Pol Cosentino

Trade, Commuting and City Structure (Job Market Paper)

Abstract: Cities are places where people commute to work and where goods are traded across space. While a large literature examines how lower commuting costs reshape cities, much less is known about within-city trade costs as a distinct force. This paper studies both channels using the construction of the Petite Ceinture railroad in nineteenth-century Paris, the world's first circular transit system, designed for both freight and passengers. Using newly digitized data on firms, population, rents, and transport networks spanning 1801 to 1906, I provide causal evidence that improved access to the railroad reshaped the spatial distribution of economic activities during this period. To quantify general equilibrium effects, I develop and calibrate a quantitative urban model in which within-city freight costs generate spatial variation in tradable goods prices, creating consumption-driven forces at the residence absent from canonical models. Counterfactuals show that removing the railroad would substantially reduce total population, consumption of tradables, and spatial specialization. Ignoring within-city freight costs leads to a 17.1% underestimation of the effects of transport infrastructure on urban structure and welfare.

[CESifo WP] [Online Appendix]

Commuting, Air Quality and Welfare 

Abstract: We study the welfare effects of public transport infrastructure investments in the Paris region, highlighting the role of local air quality improvements typically omitted by standard quantitative urban models (QUMs). We first provide reduced-form evidence that tramway expansions improve local air quality. We then develop a QUM with endogenous worker and firm location choice, transport mode choice, and local air pollution affecting amenities and productivity. Applying the model to the Grand Paris Express metro project, we find welfare gains of around 1.5%. We show that omitting the air quality channel leads to a severe underestimation of welfare gains and spatial skill sorting, given the substantial heterogeneity in pollution exposure and its valuation across skill groups.

with Pascale Champalaune

[Paper]

Revise and resubmit, The Economic Journal

Best paper by young researchers, Prix AFET - Fabrique de la Cité

Optimal Urban Transport Design in a Quantitative Spatial Model 

Abstract: Which metro design should a city adopt to maximise worker welfare? This paper develops a framework combining a quantitative urban model with a simulated annealing algorithm. The model features heterogeneous workers, multiple transport modes, and general equilibrium effects: as commuting costs fall, workers reallocate across residence and workplace and land prices adjust endogenously. The algorithm solves over one million spatial equilibria to sample (sub-)optimal metro designs. Applied to Paris in 1896, the results show that strategic metro stations are driven by location centrality, productivity, and amenities. Among the four competing 19th-century proposals, the adopted Fulgence Bienvenüe network (1896) achieves the highest welfare-efficiency score: by routing lines along Haussmann boulevards, it exploited cheap cut-and-cover construction to connect the city's most strategic locations at minimum cost.

Preliminary draft [Paper]

Boulevards and Barricades: Does Urban Structure Shape Social Unrest? 

Abstract: This paper studies how urban structure shapes social unrest. We develop a quantitative urban model in which a government trades off the economic benefits of concentrating workers against the revolutionary risk this concentration creates. We use Haussmann's renovation of Paris (1853—1870), which occurred between the 1848 revolution and the 1871 Paris Commune. Using newly digitized geo-referenced data on barricades, arrests, and firms, we show that the renovation targeted vulnerabilities revealed in 1848 by widening streets where barricades had appeared, improving military access to Places of Power, and displacing factories toward the periphery. These changes reshaped the geography of the 1871 uprising. Finally, bringing the model to the data, and comparing Haussmann's plan with thousands of counterfactual plans, we show that the preferences revealed by Haussmann's renovation put 60–80% of the weight on economic modernization and the rest on control of the city.

with Stephan Heblich and Gabriel Loumeau

Preliminary draft and slides available upon request

Beyond Neighborhood Boundaries: A Spatial Interaction Measure of Segregation

Abstract: This paper develops a flexible measure of neighborhood segregation defined by the joint distribution of match probabilities and match values across groups. The framework nests several existing segregation measures as special cases and admits a decomposition that separates the urban structure and the group size from actual residential sorting. Applying the measure to U.S. racial segregation, we show that the well-documented post-1960 decline in conventional segregation measures largely reflects urban sprawl and a shrinking in the White population share. Once these forces are netted out, residual racial sorting is flat or rising in most cities, with a clear divide between declining coastal and Northeastern metros and rising Southern and Rust Belt ones.

with Gabrielle Fack, Joachim Jarreau, Miren Lafourcade and Yanos Zylberberg

Preliminary draft soon