The Role of the European Investment Bank in Financing Renewable Energy Sources in Selected European Union Countries

Abstract

In the area of the European Union (EU) energy policy, among the entities involved in the process of financing investments in renewable energy sources (RESs), the European Investment Bank (EIB) plays a particularly important role. Therefore, the aim of the research was to identify the relationship between the EIB’s financing of RES projects and the level of energy transition, measured by the share of RES in gross final energy consumption (RE). The goal was achieved using quantitative methods and a two-way fixed-effects panel model FE (country and year), based on data from EIB, Eurostat, World Bank, OECD, EDGAR, and Our World in Data for 2012–2023. As a result of the research, it was determined that the scale of EIB financing alone does not translate into short-term growth of the RE in the examined sample (EU countries). Indeed, the effectiveness of funding depends on the regulatory and institutional context; the grid’s ability to absorb new capacities (throughput, storage, demand flexibility); and from the time horizon (delayed materialization of effects). Increasing the efficiency of converting euros into RE percentage points requires better targeting (power + grid), simplification of procedures and good financial assembly with the right allocation of risks.

1. Introduction

The confluence of energy crises, geopolitical tensions, rising societal awareness, the volatility of fossil fuel prices, and the intensification of anthropogenic CO₂ emissions has catalysed a paradigm shift from fossil fuel-based energy systems toward low-carbon and renewable energy sources. In 2022, renewable energy sources (RESs) constituted approximately 23% of the European Union’s total energy consumption. By 2023, policymakers elevated the EU’s strategic objectives, incrementally increasing the target for the share of renewable energy within gross final energy consumption from 32% to 42.5% by 2030, with aspirations to attain a 45% threshold [1]. By 2024, RES contributed nearly 50% to the electricity consumed across the European Union, whereas fossil fuels represented slightly over 25% of the overall energy mix [2].

According to a July 2025 report by PricewaterhouseCoopers [3], there has been a discernible increase in the number of enterprises undertaking climate commitments and elevating their climate ambitions. Concurrently, data from the January 2025 BloombergNEF Levelized Cost of Electricity report [4] indicate that a significant catalyst for the transition away from fossil fuels is the projected reduction in the levelised cost of renewable energy generation by as much as 11% compared to 2024. The year 2025 has the potential to mark the inaugural decline in global structural emissions, driven by the expansion of renewable energy sources. An increasingly critical element of this expansion is investments in grid infrastructure, including the development of energy storage solutions and charging infrastructure. These seek to enhance the efficiency of low-emission electricity transmission amid the rapid increase in renewable energy production [5]. Consequently, the expansion of investments in renewable energy sources (RESs) significantly contribute to the generation of new employment opportunities [6]. Finally, RESs, by providing sustainable power, contribute significantly to economic development, enhance energy security, and bolster the resilience of energy systems [7].

This process necessitates comprehensive global collaboration among governments, financial institutions, enterprises, and civil society to address critical issues such as the allocation of financial resources for the implementation of renewable energy investments [8]. However, this subject has not been extensively analysed, as most authors primarily focus their attention on examining of renewable energy sources from the perspective of the following:

(1)

the cultivation of sustainable development from both current [9,10,11,12,13,14] and future temporal perspectives [15],

(2)

assessment of renewable energy efficiency, for example (e.g.,) multi-criteria decision analysis exemplified through selected countries [16,17]; visions of energy efficiency up to the year 2050 [18]; the relationship between a nation’s level of development and its energy consumption [19]; the impact of the financial system’s efficiency on the demand for renewable energy sources (RESs) [20]; the evaluation of financing models concerning energy efficiency [12]; and the assessment of financial viability in economies characterized by high levels of pollution [21,22],

(3)

factors determining the implementation of renewable energy sources [23,24].

