POLITICAL COMPETITION AND STATE CAPACITY

EVIDENCE FROM A LAND ALLOCATION PROGRAM IN MEXICO

Leopoldo Fergusson

Horacio Larreguy

Juan Felipe Riaño

LATIN AMERICAN AND THE CARIBBEAN ECONOMIC ASSOCIATION

August 2018

The views expressed herein are those of the authors and do not necessarily reflect the views of the

Latin American and the Caribbean Economic Association. Research published in this series may

include views on policy, but LACEA takes no institutional policy positions.

LACEA working papers are circulated for discussion and comment purposes. Citation of such a paper

should account for its provisional character. A revised version may be available directly from the

author.

© 2018 by Leopoldo Fergusson, Horacio Larreguy and Juan Felipe Riaño. All rights reserved. Short

sections of text, not to exceed two paragraphs, may be quoted without explicit permission provided

that full credit, including © notice, is given to the source.

LACEA WORKING PAPER SERES. No. 0011

LACEA WORKING PAPER SERIES No. 0011 August 2018

Political Competition and State Capacity: Evidence from a Land Allocation Program in Mexico

Leopoldo Fergusson

Universidad de los Andes, Department of Economics

lfergusson@uniandes.edu.co

Horacio Larreguy

Harvard University, Department of Government

hlarreguy@fas.harvard.edu

Juan Felipe Riaño

University of British Columbia, Vancouver School of Economics

jf.riano@almuni.ubc.ca

ABSTRACT

We develop a model of the politics of state strengthening undertaken by incumbent parties that have

a comparative advantage in clientelism rather than in public good provision. The model suggests that,

when politically challenged by opponents, clientelistic incumbents may oppose investing in state

capacity. We provide empirical support for the model’s implications using policy decisions that

reflect local state capacity choices, and a difference-in-differences identification strategy that exploits

a national shock that threatened the Mexican Institutional Revolutionary Party’s hegemony in the

early 1960s with varying intensity across the various Mexican municipalities.

JEL Classification: D72, D73

Keywords: state capacity, political competition, land allocation

ACKNOWLEDGEMENTS AND FINANCIAL DISCLOSURE

We thank Jennifer Alix-Garcia, Kate Baldwin, Christopher Blair, Antonio Ciccone, Emilio Depetris,

Marcela Eslava, Camilo García-Jimeno, Ed Glaeser, Melissa Lee, Gianmarco León, Stelios

Michalopoulos, Daniel Ortega, Pablo Sanguinetti, Laura Schechter, Emily Sellars, Milan Svolik,

Andrea Tesei, Francesco Trebbi, Joachim Voth, Ebonya Washington, and seminar participants at the

Workshop on the Mexican Agrarian Reform at the University of Wisconsin, the MIT’s Political

Economy Breakfast, the Political Economy Group of LACEA, the Political Institutions Workshop at

the Barcelona Graduate School of Economics Summer Forum, the Leitner Political Economy Seminar

at Yale University, the Institut d’Economia de Barcelona at the Universitat de Barcelona, the V

Conference on the Political Economy of Development and Conflict at the Universitat Pompeu Fabra,

and the State Capacity in Comparative Perspective Conference at Harvard University for their helpful

comments. We are very grateful with Melissa Dell, who shared the Padrón e Historial de Núcleos

Agrarios (PHINA) she scraped from the Mexican Agrarian National Registry (RAN). María José

Villaseñor and her team working at Mexican National Archives, and Mateo Arbeláez Parra provided

superb research assistance. Financial support from CAF-Development Bank of Latin America is

gratefully acknowledged. A previous version of this research paper circulated as CAF Working Paper,

2015-03.

1 Introduction

A recent literature argues that state capacity is central to economic and financial devel-opment, and to political stability and democracy (Acemoglu, 2005; Acemoglu, Moscona,& Robinson, 2016; Besley & Persson, 2000, 2010; Dell, Lane, & Querubin, in press; Fearon& Laitin, 2003; Michalopoulos & Papaioannou, 2013). States with strong bureaucratic,fiscal and military capacities can provide significant shares of their societies with publicgoods, legal environments conducive to businesses and order. Yet many fragile stateslack these capacities (Acemoglu, 2005; Herbst, 2000; Michalopoulos & Papaioannou,2014). While there have been recent academic efforts to understand the sources of statestrength (Acemoglu, Robinson, & Santos, 2013; Besley & Persson, 2000, 2010; Sanchezde la Sierra, 2017), we still lack a definitive understanding of its determinants.

We fill this gap in two ways. First, we theoretically study how political incentivesaffect incumbent parties’ choices of bureaucratic state capacity. In particular, we ex-amine incumbent dominant parties that have a comparative advantage in targetingparticularistic transfers to its clients (as opposed to public good and service provision)compared to opposition parties.1 This set up is of great relevance since clientelism isubiquitous in developing countries (Baland & Robinson, 2008, 2012; Finan & Schechter,2012; Robinson, Torvik, & Verdier, 2006) and it is often the case that incumbent partiesare better positioned to engage in and enforce clientelistic exchanges (Bobonis, Gertler,Gonzalez-Navarro, & Nichter, 2017; Bowles, Larreguy, & Liu, 2017; Larreguy, 2013).2

Our model highlights that investments in local bureaucratic state capacity that reducethe cost of providing public goods and services undermine the comparative advantageof incumbent clientelistic parties. Therefore these parties have incentives to preventthese investments in areas where their dominant political position might be threatened.Since a capable bureaucracy is essential for an effective provision of public goodsand services, our model highlights a mechanism whereby incumbent politicians mayoppose state capacity strengthening.

Second, we provide empirical support to the theoretical implications of the modelby exploiting a unique policy program in México and a national shock threateningthe Institutional Revolutionary Party (PRI) hegemony in the early 1960s with varyingintensity in different municipalities. With a difference-in-differences identificationstrategy, we evaluate the differential impact of this shock on a policy choice by thenational and state PRI governments with direct implications on the future ability ofmunicipal bureaucracies to deliver public services. This policy decision originates from

1Throughout, we broadly refer to “public goods and services” as those that are not easily targeted to specific individuals orgroups in the population. These contrast with what the literature on clientelism denotes as the “particularistic” transfers, whichare goods and services that are targeted in exchange for political support (Kitschelt, 2000; Stokes, 2007; Hicken, 2011).

2Incumbent parties have greater access to resources and attract high-performing political intermediaries since they representbetter prospects. For example, see Stokes (2005) for the case of the Peronist party in Argentina, Bowles et al. (2017) for the case ofthe Unity Party in Liberia, Magaloni (2006) for the case of the Institutional Revolutionary Party in Mexico, and Larreguy, Marx,Reid, and Blattman (2016) for the National Resistance Movement in Uganda.

1

a land allocation program that redistributed more than 50% of Mexico’s agriculturalland between 1910 and 1992 (Dell, 2012; Sanderson, 1984; Torres-Mazuera, 2009).Property rights over the land were allocated in the form of ejidos to communities asa whole, which were often relocated to the locality where the land was granted. Thegovernment, therefore, shaped the location of communities since individuals weretied to the land because ejido land was not individually owned, and use rights wouldbe forfeited if the peasants moved away (de Janvry, Emerick, Gonzalez-Navarro, &Sadoulet, 2015; Fergusson, 2013). More importantly, by influencing how far awaycommunities of land petitioners were located from municipal heads, the governmentwas able to dictate the degree of future municipal bureaucratic state capacity to deliverpublic services to ejidos.

Proximity to the municipal head was a central determinant of the cost for the localstate to reach these areas to provide local public services, and consequently, of mu-nicipal bureaucratic state capacity (Herbst, 2000; Scott, 2009; Soifer, 2015).3 Severalgovernment documents highlight the distance of localities from their municipal head asone of the main barriers for local public service delivery by municipal governments anddevelopment (Baja California, 2003; Mexico, 2007; SEDESOL, 2014), while contempora-neous measures of those public services correlate negatively with the distance from themunicipality head.4 In terms of our theoretical model, the distance of allocated ejidosincreased the cost of providing public goods and services, a potentially advantageousstrategy for an incumbent clientelistic politician protecting his comparative advantage.Thus, in line with previous work emphasizing the importance of proximity to the statefor public service delivery (Henn, 2018; Herbst, 2000; Scott, 2009), we interpret thedistance from the municipal head of each ejido as a measure of local bureaucratic statecapacity choice by the national and state PRI governments, and demonstrate how thischoice across municipalities over time responded to political incentives.5

Our theoretical model features political competition between an incumbent clien-telistic party and an opposition party. The incumbent party can offer public goods andservices or particularistic goods. To capture the incumbent’s comparative advantage inclientelism, for simplicity, we assume that the opposition has no clientelistic machineand only competes by offering public goods and services. We also assume that theincumbent party can weaken bureaucratic state capacity, thus increasing the cost ofdelivering public goods and services.6

3Importantly, with the exception of the extremely few cases where new municipalities were created, the location choice ofall municipal heads predates our sample period.

4Estimates in Appendix Table A-1 show that distance to municipality heads is negatively associated with contemporarypublic service provision. We discuss later how sizable those estimates are and how they are likely to represent an underestimateof the short- and medium-term impact of the PRI’s local state-capacity decisions we study.

5We discuss later how this measure of local bureaucratic state capacity choice dominates other more reversible alternativesincluding public goods and service provision, as well as physical structures and number of bureaucrats, which are usually infre-quently and endogenously decided by local governments.

6Since we normalize the cost of particularistic transfers to one, this assumption also entails that the cost of providing public

2

The first result of the model is that the clientelistic party may benefit from increasingthe relative cost of providing public goods and services by weakening local bureaucraticstate capacity. Increasing the cost of providing public goods and services has two effectson the clientelistic party’s welfare. On the one hand, a “real-budget” effect decreasesthe resources that the clientelistic party may use to distribute particularistic transfers,which is the incumbent’s comparative advantage. On the other hand, a “relative-price”effect decreases the electoral return of the funds allocated to public goods and services,which is the opposition’s comparative advantage. When the latter effect dominates, theincumbent party opposes strengthening the local bureaucratic state capacity, since iterodes its electoral comparative advantage.

More importantly, the second result of our model, and a key testable implication,indicates that the threat of political competition can increase the clientelistic party’sinterest in weakening its ability to effectively provide public goods and services. Tomimic our empirical context, we model an increase in political competition as an erosionof the incumbent’s political machine effectiveness among the voters belonging to theclient base. An increase in political competition has two effects—a positive directand a negative indirect—on the incumbent clientelistic party’s incentive to weakenlocal bureaucratic state capacity. The direct effect of client-base erosion is that theclientelistic party is less concerned with the “real-budget” effect of making publicgoods and services delivery more costly, because particularistic transfers to clients areless effective. The indirect effect follows because equilibrium provision of public goodsand services is larger with a less effective clientelistic machine, increasing the concernsabout higher costs. Since, under reasonable assumptions over voters’ preferences,the direct effect dominates the indirect one, the incumbent clientelistic party desirespublic goods and services to be more expensive when threatened by stronger politicalcompetition. As a result, the clientelistic incumbent politician has incentives to weakenthe local bureaucratic state capacity when politically challenged.

Our empirical strategy tests whether the PRI allocated ejidos far from municipalityheads especially in areas where it expected stronger political competition, thus makingthe delivery of local public goods and services more expensive and effectively weaken-ing the bureaucratic ability to provide such goods. To that end, we exploit the fact thataround 1960, the PRI’s hegemony was threatened by increased discontent from varioussectors of the population, which reflected a significant weakening of its clientelisticmachine and was channeled through opposition parties (Bartra, 1985). The intensity ofthe agitation and the level of the consequent increase in political competition variedby municipality. We use geo-coded administrative data on the universe of allocatedlands and both election returns and an original compilation of events of social and

goods and services not only decreases with stronger local bureaucratic state capacity but does it relatively more than the cost ofclientelism, which we argue is an assumption consistent with the empirical evidence.

3

political discontent in the 1960s to measure variation in levels of political contestation.Using a difference-in-differences strategy, we test whether, relative to before its powerwas contested circa the 1960s, the national and state PRI governments granted ejidosfarther away from municipality heads in places where they faced more agitation andopposition, and consequently expected more future political contestation. This identifi-cation strategy only relies on parallel trends in our outcome of interest, and deals withconcerns of systematic differences across municipalities, such as the relative location ofthe municipal head.

Figure 1 provides a graphical intuition of our identification strategy and results.It shows two municipalities in the same state of Durango, with a roughly similararea and land available for redistribution, but that experienced different levels ofpolitical contestation in the 1960s: low in the municipality on the left plots, and highin that on the right plots. Top-level plots depict land redistribution prior to 1960 andbottom-level plots after 1960. As predicted by our theory, the ejidos allocated after 1960were significantly farther away from the municipal head where the PRI experiencedhigher levels of political contestation. Instead, those allocated before 1960 had aroughly similar distance from the municipal head in both types of municipalities. Thissuggests that political contestation effectively led to a strategic placement of ejidos faraway from municipality heads in high-competition municipalities, which would haveotherwise remained as distant to those in low-competition municipalities. Difference-in-differences estimates confirm this graphical intuition: relative to before 1960, after1960 the PRI strategically granted ejidos significantly farther away from municipalityheads in places where it expected more political contestation. Interestingly, reinforcingour interpretation, our results are significantly larger when political contestation isgenuine and not driven by “parastatal” parties only opposing the PRI in appearance.

These results survive a number of key robustness, specification and falsificationchecks. First, the presence of parallel trends in the distance of allocated ejidos acrosscontested and uncontested areas, as well as a placebo exercise before 1960, confirmthat the diverging patterns of ejido distance to municipality heads appeared onlyafter the rise of political contestation in some Mexican municipalities in the 1960s.Second, these results are not explained by differential trends in variables differingacross contested and uncontested municipalities, since they survive the inclusion ofa battery of population, geographic and climatic municipal controls interacted with apost-1960 indicator. Third, we rule out the possibility that these results are explained bymean reversion and ceiling effects by controlling for the interaction with time-varyingmeasures of the stock of allocated ejidos and the land available for allocation in eachmunicipality, respectively. Fourth, we control for state-specific trends, verifying thatthe results are not driven by differences in competition between a few states. Fifth, wecontrol for the interaction with the municipal number of ranchos and haciendas to lessen

4

the concern that our measures of political competition simply reflect the power of thelocal elites. Sixth, our results are robust to considering different measures of distancefrom ejidos to municipality heads: in levels, in logs, accounting or not for elevationterrain profile, and accounting or not for the use of roads to reach municipality heads.

We also rule out a series of alternative interpretations of our findings by consideringalternative outcomes and various heterogeneous effects. First, we show that our resultsare not explained by the allocation of more-marginal, lower-quality ejidos. Thus, thedistance of allocated ejidos from municipal heads matters for its own sake, and notbecause it correlates with land quality. Second, by confirming that there is no effecton the number of allocated ejidos, the number of beneficiaries, or the area of the allo-cated land, we are able to rule out the possibility that our results are explained by adifferential distribution of land in more competitive areas, which would be consistentwith a PRI strategy of rewarding supporters and punishing the opposition (Albertus,Diaz-Cayeros, Magaloni, & Weingast, 2016). This is also important since ejidos werea fundamental part of such a client base, and one conjecture is that the PRI simplycompensated increased competition with increased ejido allocation. Given that it didnot—and instead manipulated the distance of ejidos, thus affecting the long-term cost ofpublic goods provision—favors our emphasis on the strategic manipulation of local bu-reaucratic capacity. Third, by showing that our estimates are not larger in municipalitieswith a higher population density or social capital, or in municipalities with populousmunicipality heads, we address the concern that our empirical results simply reflect aneffort to control the revolting masses by moving them to remote areas (Campante, Do,& Guimaraes, 2017). Lastly, we dismiss concerns that the PRI’s decisions regarding thedistance of ejidos from their municipal head significantly affected other dimensions oflocal state capacity other than the bureaucratic, including coercive and fiscal capacity.

Our paper contributes to the literature on the determinants of state capacity. Severalscholars study whether and how population density and inter- and intra-state conflictshave contributed to fiscal state capacity in Europe (Tilly, 1992; Gennaioli & Voth,2015), Africa (Herbst, 2000; Sanchez de la Sierra, 2017; Thies, 2007) and Latin America(Centeno, 1997; Thies, 2005).7 We study the role that political competition plays inexplaining choices that fundamentally influence local bureaucratic capacity in contextswhere conflict did not lead to state capacity development. Recent work by Acemogluet al. (2013) and Fergusson, Robinson, Torvik, and Vargas (2014) studies politicians’incentives to avoid eliminating non-state armed actors. While our paper shares anemphasis on political incentives to sustain state fragility, we focus on the bureaucraticability to effectively provide public goods throughout the territory (Soifer, 2015), ratherthan on the monopoly of violence.

7Herbst (2000) argues that low population density limits the development of modern state institutions. Instead of takingpopulation density as given and examining its implications, our work suggests that it can be endogenous to the political economyof state formation. Specifically, it can be a by-product of the incumbent’s effort to limit the bureaucratic state capacity.

5

Our paper also speaks to the literature that highlights the negative effects that clien-telism imposes on public service delivery and the emergence of credible party platformsthat cater to wider groups of voters (Bobonis et al., 2017; Fujiwara & Wantchekon, 2013;Hicken & Simmons, 2008; Keefer, 2007; Keefer & Vlaicu, 2008). Relative to this literaturethat highlights the relative appeal of particularistic transfers in developing democracies,we emphasize the incentives of clientelistic parties in forestalling investments in thelocal bureaucratic capacity required for public service delivery.

Our paper is also related to the literature that studies the incentives of clientelisticparties to block civil service reforms (Geddes, 1994; Heredia & Schneider, 2003; Shefter,1994). While the focus of this literature is on the loss of clientelistic resources withstate-strengthening reforms, we emphasize that such reforms might also underminethe comparative advantage of the clientelistic parties by reducing the cost of providingpublic goods.

More generally, our paper is closely related to the literature that highlights thebenefits of stronger political competition for public good provision, and more generallyfor economic development (Acemoglu, Naidu, Restrepo, & Robinson, 2017; Besley,Persson, & Sturm, 2010; Galasso & Nannicini, 2011; Naidu, 2017; Nath, 2015; Solé-Ollé &Viladecans-Marsal, 2012). In contrast to this literature, we highlight that in clientelisticcontexts increased political competition might lead to lower public good provisionsince incumbent clientelistic parties might have incentives to forestall investments inthe local bureaucratic capacity required for an effective public goods and services.8

Suryanarayan (2017) also suggests that electoral incentives may lead powerful elites toweaken the state. However, rather than curtailing the bureaucratic capacity to deliverpublic goods and services, she focuses on elites’ efforts to weaken the tax infrastructureso that potential new actors with political power have limited influence over theirwealth.

Lastly, our paper contributes to the Mexican literature on the role of ejidos in sus-taining the PRI’s hegemonic position (Silva Herzog, 1959; Eckstein, 1968; Sanderson,1986), in which the closest paper to ours is that of Albertus et al. (2016), who arguethat the PRI used land allocation to reward loyalists and punish opponents. Whilethey focus on the amount of land allocated as an outcome, we focus on the spatialdistribution of such allocations to capture bureaucratic capacity choices. The Mexicanland redistribution program provides not only a novel measure that captures localbureaucratic state capacity choices but also a unique context in which to study howpolitical incentives shape such choices by incumbent clientelistic parties.

Mexico specialists argue that the redistribution of land in the form of ejidos wascentral to the consolidation of the PRI’s hegemony (e.g., Silva Herzog (1959); Eckstein

8The effects of political competition may be non-monotonic. Our findings suggest that increased competition might exacer-bate distortions on bureaucratic capacity to retain electoral advantages. However, if competition is strong enough, the clientelisticparty may be forced to change its strategy and also offer public goods as argued by Diaz-Cayeros, Estevez, and Magaloni (2016).

