Search

English Abstract: We develop a model of the politics of state capacity building undertaken by incumbent parties that have a comparative advantage in clientelism rather than in public goods provision. The model predicts that, when challenged by opponents, clientelistic incumbents have the incentive to prevent investments in state capacity. We provide empirical support for the model’s implications by studying policy decisions by the Institutional Revolutionary Party (PRI) that affected local state capacity across Mexican municipalities and over time. Our difference-in-differences and instrumental variable identification strategies exploit a national shock that threatened the Mexican government’s hegemony in the early 1960s. The intensity of this shock, which varied across municipalities, was partly explained by severe droughts that occurred during the 1950s. Spanish Abstract: Desarrollamos un modelo teorico sobre los incentivos politicos para la construccion de capacidades estatales por parte de gobernantes cuya ventaja comparativa es el clientelismo y no la provision de bienes publicos. El modelo predice que, cuando enfrentan competencia de opositores, los gobernantes clientelistas tienen el incentivo a bloquear las inversiones en capacidad estatal. Presentamos evidencia empirica que apoya estas implicaciones estudiando decisiones del Partido Revolucionario Institucional (PRI) que afectaron la capacidad estatal a traves de los municipios mexicanos y a lo largo del tiempo. Nuestras estrategias de diferencias en diferencias y variables instrumentales aprovechan un choque nacional que amenazo la hegemonia del gobierno mexicano al comienzo de la decada de 1960. La intensidad de este choque en distintos municipios estuvo en parte explicada por la sequias severas que ocurrieron en los anos 50.

Effective policy-making requires that voters avoid electing malfeasant politicians. However, informing voters of incumbent malfeasance in corrupt contexts may not reduce incumbent support. As our simple learning model shows, electoral sanctioning is limited where voters already believed incumbents to be malfeasant, while information’s effect on turnout is non-monotonic in the magnitude of reported malfeasance. We conducted a field experiment in Mexico that informed voters about malfeasant mayoral spending before municipal elections, to test whether these Bayesian predictions apply in a developing context where many voters are poorly informed. Consistent with voter learning, the intervention increased incumbent vote share where voters possessed unfavorable prior beliefs and when audit reports caused voters to favorably update their posterior beliefs about the incumbent’s malfeasance. Furthermore, we find that low and, especially, high malfeasance revelations increased turnout, while less surprising information reduced turnout. These results suggest that improved governance requires greater transparency and citizen expectations.

National audience; While social pressure is widely believed to influence voters, evidence that information passed between social ties affects beliefs, policy preferences, and voting behavior is limited. We investigate whether information about unemployment shocks diffuses through networks of strong and mostly weak social ties and influences voters in Denmark. We link surveys with population-level administrative data that log unemployment shocks afflicting respondents’ familial, vocational, and educational networks. Our results show that the share of second-degree social ties—individuals that voters learn about indirectly—that became unemployed within the last year increases a voter’s perception of national unemployment, self-assessed risk of becoming unemployed, support for unemployment insurance, and voting for left-wing political parties. Voters’ beliefs about national aggregates respond to all shocks similarly, whereas subjective perceptions and preferences respond primarily to unemployment shocks afflicting second-degree ties in similar vocations. This suggests that information diffusion through social ties principally affects political preferences via egotropic—rather than sociotropic—motives.

Malfeasance in local governments is common in developing democracies. Electoral accountability could mitigate such malfeasance, but may require media market structures that incentivise profit-maximising local media to report on incumbent malfeasance. We test this claim in Mexico, leveraging plausibly exogenous variation in the pre-election release of municipal audits revealing misallocated spending and access to broadcast media. We find that each additional local media station amplifies voter punishment (rewards) of high (zero) malfeasance by up to 1 percentage point. Local media’s accountability-enhancing effects are greater when there are fewer non-local competitors and where local outlets’ audiences principally reside within their municipality.

