Loading field map
Why do some regions grow rich while others stagnate? Why do industries concentrate in particular cities or states, and why do economic booms and busts spread unevenly across space? Regional economics emerged as a distinct field in the mid-twentieth century to answer these questions, and over the past seventy years it has developed a sequence of frameworks that alternately compete with and build on one another. The story of the field is not a simple march toward a single correct model; it is a series of debates about what drives spatial economic divergence and how best to measure it.
The earliest systematic thinking about regional economies came from location theory, a tradition stretching from Johann Heinrich von Thünen's work on agricultural land use in the 1820s through Alfred Weber's industrial location theory in the early 1900s and Walter Christaller's central place theory in the 1930s. Classical location theory asked a deceptively simple question: given transport costs, where will firms and households choose to locate? Its answer was a set of geometric models showing that economic activity arranges itself in predictable patterns around transport nodes and market centers. These models were elegant but static; they could explain the layout of a single industry or a single city, but they had little to say about how whole regions grow, decline, or interact over time.
In the 1950s, Walter Isard launched Regional Science as a deliberate attempt to overcome those limitations. Regional Science absorbed the insights of classical location theory—its transport-cost gradients and central-place hierarchies—but added tools from input-output analysis, demography, and regional accounting. Where location theory had offered a snapshot, Regional Science aimed for a moving picture: it modeled interindustry flows between regions, tracked population migration, and built the first quantitative descriptions of regional economic structure. Yet Regional Science remained largely descriptive and policy-oriented. It could map a region's economic profile, but it lacked a unified theory of why regional growth rates diverged in the first place.
The 1950s and 1960s saw three competing frameworks emerge to explain regional growth, each reacting to a different weakness in the earlier tradition.
Export Base Theory, developed by Douglas North and others in the mid-1950s, argued that a region's growth is driven by its success in exporting goods to other regions. In this view, the local economy is divided into a basic sector (industries that sell outside the region) and a non-basic sector (local services). Growth begins when external demand boosts the basic sector, which then multiplies through the local economy via spending effects. Export Base Theory was simple and empirically tractable—it gave planners a clear target (promote exports)—but it treated the region as a passive recipient of external demand and said little about internal dynamics such as innovation or labor markets.
Almost simultaneously, Gunnar Myrdal's Cumulative Causation Theory (1957) offered a starkly different picture. Myrdal argued that market forces do not automatically reduce regional inequalities; instead, they tend to amplify them. A region that gets an initial advantage—say, a new factory or a port—attracts capital and labor from poorer regions, which further strengthens its advantage and drains the periphery. Myrdal called this process "circular and cumulative causation." Unlike Export Base Theory, which saw growth as a response to external demand, Cumulative Causation emphasized self-reinforcing internal dynamics. It also directly challenged the optimistic assumption, implicit in much classical trade theory, that factor mobility would equalize incomes across space.
Neoclassical Regional Growth Theory, formalized in the 1960s by economists such as Robert Solow and Trevor Swan and extended to regions by George Borts and Jerome Stein, took the opposite position. Drawing on the standard Solow-Swan growth model, neoclassical regional theory predicted that poorer regions should grow faster than richer ones because capital is subject to diminishing returns: capital flows to regions where it is scarce and therefore earns a higher return. Over time, this convergence process should narrow regional income gaps. The neoclassical framework coexisted with Cumulative Causation in a sharp disagreement: one side saw convergence as the natural outcome of market forces, the other saw divergence as the normal result of the same forces. Both frameworks were internally consistent, but they made opposite predictions about the real world, and empirical tests in the 1970s and 1980s gave mixed results—some regions converged, others did not.
By the late 1980s, regional economics had accumulated a rich set of descriptive tools and competing hypotheses, but it remained somewhat marginal to mainstream economics. The reason was technical: most economic models assumed constant returns to scale and perfect competition, which made it difficult to model spatial concentration endogenously. Paul Krugman broke this impasse in 1991 with a paper that launched New Economic Geography (NEG).
NEG built on the older tradition of cumulative causation but gave it a rigorous microeconomic foundation. Krugman showed that when firms face increasing returns to scale (so that production is cheaper at larger scales) and transport costs are positive, manufacturing tends to concentrate in a single large market to serve the whole economy. Workers follow jobs, and jobs follow workers, creating a self-reinforcing agglomeration that Myrdal had described qualitatively but could not model formally. NEG thus revived and transformed Cumulative Causation Theory, replacing its verbal logic with a general-equilibrium model. At the same time, NEG directly challenged the neoclassical convergence story: in the presence of increasing returns, market integration could produce divergence, not convergence.
NEG also absorbed elements of classical location theory—transport costs and distance still mattered—but it moved beyond the static geometry of von Thünen and Weber to explain how the spatial structure of the economy emerges from the interaction of scale economies, trade costs, and labor mobility. For the first time, regional economics had a framework that could speak directly to mainstream trade and growth theory. NEG remains an active research program today, especially in its extensions to multiple regions, heterogeneous firms, and dynamic agglomeration.
Since the early 2000s, a new methodological school has risen alongside NEG: Quantitative Regional Economics. Where NEG focuses on building elegant theoretical models, Quantitative Regional Economics emphasizes empirical estimation and policy simulation using rich microdata. Its practitioners use techniques such as spatial econometrics, structural estimation, and large-scale computable general equilibrium models to quantify the effects of transport infrastructure, trade policy, and local labor market shocks on regional outcomes.
Quantitative Regional Economics does not reject NEG; in many ways it complements it. NEG provides the theoretical mechanisms—agglomeration, selection, sorting—that quantitative models then test and calibrate against real data. But the quantitative turn also reflects a narrowing of ambition: instead of searching for a single universal theory of regional growth, quantitative researchers build models tailored to specific policy questions, such as the impact of a new highway on employment or the welfare effects of regional trade agreements. This framework has absorbed the data infrastructure of Regional Science—input-output tables, regional accounts, spatial data—while discarding its earlier theoretical eclecticism in favor of rigorous microfoundations.
Today, New Economic Geography and Quantitative Regional Economics are the leading frameworks in the field, but they are not the only active traditions. Cumulative Causation Theory survives as a qualitative lens for understanding persistent regional inequality, especially in development economics and economic geography. Neoclassical Regional Growth Theory, while no longer dominant, still informs convergence debates in the European Union and other policy contexts where the question is whether poorer regions are catching up.
The two leading frameworks agree on several core points: that space matters for economic outcomes, that agglomeration economies are real and important, and that transport costs and trade frictions shape regional specialization. They disagree, however, on how much weight to give theory versus data. NEG theorists argue that a unified theoretical framework is essential for understanding the logic of spatial equilibrium; quantitative practitioners counter that the most useful knowledge comes from careful empirical work, even if it means setting aside general theory in favor of context-specific models. This disagreement is productive: it keeps the field from settling into a single orthodoxy and drives both theoretical innovation and empirical rigor.
Regional economics has thus moved from static location models through competing growth theories to a mature phase in which theoretical depth and empirical precision reinforce each other. The central question—why regions diverge—remains the same, but the tools for answering it have grown far more powerful.