistic diversity creates barriers to entry into communities other than one’s own language group.

In villages that are not part of the cabecera, sorting effects are unlikely given that the key decision regarding the form of governance was taken in the core town of the municipality, while communities in the periphery took that choice as given. Interestingly, our qualitative fieldwork revealed that there are actually many villages that are internally ruled by usos y costumbres but are “stuck” in party municipalities. For these villages, defecting to the nearest usos y costumbres municipality is generally not possible because municipal boundaries are rarely, if ever, renegotiated. Such decisions depend on the state congress, so the definition of the boundaries is exogenous to the choices of inhabitants in localities around them.

From a technical perspective, the credibility of this strategy of causal inference still hinges on demonstrating balance across the observations that fall under each governance type. Balance tests are conventionally done to have a sense of the degree to which the groups of localities assigned to each governance type are comparable. The standard way of doing this is to calculate difference in means tests between treatment and control groups over observable variables, excluding outcome variables that may be affected by the “treatment.”

We use Mexican 2010 census data at the locality level. Variables include demographic data, household characteristics, and economic structure. An advantage of the data we use is the level detail. For example, demographic data include a measure of the total population, disaggregated by gender and five or more subpopulation categories (children aged under 2 years, over 3 years, over 5 years, over 15 years, over 18 years, and so on). The data also disaggregate all previous combinations by gender. The same is done for example for data about ethnicity, whether people speak an indigenous language, whether they also speak Spanish or both, disaggregated by age groups and gender. The data we have are extremely rich, which allow us to perform balance tests on each one of these population subgroups rather than just on the general groups.

An important decision in this discontinuity design is to determine what constitutes a small enough neighborhood close to the border between treated and control observations. If the neighborhood is too small, the statistical test might be underpowered to detect effects, but too large a neighborhood is likely to fail to achieve balance. A quarter of the localities in the state of Oaxaca are within a 5-km band of this border, which means that the threshold will be small. We set the smallest neighborhood at 250 m from the boundary, testing balance every additional 250 meters, until reaching around 6,972 meters (a maximum band of around 14 km). Appendix B shows that these neighborhoods are well within an “optimal” bandwidth that can be estimated, taking into account a trade-off between the precision and the potential bias of the estimates.

For the sake of brevity, Tables 4 and 5 show balance tests for 500 and 2,500 m on either side of the boundary. Balance is reasonably good in small neighborhoods around the boundary. In a band of 1,000 m around the discontinuity, only one variable (male head of household) shows a statically significant difference between usos y costumbres and parties. In the wider band that includes localities within a 5-km band, in contrast, several variables are imbalanced, including, the average household size, fertility rates, the male to female ratio, and the share of population or households that can be considered Indian. This is to be expected; as the band becomes wider, the localities included in the comparison increasingly look more similar to the overall patterns of the poorly identified universe of all Oaxacan localities. In what regards political variables, it is not possible to carry out a similar balance test at the village level, simply because the level of aggregation of electoral data is higher. However, the Supplemental Appendix provides evidence, drawing from electoral precincts in 1991 and 1994, suggesting that our identification strategy is not threatened by political biases.²³

Table 4.

Balance Test at Geographic Discontinuity Between Usos and Partidos Municipalities (500 m From Border).

Variable	Parties	Usos y costumbres	Difference	p value
Population	476.2	496.8	20.5	.88
Economically active	181.8	193.4	11.6	.85
Men	235.2	246.2	11.0	.87
Women	256.5	274.4	18.0	.81
HH	119.8	128.3	8.5	.81
Average HH Size	4.33	4.26	−0.07	.22
Fertility	3.11	3.14	0.03	.56
Male/female ratio	0.964	0.961	−0.004	.81
Male head of HH	0.786	0.765	−0.020	.01*
Age <12 years	0.406	0.408	0.002	.68
Age 12-18 years	0.143	0.141	−0.003	.49
Age >65 years	0.124	0.124	0.0002	.97
Married	0.543	0.544	0.001	.84
Catholic	0.853	0.847	−0.006	.68
Indian	0.359	0.359	−0.0007	.97
Indian HH	0.431	0.456	0.024	.39
Employed	0.967	0.959	−0.007	.27
Illiteracy	0.230	0.229	−0.001	.88
Incomplete basic	0.449	0.446	−0.003	.79
Localities (n)	456	432	888

Balance test for discontinuity of 1 km. HH = households.

