📈 economics

Endogenous Spatial Deformation and Agglomeration Shadows: A One-Dimensional Model with Infrastructure-Induced Metric Change

This paper develops a continuous-space New Economic Geography model on a unit circle to demonstrate how endogenous infrastructure-induced metric deformation creates "agglomeration shadows" and establishes a critical trade-cost threshold beyond which concentrated, low-frequency investment strictly outperforms uniform investment in maximizing welfare.

Original authors: Luoluo Gu, Qi Xiao

Published 2026-06-29✓ Author reviewed ⓘ

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Luoluo Gu, Qi Xiao

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine the world's economy as a giant, flexible rubber band stretched into a perfect circle. This circle represents a continuous ring of cities and towns connected by a single, endless road. In this model, people and businesses can move anywhere along this ring, but moving isn't free; it costs energy and money to travel, just like friction on a rubber band.

This paper asks a simple but tricky question: If a government decides to pour a massive amount of money into improving the road at just one specific spot (a "hub"), does that help the whole circle, or does it create winners and losers?

Here is the breakdown of the paper's findings using everyday analogies:

1. The Setup: The Rubber Band and the "Landlord"

Usually, economic models assume that if you improve a road, everyone benefits equally. But this paper introduces a twist: Infrastructure changes the shape of the world.

The Rubber Band: Think of the economy as a rubber band. When the government builds a super-fast highway or a high-speed rail line at one point, it doesn't just make travel faster; it effectively "stretches" the rubber band so that the hub feels physically closer to everywhere else.
The Landlord: The paper uses a clever trick to keep the math working. Instead of assuming land is owned by absent landlords who take money out of the economy, it assumes every local landlord lives there and spends all their rental income on local goods. This keeps the money circulating within the circle, ensuring the math stays balanced.

2. The Mechanism: The "Black Hole" Effect

When the government invests heavily in one spot (let's call it the Hub), two things happen:

The Suck: The Hub becomes a "gravity well." Because travel costs drop dramatically from the Hub to everywhere else, businesses and workers flock there. They want the easy access to markets.
The Shadow: As people and businesses rush to the Hub, the rest of the circle (the Shadow) gets left behind. The paper finds that this isn't just a small dip; the "Shadow" can cover about 65% of the entire circle. In this zone, population density drops significantly, and the area feels "hollowed out."

The Visual: Imagine a trampoline. If you place a heavy bowling ball (the investment) in the center, the fabric curves down. The ball sinks deep (the Hub), and the fabric around it slopes upward and away. The area far from the ball is now "higher" and harder to reach, even though the total amount of rubber hasn't changed.

3. The Big Surprise: When Concentration is Good vs. Bad

The most important finding of the paper is that concentrated investment is not always bad, and uniform investment is not always good. It depends on two specific conditions:

Condition A: How "Manufacturing-Heavy" is the Economy?

Low Manufacturing (The "Dispersion" Regime): If the economy is mostly services or agriculture, spreading money evenly is best. Pouring it all into one spot creates a "Black Hole" that sucks resources away from everyone else, making the total welfare (happiness/wealth) of the group go down.
High Manufacturing (The "Resonant" Regime): If the economy is heavily based on manufacturing (making things), the rules change. In this case, concentrating investment in a single hub can actually make the whole group richer than spreading it out. The "agglomeration" (clustering) creates such strong benefits (like shared suppliers and cheaper prices) that they outweigh the pain of the "Shadow."

Condition B: How "Friction-Heavy" is the Road?

Low Friction (Small, Integrated Cities): If the road is already smooth and short (like a small city), concentrating investment is bad. It just distorts a well-functioning system.
High Friction (Large, Long Corridors): If the road is very long and travel is naturally difficult (like a massive island loop or a huge regional ring), concentrating investment works. The paper calculates that for this to happen, the cost of traveling from one side of the ring to the other must be roughly 2.4 times higher than the cost of traveling locally.