6

(1968); Sanderson (1986)). Several authors also contend that the allocation of ejidos, inwhich individuals lacked property rights over the land they worked, created a politicaldependence that enabled the PRI to maintain a loyal political clientele (Albertus et al.,2016; Sabloff, 1981), while Larreguy (2013) shows that the PRI’s clientelistic networksstill operate in ejidos in states where the PRI controls the state government. However,in contrast to the literature that emphasizes the potential manipulation of the numberof ejidos allocated, we focus on their allocation across space. Our theory builds on theidea that placing ejidos far from municipality heads allowed the PRI to weaken localbureaucratic state capacity to provide public goods and services to limit the rise ofopposition parties with a comparative advantage in such provision.

The rest of the paper is structured as follows. In Section 2 we discuss the historicalbackground, placing particular emphasis on the land redistribution program and theshock that threatened the PRI’s hegemony in the early 1960s, which are central to ourempirical exercise. We then outline the model and its main predictions in Section 3.Section 4 summarizes our empirical strategy and data sources. We present our empiricalfindings and robustness checks in Section 5. In Section 6, we rule out most plausiblealternative explanations to our findings. Section 7 concludes.

2 Background

We describe the Mexican land allocation program, highlighting its origins, character-istics and political manipulation by the state and federal PRI governments. We alsodiscuss how the allocation of ejidos shaped the spatial distribution of individuals, andthus influenced the ability of municipal bureaucracies to provide public goods andservices. We then discuss the unprecedented social and political unrest that the PRIfaced in the 1960s, how it reflected the weakening of its clientelistic machine, how itchallenged the PRI’s hegemonic position, and how the party dealt with the insurgentsand political opposition.

2.1 The land redistribution program

A long history of land dispossession fueled the agrarian discontent that contributedto the Mexican Revolution in the early 20th century. Land distribution was thusat the center of the revolution and of Mexico’s 1917 constitution; Article 27 of theconstitution and subsequent legislation set up a process of land redistribution thatpersisted throughout the century. Land distributed to peasant communities in the formof ejidos was designated communal property, and therefore could not be sold, rented, orused as collateral for credit. Members of the community typically enjoyed inheritable(but otherwise non-transferable) use rights over specific plots that would be lost in

7

the event of an extended absence. Redistribution ceased in 1992 when a program thatentailed privatization and the possibility of the commercialization of redistributed landbegan.

Communities could put forward different types of land petitions. To receive newland, they could request land grants (dotaciones) or to have their land restituted (restitu-ciones). Over time, there were changes regarding: which communities were eligible toreceive land in its various forms, the size of plots granted, the definition of propertythat could be subject to expropriation, and the rights of private landowners to appealland reform decisions (Sanderson, 1986, Ch. 3). Importantly, none of these changeshappened around 1960.

We restrict our analysis to new land endowments (dotaciones), for several reasons.First, granting authorities had direct influence over the location of new land endow-ments, but not of land restitutions that had to be necessarily located where land hadbeen usurped. Second, new land grants typically involved relocating communities. Fi-nally, dotaciones also constituted the bulk of the reform. Appendix Figure A-2 comparesthe importance over time of new land endowments and restitutions, both in termsof the number of approved petitions and the hectares involved. Clearly, restitutionswere a marginal part of the land reform. This was partly because communities neededlegal documentation to prove prior expropriation and ask for land restitution. Thisdiscouraged applications for restitutions that, indeed, were the least likely of all grantsto be resolved in favor of the petitioner village (Sanderson, 1986, p. 88).

Land allocation only appeared to be driven by peasant demands. As noted, thecomparative failure of restitution processes forced petitioners to instead request newland endowments, over which authorities had more discretion. Rural communitiesinitially had to go through a cumbersome and bureaucratic process to request landfrom the state governor, who could either reject or conditionally approve their petitions.Final approval was granted first by the National Agrarian Commission, and ultimatelyby the president.9 The resulting highly centralized system gave the regime discretionover when and where to allocate land. Albertus et al. (2016), for instance, claim thatland distribution was strategically manipulated, and was higher during election yearsand in areas with a greater potential for social unrest.

Ejidos became key to the party’s dominance via clientelistic policies. The lack ofindividual property rights made peasants highly dependent on the PRI regime asthe only source of agricultural credit, investments, and technical assistance (Albertuset al., 2016). Moreover, legislation established the “democratically elected" office ofthe comisariado ejidal to administer each communal land, which included accessingand distributing government programs to the peasants in their communities. This

9We do not have information on rejected petitions, only on those finally approved. Sanderson (1986) reports that 80% ofpetitions that reached the presidency were approved. However, this figure misses potential earlier rejections at the state level andit is not available at the level of aggregation and frequency we utilize.

8

internal organization, together with the PRI’s corporativist apparatus, facilitated thedevelopment of long-lasting clientelistic networks in communal lands (Larreguy, 2013;Sabloff, 1981).

The PRI’s decisions about where to distribute new land endowments had importantlong-term consequences for local bureaucratic capacity since migrations from ejidoswas infrequent. Once individuals were located far from municipality heads, theybecame “tied” to their land, and thus unlikely to migrate. Comparing ejidatarios toprivate peasants with similar small units of production, Yates (1981, p. 151) notesthat “in recent times there has been observed a sharp reduction in the number ofthese [small private] units as their owners have sold out and migrated to the cities,whereas ejidatarios on equally small units being prohibited from selling, are much lessmobile).”10 Consistent with Yates (1981), de Janvry et al. (2015) show that householdsthat obtained property right certificates over the land they historically worked, thanksto the titling program started in 1992, were 28% more likely to have a migrant member.

2.2 The 1960s threat to PRI’s hegemony and its response

The PRI’s power was essentially uncontested from the late 1920s to the late 1950s.However, the country’s vibrant post-revolution economic growth reached its limits inthe late 1950s, which were characterized by general social discontent and protests fromthe main sectors of society previously under control of the PRI’s clientelistic machine:industrial workers, students, teachers, and peasants. This discontent was channeledinto organized political opposition, which represented an important threat to the PRI’shegemony in many areas of the country.

The rural sector was hit particularly hard by the economic crisis throughout the1950s. International prices of Mexican commodities collapsed, and there was an overallstagnation of agricultural production. From the late 1950s until well into the 1960s,peasant movements organized throughout Mexico, but particularly in the states ofBaja California, Morelos, Nayarit, Sinaloa, and Sonora. Peasants defied the leaders ofthe National Peasant Union (CNC), which was the PRI’s arm in the rural sector, byorganizing parallel structures of representation to channel demands for employment,better wages, and increased land redistribution (Bartra, 1985). The PRI government’sresponse was often to send the army to help the local police disband the rural protestsand incarcerate the insurgent leaders. The murdering of peasant leaders was also notuncommon. Land invasions became frequent, and many peasant organizations turnedinto guerrilla groups (Bellingeri, 2003; Herrera Calderón & Cedillo, 2012).

While peasants mobilized in rural areas, industrial workers and teachers actively

10Yates (1981) later explains the reasons for this phenomenon: “Theoretically, [the ejidatario] is free to leave whenever hewishes, but in practice he is a prisoner tied to his land, because, if he left, the ejido would give him no compensation for improve-ments he may have achieved through years of hard work. He is forbidden by law to rent his land, even to another member of hisown ejido.”

9

engaged in protests and strikes in urban centers against government control over andcooptation of their unions and leaders (Herrera Calderón & Cedillo, 2012). They alsodemanded better wages and working conditions, and union democracy and autonomyfrom the state. As with the peasants, the government usually repressed protesters andincarcerated their leaders. During the 1960s students emerged as major political actors(Herrera Calderón & Cedillo, 2012). Their movements proliferated in at least one-thirdof the nation, particularly in the states of Chihuahua, Guerrero, Jalisco, Michoacán,Nuevo León, Puebla, Sinaloa, Sonora and Tabasco, and Mexico City. Students fought foracademic and institutional reforms, broader access to higher education, improvementsin infrastructure, intellectual freedom, and against unpopular administrators. Studentstrikes were also frequent, and were often repressed by the army. The massacre ofTlatelolco in 1968, at which hundreds of students were murdered, was a landmarkevent in Mexico’s history.

The political opposition absorbed the social discontent (Bartra, 1985). In the early1960s, the PRI started to face strong threats in several gubernatorial and municipalraces (Bezdek, 1973; Lujambio, 2001). At the gubernatorial level, at least six highlycompetitive elections occurred in the late 1950s and 1960s in Baja California, Chihuahua,Nayarit, San Luis Potosi, Sonora and Yucatan. Fraudulent methods largely counteredpolitical opposition. Bezdek (1973) provides extensive accounts of various forms offraud that were central to the PRI’s response to the increased political competition. Asa result, despite the increased political competition, the opposition won in only 17 outof approximately 2,400 municipalities, and in one of the 31 states that held elections(Bezdek, 1973; Lujambio, 2001).11

While the PRI’s fraud prevented the increase in political opposition from materi-alizing in electoral competition in the short-term, the threat of electoral competitionpersisted. We later document that the number events of social and political discontentin the 1960s, which come from original data we collected from the universe of articlesof two Mexican newspapers—Excelsior and El Universal—between 1960 and 1969, corre-lates strongly with the electoral competition faced by the PRI in municipal electionsduring the 1980s.

To summarize, Mexico’s land redistribution program was a central policy of the PRIregime, and allowed the PRI to consolidate its power immediately after the revolution.The PRI further manipulated land redistribution to hold on to power when it beganto be politically challenged in the 1960s due to the unprecedented social unrest andinefficiency of the PRI’s clientelistic machine to control the voter under its influence,which was channeled through opposition parties. Importantly, the PRI’s decisionsabout where to relocate the communities of land petitioners had important long-term

11Three were state capitals: San Luis Potosi in the state of San Luis Potosi; Hermosillo in the state of Sonora and Merida inthe state of Yucatan. The winning opposition mayors ran in the subsequent gubernatorial elections.

10

consequences for local bureaucratic state capacity. With this background in mind, wediscuss a model of state building that focuses on the incentives that clientelistic partiesface when their power is contested.

3 A simple model of state building and political competition under

clientelism

We develop a simple model in the spirit of Robinson et al. (2006) and Robinson andVerdier (2013) to study the incentives that clientelistic parties face concerning bureau-cratic state capacity choices.

3.1 Setup

We consider a society in which an incumbent clientelistic (C) and an opposition non-clientelistic (NC) party compete for the rents from office R by splitting an exogenouslygiven budget T in particularistic transfers (τ) and public goods and services (g). Thenumber of voters is normalized to 1 and there are two types of voters. An exoge-nously given α share of voters—which we denote as clients (c) —are embedded in theclientelistic networks of the incumbent party, and thus it can target more efficientlyparticularistic transfers to those voters. The remaining 1− α share of voters—which wedenote as non-clients (nc)—can potentially benefit from particularistic transfers fromthe incumbent politician but he cannot target them as efficiently. To capture that theincumbent has a comparative advantage in clientelism, we assume that NC is unableto provide particularistic transfers to voters, and is thus restricted to allocating allresources to delivering public goods and services.12

The assumption that incumbent clientelistic parties have a comparative advantagein clientelism is central to the predictions of our model; otherwise, all our results flip.However, as highlighted earlier, it is often the case that incumbent clientelistic partiesare better positioned to engage in and enforce clientelistic exchanges, and thus enjoysuch a comparative advantage. Even more importantly, there is also overwhelmingevidence that this is also the case in our empirical application (Larreguy, 2013; Magaloni,2006).13

The budget constraint can generally be written as:

Pg(s)g + ατc + (1− α)τnc = T, (1)

12While this might seem a stark assumption, alternatively, we could have assumed a lower efficiency of the clientelisticmachine of the non-clientelistic party, e.g., considering the case that βC

clients > βNCclients, and obtained qualitatively similar results.

We abstract from commitment issues and assume that particularistic transfers can be credibly targeted to particular individualsin order to keep the discussion as simple as possible.

13While this is not testable in Mexico’s 1960s—there were close to no governments under opposition parties—Larreguy (2013)uses a difference-in-differences strategy and shows that, for the period 1994-2012, municipalities with a stronger presence of thePRI’s clientelistic networks experienced a significantly lower provision of public schooling under PRI governors.

11

where Pg(s) is the cost of providing public goods and services, and we have normalizedthe cost of particularistic transfers to one. We consider the case where Pg(s) is adecreasing function of the bureaucratic state capacity level s, P′g(s) < 0. The debate onhow to best conceptualize and measure state capacity is old but still active. While somescholars emphasize coercive capacity of the state (particularly its ability to monopolizethe use of violence), others (including us) emphasize its bureaucratic capabilities tolevy taxes and provide public goods and services. Due to the empirical context weconsider to test our theoretical argument, we focus on the role of bureaucratic statecapacity to provide public goods and services throughout the territory more efficientlyand at lower cost (Soifer, 2015).14 However, none of our theoretical results or theirinterpretation depend crucially on taking a stand on this debate.

Importantly, since we have normalized the cost of particularistic transfers to one,P′g(s) < 0 not only implies that bureaucratic state capacity reduces the cost of providingpublic goods and services, but that it does it relatively more than the cost of clientelism,which merits further justification. While this assumption is generally intuitive, theremight be the concern that, as with the case of local public good and service provision,the cost of particularistic transfers might also depend on local bureaucratic state ca-pacity. However, the decentralized nature of clientelism, which is mediated by localintermediaries—commonly known as brokers—based in their communities, and thetypes of goods used in clientelistic exchanges (e.g., construction materials, fertilizers,food, medicines), make it significantly less dependent on local bureaucratic state ca-pacity. Moreover, even when candidates support brokers through public employment(Robinson & Verdier, 2013), these are expected to work from their communities orare ghost workers that only show up to work for their paycheck, and consequentlyweaker local bureaucratic state capacity should not affect them as much, relatively tothe public officials in charge of delivering public goods and services.15 Also, since theyare selected on political rather than merit-based grounds, they are likely to be lowerquality and less able to effectively provide public goods and services themselves.

We denote the utility that the α share of clients and the 1− α share of non-clientsreceive from particularistic transfers and public goods and services, respectively, as:

Uc = βcτc + u (g) , and

Unc = βncτnc + u (g) ,

14While our model could easily be extended to consider the state’s ability to collect taxes, this would not affect the compara-tive advantage of clientelistic parties. Therefore, we do not expect it to affect the qualitative predictions of the model. Moreover,later we dismiss concerns that the PRI’s decisions regarding the distance of ejidos from their municipal head significantly affectedcoercive and fiscal state capacities.

15Importantly, this is the case in our empirical application. For example, the brokers that operate in ejidos are community mem-bers “democratically elected” to the office of comisariado ejidal and receive no formal payment from the government (Larreguy,2013). Moreover, while in Appendix Table A-1 we show that distance to municipality heads is negatively associated with contem-porary public good and service provision, qualitative accounts indicate that clientelism is much more prevalent in isolated andrural areas of Mexico (Magaloni, 2006).

12

where the utility from public goods and services u(g) is increasing and concave, u′(·) >0 and u′′(·) < 0, and, for simplicity, we assume that the utility from particularistictransfers is linear with βclients > βnon−clients, which captures that the incumbent’sclientelistic machine is much better at targeting and enforcing transfers to clientsthan non-clients. In line with Lindbeck and Weibull (1987), all voters also receive anidiosyncratic ideological shock σi and a general perceived-valence shock δ, both towardthe non-clientelistic party. Both shocks are uniformly distributed with a density of 1and centered at 0.

3.2 Characterization

Given the policy vectors (gC, τCc , τC

nc) and (gNC) proposed by the clientelistic and non-clientelistic parties, respectively, clients support the incumbent clientelistic party aslong as

uc

(gC)+ βcτC

c > u(

gNC)+ σi + δ.

Similarly, non-clients support the incumbent party as long as

unc

(gC)+ βncτC

nc > u(

gNC)+ σi + δ.

Integrating first over σi and then over δ, the winning probability of the incumbent partyis given by

ΠC =12+ αβcτC

c + (1− α)βncτCnc + u

(gC)− u

(gNC

). (2)

Notice that the incumbent clientelistic party enjoys an electoral advantage thanksto its ability to target particularistic transfers to clients. As consequence, the extent ofpolitical competition faced by the clientelistic party is inversely related to βc, whichcaptures the efficiency of the clientelistic machine among clients.16

We then consider the interaction between the incumbent clientelistic party and theopposition party. The latter faces a trivial optimization problem and allocates all theavailable budget to public goods and services by setting gNC∗ = T/Pg(s). The formerinstead maximizes its expected payoff (ΠC × R) by solving the following problem:

maxg,τc,τnc

(12+ αβcτc + (1− α)βncτnc + u(g)− u(gNC∗)

)R,

subject to the budget constraint in equation (1).

16We will see that in equilibrium the incumbent party does not target any transfers to non-clients, and thus βnon−clients playsno role in determining the political competition faced by the clientelistic party. βclients is exogenously given and, while endoge-nizing it might be of theoretical interest, we consider an exogenous change to it in our empirical application.

13

Focusing on an interior optimum,17 the first-order condition,

u′(

gC∗)= Pg(s)βc, (3)

indicates the optimal level of public goods and services for the clientelistic party.Moreover, since βc > βnc, τC∗

nc = 0, and putting together equation (3) with the budget

constraint in equation (1), τC∗c =

T−Pg(s)gC∗

α . Note that g is decreasing in β since,

from the first-order condition, ∂gC∗

∂β =Pg(s)

u′′(gC∗)< 0. Intuitively, with a more efficient

clientelistic machine among clients, particularistic transfers become more attractive forthe clientelistic party.

3.3 Predictions

We next consider the incentives of the incumbent party to invest in bureaucratic statecapacity and derive the model predictions. For simplicity, we assume that investmentsare costless but adding an additively separable cost would not affect the predictionsof the model. Consider the model’s first proposition. Increasing bureaucratic statecapacity may increase or decrease the clientelistic party’s payoff.

Proposition 1. Bureaucratic state capacity and the clientelistic party’s payoffThe clientelistic party’s payoff may be increasing or decreasing in bureaucratic state capacity s.

Proof. The simple differentiation of the clientelistic party’s winning probability in (2)implies

∂ΠC

∂s=

[−gCβ + u′

(T

Pg(s)

)T

P2g (s)

]P′g ≶ 0.

The expression for ∂ΠC

∂s in Proposition 1 shows that an increase in s, and the conse-quent fall in Pg(s), produces two opposite effects: a “real-budget” effect and a “relative-price” effect. The “real-budget” effect is due to an increase in the resources that theclientelistic party may use to transfer benefits to supporters. Since the ability to doso represents a clientelistic incumbent party’s comparative advantage, this first effectstrengthens its electoral prospects and provides incentives to bolster bureaucratic statecapacities. The opposite, “relative-price,” effect—which is caused by a reduction inthe cost of providing public goods and services—increases the public goods and ser-vices that the opposition party may provide, which hurts the incumbent’s electoral

17We assume that limg→0u′ (g)→ ∞ and that u′(T/Pg(s)

)< Pg(s)β so that the interior condition holds.

14

prospects.18 The overall impact of an increase in bureaucratic state capacity on theclientelistic party’s payoffs, therefore, depends on which of these two effects dominates.While this depends on the value of the various model parameters, our empirical appli-cation focuses on the role of electoral competition, which we examine more closely inthe next proposition.

Proposition 2. Electoral competition and bureaucratic state capacity buildingConsider an increase in the extent of electoral competition faced by the clientelistic party,captured by a decrease in β. The clientelistic party is more likely to support a reduction inbureaucratic state capacity s as a result of this increase in competition if and only if ρ > 1,where ρ is the relative risk aversion coefficient of u(g). Formally, ∂2Π

∂s∂β > 0 ⇐⇒ ρ > 1.

Proof. Recall that ∂gC

∂β =Pg(s)

u′′(gC). Substituting Pg(s) from (3) and using the definition of

ρ = − gu′′(g)u′(g) , β

∂gC

∂β = −gC/ρ. Substituting this in the cross derivative

∂2Π∂s∂β

= −P′g

(gC + β

∂gC

∂β

),

and simplifying, we obtain the stated result.