Meanwhile, the role of banks in the process of financing renewable energy sources is growing. Banks are becoming an indispensable part of the capital ecosystem in the context of the transition to a low-carbon economy [25]. A good illustration of this phenomenon is the example of the European Investment Bank (EIB) (The legal basis for the bank’s operation is set out in the regulations—Articles 308 and 309 of the Treaty on the Functioning of the European Union (hereinafter referred to as TFEU) [26]. Detailed arrangements for the functioning of the EIB are laid down in Protocol (No 5) on the Statute of the European Investment Bank [27] and Protocol (No 28) on economic, social and territorial cohesion, annexed to the Treaty on European Union [28]). By supporting climate action, the EIB is becoming the European Union’s (EU) climate bank [29]. Therefore, it is interesting from a cognitive and application point of view to identify the relationship between EIB financing of RES projects and the level of energy transition, measured by the share of RES in gross final energy consumption (RE) in EU countries.

This study provides responses to the following research questions:

• What is the level of financing provided by the EIB for RES projects caused by the increase in the share of RES in gross final energy consumption in EU countries?
• What factors determine the effectiveness of EIB financing for RES projects in EU countries?

This inquiry is poised to engage a broad spectrum of stakeholders, including representatives of European Union institutions, national and regional public administrative bodies, banking sector professionals, system operators—namely Transmission System Operators (TSOs) and Distribution System Operators (DSOs)—energy enterprises, renewable energy developers, and academic researchers.

2. Literature Review

2.1. RES in EU Energy Policy—Main Legal Regulations

The European Union can play a key role in promoting renewable energy sources, by treating them as the foundation of its energy and climate policy [30]. Regulations concerning investments in RES constitute an important part of the legal framework for EU energy policy. These regulations aim to include all types of energy obtained from non-fossil fuels in the concept of renewable sources [31]. According to Article 2 of Directive 2001/77/EC of the European Parliament [32], renewable energy sources are those other than fossil fuels: wind energy, solar radiation, geothermal energy, wave and tidal energy, energy from hydroelectric power plants, biomass, and gas from landfills and sewage treatment plants. Biomass means the biodegradable fraction of products, waste, and residues from agriculture (including vegetal and animal substances), forestry, and related industries, as well as the biodegradable fraction of municipal and industrial waste.

In the face of escalating challenges related to climate change and the imperative to ensure energy security, the EU has adopted a series of directives, regulations, and strategies aimed at increasing the share of RES within the energy mix. Merely promoting renewable energy sources without the establishment of appropriate legal frameworks would have little chance of effective implementation on an international scale [33]. The adoption of Directive 2009/28/EC (RED I) [34] marked a significant breakthrough in the approach to sustainable development and renewable energy within EU member states. According to this directive, EU member states were obligated to achieve a 20% share of renewable energy in their final energy consumption by 2020. Building upon the objectives set forth in RED I, Directive (EU) 2018/2001 (RED II) [35] was enacted in 2018, setting ambitious targets for 2030, requiring that at least 32% of the energy consumed in EU countries originates from renewable sources, while ensuring that member states can attain this goal in a cost-effective manner. The efforts concerning the EU’s renewable energy policy goals culminated in the adoption of Directive (EU) 2023/2413 (RED III) in 2023 [36]. This directive articulates even more ambitious climate objectives—aiming to increase the share of energy from renewable sources in the European energy mix to 42.5% by 2030, with a subsequent goal of raising this indicator to 45%. The objective also encompassed encouraging investment in renewable energy production, while simultaneously affording member states the opportunity to retain their energy sovereignty [33,37,38].

When analysing the legal framework governing RES regulation, it is also essential to draw attention to the provisions of the document concerning the European Green Deal, published on 11 December 2019, in the form of the European Commission communication to the European Parliament, the European Council, the Council, the European Economic and Social Committee, and the Committee of the Regions [39]. The implementation of the Green Deal is designed to impact all sectors of the economy, which must adapt to the EU’s environmental and sustainable development objectives [40,41]. However, the ultimate success of this strategy will depend on the resolve of the EU member states in its execution. It is important to remember that EU institutions merely define the goals and overarching directions; the responsibility for taking specific implementation actions rests with the governments of individual member states [42,43].

