It is claimed that Democratic college professors cause their students to become more liberal. In this paper, we study whether there is scope for professors to do so. We link voter registration data to salary records from 33 state flagship universities to document the partisanship of college faculty and use transcript data from one flagship university to study the causal effect of instructors on student partisanship. College faculty are more likely to identify as Democrats than the country as a whole, especially in the humanities and social sciences. Students are mostly liberals when they enter college, become more liberal regardless of their major, and sort to courses where instructors share their political ideology. Exploiting plausibly random variation in when instructors teach a given course, we do not find a relationship between faculty partisanship and changes in student partisanship, and are able to rule out even small liberalizing effects, including when we restrict to courses in the humanities and social sciences. To understand these results, we study variation in the frequency of left-leaning course topics. Liberal-oriented topics are not featured at higher rates when departments have more Democrats or when Democratic instructors teach a given course, meaning student sorting leaves little room for indoctrination.

We study how coca production affects the economic outcomes of Peruvian areas exposed to it. To address endogeneity concerns, we use geographic variation in coca suitability combined with exogenous time variation in coca prices generated by eradication policies in Colombia. We combine night-time light intensity, coca density maps from the UNODC, and geolocated household survey data. We find suggestive evidence that areas exposed to coca production experienced increased economic activity from 2003 to 2017.

This paper investigates the (in)efficiency of humanitarian food aid in refugee economies, using the Rohingya refugee crisis as a case study. We exploit the quasi-natural experiment provided by the World Food Program (WFP)'s transition from in-kind food aid to e-vouchers, which can only be redeemed at WFP stores within camps. The results show households engage in food aid trade (bartering and reselling) under both aid regimes, with e-voucher holders trading significantly less.

We design and implement an in-the-field survey to understand the prevailing social norm of hosts with respect to refugees by eliciting first and second order normative beliefs in the Rohingya refugee-receiving district of Cox's Bazar, Bangladesh.

Obtaining representative information on hosts and displaced populations in a single survey is not straightforward. This paper demonstrates the value of combining traditional and nontraditional sampling frames, geospatial information, and listing exercises to design a representative survey of hosts and Rohingya displaced populations in Cox's Bazar, Bangladesh. The paper applies innovative segmentation techniques using geospatial data to delimit enumeration areas in the absence of updated cartography. The paper also highlights the importance of listing exercises to inform stratification decisions and update population counts.

Bangladesh has continued to make remarkable progress in reducing poverty since 2010. In some regards, poverty reduction has continued in a manner consistent with the previous decade. However, important differences emerge when trends are examined more closely. Poverty rates in the poorer West and richer East converged until 2010, then diverged, as poverty reduction in the poorer Western divisions again started to lag. This paper uses decomposition analysis to examine the changing nature of poverty reduction from 2005 to 2010 and 2010 to 2016, with four key insights on fertility, education, structural change, and agricultural conditions.

Bangladesh has documented consistent reductions in poverty since 2000 and has also seen considerable transformation in the sector and location of economic activities. This paper exploits variation in sectoral growth and migration across districts and time to examine whether spatial variation in sectoral growth patterns can explain spatial variation in poverty reduction. We find that reductions in poverty were largest in places where agricultural output growth was highest and where industrial growth was highest, and that poverty reduction was greater in districts sending larger numbers of international migrants.