The intuition for this result is the following. An increase in electoral competitionfaced by the clientelistic party does not change the behavior of the non-clientelistic party.Thus, the “relative-price" effect of a reduction in s and the associated increase in Pg(s)—the decrease in public goods and services offered by the opposition—is unchanged.However, an increase in electoral competition faced by the clientelistic party affectsdirectly and indirectly the “real-budget” effects of a reduction in s—fewer resources areavailable for particularistic transfers. Directly, the cost of having fewer resources forparticularistic transfers is lower with a more inefficient clientelistic machine. Indirectly,gC∗ increases when β falls, which increases the “real-budget” cost of a reduction in s.As long as the direct effect is dominant, the clientelistic party prefers lower bureaucraticstate capacity when it faces more electoral competition.

Proposition 2 states that this occurs if and only if ρ > 1, or in other words, whenthe utility from public goods and services exhibits sufficiently strong diminishingmarginal returns. The key observation is that when this is the case, as we note above,the clientelistic party provides fewer public goods and services because their marginalutility is lower. As a consequence, the indirect effect is not very large. Thus, the directeffect dominates, and the clientelistic party, when faced with more competition, prefersto strategically reduce bureaucratic state capacity.

18This reduction in cost also increases the amount of public goods and services the clientelistic party may provide. However,according to the envelope theorem, the impact of an increase in s on the clientelistic party’s winning probability via the change ingC is negligible. Note that the envelope condition does not hold for the opposition party since it faces a corner solution.

15

When ρ < 1, the reverse occurs, and electoral competition contributes to strengthen-ing bureaucratic state capacity. In this scenario, contesting the power of the clientelisticparty also creates the conditions for clientelism to gradually erode, as an increase ins and an associated fall in Pg(s) leads to a decrease in the provision of particularistictransfers. More worrisome (but perhaps also more interesting) is the case of ρ > 1,when contesting the power of the clientelistic party instead induces it to cling to itsclient base and its strategy of delivering particularistic goods by making public goodsand services more expensive through a weaker bureaucratic state capacity. As we willsee below, this was the case in Mexico.

Assessing whether ρ > 1 is not feasible in our historical empirical context due to thelack of data and variation in incumbency, and to our knowledge, there are no measuresof ρ > 1 for public goods and services in the experimental and development literature.However, we compute estimates of ρ by exploiting the fact that β

∂gC

∂β = −gC/ρ, and

using estimates of the average β, ∂gC

∂β and gC from Larreguy (2013), who studies howthe effect of the PRI’s incumbency incumbency on education supply changes acrossmunicipalities with varying strength of the PRI’s clientelistic networks. Calibrationsconsidering the number of schools, teachers or students indicate that ρ is comfortablyabove the unity.19

We end by emphasizing that the PRI could have responded to a surge in competition(i.e., an erosion in the base of clients α) by increasing ejido allocation in order to producenew clients. However, as discussed before, the historical episode that we exploitconsiders a much more generalized decrease in the effectiveness of the PRI’s clientelisticmachine and the consequent erosion of PRI support. Moreover, though they wereimportant, increasing the number of land plots allocated or beneficiaries is likely tohave had a modest effect on the base of clients, since land petitioners were likelyto fall under the PRI’s corporatist apparatus anyway. Instead, increased distance ofallocated lands from municipal heads limited the bureaucratic capacity of the state toprovide public goods and services to entire communities and increased the burdenon the already weak local bureaucracies, which significantly damaged the electionprospects of opposition parties operating there. More importantly, as we show below,there was no differential increase in the extent of land allocation in more competitivemunicipalities after the 1960s. This allows us to rule out these and other relatedalternative interpretations whose formalization we omit.

19We measure β considering the mean municipal share of ejidos, 0.234. We proxy for g using the municipal mean of schools,

teachers, and students, which are respectively given by 1.276, 8.343, and 191.6. ∂gC

∂β is respectively given by -1.276, -8.343 and-191.6 for schools, teachers and students.

16

4 Methods and data

4.1 Empirical strategy

A key implication of our model is that clientelistic parties should choose weakerbureaucratic state capacity when the efficiency of their clientelistic networks weakensand they are more likely to be challenged by an opponent. To test this prediction, weexamine whether the PRI allocated ejidos further away from their municipality heads inmunicipalities where it was politically threatened.

Given the usual lack of historical data on local bureaucratic capacity over time, anatural indirect outcome with which to explore this relationship is provision of publicgoods and services. However, the use of such an alternative outcome is problematicfor our purposes for several reasons. First, municipal-level data on public goods andservices (e.g., the percentage of households with access to piped water, sewage orelectricity) are only available in census years (approximately every ten years), andwe have not been able to retrieve it for most censuses prior to 1960, which makes ouridentification strategy impossible. Second, the use of municipal public good provision(or bureaucratic presence, if such data were available) as an outcome exacerbatespotential endogeneity concerns. In particular, it may be that more developed areas havemore public goods and services (or more public functionaries) for reasons other thanimproved local bureaucratic state capacity, as well as differential patterns of politicalcompetition.

We instead use a novel measure that captures municipal bureaucratic state capacitychoices: the distance of the ejidos from their municipality heads. This measure hasseveral advantages. First, this distance is an important determinant of the local bureau-cracy’s ability to provide the inhabitants of the newly allocated ejidos with public goodsand services, and as such, captures local bureaucratic state capacity. Many governmentdocuments point at the distance of localities from their municipal as one of the mainbarriers for public good provision and development. For example, Baja California(2003, p. 19) highlights that

“The greatest difficulties (...) concern the distribution of a large share of itspopulation in rural localities under (...) 2,500 inhabitants, many of whichare dispersed in areas of rough topography and distant from highwaysand roads, far away from the power-generating centers and municipalityheads.”

Similarly, SEDESOL (2014, p. 18) indicates that

“The interaction between the rural population and its municipal heads isvery limited due to the poor conditions of the roads, which directly affects

17

their access to basic services, health, education or the municipal governmentoffices where these services are requested.”

It is then not surprising that, the 2007-2012 National Development plan emphasizesthat

“it is necessary to provide communities with more solutions when it comesto education, health, nutrition and housing, as wells drinking water, sewage,electricity, roads and even telephone lines to bring their populations farfrom municipal heads closer (Mexico, 2007).”

To reinforce this point, Appendix Table A-1 uses locality-level outcomes fromthe 1990 and 2000 Mexican censuses to show that distance to municipality heads isnegatively associated with the contemporary provision of public goods by municipalgovernments, as captured by the share of households with piped water connections,drainage, or electricity.20 This is true both in the full set of Mexican localities (panels Aand C) and in a subsample of localities that intersects with ejidos (panels B and D). Thesize of the coefficients is economically meaningful. For instance, Panel A of AppendixTable A-1 indicates that a one-standard-deviation increase in distance (21.6 km, seeTable 1) is associated with a 7.8 percentage point drop in the share of households withelectricity (21.6×−0.0036) in 1990. Since the share of households with electricity in1990 equals 42.3%, this is quite a sizable drop of 18.39% relative to the mean share ofhouseholds with electricity. The corresponding figures for the share of households withpiped water and drainage are 14.36% and 26.39%, respectively. Moreover, while allthese drops in local public goods are sizable, they likely constitute underestimates ofthe short- and medium-term impact of the PRI’s local state-capacity decisions we studydue to remedial policies undertaken over time, as the relatively weaker results usingthe 2000 census data indicate.

The second advantage of our measure is that, compared to public goods provision(and even other measures of local bureaucratic state capacity), the distance of the newejidos from their municipality heads captures a permanent choice of local state capacity,given the inhabitants’ lack of geographical mobility (de Janvry et al., 2015; Yates, 1981).Third, since the decision of whether and where to allocate ejidos was solely under thecontrol of the national and state PRI governments, our measure addresses some of theendogeneity issues concerning the provision of municipal public goods and services,which were largely determined by local governments. Lastly, our novel measure that

20We exclude outcomes, such as education- and health-related ones, since municipal governments are not responsible fortheir provision. More specifically, we run the following regression:

Share Public Good`,m = α+δ ·Distance`,m + ηm + ε`,m,

where Share Public Good`,m is the share of households in locality ` of municipality m with either piped water, sewage, or elec-tricity. Distance`,m is calculated as the distance between the centroids of locality ` and municipality head of m, and ηm denotesmunicipality fixed effects.

18

captures local state capacity choices is available from the 1910s to the 1990s and variesyearly, which allows us to implement an identification strategy that addresses theremaining concerns regarding the endogeneity of local political competition.

To deal with endogeneity concerns regarding municipal political competition, ouridentification strategy exploits the national shock that threatened the PRI’s hegemonyin the early 1960s with varying intensity across municipalities. Our difference-in-differences strategy exploits this plausibly exogenous variation in the extent of politicalcompetition. Specifically, we conduct a within-municipality analysis where we testwhether—relative to land allocation patterns before its power was contested circa the1960s—the PRI granted ejidos farther away from municipality heads in areas where itfaced more political opposition. Our baseline specification is:

Distancee,m,t = α+β · Post1960e,m,t+

γ · (Post1960e,m,t × Political Competitionm) + ηm + prest + εe,m,t,(4)

where the dependent variable is the distance from ejido e to the municipality m head inyear t, while Post1960e,m,t is a dummy variable that equals 1 if ejido e was created after1960, Political Competitionm is a measure of political competition, ηm are municipalityfixed effects, and prest is a full set of time fixed effects identifying the presidentialperiod in which ejido e was created. ηm deals with systematic differences across munici-palities and prest with the fact that some presidents engaged in significantly more landredistribution than others, which could have led to the distribution of more isolatedejidos at times. We cluster errors at the municipality level.

We consider three different measures of Political Competitionm in municipal racesfor mayor:

a) Vote dispersion: 1−∑ni=1 p2

i , with pi equal to the vote share of each of n partiesthat run in the considered municipal races

b) Opposition vote share: 1− Votes for PRITotal votes

c) Opposition ever won:

0 if the PRI won every election in

the considered municipal races,1 otherwise

To calculate these political competition measures, we use only municipal electoraldata from the 1980s for two reasons. First, while some municipal electoral results areavailable for the 1970s, these are not complete, which leads to the concern that theiravailability is systematically correlated with the level of electoral competition. Second,the 1960s and 1970s were characterized by all sorts of electoral fraud, which we also

19

expect to be associated with the electoral competition faced by the PRI. After the 1977electoral reform, which paved the way for multiparty competition and cleaner elections,electoral figures are both fully available and much more reliable (Klesner, 1993).21

For the case of vote dispersion and opposition vote share—our first two measures ofpolitical competition—we consider their averages over all municipal elections duringthe 1980s. Similarly, we consider whether the opposition won a municipal electionduring the 1980s to define our third measure of political competition. By consideringpolitical competition measures that incorporate several municipal election results, wereduce potential noise coming from particular unusual elections. In our robustnesschecks in Section 5, we show that our qualitative results are unaffected when we insteadrestrict our sample to the first municipal election in the 1980s.

A natural potential concern is that municipal electoral competition in the 1980s is abiased proxy for the threat that the PRI faced in the early 1960s, since it is endogenousto some of the same mechanisms emphasized by our theory. Yet, if anything, we arguethat this should bias our estimates against us finding any results: our theory suggeststhat the PRI should have allocated ejidos farther from municipality heads where itspower was most threatened in the 1960s, which in turn should have lessened futurepolitical competition, particularly in the 1980s. Additionally, the likelihood that such anendogeneity might affect our measures of municipal electoral competition in the 1960scould vary across the different measures. However, we show that our results are robustacross of all of them.

Furthermore, as an additional robustness check we construct an original dataset onevents of social and political discontent in the 1960s using data from the universe ofarticles of two Mexican newspapers— Excelsior and El Universal—between 1960 and1969. These represent a proxy for political contestation that is not influenced by thechanging patterns of allocation of ejidos that resulted from the national shock on thePRI’s hegemony, and this is especially so when we additionally exclude rural events. Wethen use the number of such events to instrument for the political competition variablesthat we use in our difference-in-differences strategy, focusing merely on the extent ofpolitical competition in the 1980s that was driven by social and political discontent inthe 1960s. Due to possible concerns regarding the violation of the exclusion restrictionof our instrumental variable strategy (i.e., discontent might have affected the locationof newly allocated ejidos through an alternative mechanism other than its impact onpolitical contestation), we also report reduced form estimates where 1960s events ofsocial and political discontent are used directly as proxy for political competition.

Assuming that our measures of municipal electoral competition in the 1980s—whether instrumented or not— or the number of municipal protests are unbiased

21Notably, the new electoral legislation shortened the deadline for delivering the results of each polling station from oneweek to 24 hours for urban precincts and 72 hours for rural ones, which limited the chances of manipulating the results.

20

proxies for the electoral competition that the PRI expected to face at the municipallevel the 1960s, there could still be the concern that such measures are endogenous.However, our difference-in-differences identification strategy addresses this concernsince it does not exploit increased political contestation in particular municipalitiesover time, which is likely to be endogenous, but a national-level shock that threatenedthe PRI’s hegemony differently throughout the country. Accordingly, we show thatbefore the 1960s, there are no differential trends in our measure of local state capacitychoice across places with varying political competition.

One last potential concern regarding the interpretation of our empirical results isthat the national-level shock that threatened the PRI’s hegemony in the 1960s variedin different areas due to municipality characteristics that are correlated with electoralcompetition. We carefully address this concern by taking advantage of the richnessof our data. We first make sure that our results are not driven by predeterminedmunicipality characteristics potentially correlated with electoral competition, includingdifferences in population, geography, climate, and strength of rural elites. Second, werule out the possibility that our results are explained by differences in the stock ofallocated ejidos or agricultural land available for distribution. Additionally, we discussa number of possible alternative interpretations of our results and present a series ofexercises to address them by studying other outcomes and potential heterogeneouseffects.

4.2 Data sources and summary statistics

Our empirical analyses require data from a variety of sources. We use data on the spatiallocation of localities and municipality heads and municipal public service coveragefrom the 1990 and 2000 censuses conducted by the Instituto Nacional de Estadísticay Geografía (INEGI).22 We use the location of ejidos and their mapping to localitiesfrom Mexico’s land certification program, the Programa de Certificación de DerechosEjidales y Titulación de Solares, or PROCEDE. The number of beneficiaries at the timeof allocation, area originally allocated, and allocation date of each ejido come fromthe Padrón e Historial de Núcleos Agrarios (PHINA).23 The electoral data are from theBANAMEX-CIDAC electoral database.24

All distance of ejidos from their municipality head refer to the population-weighteddistance of the ejido localities from the municipality head (See Appendix Figure A-1for details on this computation).25 Our baseline specification considers the minimumEuclidean distance of the ejido localities from the municipality head. In our robustness

22http://www.inegi.org.mx/est/contenidos/Proyectos/ccpv/cpv2000/23The data were scraped from http://phina.ran.gob.mx/phina2/ by Melissa Dell, who generously shared it with us.24http://www.cidac.org/eng/Electoral_Database.php25We use population figures from the 2000 Census, once all ejidos were allocated.

21

checks, we consider two alternatives to this distance measure. First, we account forelevation terrain profile by penalizing our baseline distance computed when there arechanges in altitude in the straight path that connects the locality and the municipalityhead. Second, we account for the use of roads to reach the municipality head. The traceof those roads comes from the Digital Chart of the World of 1992 and the overall distanceof each locality from its municipality head is computed adding up two different figures.First, the Euclidean distance from the locality to the closest point in a road that leads tothe municipality head, and second, the length of the segment that connects such pointto the municipality head following the road path.

Our main electoral competition variables rely on the vote shares of the PRI andopposition parties. In additional exercises, we also classify the opposition as ‘friendly’or ‘unfriendly’ to the PRI; friendly parties are those classified as “parastatal” partiescontrolled by the state and only opposing the PRI in appearance (Molinar & Weldon,1990; Peiro, 1998). The classification of each party listed in our database is shown inAppendix Table A-2.

We construct ejido-level measures of climate and geography (e.g., altitude, area,rainfall, soil humidity) using corresponding data from INEGI.26 We also use informationabout the land quality of the allocated ejidos from two different sources. First, we use theinherent land quality index database reported by the U.S. Department of Agriculturethat rates soil resilience and performance around the world based on climate andgeological factors.27 These two dimensions on a three-level scale (low, medium andhigh resilience and performance) comprise a nine-level land quality index, rangingfrom the best type with high performance and resilience (class 1) to the worst type,with low performance and resilience (class 9).28 To interpret this classification as a landquality measure ranging from 1 to 9, we recalculate so that higher values indicate higherland quality. Second, we construct a soil quality measure using data from the UN Foodand Agriculture Organization (FAO) that takes into account the major environmentalconstraints and opportunities for agricultural production.29 The soil quality measure isa seven-level scale, which we turn into a dummy variable for ease of interpretation.30

We also use INEGI’s historical catalog of localities to construct several variables:municipal log population in 1900 and 1960, municipality head population in 1960, andnumber of ranchos and haciendas.31 We also construct an index of municipal socialcapital using data from the 1994 Mexican directory of civil organizations (Secretaría de

26http://www.inegi.org.mx/geo/contenidos/topografia/default.aspx27http://www.nrcs.usda.gov/wps/portal/nrcs/detail/soils/use/?cid=nrcs142p2_05401128See http://www.nrcs.usda.gov/wps/portal/nrcs/detail/soils/edu/college/?cid=nrcs142p2_05402929http://data.fao.org/map?entryId=c1f62b50-88fd-11da-a88f-000d939bc5d8&tab=metadata30Specifically, we code the first five categories of the scale (1, too cold/dry; 2, low suitability; 3, unreliable rain; 4, slope higher

than 30 degrees; 5, degraded), which capture soil of poor quality, as a 0, and the last two categories (6, medium/low rain-fedpotential; 7, high rain-fed potential), which capture soil of good quality, as a 1.

31We accessed the data from http://www.inegi.org.mx/geo/contenidos/geoestadistica/catalogoclaves.aspx

22

Gobernación, 1994). In particular, we consider the number of organizations of humanrights, popular fronts and peasants. Lastly, we construct the stock of agricultural landavailable for distribution using data from PHINA and the 2007 Agricultural Census.32

As noted, we also collected original information on social protests from two Mexicannewspapers— Excelsior and El Universal. We chose these newspapers since they werethe only two of national coverage known to be relatively uninfluenced by the nationalgovernment. We coded all news stories on protests, strikes, demonstrations, riots andmarches from the universe of articles published between 1960 and 1969, and computedthe (log) number of events reflecting social and political discontent during that periodby municipality, both including and excluding rural events. Appendix Figure A-3shows the distribution of all events per year, and Appendix A.2 presents additionaldetails on the construction of this variable.

Table 1 shows summary statistics for all the variables used in our empirical analyses.There is significant variation in our baseline distance of ejidos to their municipality heads(mean of 18.8 km and standard deviation of 20.3). The PRI’s dominance is clear in ourpolitical competition variables, although there is also important variation. On average,the vote dispersion variable equals 0.206 (standard deviation of 0.165). Vote dispersionequals 0 when one party gets all the votes, and 1− (1/n) when n parties equally splitthe vote. Thus, political competition is far from the two-party case. Consistent with thisfigure, the average opposition vote share is close to 16% (standard deviation of 14%),and the opposition won at least one election during the 1980s in 11% of municipalities.When measuring these variables using data from the first election in 1980, there ison average less competition, although the variation across municipalities is similar interms of vote dispersion and opposition vote shares.

Figure 2 plots the frequency of the allocation of ejidos over time. In spite of the well-known peak in ejido allocation that occurred during the Lázaro Cárdenas administration(1934–40), land reform was active with close to 1,000 ejidos granted every quinquenniumuntil the end of the century.