2.2. Banking Services for Renewable Energy Financing

Among the entities involved in financing investments in RES, banks warrant particular attention. Their pivotal role in the implementation of RES projects stems from several compelling reasons. Firstly, banks tend to be more actively engaged than other lenders, especially in financing technologies characterized by higher risk profiles. This includes offshore wind energy—which involves significant construction and grid access risks—and biomass, which presents substantial feedstock-related risks [44]. Secondly, in high-income countries, there is a greater demand for renewable energy, which consequently leads to increased banking activity in providing debt capital, issuing credit guarantees, distributing public funds, and offering specialized consultancy services related to RES projects and the transition to a low-carbon economy [45]. Thirdly, banks with an environmentally conscious investment approach are willing to invest in ecological projects despite the potential for lower returns and a more moderate risk of insolvency [46]. For these financial institutions, assessing the investment risks associated with RES is of paramount importance. Research onto this subject has been conducted by Baldauf and Jochem [47], who proposed a conceptualization of an investment model for power plants. In this model, energy investors interact within an imperfect and decentralized market network comprising credits, deposits, and equity in projects. The analyses indicate that the diversification of project financing strategies and corporate funding significantly influences the pace of transition to RES [47]. Additionally, banks are concerned with risks related to fluctuating costs of financing both fossil fuels and renewables. Internationally, the results of these studies have been summarized in a report by the Energy Transition Risk and Cost of Capital Project (ETRC) [48].

Despite the pivotal role that banks play in financing RES, this issue remains underexplored from a research perspective. Existing analyses predominantly focus on evaluating the following:

• The social shareholder plan—Consumer Stock Ownership Plan (CSOP)—a novel consumer ownership model in the renewable energy sector. Studies, including those conducted by Lowitzsch, have compared CSOP with traditional RES financing models. As a result of these investigations, recommendations have been made to implement innovative organizational and contractual solutions that would facilitate the integration and scaling of RES investments, wherein consumers would serve as co-owners [49]. The analyses suggest that CSOP represents a low-threshold financing method, enabling individuals, particularly low-income households, to invest in renewable energy projects [50].
• Effectiveness of financial mechanisms. Research concerning Poland, the Netherlands, and the United Kingdom—primarily driven by fossil fuels—has evaluated the evolution of RES financing, taking into account wind and solar subsidies (grants, awards, crowdfunding, community bonds, ventures, and social investments). These studies demonstrated that the diversity of financial instruments and the increasing volume of funding correlate strongly with global decarbonization efforts and climate change mitigation [51]. Conversely, Mazzucato and Semieniuk have examined how different types of financing influence the pathways of RES deployment, considering specific technological categories and associated risks. Their findings indicated that various financial actors contributed to distinct technological development trajectories. Some entities maintained balanced portfolios, while others were markedly focused on particular technologies. Additionally, these actors differed in their orientation towards high- or low-risk technologies, with private entities significantly favouring low-risk options more than public institutions [52]. Among financial instruments, the issue of determinants influencing green bond issuance during the years 2017–2023 has been explored by Ozyesil and Tembel. Their research concentrated on the impact of interest rates, renewable energy capacity, carbon emission reductions, and GDP on green bond issuance [53].
• Barriers and incentives in the financing strategy for the transition to RES. Research conducted by scientists such as Qadira, Al-Motairi, Tahir, and Al-Faghih has demonstrated that the funding gap can be mitigated through the active involvement of financial institutions in supporting the public—particularly those interested in investing in RES—by offering preferential loans or facilitating the development of community financing platforms and crowdfunding initiatives. The reluctance of individual and corporate investors, as well as energy consumers, to adopt RES has been largely influenced by a lack of public awareness regarding the benefits of renewable energy, as well as misconceptions about the associated costs, including instalment payments and operational expenses [54].

