Conference Program

Causal Inference and the Study
of Conflict and State Building
Princeton University
Friday May 18, 2012
Corwin Hall Room 127

Breakfast: 8.30am

Introductory Remarks: 9.00 – 9.05 am  Kosuke Imai (Princeton) 
Panel 1: 9.05 – 10.35 am
Title: “Patterns of Corruption and Drug-related Crime in Mexican Politics”

Ana de la O (Yale)
Looking Beyond the Incumbent: The Effects of Exposing Corruption on Electoral Outcomes (with Alberto Chong, Dean Karlan, and Leonard Wantchekon)

Does information about rampant political corruption increase electoral participation and the support for challenger parties? Democratic theory assumes that offering more information to voters will enhance electoral accountability. However, if there is consistent evidence suggesting that voters punish corrupt incumbents, it is unclear whether this translates into increased support for challengers and higher political participation. We provide experimental evidence that information about copious corruption not only decreases incumbent support in local elections in Mexico, but also decreases voter turnout, challengers' votes, and erodes voters' identification with the party of the corrupt incumbent. Our results suggest that while flows of information are necessary, they may be insufficient to improve political accountability, since voters may respond to information by withdrawing from the political process. We conclude with a discussion of the institutional contexts that could allow increased access to information to promote government accountability.


Vidal Romero (ITAM)
The Enemy at Home: Exploring the Social Roots of Criminal Organizations in Mexico (with Beatriz Magaloni and Alberto Díaz-Cayeros)

There is anecdotic evidence of a growing presence of drug trafficking organizations (DTOs) in Mexican society and government organizations. This is a quite worrisome situation. To the degree that criminal organizations become increasingly rooted in society, citizens may voluntarily or coercible side with the criminals, making it much harder for State forces to effectively fight DTOs. How embedded in Mexican society are criminal organizations? What explains the variation in embeddedness of criminal organizations across localities? Do organized crime societal networks are constructed using coercive or benevolent strategies to the population? What are the effects on support for the fight against organized crime of DTOs’ embeddedness in society?
At this moment, the literature has not fully established the degree of encroachment of criminal organizations in Mexican society, the institutional mechanisms through which their influence is used, and the circumstances under which we should expect criminal organizations to be coercive or benevolent to citizens. Reliable survey data on the topic is scarce, in good part because the issues to be measured are highly sensitive. Either because of fear or sincere support, citizens tend not to provide information about their direct relationship with DTOs or about the presence of these organizations in society. This circumstance generates, for instance, that places with strong control of DTOs are observationally equivalent to places in which there is no presence of DTOs, since data from public opinion surveys should report in both cases an absence of DTOs’ activity. Measuring DTOs presence also generates severe problems at the ground level. Interviewers are set in significant danger when directly asking about criminal organizations’ activities.
To overcome this sort of measurement problems, we implement a series of list and frame experiments embedded in two rounds of the Survey on Public Safety and Governance, which is co-coordinated by the Office of the Mexican Presidency, Stanford University, the Center for U.S.-Mexican Studies at UCSD, and ITAM.

Disscussant: Carlos Velasco Rivera

Panel 2: 10.50 – 12.20 pm
Title: “Observational Studies and Causal Inference: Lessons from the Afghanistan War"

Jason Lyall (Yale)
Dynamic Coercion in Civil War: Evidence from Air Operations in Afghanistan

Can insurgents be deterred during civil war or, alternatively, are they essentially undeterrable, having already decided ex ante to pay the costs of taking up arms? Despite an enormous literature on interstate deterrence, there are surprisingly few studies of whether the lessons of deterrence translate to the civil war battlefield. I draw on a new dataset of airstrikes and shows of force---where bombing runs are simulated but no weapons are released---in Afghanistan (2006-11) to examine how insurgents respond to (threatened) coercion. Matching via a relational database is used to pair bombed and threatened villages with similar locations to isolate the causal effects of both coercion and threats on the location and timing of insurgent attacks. I also examine whether the effects of airstrikes are conditional on the nature of civilian casualties they inflict and underlying population characteristics such as ethnicity or prior exposure to violence.


