Side Effects of Immunities: African Slavery in the US South 
(last version: pdf)
Why was slavery concentrated in the South of the United States? And, why were certain African populations extensively raided to be enslaved in those regions? The novel empirical evidence presented in this paper reveals that (i) malaria was a major determinant for the diffusion of African slavery in the southern United States and (ii) malaria-resistance made sub-Saharan Africans especially attractive for employment in these regions. I first document that African slavery was more widespread in those counties of the United States subject to higher malaria exposure. Moreover, I show that the introduction of a particularly virulent, and previously absent, strain of malaria into the US caused a large and rapid increase in the prevalence of African slave labor. Finally, by looking at the historical prices of African slaves in the Louisiana slave market, I .find that the more malaria-resistant slaves, born in the most malaria-ridden regions of Africa, commanded higher prices.

Previous Version: pdf

Malaria and Civil Violence: A Disaggregated Analysis  
(with Matteo Cervellati, Uwe Sunde and Simona Valmori)
Using geo-referenced data for Africa we investigate the role of temporary health shocks, related to the risk of malaria epidemics, for civil violence. Following the epidemiological literature, we identify climate conditions that exhibit an elevated risk for malaria outbreaks at the cell-year and cell-month level. Identification exploits the contrast of the occurrence of these conditions in cells with low malaria exposure and cells with high malaria exposure. In the low exposure areas, the population has lower genetic and acquired immunities, relative to cells with no or very high exposure. Consequently, it is hypothesized that the impact of suitable weather conditions on the likelihood of malaria outbreaks is amplified. Due to the short incubation time, malaria outbreaks can be identified at high temporal resolution. Exploiting within-cell variability both at the yearly and monthly frequencies, the analysis can account for unobserved heterogeneity by including a rich set of time and cell controls, including cell*year specific determinants of civil conflicts and seasonal patterns within the year. The results suggest that the occurrence of unusually suitable conditions for malaria outbreaks in cells with low malaria exposure significantly increases the likelihood of civil violence compared to cells with high malaria exposure. 

Bite and Divide: Malaria and Ethnolinguistic Diversity  
(with Matteo Cervellati and Giorgio Chiovelli)
We test the hypothesis that long term exposure to malaria is an important historical determinant of contemporary ethno-linguistic diversity. Using cell-level disaggregated data, we show that in areas where malaria incidence was higher we observe a larger number of ethnic and linguistic groups. Moreover, we document that the effect is already present when looking at the spatial distribution of pre-colonial African societies. To exclude the likelihood that the relationship is driven by some omitted features of tropical areas, since malaria was not present in the Americas before European settlement, we show that there is no effect of malaria on diversity when looking at the spatial distribution of pre-colonial American societies. In order to test that malaria exposure shaped the patterns of location (and isolation) of communities and limited the admixing with neighboring communities, we show that individuals in areas with high malaria exposure identify more strongly with their ethnic group and are less prone to out-group marriages. Finally, we document that populations from highly malaria-infested areas tend to live in segregated and more ethnically homogenous locations.



The Resource Curse in the Long-Run
(with Scott F. Abramson)  

This paper examines the long term impact of natural resource extraction. Focusing on the effect of coal, a resource crucial to processes of European industrialization, we estimate a substantial negative relationship between nineteenth century mining and contemporary levels of economic development.  As an exogenous source of variation in coal-extraction activities, we exploit variation in the underlying geological conditions of the region, which affect the likelihood of having coal-beds. Importantly, we find that there is no negative effect of coal extraction in the first half of the nineteenth century and that the negative effect arises only from the 1970s. We then examine a large number of conjectured mechanisms linking natural resources to lower levels of development and find no evidence of a political “resource curse” that operates through corruption, institutional quality, or conflict. Instead, we find that nineteenth-century coal mining has a persistent effect on present-day incomes by shaping the long-run structure of production; the economies of those regions where coal extraction took place continue to have lower levels of human capital and fewer workers employed in high-tech and scientifically oriented sectors.