2.3. European Investment Bank—The EU’s Climate Bank

The role of the EIB in financing RES remains largely underexplored within the realm of scientific research, despite the bank’s strategic aim to position itself as a leader in this domain. The activities of the EIB are primarily documented through reports, which serve as the main source of information. These reports typically detail initiatives undertaken by the EIB Group to support climate action and sustainable environmental practices, as well as economic and social cohesion within the European Union. Such investments play a crucial role in assisting both public and private sector organizations in adapting to the impacts of current and future climate change [55].

Based on the current review of the scientific literature, three primary research trajectories concerning the EIB can be discerned:

• The Transformation of the EIB. Researchers have focused on analysing the evolution of the EIB’s role, particularly its shift from marginal consideration of ecological issues to strategic environmental financing initiatives beginning in the 1990s. The year 1973 proved to be pivotal, as it marked the recognition of ecology as an eligible lending objective [56]. The process of transforming the EIB into a climate-oriented financial institution has also been examined through the lens of quantum trajectories, drawing on studies of Erwin Schrödinger’s concept of a “quantum leap.” Findings indicate that this “leap” was rapid and occurred during the formative period of the European Union’s climate policy framework, thereby reinforcing the EIB’s institutional and market position [57]. Furthermore, the evolution of the EIB’s activities has been subjected to a multidisciplinary analysis integrating history, economics, law, and political science, highlighting the increasing significance of this bank amid global economic challenges and climate change [58].
• The Banking System in Support of Climate Change Initiatives. The EIB activities aimed at addressing climate change have predominantly constituted a component of broader analyses concerning the banking system’s role. Research conducted between 2009 and 2016 across European Union member states indicates, for example, that global financial investments in RES remain substantially below their potential, primarily due to market barriers and perceptions of elevated risk, which discourage private sector investors. In this context, the financial support provided for RES through EIB loans was viewed positively [59]. Additionally, researchers emphasized that EIB’s efforts should be complemented by a broader transformation towards “green” fiscal, industrial, and monetary policies, which would facilitate a fundamental shift in the economy and society to effectively confront the climate crisis [60]. The analyses also emphasized that in the transition towards a low-emission economy, the EIB and development banks cannot limit themselves to merely providing better, more environmentally friendly, and inclusive loans [61]. The issue of selecting appropriate financial instruments employed by the EIB in the implementation of renewable energy investments was further explored, notably by Spielberger, who focused on evaluating EU regulations concerning green bonds, in which the EIB served as the issuer. Through these activities, the bank accumulated widely recognized expert knowledge and expanded networks within the community dedicated to sustainable finance [62].
• The Banking System during the COVID-19. The research primarily focused on formulating recommendations for the EIB actions amid the public health crisis caused by COVID-19. It specifically addressed increasing the EIB’s own capital contributions and concentrating efforts directly on the final beneficiaries of this support, including sectors such as small and medium enterprises (SMEs) [63].

3. Methodology

The aim of the research is to identify the relationship between the financing of RES projects by the EIB and the level of energy transition, measured by the share of RES in gross final energy consumption (RE). At the same time, the level of energy transition was determined in accordance with the Eurostat methodology adopted under Directive (EU) 2018/2001 of the European Parliament and of the Council of 11 December 2018 on the promotion of the use of energy from renewable sources (Official Journal of the European Parliament and of the Council of 11 December 2018 on the promotion of the use of energy from renewable sources) [36]. The level of energy transition (indicator RE) has been defined as the share of energy from renewable sources in gross final energy consumption. We measure RE as defined by UNECE (United Nations Economic Commission for Europe) [64] and World Bank [65]. This indicator is a key measure of progress in the implementation of the European Union’s energy and climate policy and compliance with the objectives of the RED II Directive (2018/2001/EU) [35].

To achieve our research goal, we adopted a two-way fixed-effects panel model FE (country and year) and selected the appropriate indicators and control variables to address two key research questions:

• What is the level of financing provided by the EIB for RES projects caused by the increase in the share of RES in gross final energy consumption in EU countries?
• What factors determine the effectiveness of EIB financing for RES projects in EU countries?

