Regional Economic Impacts of the Just Energy Transition: Lessons for Coal Regions
Source: arXiv:2607.09589 · Published 2026-07-10 · By Imke Rhoden, Jae-Hyuck Lee
TL;DR
This paper addresses the complex regional economic consequences of coal phase-out, a critical aspect of the global energy transition. Coal-dependent regions face structural challenges including persistent unemployment despite some GDP per capita growth, caused by structural dependencies on mining and power generation. Utilizing panel data from EU NUTS 2 regions over 2000-2022 with fixed effects models, the authors find coal regions carry a consistent unemployment premium of approximately 1.1 percentage points, while their GDP per capita grows about 0.2 percentage points faster annually than non-coal regions. This paradoxical combination evidences a "hollowing-out" process where population decline via outmigration and early retirement raises per-capita output metrics, but conceals worsening labor market conditions and social challenges.
Their spatial analysis shows strong geographic clustering of coal regions with economic vulnerability, underscoring the necessity of coordinated, regionally tailored transition policies rather than one-size-fits-all approaches. The paper further contrasts European transition experiences with South Korea's rapid coal phase-out plan, emphasizing the urgent need for proactive and credible national support before plant closures to enable smooth regional adjustments. Overall, the study highlights the multi-dimensional and long-term nature of coal transition challenges, advocating for integrated social, fiscal, and institutional measures from early stages to avoid exacerbating regional inequality.
Key findings
- Coal regions have roughly 1.1 percentage points higher unemployment rates than non-coal regions within the same countries (Model M4).
- Coal regions experience about 0.2 percentage points faster annual GDP per capita growth compared to non-coal regions (Model M3), indicative of a hollowing-out effect.
- Coal regions' GDP per capita remains persistently below national averages, by roughly 2,500 purchasing power standard (PPS) units as of 2022, despite convergence.
- Direct plus indirect coal employment multipliers range from 0.4 to 1.2 additional jobs per direct coal job, causing broader regional economic exposure.
- Early retirement policies reduce headline unemployment but do not aid in creating sustainable new employment opportunities.
- Infrastructure investments have stronger evidence of durable positive impact than firm-level subsidies, which risk inefficient short-term job creation.
- Proactive transition policy initiated before plant closures yields substantially better regional economic outcomes than reactive policy.
- Strong geographic clustering of coal regions shows spatial lock-in effects requiring coordinated multi-level governance and local input.
Methodology — deep read
The study uses a quantitative empirical approach analyzing panel data of European Union NUTS 2 regions from 2000 to 2022. The data include regional gross domestic product (GDP) per capita measured in purchasing power standards (PPS), unemployment rates of the population aged 20-64, and population density as a control for urbanization.
Coal regions are identified by presence of active or recently closed mining operations in 2018, yielding a core group of 38 regions. Regions with only coal power plants but no mining (67 regions) serve for robustness checks. UK regions are excluded from analysis after 2020 due to Brexit.
The key econometric model is a two-way fixed effects regression with year and country fixed effects, clustered standard errors at the NUTS 2 level. Region fixed effects cannot be included because coal region classification is time-invariant. This approach isolates within-country differences between coal and non-coal regions over time.
The main regression equations model log GDP per capita, annual GDP growth rates, and unemployment rates as dependent variables, with a binary coal region indicator as the main explanatory variable. Augmented models control for population density and initial GDP to check robustness. Spatial econometric models (spatial autoregressive and spatial error) are also implemented to address spatial autocorrelation.
The identification strategy relies on comparing regions within the same countries to mitigate cross-country heterogeneity bias, but the authors acknowledge potential confounding from unobserved regional characteristics correlated with coal specialization.
Data provenance comes primarily from Eurostat databases on regional GDP and unemployment statistics. The empirical work controls and robustness checks include excluding power-only coal regions, restricting periods, and testing alternative standard error specifications.
Additionally, the paper reviews qualitative policy lessons from European coal transitions, including programs in Germany, the UK, and Poland, and draws parallels to South Korea’s Chungnam province situation.
