Skip to main content
Open Access Publications from the University of California

The CEGA Working Paper Series showcases ongoing and completed research by CEGA staff, affiliates, visiting fellows, and CEGA-supported project publications authors. CEGA Working Papers employ rigorous evaluation techniques to measure the impact of large-scale social and economic development programs, among other research designs, and are intended to encourage discussion and feedback from the global development community.

Cover page of Political Power, Elite Control, and Long-Run Development: Evidence from Brazil

Political Power, Elite Control, and Long-Run Development: Evidence from Brazil


This paper analyzes how changes in the concentration of political power affect long-run development. We study Brazil’s military dictatorship whose rise to power dramatically altered the distribution of power of local political elites. We document that municipalities that were more politically concentrated prior to the dictatorship in the 1960s are relatively richer in 2000, despite being poorer initially. Our evidence suggests that this reversal of fortune was the result of the military’s policies aimed at undermining the power of traditional elites. These policies increased political competition locally, which ultimately led to better governance, more public goods, and higher income levels.

Cover page of Identifying Psychological Trauma among Syrian Refugee Children for Early Intervention: Analyzing Digitized Drawings using Machine Learning

Identifying Psychological Trauma among Syrian Refugee Children for Early Intervention: Analyzing Digitized Drawings using Machine Learning


Nearly 5.6 million Syrian refugees were displaced by the country’s civil war, of which 50% percent are children. Given the heightened risks of psychological distress for this population it is critical to efficiently and accurately assess well-being for this population for intervention. A digital analysis of features in children’s drawings potentially represents a rapid, cost-effective, and non-invasive method for collecting individual and aggregate data on children’s mental health. Using data collected from free drawings and self-portraits from over 2,500 Syrian refugee children in Jordan across two distinct datasets, we use regression and Lasso machine-learning techniques to understand the relationship between exposure to violence and different measures of psychological trauma. Our results suggest that individual drawing characteristics are strongly correlated with validated measures of psychological trauma and past exposure to violence, with child mental health declining with increased exposure to violence and improving with resettlement in host communities. These results serve as a proof-of-concept for the potential use of children’s drawings as a diagnostic tool in human crisis settings.

Cover page of Should Consumption Sub-aggregates Be Used to Measure Poverty?

Should Consumption Sub-aggregates Be Used to Measure Poverty?


Frequent measurement of poverty is challenging, as measurement often relies on complex and expensive expenditure surveys that try to measure expenditures on a comprehensive consumption aggregate. This paper investigates the use of consumption “sub-aggregates” instead. The use of consumption subaggregates is theoretically justified if and only if all Engel curves are linear for any realization of prices. This is very stringent. However, it may be possible to empirically identify certain goods that happen to have linear Engel curves given prevailing prices, and when the effect of price changes is small, such a sub-aggregate might work in practice. The paper constructs such linear sub-aggregates using data from Rwanda, Tanzania, and Uganda. The findings show that using sub-aggregates is ill-advised in practice as well as in theory. This raises questions about the consistency of the poverty tracking efforts currently applied across countries, since obtaining exhaustive consumption measures remains an unmet challenge.

Manipulation-Proof Machine Learning


An increasing number of decisions are guided by machine learning algorithms. In many settings, from consumer credit to criminal justice, those decisions are made by applying an estimator to data on an individual’s observed behavior. But when consequential decisions are encoded in rules, individuals may strategically alter their behavior to achieve desired outcomes. This paper develops a new class of estimator that is stable under manipulation, even when the decision rule is fully transparent. We explicitly model the costs of manipulating different behaviors, and identify decision rules that are stable in equilibrium. Through a large field experiment in Kenya, we show that decision rules estimated with our strategy-robust method outperform those based on standard supervised learning approaches.

The Impact of Cleft Lip/Palate and Surgical Intervention on Adolescent Life Outcomes: Evidence from Operation Smile in India


Cleft Lip/Palate (CLP) is a congenital orofacial anomaly appearing in approximately one in 700 births worldwide. While in high-income countries CLP is normally addressed surgically during infancy, in developing countries CLP is often left unoperated, potentially impacting multiple dimensions of life quality. Previous research has frequently compared CLP outcomes to those of the general population. But because local environmental and genetic factors both contribute to the risk of CLP and also may influence life outcomes, such studies may present a downward bias in estimates of both CLP status and restorative surgery. Working with the non- profit organization Operation Smile, this research uses quasi-experimental causal methods on a novel data set of 1,118 Indian children to study the impact of CLP status and CLP correction on the physical, psychological, and social well-being of Indian teenagers. Our results indicate that adolescents with median-level CLP severity show statistically significant losses in indices of speech quality (-1.55), academic and cognitive ability (-0.43), physical well-being (-0.35), psychological well-being (-0.23), and social inclusion (-0.35). We find that CLP surgery improves speech if carried out at an early age, and that it significantly restores social inclusion.