In this study, the effectiveness of EIB financing is understood as the translation of expenditures (asinh(EBI/λ)) into a change in the share of RES in gross final energy consumption (RE), with control of $\ln \left(\mathrm{G}\mathrm{D}\mathrm{P}\mathrm{p}{\mathrm{c}}_{\mathrm{P}\mathrm{P}\mathrm{P}}\right)$, EP and PI and considering implementation and connection delays. This means that not only the amount of support is assessed, but also the ratio of this amount to the actual increase in RES production capacity in a given country and period. This definition is based on the approach presented in the ECA’s Special Report No 08/2018 (2018) [66] and Bölük’s and Kaplan’s empirical analysis on the effectiveness of support instruments in EU countries and Turkey [67] to assess the relationship between the financing structure and the increase in installed capacity in RES (taking into account cost-effectiveness and institutional differences between countries). For the purposes of this analysis, this approach has been adopted, to focus on the impact of EIB financial support on the change in the share of RES in the energy mix of the EU countries studied.

The following research hypotheses were formulated in this study:

• The level of financing provided by the EIB for renewable energy projects does not influence the increase in the share of renewable energy in the gross final energy consumption within EU member states.
• The effectiveness of EIB funding for renewable energy projects in EU countries is determined by non-financial factors.

The analyses were carried out taking into account secondary data concerning [44,45,52,68].

(1)

RE—share of RES in gross final energy consumption in percentage points (p.p.);

(2)

EBI—financing at constant prices (EUR million);

(3)

$\ln \left(\mathrm{G}\mathrm{D}\mathrm{P}\mathrm{p}{\mathrm{c}}_{{\mathrm{P}\mathrm{P}\mathrm{P}}_{\mathrm{i}\mathrm{t}}}\right)$—log GDP (Gross Domestic Product) per capita PPP (Purchasing Power Parity) expressed in USD;

(4)

EP—energy prices for final users (EUR/kWh);

(5)

PI—Policy/regulatory intensity index (unitless). A synthetic index summarizing the intensity of policies and regulations supporting RES; higher values indicate a stronger enabling environment;

(6)

EC—CO₂ emissions per capita (t)—available in the database, not included in the M1–M5 models presented in the output part.

We use the inverse hyperbolic sine transformation asinh(EBI/λ) to stabilise the distribution of EIB financing and to obtain log-like interpretation in the presence of zero values. The IHS function is defined at zero and approximates the natural logarithm for large values, which makes it suitable for zero-inflated and heavy-tailed financial data [69,70,71].

The data were obtained from five key databases, which made it possible to conduct research both in terms of territory and time. The list of these databases includes the following:

• Eurostat [72] (data from the area: 1, 3, 4).
• EIB data on financing of RES projects, obtained directly from the bank, taking into account the following: (1) technological structure: hydropower, onshore and offshore wind energy, solar energy (photovoltaics, PV), geothermal energy, solar energy—CSP (concentration of solar energy), biomass/biogas; (2) geographic location. The list required the manual transformation of the source data, cleaning and aggregation at the country level in individual years, which allowed the obtaining of a unique database thanks to which input data were generated and charts were prepared for each country under study (included in the analytical part of the study) (data from the area: 2).
• World Bank [73] (data from the area: 3).
• OECD (Organisation for Economic Cooperation and Development), /Energy Policy Tracker [74] (data from the area: 4).
• EDGAR (Emissions Database for Global Atmospheric Research) database [75] (data from the area: 5).
• Our World in Data [76] (data from the area: 6).

The data were collected for the years 2012–2023. The adoption of the above research period made it possible to achieve the research objective with the possibility of expanding the interpretative layer of the results from changes resulting from both the energy transition after the financial crisis and the implementation of the EU climate and energy policy (among others RED I/II, European Green Deal), as well as considering the period preceding Russia’s invasion of Ukraine and the period of Russia’s invasion of Ukraine.