4. The "Critical Band"

The paper identifies a specific "sweet spot" (a critical band) where the economy is just manufacturing-heavy enough and the road is just long/friction-heavy enough that concentrated investment beats uniform investment.

Below this band: Spread the money out.
Inside this band: Build one massive hub.
Above this band: The system might break down entirely (people might leave the "Shadow" areas completely, leaving them empty).

5. The Takeaway for Policymakers

The paper argues that you cannot apply a "one-size-fits-all" rule to infrastructure.

If you are building a highway around a small, well-connected city, spread the money evenly. Don't pick a favorite spot; it will hurt the neighbors.
If you are building a massive transport ring around a huge, difficult-to-traverse region (like a large island or a continent-spanning corridor) and the region is full of factories, go big on one hub. The "gravity" of that hub will pull the whole economy forward, even if it leaves a large "shadow" behind.

In short: The paper proves that infrastructure doesn't just change travel times; it physically reshapes the economic landscape. Sometimes, making one place "super-connected" is the only way to make the whole ring work, but only if the ring is big enough and the economy is industrial enough to support the resulting "gravity well."

Technical Summary: Endogenous Spatial Deformation and Agglomeration Shadows

Problem Statement
The paper addresses a critical gap in quantitative spatial economics: the inability of existing frameworks to analyze how infrastructure investment fundamentally alters the spatial metric (effective distances) itself. Dominant models, such as those by Allen and Arkolakis, treat trade costs and distances as exogenous primitives. Consequently, they can evaluate outcomes given a cost matrix but cannot determine the equilibrium consequences when investment endogenously deforms that matrix. This limitation obscures the potential for concentrated infrastructure bets (e.g., data centers in specific provinces, ring railways) to create "agglomeration shadows"—zones of depopulation surrounding a hub—rather than uniformly uplifting regions. The paper seeks to provide a tractable general equilibrium framework where infrastructure investment is a policy parameter that deforms the effective economic metric, deriving the analytical equilibrium consequences of this deformation.

Methodology
The authors develop a continuous-space New Economic Geography (NEG) model on the unit circle $S^1$ , representing a closed transport corridor (e.g., a ring railway or coastal loop). The model integrates several key theoretical components:

Preferences and Production: The economy features Dixit-Stiglitz monopolistic competition with CES preferences (elasticity of substitution $\sigma$ , manufacturing share $\mu$ ).
Housing and Microfoundation: To close the spatial macroeconomic identity without an explicit agricultural sector, the model adopts a "local landlord" microfoundation (extending Helpman, 1998). Local landlords own immobile land, recycle rental income into manufacturing demand, and ensure that aggregate manufacturing expenditure equals gross local labor income pointwise.
Endogenous Metric: Infrastructure investment $I(x)$ strictly and convexly reduces the effective friction function $g(I) = \theta/(K_0 + I)$ . Trade costs are the exponential of the line integral of this friction along the connecting arc. This creates a fixed-point problem where the trade cost matrix depends on the investment profile, which in turn determines the equilibrium population distribution.
Analytical Approach: The authors sidestep the complexity of a fixed-point-within-a-fixed-point by treating the investment profile as an exogenous policy parameter. They employ Fourier-mode analysis on the periodic domain $S^1$ to decompose the linearized equilibrium system. This allows for the derivation of closed-form break-point conditions, mode-coupling coefficients, and a Fourier-diagonalized welfare Hessian.
Operator-Theoretic Framework: Existence and local uniqueness of the symmetric equilibrium are established by recasting the system as a strict contraction mapping on a quotient Banach space ( $C(S^1)^2 / \sim$ ), where the contraction rate is governed by the second eigenvalue of a Jacobian matrix.