5 Results

5.1 Main results and falsification test

We begin by graphically exploring our basic hypothesis as embedded in our baselinespecification in equation (4) together with the validity of our key identification assump-tion. Figure 3 illustrates the spatial distribution of electoral competition (upper panel)and the change in the average distance of allocated ejidos after 1960 compared to before1960 (lower panel). The correlation is visually apparent: more competitive areas in the

32http://www.inegi.org.mx/est/contenidos/proyectos/Agro/ca2007/Resultados_Agricola/

23

upper map (darker areas) tend to coincide with a stronger increase in the distance ofejidos from the municipal head in the lower map. We examine this correlation moresystematically in Figure 4. To construct this figure, we run a regression analogousto our baseline specification in equation (4) in which we interact each competitionmeasure with a full set of quinquennium dummies qt (where q1915 equals 1 if the ejidowas allocated from 1915 to 1919, q1920 equals 1 if it was allocated from 1920 to 1924, andso on). When running this regression, as with all subsequent tables, we standardize thecompetition measures for ease of interpretation of the coefficients.33

Figure 4 plots the resulting coefficients for the interactions between competitionand the quinquennial dummies (1915 is the omitted quinquennium).34 There are threegraphs, one for each of our competition measures: vote dispersion, opposition voteshare, and whether the opposition won an election in the 1980s. The results supportboth the validity of our identification assumption and our hypothesis. Before 1960,when the PRI’s power was not challenged, the interaction coefficients are close to zeroand statistically indistinguishable from those of the 1915 quinquennium. Therefore,prior to 1960, the distance from the allocated ejidos to their municipality heads trendedtogether in places with high or low political competition. However, starting in the 1960s,the interaction coefficients are positive and statistically different from those of the 1915quinquennium. This indicates that, after 1960, there is a significant differential increasein the distance between ejidos and their municipality heads in more competitive mu-nicipalities. Figure 4 thus confirms that the 1960s marked a stark change in the spatialpatterns of ejido allocation across municipalities with varying political competition.

Having graphically confirmed the validity of our identification assumption, PanelA of Table 2 presents the results of our baseline specification in equation (4) for ourbaseline measure of the ejido distance from the municipality head, which simply takesinto account the minimum Euclidean distance. There are three columns, one for each ofour competition measures: dispersion in Column 1, opposition vote share in Column 2,and whether the opposition ever won in Column 3. We follow this structure consistentlyin the other table panels, as well as in tables that follow. We find that the interactionterm between political competition and the post-1960 dummy—γ in equation (4)—ispositive and statistically different from zero, which again confirms that, after 1960, morecompetitive areas experienced a relative increase in the distance of the newly createdejidos from their municipality heads. The effects are non-negligible in size, and arevery precisely estimated and consistent across our measures of political competition. Aone-standard-deviation increase in vote dispersion leads to an approximately 2.83 km

33As in our baseline specification in equation (4), the regression includes municipality and time fixed effects. As we do in Table6 of the robustness checks below, we also add the interaction of the quinquennial dummies with a host of population, geographicand climatic municipal controls to make sure these patterns are not driven by trends based on other municipal characteristics thatare correlated with competition.

34The analogous figure for the log of the events of social and political discontent that we use as our baseline instrumentalvariable is in Appendix Figure A-4.

24

increase (recall that the competition measures are standardized) in the distance of ejidosfrom their municipality head after 1960, which is about 15% of the sample average. Thecoefficients for the opposition vote share and opposition ever winning imply a roughlysimilar effect, with a one-standard-deviation increase translating into an approximately3.3 and 3.16 km rise, respectively. Panels B and C of Table 2, respectively, indicate thatthe effects are virtually unchanged when accounting for elevation terrain profile or theuse of roads to reach municipality heads when computing the ejido distance from themunicipality head.

Returning to our identification assumption, we can further verify the validity ofthe parallel trends assumption observed in Figure 4 by formally testing for it, andby conducting a placebo analysis in Table 3. Panel A replicates the results in PanelA of Table 2. Panel B shows the results where we add indicators for 5-, 10- and 15-year leads to the baseline specification in equation (4). None of the indicators is eitherstatistically significant or sizable. Panel C introduces the results of a placebo analysiswhere we drop all the ejido allocations after 1960 and estimate the interaction of each ofour measures of political competition in the 1960s and an indicator for post 1935, whichcorresponds to the median year in the ejido allocations prior to 1960. The interactionswith the post-1935 indicator are precisely estimated zeros, and of an order of magnitudesmaller than those in Panel A, for all of the political competition measures. Results inPanels A and C of Table 3 then confirm that, as we observe in Figure 4, prior to 1960,the distances of the allocated ejidos from their municipality head exhibit parallel trendsacross municipalities with varying political competition.

Before presenting a detailed discussion of our robustness checks and additionalresults, we briefly note three less substantial (but nonetheless important) ones thathave been relegated to the Appendix. First, in Appendix Table A-3 we run our basicspecification and robustness checks using only the first municipal election in the 1980sto construct the competition measure. As noted, to avoid noise coming from unusualelections, we prefer our baseline measure relying on averages for all elections in the1980s. However, Table A-3 shows that our baseline results and robustness checks aresimilar when using this alternative way of measuring competition.35

Second, Appendix Table A-4 reports robustness checks considering the logarithm ofthe distance of allocated ejidos to their municipality head as an outcome rather than thelevel. This robustness check is of particular importance since, if municipalities withvarying political competition had different outcomes prior to the 1960s, the findingscould be sensitive to the transformation of the outcome variable. In particular, con-sidering absolute or proportional changes in the distance of allocated ejidos to their

35When we use whether the opposition won as the competition measure, the coefficient falls and is not significant in thesimplest baseline specification. This result is perhaps unsurprising, since such a crude approach is possibly the noisiest measureof competition, especially when relying on a single election. Moreover, when we add additional controls and as a result gainprecision, the increase in distance in more competitive places is typically statistically significant even using this measure.

25

municipality heads could yield different answers. However, the results in AppendixTable A-4 indicate that our findings are robust to a logarithmic transformation of theoutcome variable. This table also quickly summarizes the size of our effects. Notice thatthe estimated coefficients very consistently imply an effect of about a 5− 10% increasein the distance of allocated ejidos (relative to the sample mean) for a one-standarddeviation increase in the competition measures.

Third, Appendix Tables A-5 and A-6 show our main results and robustness checkswhen, respectively, accounting for elevation terrain profile or the use of roads toreach municipality heads when computing the distance of the ejido localities fromthe municipality head. The results are virtually unchanged when considering thesealternative ways of computing such distance as outcomes. If anything, the magnitudesare somewhat larger and we gain statistical significance.

Finally, though we have not reported it in the Appendix or main text to savespace, we also verified that our main results are not driven by any single state. Morespecifically, we estimated our baseline model, dropping each Mexican state from thesample one by one. In all cases, we found a significant interaction term between thepost-1960 indicator and our competition measures.

5.2 Friendly and unfriendly opposition

An additional telling exercise comes from the investigation of differential effects asa function of the nature of the opposition faced by the PRI. As described in Section4.2, some of the opposition parties were friendly to the PRI. These parties are oftenreferred to as “parastatal,” as they were presumably controlled by the state but servedthe purpose of presenting an image of political diversity and openness, and potentiallyprevented the development of true competition. Presumably, the development of suchparties was particularly important in places where the PRI expected some competition.Thus, we still expect a positive interaction between the presence of friendly oppositionparties and the distance of ejidos from municipal heads after 1960. However, since theseparties were not as threatening to the PRI’s hegemony, their effect could have beensomewhat muted.

Table 4 presents our baseline specification where we use the friendly and unfriendlyopposition vote shares, separately, as competition measures. For reference, Column1 shows our baseline result treating all opposition groups equally when computingthe competition measure. In Column 2, where we use the vote share of friendlyopposition parties, we find a significant but much smaller effect on the interaction of1.19 km. Column 3 finds that, when focusing on unfriendly parties, the coefficient onthe interaction is almost three times as large (2.91) as that of friendly parties. Finally,Column 4 includes the vote share of friendly and unfriendly parties separately in the

26

same regression. Both interaction terms are statistically significant, and we confirm alarger effect of competition of unfriendly parties’ relative to unfriendly parties’ voteshares (3.04 versus 1.45 km). A test for the inequality of these coefficients shown in thelower panel of the table is short of statistical significance at conventional confidencelevels (p = 0.15).

Table 4 thus suggests that the unfriendly opposition had a bigger effect on the PRI’slocal state capacity decisions, which is in line with our expectations. In what follows, tostack the deck against us (and to avoid making our results susceptible to the specificclassification of friendly and unfriendly parties), we continue to treat all oppositionparties equally when computing our political competition measures.

5.3 Omitted variables

We conduct additional exercises to rule out potential alternative mechanisms that aredriving our empirical results. We start by addressing the concern that our estimatesare driven by municipal factors other than electoral competition. In particular, weaddress the concern that our estimates might reflect the effect of omitted municipalitycharacteristics that correlate with electoral competition, which independently affect thedistance of the allocated ejidos from their municipal head starting in the 1960s. We testfor the relevance of this concern with the following specification:

Distancee,m,t = α + β · Post1960e,m,t + γ · (Post1960e,m,t × Political Competitionm)

+ ∑i

δm

(Post1960e,m,t × Xi

m

)+ ηm + presr + εe,m,t, (5)

where Xim is a set of (predetermined) municipal characteristics. Since the set of variables

Xim must be exogenous, we focus on geographic and climatic municipal variables that

could potentially both correlate with electoral competition and affect the distance ofallocated ejidos from their municipal head. These variables include area, historical pop-ulation, average rainfall and rain variability, soil humidity and its variability, averagealtitude and its variability (ruggedness), as well as soil quality. Importantly, we do notinclude any ejido-level controls since these might be endogenous outcomes.

We first assess whether those predetermined municipal variables are correlatedwith electoral competition at the municipality level. These associations are examined inTable 5, which considers specifications in which we run political competition measureson those municipal characteristics. We effectively observe that several variables arecorrelated with political competition. Thus, it is possible that our estimates regardingthe effect of political competition on bureaucratic state capacity choices are drivenby other reasons. Note, however, that if other reasons are driving those estimates,their effect should also manifest only from the early 1960s onward, which we find less

27

plausible.The specification in Table 6 directly assesses the extent of this concern, and presents

the results of the specification of equation (5), in which we control for the interac-tion of the post-1960 dummy with all the above-mentioned predetermined municipalgeographic and climatic characteristics. The results indicate that some of these charac-teristics have an influence on the distance of the allocated ejidos from their municipalityhead after 1960. In particular, the interactions of the post-1960 dummy with the areaand average rainfall are statistically significant across all the specifications consideringdifferent measures of electoral competition.

The signs of these significant interactions are also as expected. For instance, theejidos allocated after 1960 are particularly distant from municipality heads in largermunicipalities, which may reflect the availability of distant land. In municipalitieswith higher average rain, land was also allocated in more distant places after 1960.Since rainfall is an important determinant of land quality, this may indicate that inthose municipalities the land closest to the municipality heads was already taken, andnew ejidos could only be allocated in more distant places. A similar reasoning mayexplain the negative sign (which is short of being significant at conventional levels) ofthe interaction with rain variability in Columns 2 and 3.

However, the most relevant finding in Table 6 is that none of these potential con-founding variables can account for our main results. Once we control for the interactionwith all the climatic and geographic characteristics, the coefficients on the interactionswith each of the different electoral competition measures remain statistically signifi-cant and with very similar magnitudes to those reported in Table 3. These findingsdemonstrate that our estimates do not appear to be driven by other previously omit-ted drivers of the distance of ejidos from their municipality heads that correlate withpolitical competition.

5.4 Mean reversion and ceiling effects

Another potential concern is that our estimates simply reflect mean reversion or ceil-ing effects. For example, more land could have been allocated in more competitivemunicipalities initially. Alternatively, these municipalities could have had less landcloser to the municipal head available for redistribution. If either of these two situationswere the case, over time there would have been less land close to municipality headsavailable for redistribution in contested municipalities. Consequently, our results couldbe explained by differences in the land available for redistribution over time ratherthan by local bureaucratic state capacity choices by the PRI.

To empirically address these potential concerns, we control for mean reversionand ceiling effects by running a specification analogous to equation (5), where Xm is

28

a measure of either the stock of allocated ejidos or the stock of agricultural land stillavailable for redistribution (but not yet redistributed at time t of the creation of ejido e inmunicipality m). By including the interaction with the stock of allocated ejidos, we canaddress whether our results are driven by mean reversion. By including the interactionwith the stock of agricultural land available for redistribution, we can confirm that ourestimates are not the result of ceiling effects.

Panel A of Table 7 reports the specification in which we include the number of ejidosthat had been granted in the municipality from 1914 to year t− 1 and its interaction withthe post-1960 indicator. The coefficients of the interaction between political competitionand the post-1960 dummy remain not only significant but also similar in size to thosereported in Table 3 across all measures of political competition.36 Additionally, theinteraction between the stock of allocated ejidos and the post-1960 indicator is smalland often insignificant.

Panel B of Table 7 shows the estimates of the specifications that address the relatedconcern of ceiling effects. We estimate the stock of agricultural land available fordistribution at time t in municipality m by subtracting the stock of land allocatedfrom 1914 to year t − 1 from all the agricultural land available for redistribution.37

The estimates indicate that, even though in municipalities with a larger stock of landavailable for redistribution the ejidos granted after 1960 were more distant from theirmunicipality heads, this cannot account for the significance of our estimates. Whilethe size of the estimates of interest does drop, they remain sizable and statisticallysignificant.38

We also document in Table 10, which is discussed in section 6.2 below, that the dif-ferences in allocation patterns across municipalities with varying levels of competitionis restricted to the distance of allocated ejidos from their municipality heads, rather thanthe number of ejidos, as we would expect with mean reversion or ceiling effects. Overall,these findings lessen the concern that our results might be capturing mean reversion orceiling effects.

5.5 State politics

Since much of Mexican politics, and certainly the granting of ejidos, was determinedat the state level, one concern is that our results are driven by a few states exhibiting

36We obtain very similar results if instead Xm is a measure of the number of allocated ejidos in 1959.37More specifically, using the INEGI’s 2007 Agricultural Census and the PHINA’s records of land granted, we calculate the

stock of land available for redistribution as:

LandAvailablemt = Agricultural land 2007m − Stock of land granted since 1914m,t−1,

where Agricultural land 2007m is all of the agricultural land, and thus the potential land available for redistribution, andStock of land granted since 1914mt includes the accumulated outright grants, restitutions and enlargements. Ideally, we wouldwant to control for the availability of land for redistribution within a certain distance of the municipal head but such data is notreadily available.

38We obtain very similar results if instead Xm is a measure of the stock of agricultural land available for distribution in 1959.

29

distinct patterns in the distance of the allocated ejidos to their municipal heads afterthe 1960s. To address this potential concern, we include a series of state-specific timecontrols. Panel A of Table 8 adds state-specific cubic time trends in addition to theinteraction of each state fixed effect with the post-1960 dummy. In this specification,identification comes from variation in electoral competition across municipalities withinthe same state, and not from comparisons of municipalities across states. Estimatesin Panel A of Table 8 indicate that the results are essentially unchanged: the statisticalsignificance and size of the coefficients are in line with those in Table 3.

5.6 Endogeneity of our competition measures

As mentioned above, a potential concern about using election outcomes from the 1980sas proxies for political competition in 1960s is that these are endogenous to the processof ejido allocation. To address this concern, we use the (log) number of events reflectingsocial and political discontent—protests, strikes, demonstrations, riots and marches—from 1960 to 1969 at the municipality level both to instrument for political competitionin the 1980s and separately as a proxy for political competition in the 1960s.39

By using events reflecting social and political discontent as an instrument for ourmeasures of electoral competition in the 1980s, we are able to focus exclusively onthe extent of political competition in the 1980s that is driven by social and politicaldiscontent in the 1960s. Panel D of Table 9 reports the first stage. The coefficientsfor the instrument are positive and statistically significant for our three measures ofcompetition confirming that, as expected, social and political discontent during the1960s accounts for political competition during the 1980s. The partial F-statistic forour two measures of political competition that capture significant variation acrossmunicipalities—voter dispersion and opposition vote share—is comfortably above 20,which suggests a strong first stage.40

Columns 1-3 of Panel A report the instrumental variable (IV) estimates for ourbaseline measure of the ejido distance from the municipality head. These IV resultsare qualitatively similar to the baseline OLS results. However, the point estimates aresomewhat larger. For instance, IV results in column 1 suggest that a one-standard-deviation increase in vote dispersion led to a 5.35 km increase in the distance of thenewly allocated ejidos from their municipality after 1960, representing a slight increasefrom the 2.83 km suggested by the OLS estimates.41

39In Figure A-4, we present a graphical analysis analogous to that of Figure 4, plotting the coefficients for the interactionsbetween discontent and quinquennial dummies in a regression for ejidal distance to municipality heads. Just as when we use the1980s competition measures, prior to 1960 the distance from the allocated ejidos to their municipality heads trended together inplaces with high or low protest incidence. However, starting in the 1960s, they diverge and distance increases in municipalitiesfacing more discontent.

40The instrument is somewhat weak when using the indicator that the opposition has ever won in the municipality (column3), a measure of political competition that does not vary as much across municipalities.

41Due to the weak instrument problem anticipated earlier, the IV estimates in column 3 are sizable and a lot larger than thecorresponding OLS estimates.

30

The reduced-form estimates in column 4 of Panel A, which address possible concernsregarding the violation of the exclusion restriction of our instrumental variable strategy(i.e., discontent might have affected the location of newly allocated ejidos through analternative mechanism other than its impact on political contestation), also indicate apositive and significant effect of social and political discontent on the distance of thenewly allocated ejidos from their municipality after 1960. A one-standard-deviationincrease in (log) number of related events led to a 2.21 km increase in distance. PanelsB and C of Table 9, respectively, indicate that the IV estimates effects are similar whenaccounting for elevation terrain profile or the use of roads to reach municipality headswhen computing the ejido distance from the municipality head. The results in AppendixTables A-7 and A-8 further confirm that the set of robustness checks we consider forthe baseline OLS regressions lead to similar conclusions when also implementing thisIV approach.

Lastly, we deal with the potential concern that our instrument is subject to reversecausality since the allocation of distant ejidos causes instead discontent. While we findthis possibility unlikely a priori since the estimated effects are persistent but protestsare concentrated in the 1960s, we deal with this concern by replicating our analysisexcluding rural events of discontent in Appendix Table A-9. Results are very similar toour baseline findings in Table 9. Altogether, these findings then suggest that our resultsare not reflecting endogeneity in our measures of electoral competition.

In short, the baseline results and robustness checks conducted so far suggest that theestimates of the effect of political competition on the distance of the allocated ejidos totheir municipality head are not explained by differential pre-trends, omitted variablesthat correlate with competition, mean reversion or ceiling effects, by patterns specific toa few states, or by endogeneity in our measures of electoral competition. While thisis reassuring, in the next section we present additional exercises that deal with a fewalternative mechanisms that could be driving our estimates of interest.

6 Ruling out alternative interpretations

In this section, we deal with a few alternative mechanisms that could explain ourresults. We focus on mechanisms, other than an effort to weaken the local bureaucraticstate capacity as a strategic response to increased electoral contestation, that could alsoexplain our main empirical results. First, municipalities where the PRI faced strongerpolitical competition may be those in which the individuals who had to give away theland for redistribution resisted the most. Given this resistance, the PRI could not grantcommunities of petitioners land allocations close to the municipality head, and thushad to allocate land farther away. Sinkler (2014) effectively argues that the state issued

31

fewer land grants to peasants where commercial farmers were more powerful.Another possibility is that stronger political competition led the PRI to increase land

grants to appease opposition, which in turn forced the PRI to allocate more-marginal,lower-quality land located farther from municipality heads. Moreover, land qualitycould have a direct impact on the delivery of public goods and services, which couldalso explain the findings in Table A-1, which we interpret as evidence that the ejidosdistance to the municipality head is an important driver of the cost of providing publicgoods, and thus of local bureaucratic state capacity.