The 2012–2023 horizon ensures comparability of Eurostat and EIB definitions/classifications and spans the RED II/RED III implementation period. Earlier years exhibit weaker methodological consistency, while post–2023 series were partially incomplete at the time of analysis.

These analyses were carried out on the example of selected EU Member States:

(1)

Germany (rapid switch to renewable energy),

(2)

France (high level of nuclear energy),

(3)

Spain (previously no dependence on Russian gas, diversification of energy sources),

(4)

Italy (no dependence on Russian gas),

(5)

Poland (dominant share of energy production from coal).

The selection of the five largest EU economies (DE/FR/ES/IT/PL) aligns coherently with our previous research framework, wherein the same core sample was employed (including Norway as an additional comparator) [77]. In the present study, Norway is deliberately excluded due to the absence of comprehensive, comparable datasets (EIB/Eurostat) spanning the years 2012–2023, which ensures consistency in variable definitions and the comparability of estimations. The chosen quintet reflects a diverse array of energy mixes and encompasses key segments of the EU market, thereby enhancing the informational value of the comparative analysis. The primary objective is a rigorous examination of the core market dynamics rather than a formal extrapolation to the entire European Union; such limitations are explicitly acknowledged in the section dedicated to study constraints.

The selection of research objects was deliberate, and considered the following criteria:

(1)

the diversity of leading energy sources,

(2)

geographical location and access to energy resources,

(3)

the impact of emission-based energy sources on final energy costs [77].

The research goal was achieved by the method of collecting, processing, and presenting data.

The data collection method included a review of the literature on the subject in the field of banking services for RES projects in the context of the EU climate policy, with particular emphasis on the assessment of the state of legislation in this area.

The data processing methodology encompasses a quantitative analysis utilizing descriptive statistics alongside a panel data model with fixed effects (Fixed Effects, FE) [78,79]. This approach facilitates the control of unobserved, time-invariant characteristics of countries as well as common temporal factors. The FE model presupposes that each unit (e.g., country) possesses its own time-invariant component, denoted as ${\propto }_t$, while each year is associated with a common temporal effect ${\delta }_t$. Such a framework effectively eliminates the influence of unobservable factors that could potentially distort estimation results, thereby enhancing the precision of statistical inference in panel data analyses. The model is formally expressed as:

$$Y_{it}={\alpha }_i+{\beta }_1{X}_{1,it}+{\beta }_2{X}_{2,it}+ \dots +{\beta }_k{X}_{k,it}+{\epsilon }_{it}$$

After substituting the explanatory and explanatory variables, the formula will take the form:

$${RE}_{it}={\alpha }_i+{\delta }_t+{\beta }_0\mathrm{asinh}\left({\displaystyle \frac{{EBI}_{it}}{\lambda }}\right)+{\beta }_1\ln \left(GDPp{c}_{{PPP}_{it}}\right)+{\beta }_2{EP}_{it}+{\beta }_3{PI}_{it}+{\epsilon }_{it}$$

where:

RE_it (Y_it)—share of RES in gross final energy consumption (p.p.);

${\alpha }_i$—country fixed effects;

${\delta }_t$—year fixed effects (two-way FE);

${\beta }_t$—regression coefficients (effects of EIB funding, GDP, energy prices, and policy index);

$\mathrm{asinh}\left({\displaystyle \frac{{EBI}_{it}}{\lambda }}\right)$(X_1,it)—transformation of EIB financing (EUR million) used in estimates;

λ—the median of positive EBI values within the sample; scaling ensures unit consistency and causes the Inverse Hyperbolic Sine (IHS) function to approximate a logarithmic transformation for large values, while remaining well-defined at zero;

$\ln \left(GDPp{c}_{{PPP}_{it}}\right)$(X_2,it)—log GDP per capita PPP;

EP_it (X_3,it)—energy prices for final users (EUR/kWh);

PI_it (X_4,it)—policy/regulatory intensity index (unitless);

${\epsilon }_{it}$—a random component.

