with Daniel Gallardo Albarrán

The literature has long studied the economic consequences of epidemics. While plague and influenza episodes have received substantial attention by researchers, the same does not apply to cholera outbreaks. We examine the economic impact of cholera outbreaks in urban Prussia during the 19th century. We have constructed a dataset containing almost one thousand cities with detailed information on cholera mortality and population growth, which we use to proxy for economic development. Our results show that cities experiencing an additional one-percentage-point increase in cholera mortality grew ca. 0.5 percent less right after an epidemic. This size effect is representative of the three pandemic periods we consider, namely 1831-1837, 1848-1859 and 1865-1874. In addition, we find that post-epidemic growth recovery was rather fast during the second and fourth pandemics, but somewhat sluggish during the third one. We consider three recovery mechanisms-- access to transport infrastructures, prevalence of rural serfdom and degree of industrialization-- and find only partial support for the idea that access to the transport network helped mitigate some of the worst effects of the epidemics. The same applies to the link between rates of labour emancipation and post-epidemic growth. Most importantly, we find that industrialized cities suffered comparatively more from labour scarcity than less industrial ones. The skilled labour they demanded was relatively scarce, particularly during several mass waves of Transatlantic migration that coincided with the cholera years.

This paper presents the first comprehensive quantitative account of epidemic cholera in 19th-century Germany. Using a new dataset based on archival sources, it documents nearly half a million cholera deaths, along with outbreak timing and population at risk, across 2,685 cities and 852 rural counties within the 1871 German Empire. I document five stylized facts: First, cholera was primarily an urban disease, with city death rates averaging 3.5 times higher than in rural areas. Second, mid-sized cities (1,000-3,000 inhabitants) were the most severely affected. Third, cholera's geographic epicenter focused on the less developed North-East territories (Central Poland), but shifted South-West over time. Fourth, outbreaks spread more rapidly across regions and within cities over time, despite declining overall mortality. Fifth, local epidemics converged in severity across locations but became more spatially clustered over time. Understanding these complex patterns requires analysis of cholera’s interaction with dominant trends of 19th-century Western development, including public health reforms, urbanization, market integration, and political change. While the rich cholera historiography has long recognized these links, it merits greater attention from quantitative social scientists, including economic historians. Datasets like this one are the foundation for that engagement.

This study of Berlin’s nineteenth-century hydrological reconfiguration challenges the popular narrative that Europe’s recurrent cholera epidemics after 1831 motivated an efficient and scientific path to safe tap water supplies and sanitary sewers. Analysis of contemporary discourse reveals that the causal link between cholera and water infrastructure reform is weak and ambiguous, with aesthetic and industrial concerns taking precedence. Far from prompting effective action, miasmatic frameworks and the limited proto-epidemiological tools of the pre-bacteriological era led to inefficient and sometimes counterproductive approaches, potentially worsening urban mortality. Understanding Western Europe’s sanitary past provides important insights into the ongoing struggle for urban sanitation in today’s developing world. 

with Thilo N. H. Albers

Historical city directories are rich sources of micro-geographic data.  They provide information on the location of households and firms and their occupations and industries,  respectively.  We develop a generic algorithmic workflow that converts scans of them into geo- and status-referenced household-level  data  sets.   Applying  the  work  flow  to  our  case  study,  the  Berlin 1880 directory, adds idiosyncratic challenges that should make automation less attractive.  Yet, employing an administrative benchmark data set on household counts, incomes, and income distributions across more than 200 census tracts, we show that semi-automatic referencing yields results very similar to those from labour-intensive manual referencing. Finally, we discuss potential applications in economic history and beyond.

with Matthias Eckardt and Nikolaus Wolf

Temporary shocks can have persistent effects on the distribution of economic activity across urban space, suggesting considerable “stickiness” in adjustment to spatial equilibrium. I examine the long-run effect of World War II air raids on Berlin’s contemporary population density profile, exploiting random variation in the extent of damage caused at the scale of street blocks, while controlling for a flexible, secular time trend in the citywide density gradient. 70 years after the end of the war, blocks affected by irreparable damage feature significantly lower population density, while repairable damage had no lasting effect. These findings are consistent with a vintage effect, where large fixed costs associated with the construction of new buildings retard adjustment of the housing supply to growing demand for low-density structures.

with Daniel Gallardo Albarrán

European countries paved the way for modern economic growth during the nineteenth century with large-scale reforms facilitating human capital accumulation. The literature has investigated the drivers of this process by looking at the role of elites in broad education investments. However, less attention has been devoted to understanding the provision of infrastructures promoting workers' health, a key component of human capital at the time. This paper tests the hypothesis whether capital-labour production complementarities in the production process compelled economic and political elites to invest in public health goods. We relate the extraordinary and sudden decline in working-age population caused by the 1866 cholera epidemic in Germany to subsequent construction of piped water supplies. Our results show that cities affected by the epidemic were approximately 6 percentage points more likely to build waterworks compared to cities without the epidemic. An instrumental variable approach supports the causal interpretation of our results. In addition, we find that capital-skill complementarities were an important driver of investment in health infrastructures. We show that commercial cities, employing more valuable workers, exposed to cholera were more likely to build waterworks. Also, shocked places with a greater share of literate workers invested in these systems earlier than their counterparts with lower levels of human capital, particularly in the first two decades after the epidemic.