Another alternative interpretation of our results is that they are driven by the factthat the PRI dealt with potential insurgents by relocating them to more isolated areasthrough the allocation of ejidos. As Campante et al. (2017) suggest, the isolation of thosewho oppose incumbent regimes increases their mobilization cost, and thus reducesthe likelihood that they will show discontent or organize to challenge the regime. Arelated but somewhat more far-fetched alternative interpretation is that the PRI sentcommunities of insurgents far from municipality heads as punishment. In any case,these two last alternative interpretations suggest that the PRI should have allocatedejidos farther from municipality heads in municipalities where it faced more electoralcompetition after the 1960s.

A related additional alternative interpretation is that the PRI relocated insurgentsto more isolated areas to prevent future citizen monitoring. Stasvage (2010) arguesthat politically compact European polities were more likely to develop representativeassemblies than those with more dispersed constituents, since this allowed represen-tatives to gather more easily and citizens to monitor them more effectively. Stasvage(2010)’s argument then raises the question of whether our results could reflect a differ-ent strategic choice by the PRI: rather than locating ejidos far away to increase the costof providing public goods, they did it to undermine representative institutions and, inparticular, to hamper effective citizen monitoring.

All of the above alternative interpretations could also explain the increase in thedistance of the allocated ejidos from their municipality heads in politically contestedmunicipalities after the 1960s as the PRI’s response to political contestation. However,none of these interpretations considers that such an effect captures the PRI’s incentiveto weaken local bureaucratic state capacity in order to retain its comparative electoraladvantage. We next present a series of additional exercises to rule out these alternativeinterpretations and provide further support for our preferred interpretation. We end byreviewing and dismissing potential effects that distance of ejidos from municipal headsmight have on other dimensions of state capacity beyond the bureaucratic capabilitieswe focus on.

32

6.1 Strength of local elites

We first consider the possibility that higher resistance of landed elites in municipalitieswith stronger electoral competition could explain our results. Specifically, the concernis that the interaction between competition and the post-1960 indicator instead capturesthe omitted interaction between the strength of local rural elites and the post-1960indicator. To rule out this alternative interpretation of our findings, in Panel B of Table 8we control for the number of large landholdings in each municipality and its interactionwith the post-1960 indicator.

The estimates in Panel B of Table 8 suggest that, in municipalities with more ranchosand haciendas, ejidos were allocated farther from their municipality heads after the1960s. However, the size of the estimates is small and statistically insignificant. Moreimportantly, including the interaction of the municipal number of large landholdingsand the post-1960 indicator does not alter our coefficients of interest. Their size andstatistical significance are in line with our baseline results in Table 2. These estimatestherefore dismiss the concern that our findings are driven by stronger local elites inmunicipalities with more political competition.

6.2 Appeasing the opposition

Next, we deal with the alternative interpretation that our findings are explained bythe PRI placating the opposition through ejido allocations, which also forced it toallocate lower-quality land. To that end, we first test whether increased competitioneffectively led to the allocation of more ejidos after the 1960s. As we have emphasizedthroughout, this is also an important robustness check to verify that the PRI did notjust counteract a weakening of its clientelistic machine by simply creating more clients.Table 10 reports the results of regressions of several measures of ejido allocation onthe interaction of a post-1960 indicator and our various political competition variables.We use municipality-year as the unit of observation and measure ejido allocation indifferent ways. In particular, in Panel A we consider the number of allocated ejidos, inPanel B the number of beneficiaries, and in Panel C the total area granted. Throughoutthe specifications in Table 10 we find no support for an effect on ejido allocations. Theestimated effects on the interactions are inconsistently signed, insignificant, and oftensmall.

To further address the empirical support for this alternative interpretation, andspecifically that an increase in land allocations to appease opposition forced the PRI toallocate more-marginal, lower-quality land, we conduct our baseline specification butinstead use different measures of land quality as dependent variables. Panel A of Table11 employs an indicator that the granted land presents few constraints on agriculture,constructed using a seven-category measure of agricultural constraints from the FAO.

33

Panel B considers a nine-level index of inherent land quality from the US Departmentof Agriculture (transformed so that higher values indicate higher land quality). Theestimates across all specifications in all panels of Table 11 indicate that the land qualityof the ejidos allocated in competitive municipalities after the 1960s was not worse. Theestimates are small and statistically insignificant.

The findings of this exercise therefore suggest that our estimates do not reflect thatpolitical contestation led the PRI to increase the amount of allocated land in order toplacate the opposition, which then generated an allocation of more-marginal, lower-quality land farther from the municipality heads. Lastly, note that the absence of aneffect on land quality constitutes additional evidence that our estimates are unlikely tobe driven by the presence of stronger local landed elites in more contested municipalitieswho forced the PRI to allocate lower-quality ejidos farther from municipality heads afterthe 1960s.

6.3 Isolating insurgents and potential opposition

Finally, we assess if our findings are explained by the PRI’s relocation of potentialinsurgents to increase their costs of mobilization and to punish those who protestedagainst the regime. We also examine the related interpretation that the PRI relocatedcitizens to undermine their ability to monitor political representatives. First, notice thatif the PRI used the allocation of ejidos to isolate citizens (particularly the opposition), wewould expect that political contestation after the 1960s affected not only the distance ofthe allocated ejidos from their municipality head, but also the number of allocated ejidos.Particularly, we would expect the distribution of more ejidos, and for these to be fartherfrom municipal heads.42 Similarly, we would expect those who protested against theregime to receive more-marginal, lower-quality lands. However, as explained in detailabove, the results in Tables 10 and 11 provide no evidence of such effects on ejidoallocations.

Second, to further address the alternative interpretation that the PRI used ejido allo-cation to increase the monitoring and mobilization cost of citizens and the opposition,we test whether our estimates are larger in areas where either the monitoring capacityor the threat of insurgency and popular uprising was larger. In particular, we focus onmunicipalities with more social capital (measured by increased organizational capacity),higher population density, and more populous municipality heads. Accounting forsocial capital recognizes that some areas have enhanced organizational capabilities andindividuals with greater willingness to participate in political protests or uprisings(Moseley, 2015; Paxton, 2002; Schussman & Soule, 2005; Somma, 2010). Urbanizationand population density may also facilitate uprisings and revolution, by improving

42It is unlikely that the PRI allocated less land to punish rural insurgents who were already organizing against the PRIgovernment, and that among their demands, they were asking for more land redistribution.

34

coordination and enhancing the power of organized action, while making it harderfor the regime to disperse protesters and individually identifying them (DiPasquale &Glaeser, 1998; Glaeser & Steinberg, 2017; Wallace, 2014).

Table 12 reports the results of various specifications that estimate heterogeneouseffects by different proxies for the increased threat of insurgency or protests in a givenmunicipality. In Panel A, we include a triple interaction of the post-1960 indicatorvariable, each competition measure and social capital. Social capital is calculated asthe first principal component of the municipality’s number of human rights organiza-tions, popular fronts and peasant organizations in 1994.43 In Panel B, we consider atriple interaction with the municipality’s population density in 1960 instead of socialcapital. Lastly, in Panel C we focus on the triple interaction with the population of themunicipality head in 1960. All the estimates in Panels A and B are negative, small andstatistically insignificant. The estimates in Panel C are positive, but only statisticallysignificant for one of the electoral competition variables. More importantly, whilesomewhat smaller relative to the result in Table 3, the mean effect of the interactionsbetween the post-1960 indicator and the political competition variables remains largeand statistically significant.44

Overall, the estimates in Tables 10, 11 and 12 do not support that the allocation ofejidos far from municipality heads was intended to isolate individuals who opposed thePRI government or were likely to mobilize against it. As such, these results dismiss thealternative explanation that our findings can be explained by the PRI’s allocation ofejidos to isolate opposition and potential insurgents or protesters.

6.4 Alternative state-capacity interpretations of our distance measure

We consider the distance of ejidos from the municipality head as a measure of bureau-cratic local state capacity. However, there might be other dimensions of local statecapacity also affected by such distance. We argue that the PRI’s decisions regardingdistance of ejidos from their municipal head did not have comparable direct effects onother dimensions of local state capacity.

While taxation is a central aspect of state capacity, the distance of ejidos from theirmunicipal head did not affect the ability of the state to tax their inhabitants. Due to thelack of individual property rights over the granted land, which impeded the collectionof property taxes, the land allocation program made it much more difficult for theMexican state to generate revenues from peasants (Torres-Mazuera, 2009). However,this effect was independent from where the allocated ejidos were located. Instead, itdepended on the amount of land granted, and perhaps its quality, reflecting the extent

43The first principal component explains 70% of the variance in the data.44In all specifications of Table 12, we demean the measures of competition, social capital, population density in 1960 and

population in the municipality head in 1960 so that the double interactions can be interpreted as the corresponding effects at themean.

35

of the loss in tax base. Yet the evidence from Tables 10 and 11 indicates that neither thelevel nor the quality of allocated land responded to political contestation.

Others have emphasized the importance of the physical reach of the state as a fun-damental aspect of its capacity. For instance, Turner (1920), emphasizes the importanceof occupation of the frontier for the building of the American state. Simlarly, Scott(2009) argues that mountainous people in South-East Asia remain outside the reachof national governments, and Herbst (2000) highlights the limited geographic reachof many African states beyond their capital city and nearby surrounding areas. Thus,the allocation of ejidos farther away from municipality heads could be perceived asan attempt of the PRI to strengthen rather than weaken the capacity of the state, andspecifically its coercive reach. Nevertheless, this ignores both the process of state build-ing in Mexico and Latin American (García-Jimeno & Robinson, 2011), and contradictsthe basic patterns observed in our data. First, the Mexican state had its whole territoryunder control by the end of Lazaro Cárdenas’ presidency in 1940 (Sánchez Talanquer,2017). Second, our estimates in Appendix Table A-1 suggest that the allocation of landaway from populated areas did not lead to increased local state presence but that hadthe opposite effect.

To sum up, the exercises we conduct in this section disprove the most likely alter-native interpretations of our results. In particular, they provide evidence against thepossibility that our results are explained by alternative rationalizations that highlightthe political contestation against the PRI after the 1960s as an important driver of theincrease in the distance of the allocated ejidos from their municipality heads, but wherethe PRI’s incentives to weaken local bureaucratic state capacity play no role. They alsosuggest that other key dimensions of state capacity where not comparably affectedwith this strategic choice by the PRI. Taken together, our results suggest that the PRIdistributed ejidos farther from municipality heads where it faced stronger politicalopposition after the 1960 to deliberately weaken local bureaucratic state capacity inorder to retain its hegemony.

7 Conclusions

Although state capacity is central to economic and financial development, as well as topolitical stability and democracy, we still lack a definitive understanding of its determi-nants. A key observation in the recent literature is that, despite its benefits, investmentin state capacity cannot be taken for granted, because political incentives often pushpolitical elites to forestall, rather than encourage, a stronger state. In this paper, weexamine one such instance in the context of political clientelism. Since bureaucraticstate capacity is a key determinant of the cost of public goods provision, investments

36

in this area undermine the comparative advantage of incumbent clientelistic parties,which then have incentives to prevent state capacity strengthening in areas where theirdominant political position might be threatened.

We present a simple model capturing this mechanism and test its implications usingdata from a unique policy program in Mexico. In line with the theoretical predictions,our empirical evidence suggests that the PRI forestalled local bureaucratic state capacityby allocating communal lands far from municipality heads in areas where it expectedstronger political competition. Our estimates survive a series of robustness checks, andwe are able to rule out the most plausible alternative explanations.

In addition to helping explain the determinants of state capacity choices in contextswhere other theories fall short, our study also unveils the potentially perverse effectof political competition on economic development. In contrast to most conventionaltheories on the impact of stronger political competition—and its effect on politicalaccountability and economic development—we find that, in areas where clientelismis prevalent, more electoral competition may deter state capacity strengthening, andthus economic development. While existing work highlights the benefits of politicalcompetition for public good provision, and more generally for economic development(Acemoglu et al., 2017; Besley et al., 2010; Galasso & Nannicini, 2011; Naidu, 2017;Nath, 2015; Solé-Ollé & Viladecans-Marsal, 2012), we argue that incumbent clientelisticparties may respond to increased political competition by forestalling local bureaucraticstate capacity and, consequently public goods provision.

37

References

Acemoglu, D. (2005). Politics and economics in weak and strong states. Journal ofMonetary Economics, 52, 1199-1226.

Acemoglu, D., Moscona, J., & Robinson, J. A. (2016). State capacity and americantechnology: Evidence from the nineteenth century. American Economic Review:Papers & Proceedings, 106(61–67), 601–619.

Acemoglu, D., Naidu, S., Restrepo, P., & Robinson, J. (2017). Democracy does causegrowth. Forthcoming at the Journal of Political Economy.

Acemoglu, D., Robinson, J. A., & Santos, R. J. (2013, 01). The monopoly of violence:Evidence from Colombia. Journal of the European Economic Association, 11, 5-44.

Albertus, M., Diaz-Cayeros, A., Magaloni, B., & Weingast, B. R. (2016). Authoritariansurvival and poverty traps: Land reform in Mexico. World Development, 77(C),154–170.

Baja California. (2003). En dónde estamos. Mexicali, Baja California: Periódico Oficial delEstado de Baja California, CX(18), 10–52.

Baland, J.-M., & Robinson, J. A. (2008). Land and power: Theory and evidence fromchile. American Economic Review, 98(5), 1737–1765.

Baland, J.-M., & Robinson, J. A. (2012). The political value of land: Political reform andland prices in chile. American Journal of Political Science, 56(3), 601–619.

Bartra, A. (1985). Los herederos de Zapata: Movimientos campesinos posrevolucionarios enMéxico, 1920-1980. México: Editorial ERA.

Bellingeri, M. (2003). Del agrarismo armado a la guerra de los pobres, 1940-1974. México: ElColegio de México, Centro de Estudios Internacionales.

Besley, T., & Persson, T. (2000). The origins of state capacity: Property rights, taxation,and politics. American Economic Review, 99(4), 1218–1244.

Besley, T., & Persson, T. (2010). State capacity, conflict, and development. Econometrica,78(1), 1-34.

Besley, T., Persson, T., & Sturm, D. (2010). Political competition, policy and growth:Theory and evidence from the united states. Review of Economic Studies, 77(4),1329–1352.

Bezdek, R. (1973). Electoral oppositions in Mexico: Emergence, suppression, andimpact on political processes. PhD Dissertation, Ohio State University.

Bobonis, G., Gertler, P., Gonzalez-Navarro, M., & Nichter, S. (2017). Vulnerability andclientelism.

Bowles, J., Larreguy, H., & Liu, S. (2017). How weakly institutionalized parties monitorbrokers in developing democracies: Evidence from post-conflict liberia.

Campante, F. R., Do, Q.-A., & Guimaraes, B. (2017). Capital cities, conflict, and misgover-nance.

38

Centeno, M. A. (1997). Blood and debt: War and taxation in nineteenth-century LatinAmerica. American Journal of Sociology, 102(6), 1565-1605.

de Janvry, A., Emerick, K., Gonzalez-Navarro, M., & Sadoulet, E. (2015). Delinking landrights from land use: Certification and migration in Mexico. American EconomicReview, 105(10), 3125–3149.

Dell, M. (2012). Path dependence in development: Evidence from the Mexican revolu-tion.

Dell, M., Lane, N., & Querubin, P. (in press). The historical state, local collective action,and economic development in vietnam. Econometrica.

Diaz-Cayeros, A., Estevez, F., & Magaloni, B. (2016). The political logic of poverty relief:electoral strategies and social policy in Mexico. Cambridge: Cambridge UniversityPress.

DiPasquale, D., & Glaeser, E. L. (1998). The los angeles riot and the economics of urbanunrest. Journal of Urban Economics, 43(1), 52–78.

Eckstein, S. (1968). El marco macroeconómico del problema agrario mexicano. México: Centrode Investigaciones Agrarias.

Fearon, J. D., & Laitin, D. D. (2003). Ethnicity, insurgency, and civil war. AmericanPolitical Science Review, 97(1), 75-90.

Fergusson, L. (2013). The political economy of rural property rights and the persistenceof the dual economy. Journal of Development Economics, 103(C), 167-181.

Fergusson, L., Robinson, J. A., Torvik, R., & Vargas, J. F. (2014). The need for enemies.The Economic Journal.

Finan, F., & Schechter, L. (2012). Vote-Buying and Reciprocity. Econometrica, 80(2),863–881.

Fujiwara, T., & Wantchekon, L. (2013). Can informed public deliberation overcomeclientelism? experimental evidence from benin. American Economic Journal: AppliedEconomics, 5(4), 241–255.

Galasso, V., & Nannicini, T. (2011). Competing on good politicians. American PoliticalScience Review, 105(1), 79–99.

García-Jimeno, C., & Robinson, J. A. (2011). The myth of the frontier. In D. Costa &N. Lamoreaux (Eds.), Understanding long-run economic growth: Geography, institu-tions, and the knowledge economy. University of Chicago Press.

Geddes, B. (1994). Politician’s dilemma: Building state capacity in Latin America. Berkeley:University of California Press.

Gennaioli, N., & Voth, H.-J. (2015). State capacity and military confict. The Review ofEconomic Studies, 82(4), 1409–1448.

Glaeser, E. L., & Steinberg, B. M. (2017). Transforming cities: does urbanization promotedemocratic change? Regional Studies, 51(1), 58–68.

Henn, S. J. (2018). Complements or substitutes: How institutional choices shape the

39

interaction between traditional leaders and the state in africa.Herbst, J. (2000). States and power in Africa: Comparative lessons in authority and control.

Princeton University Press.Heredia, B., & Schneider, B. R. (2003). The political economy of administrative reform in

developing countries. In Schneider and Heredia, eds., Reinventing Leviathan. Miami:North-South Center Press.

Herrera Calderón, F., & Cedillo, A. (2012). Challenging authoritarianism in Mexico:Revolutionary struggles and the dirty war, 1964-1982. New York : Routledge.

Hicken, A. (2011). Clientelism. Annual Review of Political Science, 14, 289–310.Hicken, A., & Simmons, J. W. (2008). The Personal Vote and the Efficacy of Education

Spending. American Journal of Political Science, 52(1), 109–124.Keefer, P. (2007). Clientelism, Credibility, and the Policy Choices of Young Democracies.

American Journal of Political Science, 51(4), 804–821.Keefer, P., & Vlaicu, R. (2008). Democracy, Credibility, and Clientelism. Journal of Law,

Economics, and Organization, 24(2), 371–406.Kitschelt, H. (2000). Linkages between citizens and politicians in democratic politics.

Comparative Political Studies, 33(6-7), 845-79.Klesner, J. L. (1993). Modernization, economic crisis, and electoral alignment in Mexico.

Mexican Studies, 9(2), 187–223.Larreguy, H. (2013). Monitoring political brokers: Evidence from clientelistic networks in

Mexico.Larreguy, H., Marx, B., Reid, O., & Blattman, C. (2016). A market equilibrium approach

to reduce the incidence of vote-buying: Evidence from uganda. Working Paper.Lindbeck, A., & Weibull, J. W. (1987). Balanced-budget redistribution as the outcome of

political competition. Public Choice, 52(3), 273-297.Lujambio, A. (2001). Democratization through federalism? The national action party

strategy, 1939-2000. In Kevin Middlebrook, ed., Party Politics and the Struggle forDemocracy in Mexico. La Jolla: University of California, San Diego..

Magaloni, B. (2006). Voting for autocracy: Hegemonic party survival and its demise in mexico.New York: Cambridge University Press.

Mexico. (2007). Plan nacional de desarrolo 2007-2012. Diario Oficial de la Federación.Michalopoulos, S., & Papaioannou, E. (2013). Pre-colonial ethnic institutions and

contemporary african development. Econometrica, 81(1), 113–152.Michalopoulos, S., & Papaioannou, E. (2014). National institutions and subnational

development in africa. Quarterly Journal of Economics, 129(1), 151–213.Molinar, J., & Weldon, J. (1990). Elecciones de 1988 en méxico: Crisis del autoritarismo.