Baseline variants (M1–M2):

M1 (basic, without delays)

$$R{E}_{it}={\alpha }_i+{\gamma }_t+{\beta }_0\mathrm{asinh}\left({\displaystyle \frac{{EBI}_{it}}{\lambda }}\right)+{\theta }_1\ln \left(GDPp{c}_{{PPP}_{it}}\right)+{\theta }_{2}E{P}_{it}+{\epsilon }_{it}$$

where:

${\theta }_t$—control outcomes: economic development, energy prices, policy intensity.

M2 (M1 + policies/regulations)

$$R{E}_{it}={\alpha }_i+{\gamma }_t+{\beta }_0\mathrm{asinh}\left({\displaystyle \frac{{EBI}_{it}}{\lambda }}\right)+{\theta }_1\ln \left(GDPp{c}_{{PPP}_{it}}\right)+{\theta }_{2}E{P}_{it}+{\theta }_{3}P{I}_{it}+{\epsilon }_{it}$$

To more accurately capture the delayed realization of funding effects, models incorporating delay distributions (M3–M5) were employed. In these models, alongside the current control variables, we examine the impact of past funding, quantified through transformation asinh(EBI/λ):

M3 (lag 1):

$$R{E}_{it}={\alpha }_i+{\delta }_t+{\beta }_{-1}\mathrm{asinh}\left({\displaystyle \frac{{EBI}_{i,t-1}}{\lambda }}\right)+{\beta }_1\ln \left(GDPp{c}_{{PPP}_{it}}\right)+{\beta }_{2}E{P}_{it}+{\beta }_{3}P{I}_{it}+{\epsilon }_{it}$$

M4 (t i t − 1):

$$R{E}_{it}= {\alpha }_i+{\delta }_t+{\beta }_0\mathrm{asinh}\left({\displaystyle \frac{{EBI}_{i,t}}{\lambda }}\right)+{\beta }_{-1}\mathrm{asinh}\left({\displaystyle \frac{{EBI}_{i,t-1}}{\lambda }}\right)+{\beta }_1\ln \left(GDPp{c}_{{PPP}_{it}}\right)+{\beta }_{2}E{P}_{it}+{\beta }_{3}P{I}_{it}+{\epsilon }_{it}$$

M5 (t, t − 1, t − 2):

$$R{E}_{it}= {\alpha }_i+{\delta }_t+{\beta }_0\mathrm{asinh}\left({\displaystyle \frac{{EBI}_{i,t}}{\lambda }}\right)+{\beta }_{-1}\mathrm{asinh}\left({\displaystyle \frac{{EBI}_{i,t-1}}{\lambda }}\right)+{\beta }_{-2}\mathrm{asinh}\left({\displaystyle \frac{{EBI}_{i,t-2}}{\lambda }}\right)+{\beta }_1\ln \left(GDPp{c}_{{PPP}_{it}}\right)+{\beta }_{2}E{P}_{it}+{\beta }_{3}P{I}_{it}+{\epsilon }_{it}$$

Dynamic variants (delays modelled via asinh(EBI/λ)) have been introduced to accurately capture the delayed manifestation of financing effects within both the investment cycle and the connection process.

The two-way fixed effects panel model (country and year) is the primary analytical tool to control for unobservable heterogeneity across countries and over time. The model considers three key aspects:

• Unobservable heterogeneity between countries—by using individual α_i constants, it is possible to capture specific features of constants for each country, such as institutional structure, energy mix model, political stability, or regulatory culture. Such features, while not directly measurable, can have a significant impact on the pace of the energy transition and the effectiveness of EIB investments.
• Temporal effects—represented by the complete set of binary variables for years (${\gamma }_t$)—capture common shocks and regulatory changes (e.g., COVID-19, RED II/Fit for 55, fluctuations in energy prices), thereby mitigating the bias introduced by omitted time-varying factors.