Significantly different from 0 at a 99% level two-tailed test.

Table 5.

Balance Test at Geographic Discontinuity Between Usos and Partidos Municipalities (2500 m From Border).

Variable	Parties	Usos y Costumbres	Difference	p value
Population	798.7	398.9	399.8	.12
Economically active	326.0	148.9	177.2	.15
Men	391.7	198.9	192.8	.12
Women	436.1	220.7	215.4	.14
HH	203.9	102.7	101.2	.14
Average HH size	4.33	4.25	−0.07	.07*
Fertility	3.09	3.21	0.11	.001*
Male/female ratio	0.959	0.956	−0.004	.71
Male head of HH	0.784	0.773	−0.011	.03*
Age <12 years	0.408	0.411	0.002	.54
Age 12-18 years	0.145	0.141	−0.003	.11
Age >65 years	0.125	0.131	0.006	.13
Married	0.545	0.551	0.006	.13
Catholic	0.848	0.849	0.001	.87
Indian language	0.362	0.393	0.031	.08*
Indian HH	0.432	0.484	0.051	.01*
Employed	0.964	0.960	−0.005	.28
Illiteracy	0.234	0.233	−0.088	.90
Incomplete schooling	0.447	0.450	−0.003	.69
Localities (n)	1,066	1,057	2,123

Balance test for discontinuity of 5 km.

Significantly different from 0 at the 99% level two-tailed test.

To obtain some insight into the empirical strategy, Figures 2 and 3 provide a visualization of the difference in the provision of drinking water and sewerage under the two structures of governance. The peculiar feature of the graphs is that localities are not ordered by a conventional metric of level of development or other attribute related to their sociodemographic makeup, but rather the distance in meters to the geographic boundary that separates usos y costumbres from political party governance. The parties are arrayed to the left according to the distance to the boundary, starting from a relatively small buffer or 3 km, whereas usos y costumbres localities are ordered to the right. The figures also include a lowess regression line that fits the provision of public services according to the distance metric.

Figure 2.

Water provision along geographic discontinuity.

Figure 3.

Sewerage provision along geographic discontinuity.

The sharp difference in provision of water and sewerage across the boundary is the geographic discontinuity we seek to estimate. To estimate a causal effect of governance types over the discontinuity, taking into account the potential threats to inference, we present our results every 250 m, until reaching 1,500 m. We then present results every 500 m until reaching 3 km (on each side). And a final maximum boundary is set according to the mean square error (MSE) optimal bandwidth (described in Appendix B). We know from Table 5 that at the threshold of 2,500 m, we start running into less balanced samples, so this strategy for the presentation of findings provides a more transparent description of the size of the governance effects as we move toward less robustly identified comparisons of means.

The dependent variable is hence the percentage of households in a given locality that do not have access either to drinking water or sewerage. In keeping with the recommendations of Keele and Titiunik (2015), we move beyond a test of a regression coefficient or a difference of means, estimating a local average treatment effect (LATE). We do this through a matching procedure, controlling for geographic location. The literature on geographic regression discontinuity warns that treatment effects may be different along the length of the border. Specifically, we calculate a LATE matching every observation with the three closest localities to the boundary, controlling for geographic space, with a polynomial of order 2 for longitude and latitude.²⁴ The estimates were calculated using the rdrobust routine in STATA (Calonico, Cattaneo, Farrell, & Titiunik, 2017).

The results are presented in Table 6. The first column in the table presents the distance to the discontinuity. The second column provides the estimate of the size of the difference in means, where usos y costumbres is the treatment. The third and fourth columns present the standard error and the t test statistic, whereas the last column indicates the number of localities included in each comparison. The first block of estimates corresponds to lack of water, and the second one to lack of sewerage. After the LATE estimates, the table also presents hypothesis tests with no matching, for each public good. Those second set of estimates can be interpreted as the summary statistics of testing hypotheses over various bands on Figures 2 and 3.

Table 6.

Effect of Traditional Governance of Lack of Public Services (Water and Sewerage).