Political parties in sub‐Saharan Africa's developing democracies are often considered to lack sufficiently sophisticated machines to monitor and incentivize their political brokers. We challenge this view by arguing that the decentralized pyramidal structure of their machines allows them to engage in broker monitoring and incentivizing to mobilize voters, which ultimately improves their electoral performance. This capacity is concentrated (a) among incumbent parties with greater access to resources and (b) where the scope for turnout buying is higher due to the higher costs of voting. Using postwar Liberia to test our argument, we combine rich administrative data with exogenous variation in parties' ability to monitor their brokers. We show that brokers mobilize voters en masse to signal effort, that increased monitoring ability improves the incumbent party's electoral performance, and that this is particularly so in precincts in which voters must travel farther to vote and thus turnout buying opportunities are greater.

An Informational Theory of Electoral Targeting in Young Clientelistic Democracies: Evidence from Senegal

We theoretically and empirically study an incomplete information model of social learning. Agents initially guess the binary state of the world after observing a private signal. In subsequent rounds, agents observe their network neighbors' previous guesses before guessing again. Agents are drawn from a mixture of learning types—Bayesian, who face incomplete information about others' types, and DeGroot, who average their neighbors' previous period guesses and follow the majority. We study (1) learning features of both types of agents in our incomplete information model; (2) what network structures lead to failures of asymptotic learning; (3) whether realistic networks exhibit such structures. We conducted lab experiments with 665 subjects in Indian villages and 350 students from ITAM in Mexico. We perform a reduced‐form analysis and then structurally estimate the mixing parameter, finding the share of Bayesian agents to be 10% and 50% in the Indian‐villager and Mexican‐student samples, respectively.

Women's exposure to gender-based and intimate partner violence (GBV and IPV) is particularly acute due to COVID-19, especially in the Global South. We test whether edutainment interventions that have been shown to successfully combat GBV and IPV when delivered in person can be effectively delivered using social (WhatsApp and Facebook) and traditional (TV) media. To do so, we randomized the mode of implementation of an intervention conducted by an Egyptian women’s rights non-governmental organization seeking to support women while accommodating social distancing amid COVID-19. We found WhatsApp to be a more effective way to deliver the intervention than Facebook, but no differences across outcomes between WhatsApp and TV dissemination. Our findings show that these media campaigns had no impact on women's attitudes toward gender or marital equality, or the justifiability of violence. However, the campaign did increase women's knowledge, hypothetical, and reported use of resources available to those exposed to GBV and IPV.

A recent literature emphasizes political economy factors behind the wave of administrative splits across the developing world. While previous studies have focused on why some groups are more likely to obtain new administrative units, they do not explain why vote-maximizing incumbents use this arguably less efficient policy in the first place. We contribute to this literature by embedding administrative splits within incumbents’ broader electoral strategy of distributive policies. We develop a model in which incumbents target local public goods to groups for whom this is a credible signal of commitment, namely, those with a history of reciprocal relationship. When incumbents face increased electoral competition, however, other groups require a stronger signal, which is emitted by the costly creation of new units that reduces the cost of future transfers to those groups. We test our theory using electoral and public goods data from Senegal and find robust support for its predictions.

How do social networks moderate the way political information influences electoral accountability? We propose a simple model in which incumbent malfeasance revelations can facilitate coordination around less malfeasant challenger parties in highly connected voter networks, even when voters update favorably about incumbent party malfeasance. We provide evidence from Mexico of this mechanism by leveraging a field experiment in a context where the provision of incumbent malfeasance information increased support for incumbent parties, despite voters continuing to believe that challengers were less malfeasant than incumbents. Combining this experiment with detailed family network data, we show that—consistent with the model—the increase in incumbent party vote share due to information provision was counteracted by coordination around less malfeasant challengers in precincts with greater network connectedness. Individual-level data further demonstrate that networks facilitated explicit and tacit coordination among voters. These findings suggest that networks can help voters coordinate around information to help remove poorly performing politicians.

