📄 other

A Spatial Overlapping Generations Model with Environmental and Human Capital Externalities: Evidence from European Regions (1995–2025)

This paper develops and empirically validates a spatial overlapping generations model to demonstrate that inter-regional knowledge spillovers and environmental externalities significantly shape European regional growth, revealing that human capital diffusion drives convergence while transboundary pollution hinders welfare, thereby supporting integrated policies that boost human capital and environmental quality.

Original authors: Sid Ahmed ZENAGUI

Published 2026-07-27

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Sid Ahmed ZENAGUI

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 or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine the economy not as a single giant machine, but as a massive neighborhood of hundreds of towns, each with its own people, factories, and schools. In this neighborhood, what happens in one town doesn't stay there; it ripples out to the neighbors. If a town builds a great library, the kids in the next town over might get smarter just by hanging out nearby. If a town dumps toxic smoke, the wind might carry that smog to the next town, making everyone there sick. This is the world of regional economics, a field that studies how money, jobs, and well-being spread across different places.

To understand how these neighborhoods grow over time, economists use a tool called an Overlapping Generations (OLG) model. Think of this as a story about a relay race where the baton is passed from one generation to the next. In this story, young people work, save money, and learn skills, while older people retire and spend what they saved. The paper also looks at human capital, which is just a fancy way of saying "how much knowledge and skill a person has," and environmental externalities, which are the side effects of our actions on the air and water that we don't pay for directly. The big question is: How do these factors mix together to make some regions rich and others poor, and can we fix the gaps?

The Great European Neighborhood Experiment

In this paper, the author, Sid Ahmed Zenagui, decides to build a super-complex simulation of the European neighborhood. He looks at about 240 different regions (called NUTS-2 regions) across Europe, tracking their growth from 1995 all the way to a projected 2025. Instead of just looking at one town in isolation, he creates a "Spatial Overlapping Generations Model." This is a mouthful, but imagine it as a giant, interconnected web where every town is linked to its neighbors.

The author's main goal is to see how three specific things interact: Total Factor Productivity (TFP) (which is basically how efficiently a region uses its tools and ideas), Human Capital (the skills and education of the people), and Environmental Quality (how clean the air and water are). He wants to know: Do these things stay local, or do they spill over into neighboring towns? And does the pollution from one town hurt the growth of the next?

What the Numbers Say: The Ripple Effect

After crunching the numbers with some very advanced math (using tools like the Spatial Durbin Model and System-GMM), the paper finds some fascinating patterns.

First, Human Capital is the superstar. The study finds that for every 1% increase in a region's human capital, the GDP per capita (the average money made by a person) goes up by about 0.73%. But here's the kicker: almost 31% of that boost doesn't just stay in the town that invested in education. It spills over to the neighbors! It's like if a town built a world-class science lab, the neighboring towns would suddenly start inventing cool things too, even if they didn't build the lab themselves.

Second, Technology (TFP) also ripples out. A 1% increase in a region's efficiency raises its own GDP by about 0.33–0.40%, but it adds another 0.18% to the GDP of neighboring regions. This confirms that knowledge travels fast across borders.

Third, Pollution is a shared problem. The paper shows that carbon emissions are bad news, not just for the town that produces them, but for everyone around them. A 1% increase in carbon emissions lowers a region's GDP by about 0.12% directly, but it also drags down neighboring regions by another 0.09%. It's a "double whammy" of bad air and bad economics. The study provides robust empirical evidence that pollution is not just a local issue; the data shows it creates a "transboundary environmental damage" that hurts the whole neighborhood.

The "Trap" and the Solution

The paper also warns about a scary scenario called an "Environmental Trap." If a region pollutes too much and doesn't have enough money or technology to clean it up, it can get stuck in a cycle where the environment gets so bad that the economy collapses. The math suggests this happens if the damage from pollution is stronger than the region's ability to clean it up.

However, there is good news. The author ran some "what-if" simulations to see what would happen if policymakers acted. The results suggest that if regions combine investing in human capital (like schools and training) with improving environmental quality (like cleaning up pollution), they could boost their long-term welfare by 4–5%.

Interestingly, this combined plan helps the poorer, rural regions the most. It seems that when you teach people new skills and clean up their air, the "ripple effect" is strongest in places that need it most. The paper suggests that standard cost-benefit analyses often miss this because they only look at the direct benefits to the town spending the money, ignoring the extra 30% boost that neighbors get for free.

The Bottom Line

This paper doesn't just say "education is good" or "pollution is bad." It uses a detailed model of 240 European regions to provide strong empirical evidence that these things are deeply connected and that they travel across borders. It shows that the best way to help a struggling region isn't just to throw money at it, but to invest in its people and its environment, because those investments will naturally spread to the neighbors, lifting the whole neighborhood up together. The findings are robust, surviving many different tests and checks, suggesting that this "ripple effect" is a real and powerful force in how our world grows.

Technical Summary: A Spatial Overlapping Generations Model with Environmental and Human Capital Externalities

Problem Statement
Regional economic disparities within the European Union have persisted despite decades of cohesion policy and single market integration. This paper addresses the mechanisms driving these disparities by integrating three interacting phenomena often treated separately in the literature: (1) differences in Total Factor Productivity (TFP) and knowledge diffusion; (2) heterogeneity in human capital endowments and absorption capacity; and (3) divergent exposure to environmental degradation. The central challenge is to model these forces within a framework that accounts for intergenerational trade-offs (demographic structure) and spatial interdependence (cross-regional spillovers), neither of which is fully captured by standard non-spatial Overlapping Generations (OLG) models or purely empirical spatial growth studies.

Methodology
The paper employs a dual approach, combining a rigorous theoretical framework with a multi-stage empirical estimation strategy.