Jake Shapiro (Princeton)
Cellular Communication and Political Contestation: Evidence from Afghanistan (with Joshua Blumenstock, Michael Callen, and Tarek Ghani)

The widespread introduction of mobile communications technology in developing countries has been widely cited as a boon for economic and political development. Technology optimists laude mobile telephony for everything from increasing market transparency, to enabling new anti-corruption measures, to facilitating protest movements against corrupt autocratic regimes. Yet technology can have a dark-side as well, by facilitating new forms of corruption and enabling criminals and insurgents to better coordinate their actions. We study the impact of changes in cellular coverage on political contestation in Afghanistan using novel data from the largest telecommunications network in the country. Drawing on variation in coverage driven by events in other regions of the country, we find that the greater communications lead to lower levels of terrorist violence, and higher levels of satisfaction with the government, but only in the most violent areas of the country.
Disscussant: Jaquilyn Waddell Boie

Lunch:12.30 pm – 1.50 pm


Panel 3:2.00 – 3.30 pm


Title: “State-building, Political Stability, and the legacy of Civil War in Africa”

Michael Gilligan (NYU)
Can Community Driven Development Programs Foster Social Cohesion after Civil War

This study will report findings from a randomized control trial in 24 previously war torn communities in Sudan. Twelve randomly chosen communities received a World Bank funded community driven development funding. Treated communities were convened and community members were asked to jointly complete “Community scorecards” about the infrastructure needs of the community. Communities were then given the infrastructure improvement of their choice from a menu of possible choices (schools, water projects, health centers among others).  Control communities were given no intervention.  Both treated and control communities were studied with standard household survey techniques and with more innovative behavioral game techniques to measure social cohesion. We are interested in whether CDD programs can enhance social cohesion according to both survey and behavioral measures in previously war torn communities. Causal identification is made possible by the randomized nature of the treatment.

Leonard Wantchekon (Princeton)
The Institutional Legacy of African Independence Movements (with Omar García-Ponce)

We investigate the long- term effect of the geography of anti- colonial insurgencies (c.1900-1960) on the nature of current democratic institutions and political behavior in Africa. We find that while rural insurgencies (e.g. Madagascar, Kenya, Cameroon, Somalia) tend to generate autocracies, urban insurgencies (e.g. Senegal, South Africa, Benin, Ghana) tend to lead to democratic institutions. This is because urban insurgencies are mass uprisings that create conditions for popular political participation and strong civil societies. In contrast, rural insurgencies are military organizations that tend to limit political rights and generate less inclusive governments. We provide evidence that the relationship between the nature of anti- colonial insurgencies and democratization causal, by using the proportion of a country covered by mountaneous terrain as an instrument for rural insurgency. The results indicate that democratization may result from. The legacy of historical events, specifically the forms of political dissent under colonial rule.

Disscussant: Scott Abramson

Panel 4:3.45 – 5.15 pm


Title: “Representation, Accountability, and the Provision of
Public Goods in Weak States”

Oeindrila Dube (NYU)
‘White Man's Burden’? Experimental Evidence on Foreigner Presence and Generosity (with Jacobus Cilliers and Bilal Siddiqi)

Can the presence of foreigners affect measured altruism by shaping expectations about foreign assistance?  We experimentally vary foreigner presence across behavioral games conducted in 60 communities in Sierra Leone, and assess its effect on player contributions. We find that foreigner presence substantially increases player generosity, by over 20 percent.  The treatment effect is larger in communities that have previously received little foreign assistance, and for players who believe that the game’s purpose is to distribute resources, either as payoffs to game participants or as aid to the community. Together, the results suggest that players display more altruism in the presence of foreigners because they expect generosity to be rewarded financially, but this expectation is lower with past experience receiving foreign assistance. This has implications for how we measure altruism, and aids our understanding of how incentives to secure foreign assistance affect behavior.

Fotini Christia (MIT)
"Following the Food Aid: Experimental Evidence on the Effect of Local Institutions on Governance Quality"

This study uses a field experiment in Afghanistan, which randomly assigned elected local institutions to examine whether the creation of such bodies leads to better governance compared to traditional elite structures.  As assessed through a food aid distribution, we find that while both traditional elites and elected ones prove similarly highly effective in targeting aid to the needy and vulnerable, there is some evidence that democratically elected institutions lower the likelihood of elite embezzlement of aid.  The effects are driven by villages where the elected elites are explicitly put in charge of the aid distribution. These villages prove better in terms of aid targeting as compared to villages where traditional elites are in charge and outperform those traditional villages in terms of lower embezzlement and higher participation. However, when such democratically elected structures are not explicitly mandated to take control over the allocation of aid resources the distribution proves more inefficient than in villages with just traditional elites leading to lower levels of participation, higher levels of nepotism, and higher levels of embezzlement. Such institutional competition is also at play when we mandate female participation. This suggests that in the context of aid delivery the existence of multiple institutional structures with no clear hierarchy can lead to competition and underperformance rather than to additional checks and balances that enhance efficiency.

Disscussant: Graeme Blair

Poster Session: 5.30 – 6.30 pm


Dinner (Witherspoon Grill): 7.00 – 9.00 pm