National audience; Political accountability may be constrained by the reach and relevance of information campaigns in developing democracies and—upon receiving information—voters' ability and will to hold politicians accountable. To illuminate voter‐level constraints and information relevance absent dissemination constraints, we conducted a field experiment around Senegal's 2017 parliamentary elections to examine the core theoretical steps linking receiving different types of incumbent performance information to electoral and nonelectoral accountability. Voters immediately processed information as Bayesians, found temporally benchmarked local performance outcomes particularly informative, and updated their beliefs for at least a month. Learning that incumbents generally performed better than expected, voters durably requested greater politician contact after elections while incumbent vote choice increased among likely voters and voters prioritizing local projects when appraising incumbents. In contrast, information about incumbent duties did not systematically influence beliefs or accountability. These findings suggest voters were able and mostly willing to use relevant information to hold politicians to account.

We examine the incentives incumbents face when creating new polling places. First, doing so improves incumbents’ ability to monitor brokers and voters by reducing the number of registered voters per polling station. Second, it reduces the distance traveled by citizens to vote, which undercuts incumbents’ ability to control the electorate via turnout buying. We evaluate this trade-off in the context of Uganda, where the incumbent significantly influences electoral administration. Drawing on rich administrative data, we leverage discontinuities in the creation of polling places to causally identify the independent effects of (1) the number of voters per polling station and (2) distance to vote on electoral outcomes. We find that decreasing (1) improves incumbent electoral outcomes, while reducing (2) worsens them. The benefits for incumbents outweigh the costs, which rationalizes recent developments to expand polling infrastructure in Uganda and elsewhere.

Lack of well-functioning formal institutions leads to reliance on social networks to enforce informal contracts. Social proximity and network centrality may affect cooperation. To assess the extent to which networks substitute for enforcement, we conducted high-stakes games across 34 Indian villages. We randomized subjects’ partners and whether contracts were enforced to estimate how partners’ relative network position differentially matters across contracting environments. While socially close pairs cooperate even without enforcement, distant pairs do not. Individuals with more central partners behave more cooperatively without enforcement. Capacity for cooperation in the absence of contract enforcement therefore depends on the subjects’ network position. (JEL C93, D86, K12, O15, O17, Z13)

Traditional leaders are often given subnational authority in developing democracies. Although ubiquitous, the electoral consequences of their power have received little attention, often due to empi...

Political brokers mobilize voters all over the world, yet little is known about what motivates them to do so. This article theorizes about two drivers of brokers' efforts: (1) incentives—monetary rewards or sanctions—and monitoring and (2) partisan attachment. We examine our theory using data on the Mexican National Educational Workers Union (SNTE), Latin America's largest union and a well-known political machine. Consistent with the role of teachers as brokers, we find that the vote share of parties supported by the SNTE machine is higher in polling stations located in schools. This effect is absent when teachers are asked to mobilize voters in support of a party for which they have no partisan attachment, and it is uncorrelated with the union's monitoring capacity. This suggests that partisan attachment, rather than incentives and monitoring, explains the SNTE's effectiveness as a political machine.

Developing democracies are experiencing unprecedented increases in primary and secondary schooling. To identify education's long-run political effects, we use a difference-in-differences design tha...

How do social networks influence and moderate electoral persuasion in developing countries? An extensive literature shows that social networks are important for understanding electoral persuasion in established democracies. At the same time, these theories might not necessarily apply to democracies in the developing world, particularly when they are characterized by clientelism, coercion, and other modes of political engagement outside of formal democratic institutions. In such contexts, networks can matter for politics in different, and sometimes unexpected, ways. In surveying the literature, this chapter dentifies three general functions of networks that are important for understanding electoral persuasion behavior in developing countries: (i) information diffusion; (ii) social persuasion; and (iii) coordination and enforcement. The chapter explores the implications of these network mechanisms by exploring the roles of both voter and politician networks.