Theoretical Framework: The authors construct a multi-region OLG model for $R$ regions over discrete time. The model features:
- Demographics: Two cohorts (young and old) with population growth rates.
- Preferences: Agents maximize lifetime utility derived from consumption, environmental quality, and human capital externalities.
- Production: A Cobb-Douglas technology where output is subject to environmental damage (pollution stock) and augmented by TFP and human capital.
- Spatial Dynamics:
  - TFP: Evolves via a spatial autoregressive process where a region's productivity depends on its own persistence and the weighted average of neighbors' productivity.
  - Human Capital: Accumulation depends on educational effort, local environmental quality, and the spatial diffusion of knowledge from neighbors.
  - Environment: Pollution accumulates from production but dissipates naturally and is reduced by abatement efforts, including transboundary spillovers from neighboring regions.
- Equilibrium: The model derives a unique interior Balanced Growth Path (BGP) and establishes conditions for spatial convergence and the existence of an "environmental trap" (where pollution overwhelms abatement).
Empirical Strategy:
- Data: An unbalanced panel of approximately 240 NUTS-2 European regions from 1995 to 2025. Variables include GDP per capita (PPP), TFP (residual), CO2 emissions, human capital index (tertiary enrollment/life expectancy), and R&D expenditure.
- Spatial Weighting: A baseline row-normalized queen contiguity matrix ( $W$ ), with robustness checks using distance-based, economic, and higher-order contiguity matrices.
- Estimation Hierarchy:
  1. Baseline OLS: With fixed effects and interaction terms.
  2. Spatial Durbin Model (SDM): To capture direct effects and spatial spillovers (lags of dependent and independent variables).
  3. IV-GMM: Using spatial lags of regressors as instruments to address endogeneity (reverse causality and measurement error).
  4. System-GMM (Blundell-Bond): To capture dynamic persistence and intergenerational capital accumulation mechanisms.
- Decomposition: Following LeSage and Pace (2009), the authors decompose coefficients into direct, indirect (spillover), and total effects.

Key Contributions

Theoretical Integration: The paper bridges the gap between OLG environmental models and spatial growth models. It extends previous two-region or non-spatial OLG models by embedding a continuum of regions with endogenous human capital and full spatial dynamics.
Analytical Derivations: The authors prove the existence and uniqueness of the BGP and derive closed-form expressions for steady-state capital, environmental quality, and welfare. Crucially, they derive expressions for spatial direct, indirect, and total effects that map directly to the SDM estimator, providing a theoretical foundation for the empirical decomposition.
Comprehensive Empirical Evidence: The study assembles a unique long-horizon dataset (1995–2025) and applies a hierarchy of estimators (OLS, SDM, IV-GMM, System-GMM) to a large sample of European regions, addressing endogeneity and dynamic persistence simultaneously.
Policy Counterfactuals: The paper moves beyond estimation to simulate policy scenarios, quantifying the welfare gains of combined human capital and environmental interventions.

Key Results

Spatial Interdependence: There is strong, statistically significant positive spatial autocorrelation in GDP, CO2, and human capital (Moran's $I \approx 0.44–0.51$ ). The spatial autoregressive parameter ( $\rho$ ) is significant ( $\approx 0.21–0.28$ ), confirming inter-regional output interdependence.
Human Capital Spillovers: Human capital generates the largest total growth effect ( $\approx 0.73$ ). Approximately 31% of this effect is indirect (spillover to neighbors), confirming that knowledge diffusion is a primary driver of European convergence.
TFP Dynamics: A 1% increase in TFP raises GDP per capita by 0.33–0.40% directly and an additional 0.18% through spatial spillovers.
Environmental Damage: CO2 emissions exert a significant negative direct effect ($-0.12$) and a robust indirect effect ($-0.09$), indicating transboundary pollution damage. The total damage coefficient is approximately $-0.21$.
Complementarity: The interaction between human capital and CO2 is positive and significant, suggesting that higher human capital partially offsets environmental damage by expanding abatement capacity.
Intergenerational Dynamics: System-GMM results confirm a persistence parameter ( $\delta \approx 0.42$ ) consistent with OLG capital accumulation. The share of the young cohort positively affects growth, while the old cohort share negatively affects it, reflecting demographic shifts in savings behavior.
Robustness: Results hold across alternative weight matrices, sub-samples (EU-15 vs. new member states, urban vs. rural), and placebo tests (randomized weights/outcomes).

Significance and Claims
The paper claims that its integrated framework provides a more accurate assessment of regional growth drivers than models ignoring either spatial spillovers or intergenerational dynamics.

Policy Implication 1: Standard cost-benefit analyses of education that ignore spatial spillovers understate the social return to human capital investment by approximately one-third.
Policy Implication 2: Environmental and human capital policies are complementary; investing in human capital enhances the marginal effectiveness of environmental regulation.
Policy Implication 3: A combined program of increased human capital investment and environmental quality improvement can raise long-run welfare by 4–5% over a 30-year horizon. This combined approach yields the largest gains for disadvantaged, rural, and lower-income regions in new member states, aligning with the goals of European cohesion policy.

The authors conclude that regional human capital investment is the most productive use of cohesion funds, not only for its direct impact but for its role in generating positive externalities and amplifying environmental policy effectiveness. Future research directions suggested include integrating fiscal federalism, studying intra-regional inequality with heterogeneous agents, and applying the framework to emerging markets.

The Great European Neighborhood Experiment

What the Numbers Say: The Ripple Effect

The "Trap" and the Solution

The Bottom Line

Technical Summary: A Spatial Overlapping Generations Model with Environmental and Human Capital Externalities

More like this