What social and institutional legacies do proxy wars leave for the groups that fought as proxies? In 1962, the Geneva Agreement declared Laos neutral, barring any foreign military presence. To circumvent this amid the perceived threat of advancing communism in Southeast Asia, the CIA covertly armed the Hmong—an ethnic group living in the highland areas that the U.S Intelligence considered militarily strategic—to fight on its behalf. Using a difference-in-differences approach, I find that this alliance came at great cost to the Hmong, as evidenced by sharp falls in education that lasted a generation. Furthermore, after the war’s end in 1975, the victorious communist government unleashed a series of retaliatory policies. Ethnolinguistic maps before and after the war reveal sharp territorial losses for the Hmong, accompanied by steep declines in their political power. These losses appear to result from sustained postwar displacement, due both to initial flight as well as state-led resettlement programs, relocating Hmong villages to more legible regions. Lastly, recent survey data suggests that Hmong refugees in the U.S. who were directly exposed to the war are significantly more assertive of both their ethnic and American identities, pointing to a possible mechanism through which state suspicion may persist over time.

This paper evaluates a centrally-planned resettlement policy in rural Laos intended to modernize livelihoods and reduce poverty among ethnic groups traditionally practicing swidden (slash-and-burn) upland rice agriculture, where households cultivate dry-rice fields on forested slopes using rotational fallowing. By regrouping houses away from mountain slopes and into nearby valleys, the policy aimed to steer these societies toward permanent wet-rice paddies, considered technologically more advanced. Using the staggered timing of treatment, I find that the policy reduced village swidden plots by 30–40% five to ten years post-resettlement. However, ethnographic studies reveal that swidden practices are deeply interwoven with kinship ties and ethnic identities, and resistance to new methods reflects more than just unsuitable terrain. Using a triple-difference strategy—comparing traditional swidden-rice and wet-rice societies across regions with varying treatment intensity before and after resettlement—I find evidence of decreased well-being among those impacted: lower schooling attainment, increased child mortality, worse child health outcomes, and reduced household wealth. Exploring mechanisms, I find suggestive evidence that the policy only marginally shifted households into wet-rice cultivation; instead, these farmers reduced fallow duration as a possible coping response to limited swidden land in the new environment.

We study the establishment of U.S. National Laboratories in the 1940s--1950s to estimate local spillovers from public research infrastructure. This setting allows us to causally identify such spillovers, for two reasons: 1) Lab sites were chosen largely for security and political reasons, rather than existing or potential innovative capability and 2) We identify runner-up locations using archival sources. We find several types of knowledge spillovers: Compared to control counties, Lab counties experience large and persistent increases in patenting by non-lab inventors; non-lab patents in the same county shift toward laboratories' research fields and cite laboratory patents more frequently. Using newly digitized county data from 1936--1970, we find sustained increases in retail sales and household income. Linked 1940--1950 Census records show wage gains for pre-existing residents who remain in lab counties, with larger effects for college-educated workers. We find that cohorts exposed to laboratory establishment during school-age years attained more education, consistent with a human-capital channel. Spillovers arise despite extensive secrecy around early nuclear research, suggesting that co-location with public R&D can generate sizable local benefits even under restricted information flows.

This paper traces the origins of inequality to state expansion. Nagaland, a tribal region in India’s northeast, was incorporated into formal state administration in a staggered manner between 1881 and 1947—first under British rule and later by the Indian state. Using variation in years since annexation, I show that longer state exposure improves average well-being in modern times. However, accompanying these changes are increases in within-village wealth inequality, suggesting emerging hierarchies among individuals in the same compact community. The results are robust to an instrumental-variables strategy based on proximity to pre-annexation British tea estates, which influenced annexation timing through security concerns. Exploring channels of persistence, I find that in areas with longer state exposure, political elites hold substantially more assets relative to their constituents' mean incomes; in tandem, these areas also exhibit an expansion of public-sector employment. This pattern suggests the gradual emergence of reciprocal norms of patronage between a new elite class and the rest of society. 

This paper examines the impact of a forced agglomeration program in Mizoram, India, in response to a guerrilla uprising. Over 500 villages were relocated into one of 105 existing villages called “Grouping Centers” (GCs), while about 120 remained ungrouped. Using a difference-in-differences approach, this study finds lasting population divergence 20-30 years later and a reduction in agricultural employment in GCs, alongside increased labor-force participation and higher likelihood of having official urban status, suggesting long-term structural change.

This paper studies the interaction of modern and traditional institutions within a social identity framework in the context of a state in Northeast India. In turn, it presents an explanation for why people identify with ethnicity. Following the entry of an ethnic party in 2011, it finds that the willingness of co-ethnic voters to side with it depends on the degree to which the electoral arena is ethnically homogeneous or not. This is done exploiting three sources of variation: state elections (homogeneous) versus parliament elections (mixed), comparing two states (treated and control), before and after the party's entry. In the mixed election, voters in the treated state converge much more strongly in their distribution of candidate vote shares, in particular, to the ethnic party's representative candidate. No such effect is observed for the homogeneous elections. As an additional shock I use the boycott of the 2024 parliamentary elections by the competing ethnic group's candidates in what would have otherwise been a mixed election. When now faced with a homogeneous set of co-ethnic candidates, voting behavior returns to baseline levels. Two main conclusions follow: First, ethnic identities may be fairly instrumental, and that status-related goals deserve as much attention as material ones. Second, rather than crowd out parochial ethnic identities, modern electoral systems may instead intensify them.

This paper investigates the degree to which a shock to the gender distribution of income within the household influences spending decisions on consumption goods. Exploiting the temporal and spatial expansion of oil palm in Indonesia, a crop that favors men in the production process, we find evidence of widening incomes between spouses as men gain, relatively speaking. Women in more oil-palm intensive areas also drop out of the labor force. Looking at the composition of household expenditure, we then find decreases in shares allocated to education, while spending shares on food and alcohol/tobacco rise. Since relative income changes likely translate to relative power shifts, as predicted by the collective household model, we interpret our findings as resulting from a reallcoation of power within the household.

This paper studies the interaction of policy and traditional institutions in India's Manipur state by comparing tribes organized around relatively egalitarian power structures with tribes where local authority and land ownership is concentrated among chiefly clans. In response to the national MGNREGA program, tribes with the egalitarian system see larger economic gains of about 15%. Their share of agricultural employment also drops by 8.8 percentage points (~10%). I then find suggestive evidence that the results are likely driven by differences in program fund allocation. In the non-chief (egalitarian) regions, program expenditure per registered worker as well as wage per worker are substantially higher. Further, of the funds used, the non-chief regions also allocate a higher share towards non-agricultural projects.