Revista Mexicana de Sociología, 52(4), 229-262.Moseley, M. W. (2015). Contentious engagement: Understanding protest participation

in latin american democracies. Journal of Politics in Latin America, 7(3), 3–48.

40

Naidu, S. (2017). Suffrage, schooling, and sorting in the post-bellum u.s. south. NBERWorking Paper # 18129.

Nath, A. (2015). Bureaucrats and politicians: How does electoral competition affect bureaucraticperformance?

Paxton, P. (2002). Social capital and democracy: An interdependent relationship.American Sociological Review, 254–277.

Peiro, I. (1998). México: crisis, reestructuración económica, social y política, 1982-1996. SigloXXI Editores.

Robinson, J. A., Torvik, R., & Verdier, T. (2006). Political foundations of the resourcecurse. Journal of Development Economics, 79(2), 447 - 468.

Robinson, J. A., & Verdier, T. (2013). The political economy of clientelism. ScandinavianJournal of Economics, 115(2), 260-291.

Sabloff, P. L. (1981). Caciquismo in post-revolutionary Mexican ejido-grant communities.Albuquerque, New Mexico: Latin American Institute, University of New Mexico.

Sanchez de la Sierra, R. (2017). On the origins of the state: Stationary bandits and taxationin eastern congo.

Sánchez Talanquer, M. (2017). Historical cleavages and the rule of law: Evidence frommexico.

Sanderson, S. (1984). Land reform in Mexico: 1910-1980. Academic Press.Sanderson, S. (1986). The transformation of Mexican agriculture. Princeton, NJ: Princeton

University Press.Schussman, A., & Soule, S. A. (2005). Process and protest: Accounting for individual

protest participation. Social forces, 84(2), 1083–1108.Scott, J. C. (2009). The Art of Not Being Governed. New Haven, Connecticut: Yale

University Press.Secretaría de Gobernación. (1994). Directorio de organizaciones civiles. México:

Secretaría de Gobernación, Subsecretaría de Desarrollo Político, Dirección General deDesarrollo Político, 1994.

SEDESOL. (2014). Secretaría de desarrollo social. Diagnóstico del Programa para elDesarrollo de Zonas Prioritarias.

Shefter, M. (1994). Political parties and the state: The American historical experience.Princeton: Princeton University Press.

Silva Herzog, J. (1959). El agrarismo mexicano y la reforma agraria. México: Fondo deCultura Económica.

Sinkler, A. (2014). Land without opportunity: Public goods investment and landmarkets in rural Mexico. Comparative Political Studies, 47(11), 1611-1636.

Soifer, H. D. (2015). State building in latin america. New York: Cambridge UniversityPress.

Solé-Ollé, A., & Viladecans-Marsal, E. (2012). Lobbying, political competition, and

41

local land supply: Recent evidence from spain. Journal of Public Economics, 96(1-2),10–19.

Somma, N. M. (2010). How do voluntary organizations foster protest? the role oforganizational involvement on individual protest participation. The SociologicalQuarterly, 51(3), 384–407.

Stasvage, D. (2010, 11). When distance mattered: Geographic scale and the developmentof European representative assemblies. American Political Science Review, 104, 625–643.

Stokes, S. (2005). Perverse accountability: A formal model of machine politics withevidence from Argentina. American Political Science Review, 99(3), 315–25.

Stokes, S. (2007). Oxford handbook of comparative politics. In C. Boix & S. Stokes(Eds.), (p. 604-27). Oxford: Oxford University Press.

Suryanarayan, P. (2017). Hollowing out the state: Franchise expansion and fiscalcapacity in colonial india.

Thies, C. G. (2005). War, rivalry, and state building in Latin America. American Journalof Political Science, 49(3), 451-465.

Thies, C. G. (2007). The political economy of state building in Sub-Saharan Africa.Journal of Politics, 69(3), 716-731.

Tilly, C. (1992). Coercion, capital and European states: AD 990 - 1992. Wiley.Torres-Mazuera, G. (2009). La territorialidad rural mexicana en un contexto de descen-

tralización y competencia electoral. Revista Mexicana de Sociología, 71(3), 453–490.Turner, F. J. (1920). The frontier in American history. New York: H. Holt and Co.Wallace, J. (2014). Cities and stability: Urbanization, redistribution, and regime survival in

china. Oxford: Oxford University Press.Yates, P. L. (1981). Mexico’s agricultural dilemma. University of Arizona Press.

42

Table 1: Summary statistics

StandardMean deviation Observations

A. Public goodsa. Census in 2000Share of households in locality with...- Piped water 0.455 0.407 107,218- Drainage 0.282 0.322 107,218- Electricity 0.674 0.391 107,218b. Census in 1990Share of households in locality with...- Piped water 0.316 0.375 97,484- Drainage 0.131 0.229 97,484- Electricity 0.423 0.422 97,484

B. Bureaucratic state capacityVarying by locality:-Distance of locality to municipality head (km) 19.152 21.604 199,391-Distance of locality from municipality head accounting for terrain elevation profile (km) 19.219 22.023 199391-Distance of locality from municipality head (km) via DCW roads 21.582 23.406 199391

Varying by ejido:-Distance of ejido to municipality head (km) 18.569 21.215 17,338-Distance of ejido from municipality head accounting for terrain elevation profile (km) 18.614 21.140 17,338-Distance of ejido from municipality head via DCW roads (km) 20.953 22.136 17,338

C. Municipal political competitionAverage of 1980s elections:-Vote dispersion 0.206 0.165 2,023-Opposition vote share 0.159 0.140 2,023- Vote share friendly opposition 0.026 0.060 2,023- Vote share unfriendly opposition 0.133 0.131 2,023

-Opposition ever won 0.114 0.318 2,023First election of 1980s:-Vote dispersion 0.146 0.194 2,023-Opposition vote share 0.115 0.167 2,023-Opposition won 0.019 0.136 2,023

D. Municipal geographical covariatesLog of municipality area (km2) 5.526 1.492 2,437Log of population in 1900 7.885 1.125 2,295Average monthly rainfall (mm) 90.62 51.987 2,437Rain variability (Standard deviation of monthly rainfall) 78.051 40.352 2,437Average soil humidity (Days) 197.406 83.098 2456Soil humidity variability (Standard deviation of soil humidity) 34.231 30.248 2,456Average altitude (m) 1,438.143 876.307 2,456Ruggedness (Standard deviation of altitude) 255.643 189.214 2,456

E. Ejido land qualityAgricultural constraints (FAO) 0.181 0.376 22,816Inherent land quality index (U.S. Department of Agriculture) 4.706 2.586 22,940

F. Variables for robustness checks

Varying by municipality and year:-Number of allocated ejidos 0.141 0.791 164,715-Stock of allocated ejidos 6.109 10.641 164,715-Number of beneficiaries of ejidos 13.477 88.551 164,715-Area granted in ejidos (m2) 375.437 6,935.555 164,715-Land grant potential (1,000 km2) 3.696 11.939 164,636

Varying by municipality:- Number of ranchos and haciendas 47.033 90.628 2,455- Social capital in 1994 (Principal component) 0 1.445 2,455- Population density in 1960 (people/km2) 64.573 345.753 2,389- Population in the municipality head in 1960 (people) 5,723.717 24,873.226 2,389- Log of number of events of social and political discontent, 1960-1969 0.388 0.764 2,412

Notes: Opposition ever won = 1 if opposition won at least one election in the 1980s. Opposition vote share = 1− PRI vote share. Vote dispersion = 1−∑i=1 p2i , where pi is the vote share of each

of the parties in the considered election. Agricultural constraints is an indicator that the land presents few constraints for agriculture. The inherent land quality index varies from 1 (low quality) to9 (high quality). Social capital in 1994 is the first principal component of the number of human rights organizations, popular fronts and peasants. The land available is calculated as the potentialagricultural land in 2007 minus the stock of allocated ejidos by year. The number of events reflecting social and political discontent are counted during the period 1960-1969 using references torelated events in two Mexican newspapers with national coverage: El Universal and Excelsior.

43

Table 2: Distance of new land granted and political competition:Baseline results for different distance measures

Baseline results, ejidos allocated from 1914 to 1992(1) (2) (3)

Competition Votedispersion

Oppositionvote share

Oppositionever won

Panel A: Dependent variable: Distance of ejido from municipality head

Post-1960 × Competition 2.83** 3.25** 3.16**(1.17) (1.41) (1.52)

R-squared 0.58 0.58 0.58

Panel B: Dependent variable: Distance of ejido from municipality head accounting forthe terrain elevation profile

Post 1960 × Competition 2.98** 3.40** 3.33**(1.28) (1.55) (1.66)

R-squared 0.54 0.55 0.55

Panel C: Dependent variable: Distance of ejido from municipality head via DCW roads

Post-1960 × Competition 3.04** 3.44** 3.41**(1.31) (1.58) (1.70)

R-squared 0.55 0.55 0.55

Observations 17,338 17,338 17,338

Notes: Robust standard errors in parentheses are clustered at the municipality level, *** p<0.01, ** p<0.05, * p<0.1.Regressions are at the ejido level. All specifications include municipality and presidential-term fixed effects. Compe-tition refers to political competition measured at the municipality level using the variable indicated in each column(see the notes to Table 1 and the main text for exact definitions). All competition measures are standardized. Distanceof ejido from municipality head in panel A refers to the population-weighted minimum Euclidean distance of the ejidolocalities from the municipality head (See Appendix Figure A-1 for details). The distance of ejido from municipalityhead in panel B accounts for terrain by penalizing the minimum Euclidean distance in panel A when there are changesin altitude in the straight path that connects the localities within the ejido and their municipality head (See AppendixFigure A-1 for details). The distance from the municipality head via DCW roads in panel C accounts for the use ofroads to reach the municipality head. The trace of those roads comes from the Digital Chart of the World of 1992 andthe overall distance of each locality from its municipality head is computed adding up two different figures. First,the Euclidean distance from the locality to the closest point in a road that leads to the municipality head, and second,the length of the segment that connects such point to the municipality head following the road path (See AppendixFigure A-1 for details).

44

Table 3: Distance of new land granted and political competition:Baseline results, test for parallel trends, and falsification exercise

Dependent variable: Distance of ejido from municipality head(1) (2) (3)

Competition Votedispersion

Oppositionvote share

Oppositionever won

Panel A: Baseline results, ejidos allocated from 1914 to 1992

Post-1960 × Competition 2.83** 3.25** 3.16**(1.17) (1.41) (1.52)

Observations 17,338 17,338 17,338R-squared 0.58 0.58 0.58

Panel B: Ruling out anticipation, ejidos allocated from 1914 to 1992

Post-1960 × Competition 2.61* 3.02* 1.90(1.44) (1.77) (1.78)

1(Year + 5 ≥ 1960) × Competition -0.68 -0.86 0.70(0.92) (1.01) (1.06)

1(Year + 10 ≥ 1960) × Competition 0.74 0.80 0.02(0.93) (1.04) (0.82)

1(Year + 15 ≥ 1960) × Competition 0.29 0.44 0.76(0.57) (0.65) (0.54)

Observations 17,338 17,338 17,338R-squared 0.58 0.58 0.58

Panel C: Falsification, ejidos allocated from 1914 to 1960, placebo 1935

Post-1935 × Competition 0.20 0.17 0.39(0.33) (0.32) (0.29)

Observations 12,575 12,575 12,575R-squared 0.57 0.57 0.57

Notes: Robust standard errors in parentheses are clustered at the municipality level, *** p<0.01, ** p<0.05, * p<0.1.Regressions are at the ejido level. All specifications include municipality and presidential-term fixed effects. Post-1960(1935) is a dummy variable that equals 1 if the ejido is granted after 1960 (1935). 1(Year +j ≥ 1960) is a dummy variablethat equals 1 if the allocation year plus j years is greater than 1960. Competition refers to political competition measuredat the municipality level using the variable indicated in each column (see the notes to Table 1 and the main text for exactdefinitions). All competition measures are standardized.

45

Table 4: Friendly and unfriendly opposition

Dependent variable: Distance of ejido from municipality head(1) (2) (3) (4)

Post-1960 × Vote share opposition 3.25**(1.41)

Post-1960 × Vote share friendly opposition 1.19** 1.45***(0.55) (0.53)

Post-1960 × Vote share unfriendly opposition 2.91* 3.04**(1.49) (1.50)

Observations 17,338 17,338 17,338 17,338R-squared 0.58 0.58 0.58 0.58

Test of inequality of coefficients in Column 4

Ho: βPost-1960 × Vote share unfriendly ≤ βPost-1960 × Vote share friendly p-valueHa: βPost-1960 × Vote share unfriendly > βPost-1960 × Vote share friendly 0.1457

Notes: Robust standard errors in parentheses are clustered at the municipality level, *** p<0.01, ** p<0.05, * p<0.1. Regressions are atthe ejido level. All specifications include municipality and presidential-term fixed effects. Post-1960 is a dummy variable that equals 1 ifthe ejido is granted after 1960. All vote shares are standardized. For the classification of friendly opposition, see Section 4.2 and AppendixTable A-2.

46

Table 5: Covariate balance

Dependent variable:Log of

municipalityarea

Log ofpopulation

in 1900

Averagemonthlyrainfall

Rainvariability

Averagesoil

humidity

Soilhumidityvariability

Averagealtitude

Ruggedness(altitude

variability)

AgriculturalConstraints

Inherentland

Qualityindex

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)

Vote dispersion 0.050 0.218*** -2.969* -2.866** 0.271 -1.080* -23.030 -20.071*** 0.017 0.127(0.042) (0.037) (1.474) (1.402) (3.343) (0.588) (23.590) (6.195) (0.015) (0.096)

R-squared 0.536 0.291 0.564 0.509 0.087 0.031 0.524 0.231 0.455 0.296

Opposition vote share 0.036 0.195*** -3.064** -2.862** -0.557 -1.155* -14.373 -17.530*** 0.012 0.123(0.039) (0.034) (1.453) (1.365) (3.211) (0.595) (23.931) (5.810) (0.015) (0.099)

R-squared 0.535 0.286 0.564 0.509 0.087 0.031 0.523 0.229 0.454 0.296

Opposition ever won -0.019 0.023 -1.919 -1.313 -0.196 -0.917 -6.366 -10.184** -0.010 0.007(0.024) (0.028) (1.242) (0.980) (2.545) (0.704) (15.660) (4.757) (0.006) (0.069)

R-squared 0.535 0.261 0.562 0.505 0.087 0.031 0.523 0.225 0.454 0.293

Observations 1,788 1,676 1,788 1,788 1,800 1,800 1,800 1,800 1,793 1,794Notes: Robust standard errors in parentheses are clustered at the state level, *** p<0.01, ** p<0.05, * p<0.1. Regressions are at the municipality level, with the dependent variable as indicated in each column title. All specificationsinclude state fixed effects. See the notes to Table 1 and the main text for exact definitions.

47

Table 6: Distance of new land granted and political competition:Controlling for differential trends based on municipal characteristics

Dependent variable: Distance of ejido from municipality head(1) (2) (3)

Competition Votedispersion

Oppositionvote share

Oppositionever won

Post 1960 × Competition 2.466** 2.919** 3.519***(1.030) (1.226) (1.302)

Additional controls:

Post 1960 × Log of municipality area 3.993*** 3.952*** 4.156***(0.952) (0.920) (0.937)

Post 1960 × Log of population in 1900 -0.841 -0.948 -0.512(0.677) (0.707) (0.544)

Post 1960 × Average monthly rainfall 0.035* 0.043* 0.034*(0.021) (0.023) (0.020)

Post 1960 × Rain variability -0.034 -0.040 -0.036(0.029) (0.031) (0.028)

Post 1960 × Average soil humidity 0.005 0.006 0.007(0.008) (0.008) (0.009)

Post 1960 × Soil humidity variability 0.008 0.009 0.011(0.017) (0.018) (0.016)

Post 1960 × Average altitude -0.001 -0.001 -0.001(0.001) (0.001) (0.001)

Post 1960 × Ruggedness 0.004 0.003 0.000(0.005) (0.005) (0.004)

Post 1960 × Agricultural constraints -0.943 -0.448 0.249(2.009) (1.924) (1.761)

Post 1960 × Inherent land quality index 0.428 0.442 0.522*(0.300) (0.302) (0.290)

Observations 16,222 16,222 16,222R-squared 0.586 0.587 0.588

Notes: Robust standard errors in parentheses are clustered at the municipality level, *** p<0.01, ** p<0.05, * p<0.1. Regressionsare at the ejido level. All specifications include municipality and presidential-term fixed effects. Post-1960 is a dummy variable thatequals 1 if the ejido is granted after 1960. Competition refers to political competition measured at the municipality level using thevariable indicated in each column (see the notes to Table 1 and the main text for exact definitions). All competition measures arestandardized. The number of observations changes relative to those in baseline regressions as some covariates are not availablefor all ejidos.

48

Table 7: Distance of new land granted and political competition:Controlling for the stock of allocated ejidos and the land available for redistribution

Dependent variable: Distance of ejido from municipality head(1) (2) (3)

Competition Votedispersion

Oppositionvote share

Oppositionever won

Panel A: Stock of allocated ejidos

Post-1960 × Competition 2.37* 2.83* 3.17**(1.22) (1.48) (1.47)

Stock of allocated ejidos 0.05 0.05 0.05(0.03) (0.03) (0.03)

Post-1960 × Stock of allocated ejidos 0.04 0.04 0.07**(0.03) (0.04) (0.03)

R-squared 0.58 0.58 0.58

Panel B: Land available for redistribution

Post-1960 × Competition 1.41* 1.61* 2.03***(0.81) (0.87) (0.65)

Land grant potential -0.53*** -0.53*** -0.54***(0.07) (0.07) (0.07)

Post-1960 × Land grant potential 0.19** 0.19** 0.19**(0.09) (0.09) (0.09)

R-squared 0.59 0.59 0.59

Observations 17,337 17,337 17,337Notes: Robust standard errors in parentheses are clustered at the municipality level, *** p<0.01, ** p<0.05, * p<0.1. All specificationsinclude municipality and presidential-term fixed effects. Post-1960 is a dummy variable that equals 1 if the ejido is granted after1960. Competition refers to political competition measured at the municipality level using the variable indicated in each column. Thestock of allocated ejidos is the sum of ejidos granted in the municipality since 1914 and up to one year before the allocation of theejido of interest. Land available for redistribution is the difference between total available agricultural land and the amount of landallocated since 1914 and up to one year before the allocation of the ejido of interest. See the notes to Table 1 and the main text for exactdefinitions. All competition measures are standardized. The number of observations changes relative to those in baseline regressionsas some covariates are not available for all ejidos. In particular, we lost one observation from the municipality of San Jacinto Amilpasin the state of Oaxaca, which does not have agricultural information in the 2007 agricultural census.

49

Table 8: Distance of new land granted and political competition:State-specific trends and strength of rural elites

Dependent variable: Distance of ejido from municipality head(1) (2) (3)

Competition Votedispersion

Oppositionvote share

Oppositionever won

Panel A: State-specific trends

Post-1960 × Competition 2.09*** 2.72*** 3.24***(0.63) (0.70) (0.94)

R-squared 0.59 0.59 0.59

Cubic state trends X X XPost-1960 × State indicator X X X

Panel B: Strength of rural elites

Post-1960 × Competition 2.36*** 2.69** 2.92**(0.91) (1.10) (1.29)

Post-1960 × Number of ranchos and haciendas 0.01 0.01 0.02(0.01) (0.01) (0.01)

R-squared 0.58 0.58 0.58

Observations 17,338 17,338 17,338Notes: Robust standard errors in parentheses are clustered at the municipality level, *** p<0.01, ** p<0.05, * p<0.1. Regressionsare at the ejido level. All specifications include municipality and presidential-term fixed effects. Post-1960 is a dummy variable thatequals 1 if the ejido is granted after 1960. Panel A includes cubic time trends interacted with state dummies and the interaction of eachstate dummy with the Post-1960 dummy. In Panel B, the number of ranchos and haciendas is the number of large landholdings, alsomeasured at the municipality level. Competition refers to political competition measured at the municipality level using the variableindicated in each column. See the notes to Table 1 and the main text for exact definitions. All competition measures are standardized.