One concrete example discussed is the contrasting transition trajectories of Lusatia, Germany versus Dytiki Makedonia, Greece, where institutional capacity differences critically impacted fund absorption, project design, and economic diversification.
The empirical results are exploratory and descriptive; causal inference is limited by the observational panel data and time-invariant coal region status. Nevertheless, the multi-method approach triangulates persistent unemployment premiums and a hollowing-out GDP growth pattern verified across multiple robustness tests.
Technical innovations
- Application of two-way fixed effects panel regression with clustered errors to isolate within-country economic differences of coal regions over 23 years.
- Incorporation of spatial econometric models addressing geographic clustering and spatial dependence in regional economic outcomes.
- Novel longitudinal quantification of the "hollowing-out" phenomenon in coal regions, revealing simultaneous GDP per capita growth and worsening labor market conditions.
- Cross-national comparative policy analysis linking empirical regional economic indicators with detailed evaluation of European and South Korean just transition programs.
Datasets
- Eurostat regional GDP and unemployment data — ~281 EU NUTS 2 regions (2000-2022)
- Alves Dias et al., 2018 coal employment dataset — coal direct and indirect jobs per region
- National and regional policy program data for Germany, UK, Poland, and South Korea (various government and research sources)
Baselines vs proposed
- Model M1 (no country fixed effects): coal regions GDP per capita ~19.2% below EU average
- Model M2 (with country fixed effects): coal region GDP per capita gap reduces to 6.0%, statistically insignificant
- Model M3: coal regions GDP per capita growth 0.23 percentage points per year faster than non-coal regions
- Model M4: coal regions have 0.87 percentage points higher unemployment rate (ages 20-64)
- Robustness models restricting to post-2013 and excluding power-only regions confirm stability of estimates
Limitations
- Coal region classification is time-invariant, limiting causal inference due to potential unobserved regional confounders.
- Unemployment statistics may underestimate labor market distress due to unregistered long-term sickness or inactivity.
- Limited municipal or subregional data granularity constrains fine-scale spatial analysis.
- Policy analysis largely draws on secondary literature and case studies without formal econometric testing of program impacts.
- South Korean regional data on employment and fiscal trends remain incomplete, limiting direct comparisons.
- Absence of distributional analysis within regions (e.g., income inequality, demographic subgroups).
Open questions / follow-ons
- How can causal identification be improved to isolate impacts of coal phase-out policies versus pre-existing regional trends?
- What specific institutional capacities most effectively enable coal regions to absorb transition funds and diversify economically?
- How do demographic shifts (age, skills) interact dynamically with economic output and labor force participation in coal transitions?
- What are best practices for integrating local stakeholder participation to enhance legitimacy and uptake of just transition programs?
Why it matters for bot defense
While this paper does not address bot-defense or CAPTCHA technologies directly, its insights have indirect relevance for bot-defense practitioners concerned with socioeconomic resilience in digitally underserved or economically vulnerable regions. Understanding how economic structural shifts and policy interventions impact social cohesion, employment rates, and population mobility can inform the design of CAPTCHAs or bot defenses that account for user accessibility and legitimacy in diverse regional contexts. Particularly, regions undergoing rapid structural transitions could see shifts in online activity patterns, account creation behaviors, or fraud risk profiles. For CAPTCHA and bot-defense engineers focused on equity and inclusive user experience, recognizing hollowing-out dynamics and regional disparities highlights the importance of adaptable, nuanced user verification systems tuned to varying socio-economic conditions. More broadly, the paper exemplifies rigorous longitudinal regional data analysis and policy evaluation frameworks that could inspire similar methodologies in cybersecurity and fraud prevention research focusing on geographically localized attack or usage patterns.
Cite
@article{arxiv2607_09589,
title={ Regional Economic Impacts of the Just Energy Transition: Lessons for Coal Regions },
author={ Imke Rhoden and Jae-Hyuck Lee },
journal={arXiv preprint arXiv:2607.09589},
year={ 2026 },
url={https://arxiv.org/abs/2607.09589}
}