How Important is the Yellow Pages? Experimental Evidence from Tanzania


Mobile phones reduce the cost of communicating with existing social contacts, but do not eliminate frictions in forming new relationships. We report the findings of a twosided randomized control trial in central Tanzania, centered on the production and distribution of a ”Yellow Pages” phone directory with contact information for local enterprises. Enterprises randomly assigned to be listed in the directory receive more business calls, make greater use of mobile money, and are more likely to employ workers. There is evidence of positive spillovers, as both listed and unlisted enterprises in treatment villages experience significant increases in sales relative to a pure control group. Households randomly assigned to receive copies of the directory make greater use their phones for farming, are more likely to rent land and hire labor, have lower rates of crop failure, and sell crops for weakly higher prices. Willingness-to-pay to be listed in future directories is significantly higher for treated enterprises.

Can Social Protection Reduce Environmental Damages?


Why do damages from changes in environmental quality differ across and within countries? Causal investigation of this question has been challenging because differences may stem from heterogeneity in cumulative exposure or differences in socioeconomic factors such as income. We revisit the temperature-violence relationship and show that cash transfers attenuate one-half to two-thirds of the effects of higher same-day temperatures on homicides. Our results not only demonstrate causally that income can explain much of the heterogeneity in the marginal effects of higher temperatures, but also imply that social protection programs can help the poor adapt to rising temperatures.

Torture as a method of criminal prosecution: Police Brutality, the Militarization of Security and the Reform of Inquisitorial Criminal Justice in Mexico


How can societies restrain their coercive institutions and transition to a more humane criminal justice system? We argue that two main factors explain why torture can persist as a generalized practice in democratic societies: weak institutional protections of the rights of criminal suspects and the militarization of policing, which leads the police to act as if their job were to occupy a war zone. With the use of a large survey of the Mexican prison population and leveraging the date and place of arrest, this paper provides valid causal evidence about how these two explanatory variables shape torture. Our paper provides a grim picture of the survival of authoritarian policing practices in democracies. It also provides novel evidence of the extent to which the abolition of inquisitorial criminal justice institutions - a remnant of colonial legacies and a common trend in the region - has worked to restrain police brutality.

Do Campaign Contribution Limits Curb the Influence of Money in Politics?


Over 40% of countries around the world have adopted limits on campaign contributions to curb the influence of money in politics. Yet, we have limited knowledge on whether and how these limits achieve this goal. With a regression discontinuity design that uses institutional rules on contribution limits in Colombian municipalities, we show that looser limits increase the number and value of public contracts assigned to the winning candidate’s donors. The evidence suggests that this is explained by looser limits concentrating influence over the elected candidate among top donors and not by a reduction in electoral competition or changes in who runs for office. We further show that looser limits worsen the performance of donor-managed contracts: they are more likely to run over costs and require time extensions. Overall, this paper demonstrates a direct link between campaign contribution limits, donor kickbacks, and worse government contract performance.

Political Quotas and Governance


Reserving political office for members of a particular, usually disadvantaged, group is a common form of political quota in many parts of the world. This has been shown to improve distributional access in favour of reserved groups, but often conjectured (and shown) to come at the cost of governance quality. We develop the first theoretical model to demonstrate the opposite possibility; a reduction in political competition - due to office being restricted to members of a pre-designated group - can improve governance. The model establishes a tight set of predictions regarding when improvements should be expected to occur, and when not. Such predictions are not yielded by alternative theories of political competition, are a priori unlikely to occur by chance, and have never been investigated in the large empirical literature on the effects of political reservations. We first show, in a Maharashtrian sample of rural villages, that governance outcomes dramatically increase under reservations. This is the first such effect documented in the literature. We then demonstrate a non-uniform pattern of improvement that lines up precisely with the predictions of the theory developed here.