We estimate the effects of one of the largest anti-vote-buying campaigns ever studied—with half a million voters exposed across 1427 villages—in Uganda’s 2016 elections. Working with civil society organizations, we designed the study to estimate how voters and candidates responded to their campaign in treatment and spillover villages, and how impacts varied with campaign intensity. Despite its heavy footprint, the campaign did not reduce politician offers of gifts in exchange for votes. However, it had sizable effects on people’s votes. Votes swung from well-funded incumbents (who buy most votes) towards their poorly-financed challengers. We argue the swing arose from changes in village social norms plus the tactical response of candidates. While the campaign struggled to instill norms of refusing gifts, it leveled the electoral playing field by convincing some voters to abandon norms of reciprocity—thus accepting gifts from politicians but voting for their preferred candidate.

We estimate the effects of one of the largest anti-vote-buying campaigns ever studied — with half a million voters exposed across 1427 villages—in Uganda’s 2016 elections. Working with civil society organizations, we designed the study to estimate how voters and candidates responded to their campaign in treatment and spillover villages, and how impacts varied with campaign intensity. Despite its heavy footprint, the campaign did not reduce politician offers of gifts in exchange for votes. However, it had sizable effects on people’s votes. Votes swung from well-funded incumbents (who buy most votes) towards their poorly-financed challengers. We argue the swing arose from changes in village social norms plus the tactical response of candidates. While the campaign struggled to instill norms of refusing gifts, it leveled the electoral playing field by convincing some voters to abandon norms of reciprocity—thus accepting gifts from politicians but voting for their preferred candidate.

Throughout much of the developing world, politicians rely on political brokers to buy votes prior to elections. We investigate how social networks help facilitate vote-buying exchanges by combining village network data of brokers and voters with broker reports of vote buying. We show that networks diffuse politically-relevant information about voters to brokers who leverage it to target voters. In particular, we find that brokers target reciprocal voters who are not registered to their party and about whom they can hear more information through their social network. These results highlight the importance of information diffusion through social networks for vote buying and ultimately for political outcomes.

The rise of broadcast and then digital media has had important political implications across the developing world. First considering independent media outlets, we review evidence showing that the media’s editorial content, revelation of information about candidates, and capacity to provide a platform for politicians can significantly shape electoral outcomes and mobilization. Unlike established democracies, the media is often used to buttress and oppose autocratic regimes. With respect to government control of the media, we review evidence of media bias, as well as its determinants and effectiveness at reducing opposition. With respect to media’s liberation potential, we examine how broadcast and internet-based technologies are—not without difficulties—providing new opportunities for facilitating dissent and change. We highlight methodological innovations, the challenges of isolating theoretical mechanisms, and avenues for future research throughout.

A preregistered meta-analysis of six field experiments finds no evidence overall that information campaigns shape voter behavior., Voters may be unable to hold politicians to account if they lack basic information about their representatives’ performance. Civil society groups and international donors therefore advocate using voter information campaigns to improve democratic accountability. Yet, are these campaigns effective? Limited replication, measurement heterogeneity, and publication biases may undermine the reliability of published research. We implemented a new approach to cumulative learning, coordinating the design of seven randomized controlled trials to be fielded in six countries by independent research teams. Uncommon for multisite trials in the social sciences, we jointly preregistered a meta-analysis of results in advance of seeing the data. We find no evidence overall that typical, nonpartisan voter information campaigns shape voter behavior, although exploratory and subgroup analyses suggest conditions under which informational campaigns could be more effective. Such null estimated effects are too seldom published, yet they can be critical for scientific progress and cumulative, policy-relevant learning.

