Publications & Accepted papers


(conditionally accepted, Journal of the European Economic Association)

This paper examines the short and long-run effects of emigration on human capital accumulation at origin. I focus on the Galician diaspora, one of the largest emigration episodes of the 20th century in Europe. I construct a database of all municipalities combining newly-digitized historical data with contemporary census and survey data, and exploit variation in pioneering emigration and a pull factor for identification. I find that while emigration depressed literacy rates at origin in the short run, its impact became positive after one decade and led to gains in human capital that persist one hundred years later. I explore several channels and provide evidence of two novel mechanisms through which emigration can raise human capital in the long run. Galician migrants funded associations that financed the construction of schools in their hometowns and diffused norms conducive to a persistent change in beliefs about the value of education.


This paper examines the evolution and determinants of migrant self-selection by gender during the age of mass migration. Using newly digitised district-level data on emigration from Portugal, we construct a literacy-based self-selection index and document three stylized facts: (i) average self-selection was positive throughout 1885-1930; (ii) men were more positively self-selected than women during 1885-1915 but similarly thereafter; (iii) aggregate self-selection measures mask substantial variation across districts and time.  Our econometric analysis shows that self-selection was negatively associated with both the size of migrant flows and the share of migrants going to Brazil, while emigration to Africa was related to a more positive selection. The decreasing importance of flows to Brazil in favor of Europe after 1915 may partly explain the vanishing of differences in self-selection between men and women.


Working Papers


This paper analyzes how internal migration may lead to the diffusion of fertility norms across distinct groups. We exploit a unique historical experiment, Indonesia’s transmigration program, which relocated millions of people from Java and Bali (i.e., transmigrants) to the outer islands in the 1980s. This program created hundreds of new villages in which transmigrants and local natives mixed, creating quasi-random exposure to different fertility norms. Relying on various population censuses, we find strong evidence of cultural diffusion from transmigrants to natives but not the other way (i.e., asymmetric transmission). On average, exposure to transmigrants holding fertility norms equivalent to one fewer child led to about 0.5 fewer children among young local native women. Our results show that transmission occurs through exposure at various levels (e.g., immediate neighbors, block, and village level) and prestige-biased transmission plays a major role (i.e., exposure to transmigrants holding prestigious and visible occupations). We explore and rule out several alternative mechanisms such as the influence of economic factors, mixing through inter-marriages, and the diffusion of information on family planning.


This paper exploits a large resettlement policy to provide causal evidence on how migrant networks and diversity affect economic outcomes in a rural setting. Between 1932 and 1941, the Italian government relocated about 4,000 families to the Pontine Marshes, a previously uninhabitable area undergoing a massive reclamation project. The settlers came from different regions of Italy and were quasi-randomly assigned to plots of land,  creating exogenous variation in neighborhood composition. Using comprehensive data on the universe of settlers, we find that larger networks (i.e., neighbors from the same province of origin) increased the likelihood of long-term settlement and land purchase. This effect partially operated through higher productivity facilitated by cooperation and solidarity among network members. While greater diversity hindered economic performance, this effect is fully explained by its inverse relationship with network size. Yet, for the most isolated settlers, our results suggest that diversity may have also led to more solidarity between settlers. Finally, we explore and rule out a potential role for the diffusion of agricultural knowledge.


Migrants’ experiences abroad can shape the information they share with their families left behind, having profound implications in these communities. This paper leverages unique historical data to provide novel evidence on the content of migrants’ communications, the factors that shape them, and their potential effects at origin. Focusing on Irish migrants in North America (US and Canada), we construct a unique dataset of over 6,000 letters from the 19th and early 20th centuries. We identify key themes in the letters using Large Language Models (LLMs) and characterize the writers’ local environment using census data. First, we examine how individual and local factors -such as gender, religion, diaspora size, or discrimination- influence the presence and salience of economic, religious, and political themes. Next, we turn our attention to religion, given its prominent role in Irish identity and the anti-Catholicism faced by migrants. We exploit the staggered expansion of the Catholic Church in North America to assess its causal impact on religious content in the letters. Finally, we use full-count Irish census data to investigate whether this unexpected increase in religiosity manifested in migrants' letters contributed to a revival of the Catholic faith in migrants' home communities. 


Sports teams can be a powerful source of social and political identification, shaping individuals' beliefs and behavior. This paper empirically examines how the emotional impact of sports affects identity, preferences, and political outcomes. I focus on Football Club Barcelona (FCB), a team historically linked to Catalan identity and with huge popularity. Using survey data and exploiting the quasi-random nature of FCB performance, I find that winning major titles at the end of the season significantly strengthens Catalan identification and support for secession for several weeks. The outcomes of matches during the season have only short-lived effects, with defeats reducing identification and secessionist preferences for one day. These effects extend to electoral outcomes. Using electoral data at the municipality level from the past 50 years, I find that FCB defeats one day before an election reduce the vote share for pro-Catalan identity parties. The findings suggest that these effects are mediated by changes in euphoria and optimism, as FCB performance influences economic expectations, political prospects, and trust in government.


Work in Progress