Macro-Spatial
Economics

Motivating the course with a compelling empirical puzzle: why don’t workers move to high-wage cities? Starting from U.S. wage dispersion across cities and housing price growth, we build the spatial equilibrium model and show that spatial misallocation — driven by housing supply constraints — reduces aggregate U.S. GDP by 36% Or does it?.

Empirical facts about the spatial economy: wage dispersion, trade flows, city sizes. Two seminal models — Roback (1982) on spatial equilibrium and Krugman (1991) on agglomeration — combined into a simple quantitative framework showing how gravity equations emerge naturally from equilibrium conditions. (Allen & Arkolakis, Sections 1–3.)

The general workhorse spatial model nesting Roback and Krugman as special cases. We study existence, uniqueness, and comparative statics of spatial equilibrium, then introduce exact hat algebra — the key tool for computing counterfactuals without full calibration. (Allen & Arkolakis, Sections 4–6.)

Taking the quantitative spatial framework to data: gravity estimation in practice, exact hat algebra for counterfactual analysis, identifying structural elasticities, and policy applications including infrastructure investment, trade liberalisation, and place-based policy evaluation. Technical capstone of the Allen & Arkolakis sequence. (Sections 7 to end.)

When a city becomes more productive, how large are the welfare gains and who captures them? Urban welfare accounting from Desmet & Rossi-Hansberg, then the frontier general-equilibrium incidence results from Hornbeck & Moretti — wages, rents, inequality — with a focus on how housing tenure (renters vs. homeowners) determines who bears the costs. Closes the first arc of the course.

How do workers adjust to local labor demand shocks, and why is adjustment incomplete? Monte, Redding & Rossi-Hansberg model commuting as the fast margin and migration as the slow margin — complements, not substitutes. Notowidigdo shows why poor regions retain population despite negative shocks: housing prices and government transfers limit outmigration.

The spatial econometric toolkit for geographically structured data: spatial dependence and autocorrelation, weight matrices (Queen/Rook contiguity, distance-based), Moran’s I tests, and a taxonomy of spatial regression models (SLM, SDM, SEM). Special focus on decomposing direct vs. indirect marginal effects. Hands-on R session with sf and spdep.

Housing as the primary transmission channel between local shocks and aggregate fluctuations. Beraja, Hurst & Ospina use regional wage–employment variation to show wages are more flexible than aggregate data suggest and that demand shocks drove the Great Recession employment collapse. Mian, Rao & Sufi document the ZIP-level housing wealth collapse and its three amplification channels.

Why static spatial equilibrium is insufficient: transition dynamics, forward-looking migration, and homeownership as a capital investment. Greaney, Parkhomenko & Van Nieuwerburgh show how housing-as-asset changes welfare calculations — steady-state comparisons are not enough. Caliendo, Dvorkin & Parro provide empirical discipline: adjustment to the China trade shock takes decades.

The course ends at the live research frontier. Guaitoli, Pancrazi & Raimondo present an OLG spatial model with intergenerational congestion, building directly on the Greaney et al. dynamic framework. Fajgelbaum & Gaubert close the normative arc: what does optimal spatial taxation look like when worker sorting and agglomeration are endogenous?