Key Contributions and Results

Continuum Stability and Break-Point Condition: The paper derives a closed-form break-point condition for agglomeration: $\mu > (\sigma - 1)/\sigma$ . Unlike discrete two-region models, the continuum formulation on $S^1$ demonstrates that this condition is exact and that the symmetric equilibrium is stable if and only if the first Fourier mode ( $n=1$ ) is stable. Higher modes are strictly more stable.
Mode-Coupling Coefficients: The authors derive closed-form coefficients ( $C_{n,m}$ ) linking the shape of investment to agglomeration patterns. They show that non-uniform investment couples previously independent Fourier modes, with coupling strength determined by the difference in trade-freeness-weighted mode numbers.
Operator-Theoretic Uniqueness and Critical Slowing: The paper proves the local existence and uniqueness of the long-run equilibrium in the stable interior regime. It identifies a "critical slowing" phenomenon: as the manufacturing share $\mu$ approaches the break point, the second Jacobian eigenvalue $\nu$ approaches 1, causing the convergence rate of discrete tatonnement processes to diverge.
Agglomeration Shadows: The model characterizes "agglomeration shadows" as regions where equilibrium population density falls below the symmetric benchmark. Under concentrated investment (e.g., a Gaussian hub), the shadow extends across approximately 65% of the circle, with population density declining monotonically from the hub to a minimum at the antipodal point.
Variational Welfare and Regime-Dependent Dominance: The paper casts welfare as a functional of the investment distribution.
- First-Order Stationarity: Uniform investment is a stationary point for welfare (first variation vanishes).
- Regime-Dependent Theorem: The paper establishes a critical threshold $\mu^{**}$ $μ^{**}$ (approximately 0.6377 for baseline parameters) that divides the parameter space into two regimes:
  1. Dispersion-Dominated Regime ( $\mu < \mu^{**}$ ): Uniform investment strictly maximizes welfare. Concentrated investment creates deadweight losses comparable to small sales taxes.
  2. Agglomeration-Resonant Regime ( $\mu^{**} < \mu < \mu^*$ ): Low-frequency concentrated investment (specifically mode $n=1$ ) strictly outperforms uniform investment. In this regime, the market-access externalities of a hub outweigh the congestion and metric-convexity penalties.
Trade-Cost Decay Threshold: The reversal to concentrated investment optimality requires a sufficiently high trade-cost decay parameter $b$ (specifically $b > b^* \approx 5.24$ ). This implies the phenomenon is relevant only for large-scale corridors or those with high friction, where end-to-end trade costs exceed roughly 2.4 times the local cost.

Significance and Claims
The paper claims to provide the first analytical derivation of how infrastructure-induced metric deformation reshapes spatial equilibrium and welfare in a continuous setting. Its significance lies in:

Theoretical Rigor: Providing an operator-theoretic proof of existence and uniqueness that does not rely on external spectral radius conditions, while explicitly linking the break-point parameter to the contraction regime.
Policy Insight: Challenging the assumption that concentrated infrastructure always benefits surrounding regions. The paper demonstrates that concentrated investment can generate "losers" (the shadow region) and that the welfare optimality of concentration is conditional on structural parameters ( $\mu, \sigma, b$ ).
Methodological Innovation: Utilizing the unit circle $S^1$ to enable full Fourier decomposition, allowing for closed-form solutions for mode-coupling and welfare Hessians that are intractable in general network or 2D settings.
Empirical Relevance: Identifying a specific "critical band" of parameters where policy prescriptions reverse. The authors note that while mature service economies likely fall in the dispersion-dominated regime (favoring uniform investment), economies in earlier industrialization phases or specific high-friction corridors may fall in the agglomeration-resonant regime, where concentrated hubs are welfare-enhancing.

The paper explicitly limits its claims to the stable interior regime and local perturbations, acknowledging that free-boundary regimes (where population drops to zero) and finite-magnitude investment profiles require further nonlinear analysis. It positions its findings as a conditional, falsifiable policy boundary rather than a universal prescription for infrastructure allocation.