During the second half of the 19th century, the incidence of infectious and epidemic diseases declined dramatically across Western cities. While contemporary observers agreed that changes to the social geography of densifying cities were a fundamental driver of this urban health transition, the spread of mixed-income housing gave rise to controversial debate. Proponents of social mixing anticipated the New Urbanist case for local desegregation, expecting substantial benefits from the transmission of health-conducive norms, shared use of sanitary infrastructure, and access to informal help-in-kind networks within diverse buildings. Critics warned that mixed-income housing undermined traditional social barriers and exposes residents to a more diverse set of disease vectors, thus facilitating the spread of epidemics across all classes of urban society. While the social mixing controversy played a central role in the 19th century housing reform discourse, the interplay of increasing diversity and declining epidemic incidence during the urban health transition has not been systematically explored. 

Filling this gap, this paper provides causal estimates of the effect of social diversity on epidemic incidence from the universe of buildings in Berlin in 1866 – the year of the most lethal cholera outbreak in what was arguably 19th century Europe's most diverse urban environment. Adaption of the biological concept of species diversity and a mapping of the occupations of roughly 89,000 household heads listed in the 1866 town directory to a social stratification scale provide a building-level measure of social diversity. Conditional on a wide range of risk factors, and exploiting exogenous variation from building lots' geometric properties in a non-linear control function approach, social diversity has an ambiguous effect on cholera incidence: A one-standard-deviation increase in diversity results in a 26 percentage points higher risk of initial incidence (which is about the citywide a priori probability of incidence, 26.5 %), and a .5 percentage points drop in the case rate among affected buildings (11 % of the mean case rate, 4.5 %). These competing margins roughly cancel each other out on average, resulting in a close-to-zero overall effect. Results are robust to a variety of sensitivity checks, controls for spatial auto-correlation and alternative measures of diversity. 

A plausible interpretation is that diversity increases a residential community's probability to contract infectious diseases from outside vectors, while also providing additional means to confine further in-house spread – meaning that both critics and supporters of social mixing had a point. A causal mediation framework quantifies the importance of four potential underlying mechanisms. Most importantly, mediation analysis reveals that diversity and the number of infected neighbors living close-by strongly interact in increasing the risk of initial infection, while bearing little effect along the intensive margin. In contrast, about 10 % of the incidence-reducing effect of diversity on the intensive margin is attributable to mediation via shared sanitary infrastructure in the form of building-wide tap water access.

During the late 19th and early 20th centuries, the roll-out of tap water and sewerage infrastructure contributed to a dramatic mortality decline in Western cities. However, recent studies show that substantial health benefits only materialized after both systems were in joint operation, whereas the isolated introduction of tap water infrastructure offered little benefit and could even increase mortality before the advent of sanitary sewers. This paper offers an explanation based on a negative externality arising from spatio-temporal disparity in the roll-out of tap water and sewerage networks within a city. Access to a cheap, reliable and pressurized fresh water source allows individuals to clean up their local environment by flushing away liquid and solid waste. While this results in direct health benefits, the emission of pathogens to streets, gutters, ground water acquirers and open water bodies imposes indirect costs on neighbors located along the spatial trajectory of the resulting waste flows. As a result, city-level mortality does not necessarily decrease monotonously in the roll-out of tap water infrastructure. Sanitary sewers neutralize the externality, thus yielding social benefits on top of any direct gains for connecting households. 

I test this mechanism using a newly constructed dataset of all roughly 14,300 buildings in Berlin, observed in 1875 and 1880 – a crucial five-year period during which the city's mortality rate dropped by almost 20 %, while the share of buildings connected to the tap water and sewerage networks increased from 59 to 84 % and 0 to 55 %, respectively. Based on a digital terrain model and its implied flow direction matrix, I trace the spatial trajectory of waste flows, thus directly measuring the impact of the health externality. In a difference-indifferences PPML specification, I find that a building’s crude death rate (CDR) drops by .085 and .168 percentage points after obtaining tap water and sewerage, respectively – accounting for 13 and 26 % of the total CDR decline between 1875 and 1880. On top of these direct benefits, a one-standard-deviation drop in waste flow exposure decreases the CDR by .067 percentage points or 11 % of the total CDR decline. I show that the implied aggregate net effects are consistent with recent city-level studies. Exploiting sharp cut-offs in the roll-out of sewerage across Berlin’s main water bodies, I report consistent results from a complimentary regression discontinuity design. 

Using building-level occupational data, I explore the distribution of direct benefits and indirect costs across the income distribution. In line with long-standing claims in the qualitative literature, I find that the benefits from pre-sewerage tap water infrastructure were biased in favor of the affluent, thus widening the socioeconomic health gap. Rather than closing this gap, the roll-out of sewerage further widened urban health inequality. Based on counter-factual analyses, I conclude that there was no feasible alternative policy that would have brought about similar average health benefits while reducing both socioeconomic health inequality and exposure to waste flows. The paper concludes with a discussion of the role of private and public sanitary infrastructure provision during the Western historical mortality transition.