A large literature examining advanced and consolidating democracies suggests that education increases political participation. However, in electoral authoritarian regimes, educated voters may instead deliberately disengage. If education increases critical capacities, political awareness, and support for democracy, educated citizens may believe that participation is futile or legitimizes autocrats. We test this argument in Zimbabwe—a paradigmatic electoral authoritarian regime—by exploiting cross-cohort variation in access to education following a major educational reform. We find that educationdecreasespolitical participation, substantially reducing the likelihood that better-educated citizens vote, contact politicians, or attend community meetings. Consistent with deliberate disengagement, education’s negative effect on participation dissipated following 2008’s more competitive election, which (temporarily) initiated unprecedented power sharing. Supporting the mechanisms underpinning our hypothesis, educated citizens experience better economic outcomes, are more interested in politics, and are more supportive of democracy, but are also more likely to criticize the government and support opposition parties.

Despite its prevalence, little is known about when parties buy turnout. We emphasize the problem of parties monitoring local brokers with incentives to shirk. Our model suggests that parties extract greater turnout buying effort from their brokers where they can better monitor broker performance and where favorable voters would not otherwise turn out. Exploiting exogenous variation in the number of polling stations—and thus electoral information about broker performance—in Mexican electoral precincts, we find that greater monitoring capacity increases turnout and votes for the National Action Party (PAN) and the Institutional Revolutionary Party (PRI). Consistent with our theoretical predictions, the effect of monitoring capacity on PRI votes varies nonlinearly with the distance of voters to the polling station: it first increases because rural voters—facing larger costs of voting—generally favor the PRI, before declining as the cost of incentivizing brokers increases. This nonlinearity is not present for the PAN, who stand to gain less from mobilizing rural voters.

Effective policy-making requires that voters avoid electing malfeasant politicians. However, as our simple learning model emphasizing voters’ prior beliefs and updating highlights, informing voters of incumbent malfeasance may not entail sanctioning. Specifically, electoral punishment of incumbents revealed to be malfeasant is rare where voters already believed them to be malfeasant, while information’s effect on turnout is non-linear in the magnitude of revealed malfeasance. These Bayesian predictions are supported by a field experiment informing Mexican voters about malfeasant mayoral spending before municipal elections. Given voters’ low expectations and initial uncertainty, as well as politician responses, relatively severe malfeasance revelations increased incumbent vote share on average. Consistent with voter learning, rewards were lower among voters with lower malfeasance priors, among voters with more precise prior beliefs, when audits revealed greater malfeasance, and among voters updating less favorably. Furthermore, both low and high malfeasance revelations increased turnout, while less surprising information reduced turnout. Institutional subscribers to the NBER working paper series, and residents of developing countries may download this paper without additional charge at www.nber.org.

Background Sickle Cell Trait (SCT) has been shown to be protective against malaria. A growing literature suggests that malaria exposure can reduce educational attainment. This study assessed the relationship and interactions between malaria, SCT and educational attainment in north-eastern Tanzania. Methods Seven hundred sixty seven children were selected from a list of individuals screened for SCT. Febrile illness and malaria incidence were monitored from January 2006 to December 2013 by community health workers. Education outcomes were extracted from the Korogwe Health and Demographic Surveillance system in 2015. The primary independent variables were malaria and SCT. The association between SCT and the number of fever and malaria episodes from 2006 to 2013 was analyzed. Main outcomes of interest were school enrolment and educational attainment in 2015. Results SCT was not associated with school enrolment (adjusted OR 1.42, 95% CI [0.593,3.412]) or highest grade attained (adjusted grade difference 0.0597, 95% CI [−0.567, 0.686]). SCT was associated with a 29% reduction in malaria incidence (adjusted IRR 0.71, 95% CI [0.526, 0.959]) but not with fever incidence (adjusted IRR 0.905, 95% CI [0.709-1.154]). In subgroup analysis of individuals with SCT, malaria exposure was associated with reduced school enrollment (adjusted OR 0.431, 95% CI [0.212, 0.877]). Conclusions SCT appears to reduce incidence of malaria. Overall, children with SCT do not appear to attend more years of school; however children who get malaria despite SCT appear to have lower levels of enrolment in education than their peers.