	Distance (<)	Coefficient	SE	t test	n
No water
Exact matching (3 closest)	250	−34.23	11.16	−3.07	181
	500	−23.91	8.08	−2.96	355
	750	−20.62	6.75	−3.05	554
	1,000	−20.88	6.01	−3.48	722
	1,250	−19.42	5.43	−3.58	907
	1,500	−16.80	4.97	−3.38	1,087
	2,000	−12.62	4.32	−2.92	1,427
	3,000	−8.00	3.55	−2.26	2,112
	6,097	−5.69	2.57	−2.22	3,828
No matching	250	−14.05	5.96	−2.36	181
	500	−12.37	4.20	−2.94	355
	750	−9.33	3.41	−2.74	554
	1,000	−5.39	2.99	−1.80	722
	1,250	−5.57	2.66	−2.10	907
	1,500	−4.71	2.43	−1.94	1,087
	2,000	−5.06	2.12	−2.38	1,427
	3,000	−6.35	1.74	−3.65	2,112
	6,097	−7.12	1.29	−5.51	3,828
	All	−7.85	0.96	−8.22	7,154
No sewerage
Exact matching (3 closest)	250	−37.97	10.99	−3.46	181
	500	−10.61	8.15	−1.30	355
	750	−8.50	6.61	−1.29	554
	1,000	−9.46	5.82	−1.62	722
	1,250	−9.44	5.24	−1.80	907
	1,500	−8.77	4.78	−1.83	1,087
	2,000	−7.27	4.16	−1.75	1,427
	3,000	−2.94	3.38	−0.87	2,112
	6,972	−0.46	2.26	−0.20	3,828
No matching	250	−2.01	5.31	−0.38	181
	500	−6.61	3.76	−1.76	355
	750	−3.28	3.03	−1.08	554
	1,000	−0.46	2.67	−0.17	722
	1,250	2.01	2.38	0.84	907
	1,500	2.29	2.16	1.06	1,087
	2,000	4.54	1.88	2.42	1,427
	3,000	4.50	1.54	2.92	2,112
	6,972	6.72	1.23	5.45	4,223
	All	20.77	0.84	24.63	7,154

The hypothesis tests for the difference in the provision of public goods along the geographic discontinuity are striking. Water provision is much worse in localities that fall under the jurisdiction of political party governance than in usos y costumbres. The effect is rather large, of more than 10 percentage points in thresholds below 500 m and of around 6 points in wider bands. The results on the geographic discontinuity design support the conclusion that usos y costumbres municipalities are better than party municipalities at delivering drinking water, one of the scarcest and most valuable goods with important public health impacts. We find no significant effects on the coverage of sewerage.

That is, households living under usos y costumbres do not seem to be at a disadvantage in the provision of sewerage. This is contrary to the huge gap that was observed in the descriptive statistics. Only when the discontinuity is extended to 3,000 m or more (which as we have mentioned is less well-identified) does a difference in the provisions of sewerage emerge, but as previously discussed, this is not generated by the governance type but rather by the fact that the communities compared in that universe are no longer similar in terms of modernization and levels of development.²⁵

Conclusion

The results in this article suggest that political parties in poor indigenous municipalities are a source of elite capture. Given the long history of electoral authoritarianism in Mexico, in party municipalities, poor citizens appear to interact with local party elites according to entrenched and vertical patron–client relationships.

By contrast, the conclusion that emerges from our study is that indigenous communities, when ruled by traditional leaders, norms, and practices, fare better in the provision of essential local public goods. The article argued that the profound institutional differences among governance forms affect the provision of public goods through three mechanisms: greater local government accountability, broader civic engagement in collective decision making, and an effective system of community service that allows for widespread cooperation with activities that promote the well-being of the community. Citizens in these communities are better informed and more participatory, and hence appear to be more capable of restraining elites from capturing resources.

Our findings are relevant for the debate about indigenous autonomies, an important theme in Mexico and in many Latin American countries such as Bolivia, Guatemala, Peru, Ecuador, and Colombia. After the Zapatista rebellion in 1994 in the southern state, Chiapas, there was a renewed debate about indigenous autonomies. As part of the peace accords in San Andrés Larráinzar, the federal government committed to grant autonomy to indigenous municipalities like Oaxaca, but later refused to enact the laws arguing that democratization made the demands unnecessary and that Indian practices go against the Constitution. To the contrary, our study highlights some of the limits of democratization in the context of indigenous Mexico, where electoral competition is more prone to capture government and distribute public goods along partisan lines.

