Research

Working papers

Job Market Paper: "Long-Term Environmental Effects of the Indonesian Transmigration Program" [link]

Abstract: Deforestation in Indonesia has been attributed to a variety of causes, with some pointing to the 20th-century voluntary mass resettlement - called Transmigration Program - of Javanese and Balinese individuals to the "Outer Islands" of the country, until then sparsely populated by indigenous groups. Using the Transmigration Census of 1998 together with Census data and a novel index of ecosystem integrity, this paper estimates the environmental externalities that the Indonesian Transmigration Program created on Indonesian forests. Results show that transmigrant villages are significantly more destructive to their ecosystems, especially when the share of indigenous people allowed to join is lower. There is also evidence that villages turned to more damaging palm oil plantations to cope with the low rice productivity of randomly assigned soil quality. Ethnic fractionalization has a strong positive impact on ecosystem integrity, a surprising result in contrast to some existing literature. This research suggests that economically vulnerable communities with little knowledge about their resources have a detrimental effect on the ecosystem, and integrating migrants with the local population in a new environment is crucial to ensure the sharing of information and the sustainability of the surrounding ecosystem.

"Can Minority Language Policy Work? Evidence from Wales" [link]

Abstract: Between 50 and 90% of the approximately 7,000 languages spoken today are expected to be extinct by 2100. While legal protection and funding for minority languages have increased globally, the benefits and costs of preserving linguistic diversity are unclear and hardly quantifiable. This paper uses a policy change regarding Welsh-language instruction in schools across Wales to evaluate the impact of compulsory minority language education on the dynamic of its prevalence. I use data from the UK Census to quantify the impact of years of exposure to the language in classes on the likelihood of being a Welsh speaker. Results show that the policy significantly expanded the number of speakers, increasing the likelihood of being a speaker by about 3.6% for those who underwent Welsh education. Results are much higher for individuals in school, with a significant drop in knowledge post-secondary education. Effects are also significantly higher for females (2.5%), in areas where Welsh was spoken by a majority before the policy, and in regions where less than 10% of people spoke the language. In absolute terms, the policy created around 64,287 speakers in twenty years, and the average cost to create a new speaker is estimated at £4,112.72.

"Strategic Geographic Indications: Mitigating Deforestation and Malnutrition in East Timor" [link]

Abstract: The paper develops a theoretical model and empirical framework to design a Geographical Indication (GI) for coffee in East Timor that maximizes economic benefits for smallholders while preserving rainforest. In the context of the “coffee paradox”—where consumer prices for coffee have risen globally even as farmers’ incomes remain stagnant—shade-grown coffee production generates both a private negative externality for farmers, in the form of food insecurity, and a social positive one, namely rainforest conservation. A theoretical two-sector model is constructed to capture the land-use trade-offs between subsistence food crops, coffee, and rainforest conservation. The model shows that without intervention, farmers facing immediate food needs and limited market access over-clear forest for low-value subsistence farming, whereas a social planner would favor conserving forest and maintaining coffee plants for long-term gain. I demonstrate that a GI policy—by raising the effective price of coffee—can align farmers’ incentives with the social optimum, reducing deforestation and improving incomes without subsidies or enforcement. Using cupping score data from national coffee quality competitions (2016–2022), I identify regions and elevations consistently scoring above the specialty threshold (80 points) as prime candidates for GI status. Empirical analysis confirms that elevation is a strong predictor of coffee quality: a 100-meter rise in altitude increases cupping scores by roughly 0.2 points, and only farms above 945 meters have a better-than-even chance of producing specialty-grade coffee. A potential benefit analysis suggests that implementing a coffee GI in Timor-Leste could generate substantial income gains (over \$1,600 per participating farmer annually) while safeguarding forest cover. These findings position GI policy as a viable, low-cost strategy for sustainable rural development in least-developed countries, provided it is underpinned by evidence-based geographic delineation of quality.

Works in Progress

"Let Me Think About It: Political Parties and Delays in FEMA Disaster Declarations" (with Pedro Vitale Simon)

"Caffeine or Calories? Coffee Agroforestry and Food Security in East Timor"

"Deforestation and Air Pollution in the Indonesian Rainforest" (with Xinhui Sun)

"Estimating the WTP for Indigenous Language Conservation in the U.S." (with Amy Ando)