Loading field map
Why do economic activities cluster in some places and not others? Why do cities form, and why do some regions grow rich while nearby areas stagnate? Spatial economics is the branch of urban economics that builds formal models to answer these questions. Over nearly two centuries, the subfield has developed five major frameworks, each offering a different answer to the same puzzle. The story of spatial economics is not a simple march of progress; it is a series of debates about what drives spatial concentration—transport costs, land rents, increasing returns, or micro-level data—and how best to model it.
The first systematic attempts to explain the location of economic activity emerged in the nineteenth and early twentieth centuries. Classical Location Theory, rooted in the work of Johann Heinrich von Thünen (1826) on agricultural land use and Alfred Weber (1909) on industrial location, treated location as a problem of minimizing transport and production costs. Von Thünen showed that different crops would be grown in concentric rings around a market town depending on their perishability and transport costs. Weber extended the logic to factories, asking where a firm should locate to minimize the combined costs of transporting raw materials and finished goods. These models were elegant and intuitive, but they assumed a featureless plain, perfect competition, and fixed production technologies. They could explain simple patterns—why heavy industry locates near raw materials—but they had little to say about the internal structure of cities or the emergence of large-scale agglomerations. By the mid-twentieth century, economists increasingly saw Classical Location Theory as too narrow: it treated space as a friction to be overcome rather than as a source of economic advantage in its own right.
In the 1960s, a new generation of economists brought the tools of neoclassical price theory to bear on urban space. Neoclassical Urban Economics, launched by William Alonso's 1964 book Location and Land Use, replaced the cost-minimization framework with a general equilibrium model of land markets. Alonso's bid-rent model showed how households and firms compete for land in a monocentric city, trading off commuting costs against the desire for larger lots near the center. The result was a formal, mathematically rigorous theory of urban spatial structure: land rents decline with distance from the central business district, and different activities sort into different rings. This framework absorbed the classical concern with transport costs but embedded it in a market-clearing equilibrium. It could explain why poor households often live near the city center (they have a high marginal cost of commuting) and why offices cluster downtown. Yet the monocentric model soon showed its limits. It assumed a single, exogenously given center, ignored increasing returns, and treated cities as isolated points rather than parts of a broader system. By the 1970s, researchers were pushing against these constraints.
New Urban Economics, which flourished from the 1970s through the 1990s, did not reject the neoclassical framework so much as broaden it. Where Neoclassical Urban Economics had focused on a single, stylized monocentric city, New Urban Economics introduced multiple centers, housing market dynamics, and a much richer set of urban features. It also shifted the subfield's center of gravity from pure theory toward empirical testing. Researchers like Edwin Mills and Bruce Hamilton developed models of suburbanization and polycentric cities, while others incorporated durable housing, zoning, and local public goods. This framework coexisted with its neoclassical predecessor, often using the same bid-rent logic but applying it to more realistic settings. Its distinctive contribution was to make spatial economics empirically accountable: rather than deducing city structure from first principles, New Urban Economics asked whether the models could match observed patterns of land use, commuting, and population density. The answer was often yes, but with important caveats. The monocentric model, for instance, predicted that employment would remain concentrated downtown, yet by the 1980s most U.S. jobs were in suburbs. New Urban Economics responded by building models of edge cities and decentralized employment, but it struggled to explain why such clusters formed in the first place. That question required a different kind of theory.
New Economic Geography (NEG), launched by Paul Krugman's 1991 paper "Increasing Returns and Economic Geography," broke sharply with the frameworks that preceded it. Classical and Neoclassical models had assumed constant returns to scale and perfect competition, which made it hard to explain why economic activity concentrates in a few locations. NEG introduced two key ingredients: increasing returns at the firm level (modeled through monopolistic competition) and transport costs. In Krugman's core-periphery model, firms want to locate near large markets to save on transport costs, but workers follow firms to those markets, creating a self-reinforcing cycle of agglomeration. The result is that even initially identical regions can diverge into a rich core and a poor periphery. This framework revived the classical interest in industrial location but gave it a rigorous microeconomic foundation. It also transformed the subfield's scale of analysis: where Neoclassical Urban Economics focused on the internal structure of a single city, NEG modeled entire systems of cities and regions. Its influence has been enormous, but it has also drawn criticism. Early NEG models were highly stylized, often assuming a single industry and a flat geography. They also struggled to incorporate the detailed urban features—housing markets, commuting, local public goods—that New Urban Economics had spent decades refining. By the late 1990s, a tension had emerged between the theoretical elegance of NEG and the empirical richness of the New Urban tradition.
Quantitative Urban Economics, which has become the dominant framework since the early 2000s, is best understood as a synthesis of the two preceding traditions. It takes from New Economic Geography the emphasis on increasing returns, trade-offs between scale and transport costs, and the goal of explaining the spatial distribution of economic activity across cities and regions. From New Urban Economics, it inherits a commitment to empirical rigor, detailed urban structure, and policy relevance. The key methodological innovation is structural estimation: researchers write down a full equilibrium model of a city or system of cities, then use micro-data on wages, rents, commuting times, and firm locations to estimate the model's parameters. This allows them to quantify the forces that drive agglomeration—what share of a city's productivity premium comes from density versus worker sorting versus local amenities—and to simulate the effects of policies like transport investments or zoning reforms. Pioneered by economists such as Donald Davis, David Weinstein, and Esteban Rossi-Hansberg, Quantitative Urban Economics has become the field's workhorse. It can answer questions that earlier frameworks could only pose: How much does a new subway line raise land values? Do place-based subsidies actually attract firms? Yet the synthesis is not complete. Quantitative models are computationally demanding and often rely on strong assumptions about functional forms and market structure. Critics argue that they sacrifice the theoretical transparency of NEG for empirical fit, and that their policy predictions are only as good as the assumptions buried in the estimation.
Today, New Economic Geography and Quantitative Urban Economics are the two leading frameworks, and they coexist in a productive but uneasy relationship. They agree on the fundamental point that increasing returns and transport costs are central to understanding spatial concentration. They also share a commitment to general equilibrium modeling: both frameworks insist that changes in one location ripple through the entire system of cities. Where they disagree is on method and scope. NEG theorists prioritize analytical tractability and the discovery of general principles—the conditions under which agglomeration occurs, the role of trade costs, the possibility of multiple equilibria. Quantitative Urban Economists prioritize empirical fit and policy relevance, often building models that are too complex for closed-form solutions but can be taken to data. The two frameworks also differ in their treatment of microfoundations: NEG models typically assume a single sector and a representative firm, while quantitative models often incorporate heterogeneous firms, workers, and housing markets. A third, smaller tradition—New Urban Economics—persists in applied work on housing and land use, but its theoretical ambitions have largely been absorbed by the quantitative approach. The central debate in the subfield today is not about whether agglomeration matters, but about how to measure its causes and consequences with enough precision to guide policy. That debate is unlikely to be settled soon, and it is precisely what keeps spatial economics a lively and contested field.
Spatial economics has traveled from cost-minimization diagrams to structural estimation of city systems. Each framework left something behind: Classical Location Theory's simplicity, Neoclassical Urban Economics' monocentric focus, New Urban Economics' empirical turn, New Economic Geography's theoretical elegance, and Quantitative Urban Economics' computational ambition. None has fully replaced the others. Instead, the subfield has accumulated a layered toolkit, with each layer responding to the limitations of what came before. The result is a field that can now ask—and sometimes answer—questions that von Thünen could not have imagined: Why do some cities grow while others shrink? How do trade and migration reshape regional inequality? And what happens to a city when a factory closes or a highway opens? The answers remain provisional, but the frameworks for finding them are richer than ever.