The almost universal prevalence of political parties as vehicles for citizen representation imposes a tension between democratization and the provision of public services that are essential for development. If political parties are more prone to capture the provision of these goods and subject their distribution to political allegiance, both the quality of democracy and the proper foundations for long-term development are reduced. It seems plausible that these inherent tensions of partisan distribution of benefits are at the center of a broader crisis of representation of political parties (Mainwaring, Bejarano, & Leongómez, 2006). Support for democracy requires a minimal flow of benefits to citizens.

Our study suggests that decentralized community-led development is more likely to work where strong civic communities are already present (Putnam, 1995). Our results echo old wisdom about the importance of citizen engagement, participation, information, and overall awareness and empowerment in public affairs. Democracy captured is democracy denied, which appears to be a more serious problem in localities governed by political parties. Traditional governance practices, for some indigenous communities with strong organizational resources and a history of social collaboration, offer a way out of this serious accountability deficit that can be deployed to enhance development and better serve the poorest citizens.

Supplemental Material

ElectoralDiscontinuityAppendix – Supplemental material for Public Good Provision and Traditional Governance in Indigenous Communities in Oaxaca, Mexico

Supplemental material, ElectoralDiscontinuityAppendix for Public Good Provision and Traditional Governance in Indigenous Communities in Oaxaca, Mexico by Beatriz Magaloni, Alberto Díaz-Cayeros and Alexander Ruiz Euler in Comparative Political Studies

Supplemental Material

Replication_material_CPS_217.R2 – Supplemental material for Public Good Provision and Traditional Governance in Indigenous Communities in Oaxaca, Mexico

Supplemental material, Replication_material_CPS_217.R2 for Public Good Provision and Traditional Governance in Indigenous Communities in Oaxaca, Mexico by Beatriz Magaloni, Alberto Díaz-Cayeros and Alexander Ruiz Euler in Comparative Political Studies

Footnotes

Appendix A

Appendix B

Acknowledgements

We thank Kate Baldwin, Daniela Barba Sanchez, Allison Benton, Carles Boix, Tulia Falleti, Francis Fukuyama, Kristof Gosztonyi, Katharina Holzinger, Florian Kern, Simeon Nichter, Joan Ricart-Huguet, and Scott Rozelle for useful comments, suggestions, and discussions. Three anonymous referees provided insightful critiques and important suggestions.

Authors’ Note

Previous versions were presented at the Governance and Public Goods Provision Conference at Stanford University, May 18-19, 2012; the 2013 American Political Science Association Meeting at Chicago; the Traditional Governance and Indigenous Peoples Conference at University of Konstanz, June 15-17, 2016; and the Comparative Politics Seminar at Princeton University, April 12, 2017. All errors remain our own.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by Stanford University, the University of California Office of the President (UCOP), the Center for US-Mexican Studies at the University of California, San Diego and the William and Flora Hewlett Foundation.

ORCID iD

Alberto Díaz-Cayeros

Supplemental Material

Supplemental material for this article is available online at the CPS website .

Notes

Author Biographies

Beatriz Magaloni is professor of political science and Director of the Poverty, Violence and Governance Laboratory at the Center on Democracy, Development and Rule of Law at Stanford University.

Alberto Diaz-Cayeros is Senior Fellow at the Center on Democracy, Development and Rule of Law and the Director of the Center for Latin American Studies at Stanford University.

Alexander Ruiz Euler is the lead data scientist for the innovation team at Cuebiq. He received his PhD in political science from the University of California, San Diego.

References

Acemoglu

Reed

Robinson

J. A.

(2014). Chiefs: Economic development and elite control of civil society in Sierra Leone. Journal of Political Economy, 122, 319–368.

Angrist

J. D.

Pischke

J.-S.

(2008). Mostly harmless econometrics: An empiricist’s companion. Princeton, NJ: Princeton University Press.