We examine how campaign advertising affects electoral support. We propose a simple model where advertising disproportionately benefits non-dominant political parties, because voters are uncertain about and biased against such parties. We test this argument in Mexico, where one of the three main parties dominates in many localities. To identify the effects of exposure to campaign advertising, we exploit differences across neighboring precincts in campaign ad distribution. These differences originate from cross-state media coverage spillovers induced by a 2007 reform that equalized access to ad slots across all broadcast media. We find that, on average, ads on AM radio increase the vote shares of the PAN and the PRD, but not the previously hegemonic PRI. Consistent with our model, campaign advertising is most effective in poorly informed and politically uncompetitive electoral precincts, and against locally dominant parties of intermediate strength.

Agents often use noisy signals from their neighbors to update their beliefs about a state of the world. The effectiveness of social learning relies on the details of how agents aggregate information from others. There are two prominent models of information aggregation in networks: (1) Bayesian learning, where agents use Bayes' rule to assess the state of the world and (2) DeGroot learning, where agents instead consider a weighted average of their neighbors' previous period opinions or actions. Agents who engage in DeGroot learning often double-count information and may not converge in the long run. We conduct a lab experiment in the field with 665 subjects across 19 villages in Karnataka, India, designed to structurally test which model best describes social learning. Seven subjects were placed into a network with common knowledge of the network structure. Subjects attempted to learn the underlying (binary) state of the world, having received independent identically distributed signals in the first period. Thereafter, in each period, subjects made guesses about the state of the world, and these guesses were transmitted to their neighbors at the beginning of the following round. We structurally estimate a model of Bayesian learning, relaxing common knowledge of Bayesian rationality by allowing agents to have incomplete information as to whether others are Bayesian or DeGroot. Our estimates show that, despite the flexibility in modeling learning in these networks, agents are robustly best described by DeGroot-learning models wherein they take a simple majority of previous guesses in their neighborhood.

We estimate the effect of local media outlets on political accountability in Mexico, focusing on malfeasance by municipal mayors. We study federal grants earmarked for infrastructure projects targeting the poor, and leverage two sources of plausibly exogenous variation. First, we exploit variation in the timing of the release of municipal audit reports. Second, and moving beyond existing studies, we exploit variation in media exposure at the electoral precinct level. In particular, we compare neighboring precincts on the boundaries of media stations’ coverage areas to isolate the effects of an additional media station. We find that voters punish the party of malfeasant mayors, but only in electoral precincts covered by local media stations (which emit from within the precinct’s municipality). An additional local radio or television station reduces the vote share of an incumbent political party revealed to be corrupt by 1 percentage point, and reduces the vote share of an incumbent political party revealed to have diverted funds to projects not benefiting the poor by around 2 percentage points. We also show that these electoral sanctions persist: at the next election, the vote share of the current incumbent’s party continues to be reduced by a similar magnitude. The electoral costs of diverting resources away from the poor are especially large for the populist Institutional Revolutionary Party (PRI) party. However, we find no effect of media stations based in other municipalities.Institutional subscribers to the NBER working paper series, and residents of developing countries may download this paper without additional charge at www.nber.org.

In settings with poor formal contract enforcement, profitable investments are likely unrealized. While social closeness can mitigate contractual incompleteness, we examine how to improve the preponderance of cases where contracting parties cannot rely upon social ties. We ask if a community can enlist members to monitor transactions or punish offending parties. We conduct a laboratory experiment in 40 Indian villages, with 960 non-anonymized subjects, where we have social network data. Participants play modified sender-receiver investment games, with and without third-party monitors and punishers. We examine whether network centrality of the third party increases efficiency of interaction. Furthermore, we decompose the efficiency increase into a monitoring channel (central third parties are valuable since they may influence reputations) and an enforcement channel (central third parties may be more able to punish without fear of retaliation). Assigning a third party at the 75th percentile of the centrality distribution (as compared to the 25th) increases efficiency by 21% relative to the mean: we attribute 2/5 of the effect to monitoring and 3/5 to enforcement. The largest efficiency increase occurs when senders and receivers are socially distant, unable to maintain efficient levels autonomously. Results cannot be explained by demographics such as elite status, caste, wealth or gender. Our findings show not every member is equally well-equipped to be part of a local institution. Knowing that a central third party observes their interaction increases sender-receiver efficiency. More importantly, to be able to punish someone, the third party must be important in the community.

