Regional budget allocation and renewable energy development in Ukraine: Implications for public expenditure management

  • 4 Views
  • 0 Downloads

Creative Commons License DMCA.com Protection Status
This work is licensed under a Creative Commons Attribution 4.0 International License

Type of the article: Research Article

Abstract
Efficient regional public expenditure is critical for aligning decentralized economic development with renewable energy, energy security, and reconstruction priorities. This study aims to examine whether expenditure across selected regional budget programs is systematically associated with renewable energy development in Ukraine and whether these relationships remain robust across alternative temporal, distributional, and nonlinear specifications. The analysis uses a balanced panel of 25 Ukrainian regions for 2018–2021 and applies program-specific two-way fixed-effects models with CR2 standard errors, Benjamini–Hochberg adjustments, lagged and same-sample specifications, wild-cluster-bootstrap inference, presence–intensity decomposition, alternative transformations, winsorization, and formal quadratic tests. In the baseline capacity growth models, expenditure from local target funds (β = 0.9152, p = 0.0192) and electric transport measures (β = 0.1499, p = 0.0126) showed nominally positive associations, but neither survived multiplicity adjustment (q = 0.1054). Wild-cluster-bootstrap inference did not confirm these estimates, producing p-values of 0.4871 and 0.3597, respectively, while both programs were observed in only five region–year cases across two regions. No program coefficient was significant at the 5% level in the electricity production models; SME support produced the strongest negative estimate (β = −0.2152, p = 0.0580, q = 0.5995), whereas installed renewable capacity remained positively associated with production (β = 0.4108–0.4803, p = 0.0041–0.0220). Lagged, presence–intensity, transformed, winsorized, and nonlinear specifications provided no multiplicity-robust evidence, with formal U-test q-values no lower than 0.5789.

Acknowledgment
The authors acknowledge funding from the Swiss National Science Foundation (SNSF) [Grant No. IZURZ1_224119]. The authors bear sole responsibility for the conclusions and results of the research.

view full abstract hide full abstract
    • Table 1. Linear two-way fixed-effects estimates for annual growth in installed renewable energy capacity
    • Table 2. Linear two-way fixed-effects estimates for renewable electricity production
    • Table 3. One-year lagged TWFE estimates for renewable energy development
    • Table 4. Comparison of CR2 and wild-cluster-bootstrap inference for the principal findings
    • Table A1. Descriptive statistics
    • Table A2. Zero observations and territorial coverage of regional expenditure programs
    • Table A3. Pooled Spearman correlation matrix
    • Table A4. Within-region Pearson correlation matrix
    • Table A5. Multicollinearity diagnostics for pooled and two-way fixed-effects specifications
    • Table A6. Territorial units included in the balanced panel
    • Table B1. Detailed CR2 inference and model-fit statistics: Annual growth in installed renewable energy capacity
    • Table B2. Detailed CR2 inference and model-fit statistics: Renewable electricity production
    • Table B3. Stability of common control and year effects across the parsimonious models
    • Table B4. Evidential status of the nominally significant or weakly significant program coefficients
    • Table B5. Detailed one-year lagged estimates for annual growth in installed renewable energy capacity
    • Table B6. Detailed one-year lagged estimates for renewable electricity production
    • Table B7. Comparison of contemporaneous and one-year lagged program estimates
    • Table B8. Stability of controls and year effects in the lagged specifications
    • Table B9. Summary of wild-cluster-bootstrap inference by model family
    • Table B10. Wild-cluster-bootstrap inference for program–expenditure coefficients
    • Table B11. Same-sample comparison of contemporaneous and one-year-lagged program estimates for renewable electricity production, 2019–2021
    • Table B12. Program-presence and positive-spending-intensity models for annual growth in installed renewable energy capacity
    • Table B13. Program-presence and positive-spending-intensity models for renewable electricity production
    • Table B14. IHS-transformation robustness estimates for regional program expenditure
    • Table B15. Winsorized positive-spending robustness estimates
    • Table B16. Formal quadratic and U-shape tests for regional program expenditure
    • Conceptualization
      Serhiy Lyeonov, Nadiya Kostyuchenko, Denys Smolennikov, Inna Tiutiunyk, Oleksandr Telizhenko
    • Data curation
      Serhiy Lyeonov, Inna Tiutiunyk
    • Formal Analysis
      Serhiy Lyeonov, Inna Tiutiunyk
    • Investigation
      Serhiy Lyeonov
    • Methodology
      Serhiy Lyeonov
    • Project administration
      Serhiy Lyeonov
    • Software
      Serhiy Lyeonov, Oleksandr Telizhenko
    • Supervision
      Serhiy Lyeonov
    • Validation
      Serhiy Lyeonov, Denys Smolennikov
    • Visualization
      Serhiy Lyeonov, Nadiya Kostyuchenko
    • Writing – original draft
      Serhiy Lyeonov, Nadiya Kostyuchenko, Denys Smolennikov, Inna Tiutiunyk, Oleksandr Telizhenko
    • Writing – review & editing
      Serhiy Lyeonov, Nadiya Kostyuchenko, Denys Smolennikov, Inna Tiutiunyk, Oleksandr Telizhenko
    • Funding acquisition
      Nadiya Kostyuchenko, Denys Smolennikov, Inna Tiutiunyk, Oleksandr Telizhenko
    • Resources
      Nadiya Kostyuchenko, Denys Smolennikov, Inna Tiutiunyk, Oleksandr Telizhenko