Conferences & Seminars

Berkeley Center for Economics and Politics

Poster Session

Friday, September 11, 5:30pm-6:30pm


The Impact of Minority Representation on Racial Violence
Abhay Aneja
Haas-BPP, UC Berkeley

This paper provides evidence regarding the effects of political representation on violent crimes committed against historically marginalized minorities in India. Using data on the election of state legislators for 20 Indian states over 21 years, we examine if legislators from political parties that represent India’s lowest castes (the Untouchables as well as certain Indian tribes) can reduce the incidence of ethnic violence perpetrated against these populations. To ascertain the causal effect of political representation, we use the quasi-random variation that arises from close elections to instrument for the state-level proportion of low-caste legislators. We find that increasing state-level political representation for marginalized groups has the effect of reducing the incidence of violent crimes committed against these communities—in particular, murder, kidnapping, assault, rape, and hate crimes. We also show that a simple OLS regression comparing changing proportions of low-caste representatives across states greatly overstates the crime-reducing effect of this political representation. Our results are robust to several different specifications. We show that the reductions in ethnic violence do not appear to be the product of a potential overall crime reduction strategy adopted by representatives of low-caste political parties. Moreover, we also provide evidence on the channels through which political representation may impact crimes against low-caste populations. For example, we provide evidence that low-caste representatives may shift police practices with respect to ethnic violence against low-crime citizens, as evidenced by increasing increased rates of arrests. Moreover, we find evidence that low-caste representatives may deter low-caste violence by increasing the number of convictions for low-caste violence, as well as the speed with which such cases are processed by courts.


Rating government veterinarian quality: experimental evidence from Pakistan
Arman Rezaee
Economics, UC San Diego

Livestock agriculture accounts for twelve percent of GDP in Pakistan, and is essential for the rural poor. Artificial insemination (AI) is crucial to renewing livestock, but the AI market suffers from imperfect information about veterinarian quality. To overcome this inefficiency, we developed and implemented in a district of rural Punjab a novel cellular-based information clearinghouse, similar to  The clearinghouse measures, aggregates, and disseminates to farmers the success rate of government veterinarians in impregnating cows, an objective measure of veterinarian quality. Farmers treated with that information through a randomized control trial are 34 percent more likely than controls to return to a government veterinarian for AI. Among returning farmers, treated farmers have an 18 percentage point higher AI success rate. These benefits to rural farmers of improving service quality using the cellular network hold out hope for similar transparency enhancing interventions as the cellular network improves and becomes cheaper.

The Origin of the State: A Quantitative Test of Circumscription Theory
David Schönholzer
Economics, UC Berkeley

In a highly influential article in Science, Carneiro (1970) posits that early states were created in areas of high agricultural productivity that were circumscribed by unproductive land. This project assesses the validity of circumscription theory quantitatively. Circumscription intensity is calculated on a grid with 1/4 degree cell size (approximately 28 km at the equator) across the globe. Regression results show that circumscription intensity is a significant predictor of the locations of initial state formation, although there is substantial heterogeneity in the strength of the association across different regions. Similar archaeological evidence for chiefdoms shows no significant relationship with circumscription. The location of early states is then predicted using the regression model. This reveals some locations that are highly circumscribed but do not feature prominently in the archaeological record on early state creation.

Social Protection and Clientelism in Ghana
Rachel Strohm
Political Science, UC Berkeley

In low­ and middle­ income countries, it's commonly observed that poor people depend
on clientelistic relationships with politicians to access goods and services, while richer people are able to purchase these items on the market and opt out of clientelism if they wish. Many of these countries have recently begun to implement social protection programs which provide cash transfers to poor citizens. One implication is that raising the incomes of the poor will allow them to disengage from patronage networks. However, politicians may also be able to direct cash transfers to current supporters or use the promise of future transfers to gain new support, thus reinforcing their clientelistic networks. It is unclear ex ante which effect might predominate. In this poster, I will present the research design for an experimental study of the impact of social protection on patronage participation in Ghana.


Political Favor Exchange in Democracy
Ferenc Szucs
Economics, UC Berkeley

To investigate the mechanisms of political favor exchange we combine a model with evidence on favoritism in the advertising market in Hungary. Our model predicts favoritism and shows that it can be inefficient because politicians prefer agents with poor outside options from whom more rents can be extracted. Empirically, we show that (1) state-owned firms--but not private firms--dramatically tilt advertising under right-wing governments towards low-circulation right-wing newspaper; (2) each new government replaces the managers of these firms. Both findings are consistent with the model’s prediction on favor exchange, and finding (1) also with the prediction on inefficiency. Additional evidence shows that alternate explanations are unlikely. We also find that favors flow through indirect paths and that favoritism by different actors is synchronized, suggesting that in our context favor exchange is centralized.