In settings with poor formal contract enforcement, profitable investments are likely unrealized. While social closeness can mitigate contractual incompleteness, we examine how to improve the preponderance of cases where contracting parties cannot rely upon social ties. We ask if a community can enlist members to monitor transactions or punish offending parties. We conduct a laboratory experiment in 40 Indian villages, with 960 non-anonymized subjects, where we have social network data. Participants play modified sender-receiver investment games, with and without third-party monitors and punishers. We examine whether network centrality of the third party increases efficiency of interaction. Furthermore, we decompose the efficiency increase into a monitoring channel (central third parties are valuable since they may influence reputations) and an enforcement channel (central third parties may be more able to punish without fear of retaliation).Assigning a third party at the 75th percentile of the centrality distribution (as compared to the 25th) increases efficiency by 21% relative to the mean: we attribute 2/5 of the effect to monitoring and 3/5 to enforcement. The largest efficiency increase occurs when senders and receivers are socially distant, unable to maintain efficient levels autonomously. Results cannot be explained by demographics such as elite status, caste, wealth or gender.Our findings show not every member is equally well-equipped to be part of a local institution. Knowing that a central third party observes their interaction increases sender-receiver efficiency. More importantly, to be able to punish someone, the third party must be important in the community.Institutional subscribers to the NBER working paper series, and residents of developing countries may download this paper without additional charge at www.nber.org.

Traditional leaders are often given sub-national territorial authority in developing democracies. Though ubiquitous, the political consequences of their power has received little rigorous attention. We argue that such traditional leaders, whose power depends on the state, may be incentivized to strategically support political parties who can guarantee their survival and provide them with rents. We study this quid pro quo in the Apartheid-era Bantustans of South Africa. We show that an alignment between the state party and the chiefs maps to increased political support for the party. Further, we provide quantitative evidence consistent with chiefs acting as clientelistic brokers. Our results suggest that chiefs boost African National Congress (ANC) vote-share by 8.2 percentage points in the Bantustans. This translates into roughly 4.5% of the ANC's total vote-share, and a distortion in the national vote of 2.5 percentage points. This distortion is pivotal in determining whether the ANC is able to alter South Africa's constitution.

In the absence of strong formal institutions, social networks play a significant role in contract enforcement and in determining the scope of co-investment. Social norms and rules of thumb may dictate how friends vs. strangers decide to share an economic surplus or how popular vs. socially isolated individuals sustain cooperation with others. We shed light on the effects of network characteristics on investment decisions through a framed field experiment. Our laboratory protocol builds on a basic two-party trust game with a sender and receiver. In some treatments, we introduce third-party monitors or punishers that may or may not be identifiable by the other two participants. We find that the social network interacts with the play of the game in economically meaningful ways. First, social proximity mitigates the co-investment problem. Both senders and receivers in socially close pairs make larger transfers to each other. Second, while on average, third-party punishment decreases the size of investments made by the sender, socially important punishers are efficiency enhancing. Third, characteristics such as caste and elite status affect play, but in ways distinct from network centrality. Elites benefit from higher partner transfers, but do not use their status to increase total surplus. Finally, we use our results to provide the first assessment of institutional structures as a function of network shape. Typically, socially far judges encourage efficient behavior, while socially close judges are prone to collusion. However, high centrality judges are able to resist proximity-based collusion and restore high levels of investment.