Angrist

J. D.

Pischke

J.-S.

(2014). Mastering ‘metrics: The path from cause to effect. Princeton, NJ: Princeton University Press.

Baldwin

(2015). The paradox of traditional chiefs in democratic Africa. Cambridge, UK: Cambridge University Press.

Bardhan

(2002). Decentralization of governance and development. The Journal of Economic Perspectives, 16, 185–205.

Bardhan

Mookherjee

(2005). Decentralizing antipoverty program delivery in developing countries. Journal of Public Economics, 89, 675–704.

Benton

A. L.

(2012). Bottom-up challenges to national democracy: Mexico’s (legal) subnational authoritarian enclaves. Comparative Politics, 44, 253–271.

Benton

A. L.

(2016). How “participatory governance” strengthens authoritarian regimes: Evidence from electoral authoritarian Oaxaca, Mexico. Journal of Politics in Latin America, 8(2), 37–70.

Benton

A. L.

(2017). Configuring authority over electoral manipulation in electoral authoritarian regimes: Evidence from Mexico. Democratization, 24, 521–543.

10.

Box-Steffensmeier

J. M.

Brady

H. E.

Collier

(2008). The Oxford handbook of political methodology (Vol. 10). Oxford, UK: Oxford University Press.

11.

Calonico

Cattaneo

Farrell

M. H.

Titiunik

(2017). Rdrobust: Software for regression discontinuity designs. Stata Journal, 17, 372–404.

12.

Calonico

Cattaneo

M. D.

Titiunik

(2014). Robust nonparametric confidence intervals for regression-discontinuity designs. Econometrica, 82, 2295–2326.

13.

Chandra

(2007). Why ethnic parties succeed: Patronage and ethnic head counts in India. Cambridge, UK: Cambridge University Press.

14.

Chong

De la O.

A. L.

Karlan

Wantchekon

(2014). Does corruption information inspire the fight or quash the hope? A field experiment in Mexico on voter turnout, choice, and party identification. The Journal of Politics, 77, 55–71.

15.

Díaz Montes

. (2002). Comportamiento electoral en los municipios indígenas de Oaxaca [Electoral behavior in indigenous municipalities of Oaxaca], 1992-1998. In Dilemas de la democracia en México. Los actores sociales ante la representación política. Mexico: Centro Francés de Estudios Mexicanos y Centroamericanos and Instituto Federal Electoral.

16.

Danielson

M. S.

Eisenstadt

T. A.

(2009). Walking together, but in which direction? Gender discrimination and multicultural practices in Oaxaca, Mexico. Politics & Gender, 5, 153–184.

17.

Diaz-Cayeros

Estevez

Magaloni

(2016). The political logic of poverty relief: Electoral strategies and social policy in Mexico. Cambridge, UK: Cambridge University Press.

18.

Díaz-Cayeros

Magaloni

Ruiz-Euler

(2014). Traditional governance, citizen engagement, and local public goods: Evidence from Mexico. World Development, 53, 80–93.

19.

Díaz-Cayeros

Magaloni

Ruiz-Euler

(2015). Information, female empowerment and governance in Oaxaca, Mexico SCID (Working Paper No. 544). http://kingcenter.stanford.edu/sites/default/files/publications/544wp_0.pdf

20.

Díaz-Cayeros

Silva Castañeda

(2004). Descentralización a Escala Municipal en México: La Inversión en Infraestructura Social [Municipal scale decentralization in Mexico: Social infrastructure investment]. In Serie Estudios y Perspectivas (Vol. 15). Comisión Económica para América Latina y el Caribe (CEPAL) Chile: Santiago.

21.

Dunning

(2012). Natural experiments in the social sciences: A design-based approach. Cambridge, UK: Cambridge University Press.

22.

Eisenstadt

T. A.

(2007). Usos y costumbres and Postelectoral Conflicts in Oaxaca, Mexico, 1995-2004: An empirical and normative assessment. Latin American Research Review, 42, 52–77.

23.

Eisenstadt

T. A.

(2011). Politics, identity, and Mexico’s indigenous rights movements. Cambridge, UK: Cambridge University Press.

24.