50

Table 9: Distance of new land granted, political competition and discontent

(1) (2) (3) (4)

Competition Votedispersion

Oppositionvote share

Oppositionever won

Panel A: Dependent variable: Distance of ejido from municipality head

ReducedIV Estimates Form

Post1960 × Competition 5.35** 5.35** 14.36**(2.32) (2.33) (6.15)

Post1960 × Log of number of events of discontent 2.21**(0.99)

Panel B: Dependent variable: Distance of ejido from municipality head accounting for the terrain elevation profile

ReducedIV Estimates Form

Post 1960 × Competition 5.67** 5.66** 15.21**(2.51) (2.52) (6.68)

Post1960 × Log of number of events of discontent 2.34**(1.07)

Panel C: Dependent variable: Distance of ejido from municipality head via DCW roads

ReducedIV Estimates Form

Post 1960 × Competition 5.74** 5.74** 15.41**(2.58) (2.59) (6.79)

Post1960 × Log of number of events of discontent 2.37**(1.11)

Panel D: Dependent variable: Post1960 × Competition

First Stage

Post 1960 × Log of number events of discontent 0.41*** 0.41*** 0.15**(0.04) (0.06) (0.07)

R-squared 0.23 0.22 0.03Partial F-statistic 106.2 54.81 4.893

Observations 17,279 17,279 17,279 17,279Notes: Robust standard errors in parentheses are clustered at the municipality level, *** p<0.01, ** p<0.05, * p<0.1. Regressions are at the ejido level.All specifications include municipality and presidential-term fixed effects. Post-1960 is a dummy variable that equals 1 if the ejido is granted after 1960.Competition refers to political competition measured at the municipality level using the variable indicated in each column (see the notes to Table 1 andthe main text for exact definitions). The number of events reflecting social and political discontent protests are counted during the period 1960-1969 usingreferences to related events in two Mexican newspapers of national coverage: El Universal and Excelsior. All competition measures are standardized. Thenumber of observations changes relative to those in baseline regressions as information about protests is not available for all ejidos. See the note in Table 2and Figure A-1 for details on the computation of the various distances.

51

Table 10: Amount of new land granted and political competition:Is there an effect on the intensity of the land allocation program?

(1) (2) (3)

Competition Votedispersion

Oppositionvote share

Oppositionever won

Panel A: Dependent variable: Number of allocated ejidos

Post 1960 × Competition -0.001 0.001 0.002(0.003) (0.003) (0.003)

Observations 130,278 130,278 130,278R-squared 0.1011 0.1010 0.1011

Panel B: Dependent variable: Number of beneficiaries of ejidos

Post 1960 × Competition -0.07 0.15 0.20(0.27) (0.28) (0.22)

Observations 130,495 130,495 130,495R-squared 0.081 0.080 0.081

Panel C: Dependent variable: Area granted in ejidos

Post 1960 × Competition -4.45 -1.65 -1.17(4.98) (5.03) (4.13)

Observations 130,496 130,496 130,496R-squared 0.080 0.081 0.081

Notes: Robust standard errors in parentheses are clustered at the municipality level, *** p<0.01, ** p<0.05, * p<0.1. Regressions are atthe municipality-year level. All specifications include municipality and presidential-term fixed effects. Post-1960 is a dummy variablethat equals 1 after 1960, which is included in addition to the reported interaction term. Competition refers to political competitionmeasured at the municipality level using the variable indicated in each column. The regressions also control for the interaction ofPost-1960 with the host of population, geographic and climatic municipal controls in Table 5. See the notes to Table 1 and the maintext for exact definitions. All competition measures are standardized.

52

Table 11: Distance of new land granted and political competition:Is it land quality or distance?

(1) (2) (3)

Competition Votedispersion

Oppositionvote share

Oppositionever won

Panel A: Dependent variable: Agricultural constraints (FAO)

Post-1960 × Competition 0.003 0.005 0.007(0.006) (0.007) (0.007)

Observations 16,114 16,114 16,114R-squared 0.814 0.814 0.814

Dependent variable: Land quality index (U.S/ Department of Agriculture)

Post-1960 × Competition -0.003 0.010 -0.002(0.061) (0.062) (0.049)

Observations 16,181 16,181 16,181R-squared 0.778 0.778 0.778

Notes: Robust standard errors in parentheses are clustered at the municipality level, *** p<0.01, ** p<0.05, * p<0.1. Regressionsare at the ejido level. All specifications include municipality and presidential-term fixed effects. Post-1960 is a dummy variable thatequals 1 if the ejido is granted after 1960, which is included in addition to the reported interaction term. Competition refers to politicalcompetition measured at the municipality level using the variable indicated in each column. The dependent variable is the land qualityof each allocated ejido as measured using each of the variables in each panel title. The regressions also control for the interaction ofPost-1960 with the host of population, geographic and climatic municipal controls in Table 5. See the notes for Table 1 and the maintext for exact definitions. All competition measures are standardized. The number of observations changes relative to those in baselineregressions as some covariates are not available for all ejidos.

53

Table 12: Distance of new land granted and political competition:Adding social capital, population density and municipality head’s population

Dependent variable: Distance of ejido from municipality head(1) (2) (3)

CompetitionVote

dispersionOppositionvote share

Oppositionever won

Panel A: Social capital in 1994

Post-1960 × Competition 2.92** 3.57** 3.14*(1.36) (1.67) (1.64)

Post-1960 × Social capital in 1994 0.23 -0.07 0.13(0.94) (0.90) (0.35)

Post-1960 × Competition × Social capital in 1994 -0.39 -0.26 -0.03(0.60) (0.51) (0.35)

R-squared 0.58 0.58 0.58

Panel B: Population density in 1960

Post-1960 × Competition 3.30*** 3.58*** 3.22**(1.08) (1.26) (1.37)

Post-1960 × Population density in 1960 -0.09*** -0.09*** -0.08***(0.02) (0.02) (0.02)

Post-1960 × Competition × Population density in 1960 -0.01 -0.02 -0.02(0.02) (0.03) (0.03)

R-squared 0.58 0.58 0.58

Panel C: Population in the municipality head in 1960

Post-1960 × Competition 1.99* 2.39** 1.16(1.04) (1.18) (0.71)

Post-1960 × Population in the municipality head in 1960 0.87* 0.74 1.34***(0.46) (0.46) (0.44)

Post-1960 × Competition × Population in the municipal head in 1960 0.30 0.41 1.71**(0.59) (0.57) (0.82)

R-squared 0.58 0.58 0.58

Observations 17,338 17,338 17,338Notes: Robust standard errors in parentheses are clustered at the municipality level, *** p<0.01, ** p<0.05, * p<0.1. Regressions are at the ejido level. Allspecifications include municipality and presidential-term fixed effects. Post-1960 is a dummy variable that equals 1 if the ejido is granted after 1960. Competitionrefers to political competition measured at the municipality level using the variable indicated in each column. We demean the measures of competition, socialcapital, population density and population in the municipality head in 1960 so that the double interactions can be interpreted as the corresponding effects at themean. All competition measures are standardized.

54

Figure 1: Allocation of ejidos within two similar municipalities in Durango

Low Competition High CompetitionPr

e19

60

^

^

^

^

^

^

^

^

^

^

^

^

^^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^^

^

^

^

^

^

^

^^

^

^^

^

^

^

^^^

^^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^^

^

^ ^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^^

^

^

^

^

^

^

^^

^

^^

^

^

^

^

^^

^^

^

^

^

^

^

^^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^̂

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^^

^

^

^

^

^

^

^

^^^

^

^

^

^

^

^

^

^

^^

^^

^

^ ^

^

^̂

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^^^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^ ^

^

^^

^

^

^^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^^

^

^

^^

^

^

^^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^^

^

^

^

^^ ^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^̂

^

^^

^

^^

^^

^^

^

^

^

^

^

^

^

^^

^^

^

^

^

^

^^

^^

^

^

^

^

^

^

^^^

^̂

^

^

^ ^^

^^ ^^

^

^̂

^

^^

^̂

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^^

^

^^

^

^

^^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^^

^^

^

^

^

^

^

^

^

^

^^

^^

^

^

^

^^

^

^^

^ ^

^

^

^^

^

^

^

^

^

^

^^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^^

^^

Municipality Boundary^ Municipality Head

Ejido Allocated Pre 1960

^

^

^

^

^

^

^

^

^

^

^

^

^^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^^

^

^

^

^

^

^

^^

^

^^

^

^

^

^^^

^^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^^

^

^ ^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^^

^

^

^

^

^

^

^^

^

^^

^

^

^

^

^^

^^

^

^

^

^

^

^^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^^

^

^

^^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^̂

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^^

^

^

^

^

^^

^

^

^

^

^^

^

^

^

^

^

^

^

^^^

^

^

^

^

^

^

^

^

^^

^^

^

^ ^

^

^̂

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^^^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^ ^

^

^^

^

^

^^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^^

^

^

^^

^

^

^^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^^

^

^

^

^^ ^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^̂

^

^^

^

^^

^^

^^

^

^

^^

^

^

^

^^

^^

^

^

^

^

^^

^^

^

^

^

^

^

^

^^^

^̂

^

^

^ ^^

^^ ^^

^

^̂

^

^^

^̂

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^^

^

^^

^

^

^^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^^

^^

^

^

^

^

^

^

^

^

^^

^^

^

^

^

^^

^

^^

^ ^

^

^

^^

^

^

^

^

^

^

^^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^^

^^

Municipality Boundary^ Municipality Head

Ejido Allocated Pre 1960

Average Euclidean distance from municipality head: 32.64 km Average Euclidean distance from municipality head: 28.13 km

Post

1960

^

^

^

^

^

^

^

^

^

^

^

^

^^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^^

^

^

^

^

^

^

^^

^

^^

^

^

^

^^^

^^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^^

^

^ ^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^^

^

^

^

^

^

^

^^

^

^^

^

^

^

^

^^

^^

^

^

^

^

^

^^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^̂

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^^

^

^

^

^

^

^

^

^^^

^

^

^

^

^

^

^

^

^^

^^

^

^ ^

^

^̂

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^^^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^ ^

^

^^

^

^

^^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^^

^

^

^^

^

^

^^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^^

^

^

^

^^ ^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^̂

^

^^

^

^^

^^

^^

^

^

^

^

^

^

^

^^

^^

^

^

^

^

^^

^^

^

^

^

^

^

^

^^^

^̂

^

^

^ ^^

^^ ^^

^

^̂

^

^^

^̂

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^^

^

^^

^

^

^^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^^

^^

^

^

^

^

^

^

^

^

^^

^^

^

^

^

^^

^

^^

^ ^

^

^

^^

^

^

^

^

^

^

^^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^^

^^

Municipality Boundary^ Municipality Head

Ejido Allocated Post 1960

^

^

^

^

^

^

^

^

^

^

^

^

^^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^^

^

^

^

^

^

^

^^

^

^^

^

^

^

^^^

^^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^^

^

^ ^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^^

^

^

^

^

^

^

^^

^

^^

^

^

^

^

^^

^^

^

^

^

^

^

^^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^^

^

^

^^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^̂

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^^

^

^

^

^

^^

^

^

^

^

^^

^

^

^

^

^

^

^

^^^

^

^

^

^

^

^

^

^

^^

^^

^

^ ^

^

^̂

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^^^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^ ^

^

^^

^

^

^^

^

^

^

^

^

^

^

^

^

^ ^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^^

^

^

^^

^

^

^^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^^

^

^

^

^^ ^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^̂

^

^^

^

^^

^^

^^

^

^

^^

^

^

^

^^

^^

^

^

^

^

^^

^^

^

^

^

^

^

^

^^^

^̂

^

^

^ ^^

^^ ^^

^

^̂

^

^^

^̂

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^^

^

^^

^

^

^^

^ ^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^^

^^

^

^

^

^

^

^

^

^

^^

^^

^

^

^

^^

^

^^

^ ^

^

^

^^

^

^

^

^

^

^

^^

^

^ ^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^

^^

^

^

^

^^

^^

Municipality Boundary^ Municipality Head

Ejido Allocated Post 1960

Average Euclidean distance from municipality head: 33.22 km Average Euclidean distance from municipality head: 49.76 km

Notes: Both municipalities belong to the same state (Durango) and are similar in area and land available for redistribution. High and Lowcompetition is defined based on whether the vote share for opposition parties is above or below the median of the distribution of the voteshares.

Figure 2: Allocation of ejidos over time

190

750

1611

2295

4886

1075

568699

501

890

1335

658839

626349

66

01,000

2,000

3,000

4,000

5,000

Frequency

1915 1920 1925 1930 1935 1940 1945 1950 1955 1960 1965 1970 1975 1980 1985 1990

Notes: Number of allocated ejidos. Authors’ calculation with data fromthe Padrón e Historial de Núcleos Agrarios - PHINA. Baseline sample of munici-palities with political information data.

55

Figure 3: Opposition vote share and distance to municipality head

Notes: Municipal boundaries are in black. Opposition vote share is calculated as 1—PRI vote share. Differencein the average distance of ejidos from municipality head is calculated at the municipality level as the averagedistance from the municipality head post-1960 minus the average distance before 1960. Cutoffs in both mapscorrespond to the division of each variable into quartiles.

56

Figure 4: Political competition and distance from municipality head

Vote dispersion Opposition vote share-2

02

46

8Pa

ram

eter

est

imat

e

1920

1925

1930

1935

1940

1945

1950

1955

1960

1965

1970

1975

1980

1985

1990

-20

24

68

Para

met

er e

stim

ate

1920

1925

1930

1935

1940

1945

1950

1955

1960

1965

1970

1975

1980

1985

1990

Opposition ever won

-20

24

6Pa

ram

eter

est

imat

e

1920

1925

1930

1935

1940

1945

1950

1955

1960

1965

1970

1975

1980

1985

1990

Notes: Each graph is based on a regression of the distance of theallocated ejidos from their municipality heads on the interactionof each competition measure and a full set of quinquenniumdummies qt (q1915 equals 1 if the ejido was allocated from 1915 to1919, q1920 equals 1 if it was allocated from 1920 to 1924, and so on),with municipality and presidential-term fixed effects. Each figureplots the coefficients of the interaction terms (1915 is the omittedquinquennium) with 95% confidence intervals. All competitionmeasures are standardized.

57

A Online Appendix

A.1 Additional Robustness

Table A-1: Distance from municipality head and contemporaneous public goodand service delivery

(1) (2) (3)Share of households in locality with:

Dependent variable: Piped water Drainage Electricity

Panel A: Full set of localities in 1990

Distance of locality from municipality head -0.0021*** -0.0016*** -0.0036***(0.0003) (0.0002) (0.0006)

Observations 83,532 83,532 83,532R-squared 0.2795 0.2432 0.3444

Panel B: Localities in 1990 that overlap with ejidos

Distance of locality from municipality head -0.0017*** -0.0010*** -0.0033***(0.0003) (0.0002) (0.0005)

Observations 31,959 31,959 31,959R-squared 0.3152 0.2768 0.3903

Panel C: Full set of localities in 2000

Distance of locality from municipality head -0.0013*** -0.0024*** -0.0024***(0.0003) (0.0006) (0.0005)

Observations 107,218 107,218 107,218R-squared 0.2783 0.3643 0.2989

Panel D: Localities in 2000 that overlap with ejidos

Distance of locality from municipality head -0.0011*** -0.0018*** -0.0023***(0.0003) (0.0004) (0.0004)

Observations 41,006 41,006 41,006R-squared 0.3118 0.4255 0.3713

Notes: Cross-section of localities. All specifications include municipality fixed effects. Robust standard errors inparentheses are clustered at the municipality level, *** p<0.01, ** p<0.05, * p<0.1.

58

Table A-2: Classification of opposition parties

Party Oppositionabbreviation Name details and coalitions classification

PST Partido Socialista de los Trabajadores FriendlyPRT Partido Revolucionario de los Trabajadores UnfriendlyPRDPRT PRD + PRT UnfriendlyPRDPPSPFCRN PRD + PPS + PFCRN (Frente Cardenista de Reconstrucción Nacional) UnfriendlyPRDPMT PRD + PMT UnfriendlyPRD Partido de la Revolución Democrática UnfriendlyPPS Partido Popular Socialista FriendlyPPM Partido del Pueblo Mexicano UnfriendlyPMT Partido Mexicano de los Trabajadores UnfriendlyPFCRNPMSPPS PFCRN + PMS + PPS FriendlyPDM Partido Demócrata Mexicano UnfriendlyPCM Partido Comunista Mexicano UnfriendlyPCDP Partido del comité de Defensa Popular UnfriendlyPC Previous PCM UnfriendlyPARM Partido Auténtico de la Revolución Mexicana FriendlyPAN Partido de Acción Nacional UnfriendlyOther Votes for other parties not specified in electoral database Unfriendly

Notes: The parties listed are the full set of PRI opposition parties registered in the BANAMEX-CIDAC electoral database for municipal races in our sampleperiod for computing electoral competition (1980s). A party is classified as friendly if it is listed as ‘parastatal’ in Molinar and Weldon (1990) and Peiro (1998).

59

Table A-3: Distance of new land granted and political competition:Results using the first election in the 1980s

Dependent variable: Distance of ejido from municipality head(1) (2) (3)

CompetitionVote

dispersionOppositionvote share

Oppositionwon

Panel A: Baseline results, ejidos allocated from 1914 to 1992

Post-1960 × Competition 2.36** 2.12** 1.33(1.05) (0.91) (0.87)

Panel B: Falsification, ejidos allocated from 1914 to 1960, Placebo using post 1935

Post-1935 × Competition 0.13 0.19 0.25(0.30) (0.27) (0.17)

Panel C: Controlling for differential trends based on municipal characteristics

Post-1960 × Competition 1.815** 1.765** 1.618**(0.914) (0.805) (0.674)

Panel D: Controlling for the stock of allocated ejidos

Post-1960 × Competition 1.91* 1.79* 1.42*(1.07) (0.92) (0.85)

Panel E: Controlling for the land available for redistribution

Post-1960 × Competition 1.07* 1.07 1.04*(0.64) (0.65) (0.56)

Panel F: Controlling for state-specific trends

Post-1960 × Competition 1.39** 1.29** 1.22*(0.59) (0.61) (0.72)

Panel G: Controlling for the strength of rural elites

Post-1960 × Competition 1.87** 1.67** 1.03(0.79) (0.70) (0.70)

Notes: Robust standard errors in parentheses are clustered at the municipality level, *** p<0.01, ** p<0.05, * p<0.1. Regressions are at the ejido level. All specifications include municipality andpresidential-term fixed effects. Post-1960 (1935) is a dummy variable that equals 1 if the ejido is granted after 1960 (1935). Competition refers to political competition measured at the municipality levelusing the variable indicated in each column (see the notes to Table 1 and the main text for exact definitions). All competition measures are standardized.