Eisenstadt

T. A.

Ríos

(2014). Multicultural institutions, distributional politics, and postelectoral mobilization in indigenous Mexico. Latin American Politics and Society, 56, 70–92.

25.

Fox

Aranda

(1996). Decentralization and rural development in Mexico. San Diego: Center for US-Mexican Studies, at the University of California, San Diego.

26.

Holzinger

Kern

F. G.

Kromrey

(2016). The dualism of contemporary traditional governance and the state institutional setups and political consequences. Political Research Quarterly, 69, 469–481.

27.

Humphreys

Jacobs

A. M.

(2015). Mixing methods: A Bayesian approach. American Political Science Review, 109, 653–673.

28.

Imbens

Kalyanaraman

(2012). Optimal bandwidth choice for the regression discontinuity estimator. The Review of Economic Studies, 79, 933–959.

29.

Keele

L. J.

Titiunik

(2015). Geographic Boundaries as Regression Discontinuities. Political Analysis, 23, 127–155.

30.

Lee

D. S.

Lemieuxa

(2010). Regression discontinuity designs in economics. Journal of Economic Literature, 48, 281–355.

31.

López Bárcenas

. (2005). Elecciones por usos y costumbres en Oaxaca [Elections through usos y costumbres in Oaxaca]. In Estudios en homenaje a don Jorge Fernández Ruiz. Derecho constitucional y política (Vol. 4). Universidad Nacional Autónoma de México. Mexico: Universidad Nacional Autónoma de México.

32.

Magaloni

(2006). Voting for autocracy: Hegemonic party survival and its demise in Mexico. Cambridge, UK: Cambridge University Press.

33.

Mainwaring

Bejarano

A. M.

Leongómez

E. P.

(2006). The crisis of democratic representation in the Andes. Stanford, CA: Stanford University Press.

34.

Olken

B. A.

(2008). Direct democracy and local public goods: Evidence from a field experiment in Indonesia (Technical report National Bureau of Economic Research). Retrieved from https://www.nber.org/papers/w14123

35.

Posner

D. N.

(2005). Institutions and ethnic politics in Africa. Cambridge, UK: Cambridge University Press.

36.

Putnam

R. D.

(1995). Bowling alone: America’s declining social capital. Journal of Democracy, 6, 65-65.

37.

Recondo

(2007). La política del gatopardo: Multiculturalismo y Democracia en. Oaxaca [The politics of the leopard: Multiculturalism and democracy in Oaxaca]. Mexico: CIESAS.

38.

Ríos

(2006). Conflictividad postelectoral en los Usos y Costumbres de Oaxaca (Master’s thesis) [Postelectoral conflict in usos y costumbres in Oaxaca]. Instituto Tecnológico Autónomo de México (ITAM).

39.

Scott

J. C.

(2014). The art of not being governed: An anarchist history of upland Southeast Asia. New Haven, CT: Yale University Press.

40.

Seawright

(2016). Multi-method social science: Combining qualitative and quantitative tools. Cambridge, UK: Cambridge University Press.

41.

Spybrook

Raudenbush

S. W.

Liu

Congdon

Martínez

(2008). Optimal design for longitudinal and multilevel research: Documentation for the “optimal design.” Ann Arbor: University of Michigan.

42.

Stephen

(2005). Negotiating global, national and local “rights” in a Zapotec community. Political and Legal Anthropology Review, 28, 133–150.

43.

Stephen

(2013). We are the face of Oaxaca: Testimony and social movements. Durham, NC: Duke University Press.

44.

Stiglitz

J. E.

(2002). Participation and development: Perspectives from the comprehensive development paradigm. Review of Development Economics, 6, 163–182.

45.

Trejo

(2009). Religious competition and ethnic mobilization in Latin America: Why the Catholic church promotes indigenous movements in Mexico. American Political Science Review, 103, 323–342.

46.

Trejo

(2012). Popular movements in autocracies: Religion, repression, and indigenous collective action in Mexico. Cambridge, UK: Cambridge University Press.

47.

World Bank. (2003). World development report 2004: Making services (Work for poor people). Retrieved from http://documents.worldbank.org/curated/en/832891468338681960/World-Development-Report-2004-making-services-work-for-poor-people

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.89 MB

0.00 MB

0.69 MB