60

Table A-4: Distance of new land granted and political competition:Results using log of distance of ejido from municipality head

Dependent variable: Log of distance of ejido from municipality head(1) (2) (3)

CompetitionVote

dispersionOppositionvote share

Oppositionever won

Panel A: Baseline results, ejidos allocated from 1914 to 1992

Post-1960 × Competition 0.06*** 0.07*** 0.09***(0.02) (0.02) (0.02)

Panel B: Falsification, ejidos allocated from 1914 to 1960, placebo using post-1935

Post-1935 × Competition -0.01 -0.01 0.02(0.02) (0.02) (0.02)

Panel C: Controlling for differential trends based on municipal characteristics

Post-1960 × Competition 0.059** 0.067*** 0.089***(0.025) (0.025) (0.021)

Panel D: Controlling for the stock of allocated ejidos

Post-1960 × Competition 0.05** 0.06** 0.09***(0.03) (0.03) (0.02)

Panel E: Controlling for the land available for redistribution

Post-1960 × Competition 0.04* 0.05** 0.07***(0.02) (0.02) (0.02)

Panel F: Controlling for state-specific trends

Post-1960 × Competition 0.05** 0.06** 0.08***(0.02) (0.02) (0.02)

Panel G: Controlling for the strength of rural elites

Post-1960 × Competition 0.06** 0.07*** 0.08***(0.02) (0.02) (0.02)

Notes: Robust standard errors in parentheses are clustered at the municipality level, *** p<0.01, ** p<0.05, * p<0.1. Regressions are at the ejido level. All specifications include municipality andpresidential-term fixed effects. Post-1960 (1935) is a dummy variable that equals 1 if the ejido is granted after 1960 (1935). Competition refers to political competition measured at the municipality levelusing the variable indicated in each column (see the notes to Table 1 and the main text for exact definitions). All competition measures are standardized.

61

Table A-5: Distance of new land granted and political competition:Results using distance of ejido from municipality head accounting for the terrain

elevation profile

Dependent variable: Distance of ejido from municipality head accounting for the terrain elevation profile(1) (2) (3)

CompetitionVote

dispersionOppositionvote share

Oppositionever won

Panel A: Baseline results, ejidos allocated from 1914 to 1992

Post 1960 × Competition 2.98** 3.40** 3.33**(1.28) (1.55) (1.66)

Panel B: Falsification, ejidos allocated from 1914 to 1960, placebo using post-1935

Post-1935 × Competition 0.22 0.22 0.32(0.02) (0.02) (0.02)

Panel C: Controlling for differential trends based on municipal characteristics

Post-1960 × Competition 2.672** 3.144** 3.701**(0.025) (0.025) (0.021)

Panel D: Controlling for the stock of allocated ejidos

Post-1960 × Competition 2.51* 2.96* 3.34** *(0.03) (0.03) (0.02)

Panel E: Controlling for the land available for redistribution

Post-1960 × Competition 1.47* 1.65* 2.11*** *(0.02) (0.02) (0.02)

Panel F: Controlling for state-specific trends

Post-1960 × Competition 2.24*** 2.86*** 3.38***(0.02) (0.02) (0.02)

Panel G: Controlling for the strength of rural elites

Post-1960 × Competition 2.48** 2.80** 3.08**(0.02) (0.02) (0.02)

Notes: Robust standard errors in parentheses are clustered at the municipality level, *** p<0.01, ** p<0.05, * p<0.1. Regressions are at the ejido level. All specifications include municipality andpresidential-term fixed effects. Post-1960 (1935) is a dummy variable that equals 1 if the ejido is granted after 1960 (1935). Competition refers to political competition measured at the municipality levelusing the variable indicated in each column (see the notes to Table 1 and the main text for exact definitions). All competition measures are standardized.

62

Table A-6: Distance of new land granted and political competition:Results using distance of ejido from municipality head via DCW roads

Dependent variable: Distance of ejido from municipality head via DCW roads(1) (2) (3)

CompetitionVote

dispersionOppositionvote share

Oppositionever won

Panel A: Baseline results, ejidos allocated from 1914 to 1992

Post-1960 × Competition 3.04** 3.44** 3.41**(0.02) (0.02) (0.02)

Panel B: Falsification, ejidos allocated from 1914 to 1960, placebo using post-1935

Post-1935 × Competition 0.14 0.13 0.36(0.02) (0.02) (0.02)

Panel C: Controlling for differential trends based on municipal characteristics

Post-1960 × Competition 2.754** 3.182** 3.780**(0.025) (0.025) (0.021)

Panel D: Controlling for the stock of allocated ejidos

Post-1960 × Competition 2.55* 2.98* 3.43**(0.03) (0.03) (0.02)

Panel E: Controlling for the land available for redistribution

Post-1960 × Competition 1.49* 1.64* 2.16***(0.02) (0.02) (0.02)

Panel F: Controlling for state-specific trends

Post-1960 × Competition 2.18*** 2.79*** 3.44***(0.02) (0.02) (0.02)

Panel G: Controlling for the strength of rural elites

Post-1960 × Competition 2.54** 2.83** 3.16**(0.02) (0.02) (0.02)

Notes: Robust standard errors in parentheses are clustered at the municipality level, *** p<0.01, ** p<0.05, * p<0.1. Regressions are at the ejido level. All specifications include municipality andpresidential-term fixed effects. Post-1960 (1935) is a dummy variable that equals 1 if the ejido is granted after 1960 (1935). Competition refers to political competition measured at the municipality levelusing the variable indicated in each column (see the notes to Table 1 and the main text for exact definitions). All competition measures are standardized.

63

Table A-7: Distance of new land granted and political competition:IV Results using the log of number of events of discontent

as an instrument for political competition

Dependent variable: Distance of ejido from municipality head(1) (2) (3) (4)

ReducedEstimates IV Estimates Form

CompetitionVote

dispersionOppositionvote share

Oppositionwon

-

Panel A: Baseline results, ejidos allocated from 1914 to 1992

Post1960 × Competition 5.35** 5.35** 14.36**(2.32) (2.33) (6.15)

Post 1960 × Log of number of events of discontent 2.21**(0.99)

Observations 17,279 17,279 17,279 17,279R-squared 0.57 0.58 0.54 0.04

Panel B: Falsification, ejidos allocated from 1914 to 1960, Placebo using post 1935

Post 1935 × Competition -1.52 -1.56 -3.35(1.09) (1.13) (3.21)

Post 1935 × Log of number of events of discontent -0.51(0.33)

Observations 12,568 12,568 12,568 12,568R-squared 0.56 0.56 0.56 0.02

Panel C: Controlling for differential trends based on municipal characteristics

Post 1960 × Competition 3.90* 4.04* 8.94**(2.18) (2.26) (4.13)

Post 1960 × Log of Number of events of discontent 1.47*(0.84)

Observations 16,224 16,224 16,224 16,224R-squared 0.58 0.58 0.58 0.06

Panel D: Controlling for the stock of allocated ejidos

Post 1960 × Competition 3.96 4.02 8.13(2.89) (2.88) (5.22)

Post 1960 × Log of number of events of discontent 1.55(1.17)

Observations 17,279 17,279 17,279 17,279R-squared 0.58 0.58 0.57 0.04

Panel E: Controlling for the land available for redistribution

Post 1960 × Competition 2.06 2.10 6.18(1.65) (1.70) (4.43)

Post 1960 × Log of number of events of discontent 0.82(0.66)

Observations 17,279 17,279 17,279 17,279R-squared 0.59 0.59 0.59 0.07

Panel F: Controlling for state-specific trends

Post 1960 × Competition 3.10* 3.27* 7.29*(1.59) (1.69) (3.87)

Post 1960 × Log of number of events of discontent 1.01*(0.60)

Observations 17,279 17,279 17,279 17,279R-squared 0.59 0.59 0.59 0.07

Panel G: Controlling for the strength of rural elites

Post 1960 × Competition 4.22** 4.43** 11.70**(1.89) (2.03) (5.51)

Post 1960 × Log of number of events of discontent 1.67**(0.76)

Observations 17,279 17,279 17,279 17,279R-squared 0.58 0.58 0.56 0.04

Notes: Robust standard errors in parentheses are clustered at the municipality level, *** p<0.01, ** p<0.05, * p<0.1. Regressions are at the ejido level. All specifications include municipality and presidential-term fixed effects. Post-1960 (1935) is a dummy variable that equals 1 if the ejido is granted after 1960 (1935). Competition refers to political competition measured at the municipality level using the variableindicated in each column (see the notes to Table 1 and the main text for exact definitions). The number of events of social or political discontent are counted during the period 1960-1969 using references torelated events in two Mexican newspapers with national coverage: El Universal and Excelsior. All competition measures are standardized.

64

Table A-8: Distance of new land granted and political competition and events ofdiscontent: robustness of first-stage estimates

Dependent variable: Post 1960 × Competition(1) (2) (3)

CompetitionVote

dispersionOppositionvote share

Oppositionever won

Panel A: Baseline results, ejidos allocated from 1914 to 1992

Post 1960 × Log of number of events of discontent 0.41*** 0.41*** 0.15**(0.04) (0.06) (0.07)

Observations 16,999 16,999 16,999R-squared 0.23 0.22 0.03Partial R-squared 0.219 0.213 0.0300F 9.204 5.596 2.271Partial F 106.2 54.81 4.893

Panel B: Falsification, ejidos allocated from 1914 to 1960, placebo using post-1935

Post 1935 × Log of number of events of discontent 0.33*** 0.32*** 0.15**(0.04) (0.04) (0.08)

Observations 12,253 12,253 12,253R-squared 0.15 0.14 0.03Partial R-squared 0.151 0.144 0.0296F 9.255 6.573 1.597Partial F 80.75 53.26 3.915

Panel C: Controlling for differential trends based on municipal characteristics

Post 1960 × Log of Number of events of discontent 0.38*** 0.36*** 0.16**(0.04) (0.05) (0.07)

Observations 15,963 15,963 15,963R-squared 0.28 0.29 0.06Partial R-squared 0.179 0.167 0.0319F 9.917 5.986 5.824Partial F 85.70 56.58 5.400

Panel D: Controlling for the stock of allocated ejidos

Post 1960 × Log of number of events of discontent 0.39*** 0.39*** 0.19***(0.04) (0.05) (0.07)

Observations 16,999 16,999 16,999R-squared 0.23 0.22 0.04Partial R-squared 0.167 0.158 0.0365F 8.626 5.213 2.888Partial F 96.10 56.85 8.065

Panel E: Controlling for the land available for redistribution

Post 1960 × Log of number of events of discontent 0.40*** 0.39*** 0.13**(0.04) (0.06) (0.07)

Observations 16,999 16,999 16,999R-squared 0.24 0.24 0.06Partial R-squared 0.201 0.192 0.0218F 10.69 6.902 2.416Partial F 94.93 50.21 4.001

Panel F: Controlling for state-specific trends

Post 1960 × Log of number of events of discontent 0.36*** 0.34*** 0.14**(0.03) (0.04) (0.06)

Observations 16,999 16,999 16,999R-squared 0.41 0.44 0.23Partial R-squared 0 0 0F 150812 22608 181179Partial F 110.7 73.22 5.043

Panel G: Controlling for the strength of rural elites

Post 1960 × Log of number of events of discontent 0.40*** 0.38*** 0.14**(0.04) (0.05) (0.07)

Observations 16,999 16,999 16,999R-squared 0.23 0.24 0.04Partial R-squared 0.193 0.176 0.0240F 9.553 6.480 2.179Partial F 103.8 66.57 4.520

Notes: Robust standard errors in parentheses are clustered at the municipality level, *** p<0.01, ** p<0.05, * p<0.1. Regressions are at the ejido level. All specifications include municipality andpresidential-term fixed effects. Post-1960 is a dummy variable that equals 1 if the ejido is granted after 1960. Competition refers to political competition measured at the municipality level using thevariable indicated in each column (see the notes to Table 1 and the main text for exact definitions). The number of events of social or political discontent are counted during the period 1960-1969 usingreferences to related events in two Mexican newspapers with national coverage: El Universal and Excelsior. All competition measures are standardized.

65

Table A-9: Distance of new land granted, political competition and non-rural discontent

(1) (2) (3) (4)

Competition Votedispersion

Oppositionvote share

Oppositionever won

Panel A: Dependent variable: Distance of ejido from municipality head

ReducedIV Estimates Form

Post 1960 × Competition 4.40** 4.33** 9.60**(2.12) (2.10) (4.22)

Post1960 × Log of number of events of discontent 1.98**(0.99)

Panel B: Dependent variable: Post1960 × Competition

First Stage

Post 1960 × Log of number of non-rural protests 0.45*** 0.46*** 0.21***(0.05) (0.06) (0.08)

R-squared 0.222 0.223 0.0461Partial F-statistic 96.16 54.08 6.682

Observations 17,279 17,279 17,279 17,279Notes: Robust standard errors in parentheses are clustered at the municipality level, *** p<0.01, ** p<0.05, * p<0.1. Regressions are at the ejido level. All specifications includemunicipality and presidential-term fixed effects. Post-1960 is a dummy variable that equals 1 if the ejido is granted after 1960. Competition refers to political competition measuredat the municipality level using the variable indicated in each column (see the notes to Table 1 and the main text for exact definitions). The number of events reflecting social andpolitical discontent protests are counted during the period 1960-1969 using references to related events in two Mexican newspapers of national coverage: El Universal and Excelsior.We exclude events that are related to agriculture, ejidos or peasants movements. All competition measures are standardized. The number of observations changes relative to thosein baseline regressions as information about protests is not available for all ejidos. See the note in Table 2 and Figure A-1 for details on the computation of the various distances.

66

Figure A-1: Spatial distribution of ejidos and computation of distances

Panel A: Example of location and distribution of main geographical features in the administrative data

Panel B: Computation of distances of ejido from municipality head

Consider a hypothetical municipality similar to those presented in Panel A, with ejidos that mayinclude multiple localities. This municipality has one ejido (E) with two localities: L1 and L2. Eachlocality has a number on inhabitants given by Population(L1) and Population(L2), respectively.Let d1 and d2 denote the distances of these localities form the municipality head. We computedifferent measures of d1 and d2 depending on whether or not they account for terrain and roads asillustrated in the following figures:

d1

d2

L1

L2

E

Mun Head

d1

d2

L1

L2

E

Mun Head

L1

L2

E

Mun Head

Road

d2

d1

Option 1: Minimum Euclidean distance Option 2: Minimum distance accountingfor terrain elevation

Option 3: Minimum distance via DCWroads

Using each of these options we defined the distance of ejido (E) from the municipality head as:

d(E, Mun Head) = d1

(Population(L1)

Population(L1)+Population(L2)

)+ d2

(Population(L2)

Population(L1)+Population(L2)

).

In other words, it is the population-weighted average distance form the municipality head to thelocalities within ejido E.

Notes: The distance from a locality to the municipality head accounting for elevation terrain profile (Option 2) penalizes the minimum Euclideandistance (Option 1) when there are changes in altitude between them. The distance via DCW roads (Option 3) accounts for the use of roads toreach the municipality head. The trace of those roads comes from the Digital Chart of the World of 1992 and the overall distance of each localityfrom its municipality head is computed adding up two different figures. First, the Euclidean distance from the locality to the closest point in aroad that leads to the municipality head, and second, the length of the segment that connects such point to the municipality head following theroad path.

67

A.2 Coding of events of social and political discontent during the 1960s

To measure social and political discontent during the 1960s, we relied on all issuesof Mexico’s two main newspapers, Excelsior and El Universal, from January 1st, 1960to December 31st, 1969. We searched on the articles’ title, subtitle, and main text toidentify all news about protests, strikes, demonstrations, riots and marches for everymunicipality.

When news do not mention a particular location or when they refer to national orstate-level protests, we err on the conservative side and avoid inputting any values tocovered municipalities. If instead a given municipality (or municipalities) are listed,we then coded the corresponding location as affected by the protest.

The following words were used to identify news articles about social and politicaldiscontent:

• Protestas (protests) and the n-gram “protest*”

• Huelgas (strikes) and the n-gram “huelg*”

• Manifestaciones (demonstrations) and the n-gram “manifesta*”

• Disturbios (riots) and the n-gram “Disturbio*”

• Marchas (marches) and the n-gram “March*”

Each of the resulting news articles where then verified to identify the municipality ofoccurrence.

Figure A-3 shows the distribution of events of social and political discontent overtime. The most common words in the resulting set of articles are the following (exclud-ing common Spanish expressions and distinguishing capital letters):

Freq Word Freq Word Freq Word Freq Word Freq Word1749 huelga 229 Campesinos 147 Ciudad 109 quienes 93 Tijuana

851 contra 219 Agenda 141 líder 107 entidad 93 labores656 campesinos 210 aumento 140 general 107 federal 92 secretario556 trabajadores 208 ciudad 137 Veracruz 107 nuevo 91 Denuncian435 tierras 206 Obrera 132 Acapulco 106 intervención 91 comercio413 Sindicato 196 problema 132 Estados 106 movimiento 91 médicos355 estudiantes 195 Universidad 131 empresas 105 Juárez 91 Morelos354 Trabajadores 190 obreros 130 agitación 104 mitin 90 textiles334 conflicto 181 Puebla 126 industria 103 Industria 89 compañía328 maestros 181 agua 125 Durango 102 municipio 88 Aviación325 Nacional 178 Unión 123 policía 102 impuestos 88 capital319 contrato 172 denuncian 123 zona 101 pagos 87 ejidal315 gobernador 171 piden 122 personas 100 salarios 86 nacional312 gobierno 166 San 121 manifestación 98 descontento 86 fábrica300 estados 164 país 121 Estudiantes 98 está 85 Confederación290 ejidatarios 162 sindicato 121 terrenos 97 Ejidatarios 85 dirigentes290 empresa 161 revisión 120 Compañía 97 Presidente 85 demandas283 protesta 159 Mexicana 118 estudiantil 97 Junta 85 hambre272 paro 158 situación 117 frente 96 evitar 84 escuelas269 grupo 157 colectivo 117 debido 95 telefonistas 84 agrario261 México 157 líderes 114 República 95 comerciantes 84 región258 autoridades 152 problemas 112 alcalde 95 local 84 quejan249 parte 149 pláticas 112 textil 95 años 83 empleados247 presidente 148 falta 111 población 94 Federación 83 ejidales241 Huelga 148 apoyo 109 servicio 94 Piden 82 Maestros240 municipal 147 Guerrero 109 denuncia 93 Secretaría 80 paros

68

Figure A-2: Evolution of new land endowments, and restitutions

8

42

57

38

15

9

2 2 1 1 3 1 0 1 3 0020

0040

0060

0080

00

1915

1920

1925

1930

1935

1940

1945

1950

1955

1960

1965

1970

1975

1980

1985

1990

Number of Events

010

000

2000

030

000

4000

050

000

1915

1920

1925

1930

1935

1940

1945

1950

1955

1960

1965

1970

1975

1980

1985

1990

Log(Hectares)

Dotaciones (Endowments) Restituciones (Restitutions)

Notes: The number of events refers to the number of approved petitions. Authors’ calculation with data from the Padróne Historial de Núcleos Agrarios

Figure A-3: Number of social and political events reflecting discontent per year

568

756

445

392 389

306

409386

533

276

020

040

060

080

0To

tal r

efer

ence

s to

pro

test

s

1960 1961 1962 1963 1964 1965 1966 1967 1968 1969Year

Notes: Total number of social and political events reflecting discontent peryear as reported in news articles referring to protests, strikes, demonstra-tions, riots and marches (excluding national and state level protests notspecifying municipality of occurrence). Authors’ calculation with newsfrom Excelsior and El Universal.

69

Figure A-4: Social and Political discontent and distance from municipality head

-50

510

Para

met

er e

stim

ate

1920

1925

1930

1935

1940

1945

1950

1955

1960

1965

1970

1975

1980

1985

1990

Notes: Each graph is based on a regression of the distance of the allocated ejidosfrom their municipality heads on the interaction of the (log of) the number of eventsreflecting social and political discontent protests and a full set of quinquenniumdummies qt (q1915 equals 1 if the ejido was allocated from 1915 to 1919, q1920equals 1 if it was allocated from 1920 to 1924, and so on), with municipality andpresidential-term fixed effects. Each figure plots the coefficients of the interactionterms (1915 is the omitted quinquennium) with 95% confidence intervals. Thenumber of events reflecting social and political discontent protests are countedduring the period 1960-1969 using references to related events in two Mexicannewspapers of national coverage: El Universal and Excelsior. The measure of socialand political discontent is standardized.

70