The Fixed Effects (FE) model presupposes that each individual unit possesses an unobserved, time-invariant component. Estimation predicated on within-unit variability mitigates the influence of omitted, time-invariant factors:

• Eliminate the influence of unobserved constant variables over time that may be correlated with regressors (explanatory variables).
• Analysing dynamic changes occurring in units (e.g., EU countries) regardless of their individual characteristics.
• Increasing the accuracy of statistical inference, especially in research based on observational data of a panel nature.

The correctness of the assumptions of the model will be carried out as a result of the application of a series of diagnostic tests: Pesaran CD test, VIF indicators, and AIC/BIC.

The fundamental inference is grounded in Driscoll–Kraay standard errors (L = 2) [80]. The primary objective of the Pesaran CD test is to assess the null hypothesis of cross-sectional independence in the residuals of panel models. Under the null hypothesis (H₀), the CD statistic approximately follows a standard normal distribution, N(0,1). Decisions are based on the absolute value |CD| and the corresponding p-value. Given the modest sample size (small N) and moderate time dimension (T), the power of the test is inherently limited; consequently, inference primarily hinges on Driscoll–Kraay standard errors (L = 2) as the main approach. The comparisons were conducted on identical samples, including variants with lagged variables [81].

The Variance Inflation Rate (VIF) for the X_j regressor is defined as VIF_j = 1/(1 − R_j²), where R_j² is derived from the auxiliary regression of X_j to the other explanatory variables. VIF values between 1 and 5 should usually not raise objections. Values above 10 are sometimes considered a signal of problematic collinearity. In models with fixed effects, VIF will be computed for continuous variables. In FE models, VIF calculations are performed solely for continuous variables; binary (dichotomous) variables—such as country or year fixed effects—are not subjected to VIF assessment [82].

The information criteria link model is fit with a complexity penalty: AIC = −2·ln(L) + 2k, BIC = −2·ln(L) + k·ln(N), where L is the maximum value of the confidence function, k is the number of parameters, and N is the number of observations. A comparative rule is that lower AIC/BIC values will indicate a better trade-off between fit and complexity. Cautious comparisons will be most reliable when models are estimated on the same sample; in delayed shots (shorter T) [83,84].

Based on the results of diagnostic tests, it is possible to confirm the validity of the Driscoll–Kraay estimator, which provides resistance to heteroskedastic, autocorrelation, and spatial correlation in panel data with a small number of units and a long observation period. The estimation was carried out in the STATA/19 environment (Timberlake Consultants), and the database was prepared based on Excel (Office 365).

The method of data presentation included a graphical approach in the form of tabulars and graphs.

The analyses have been carried out in a normative approach, which makes it possible to determine the conditions necessary to design new solutions in the sphere of finance [85].

4. Results

In accordance with the construction of indicators and the selection of control variables in the panel regression model with fixed effects adopted in the methodological part, the aim was to obtain answers to two research questions:

• What level of financing provided by the EIB for RES projects causes an increase in the share of RES in gross final energy consumption in EU countries?
• What factors determine the effectiveness of EIB financing for RES projects in EU countries?

In the study, it was assumed that the dependent variable is RE (in p.p.). The scale of financing is measured in EIB amounts (EUR million), and, due to the skewness of the distribution and the presence of zeros, the asinh(EBI/λ) transformation has been applied. The control vector includes $\ln \left(\mathrm{G}\mathrm{D}\mathrm{P}\mathrm{p}{\mathrm{c}}_{\mathrm{P}\mathrm{P}\mathrm{P}}\right)$, final energy prices (EP, EUR/kWh), and the Policy Intensity Index (PI). All models are estimated with bidirectional fixed effects (country and year). The dynamic variants (M3–M5) introduce EIB delays (t − 1, t − 2) to capture the investment and connection cycle. The inference is based on Driscoll–Kraay standard errors (L = 2), which are robust to heteroskedasticity, autocorrelation, and cross-sectional dependence in short panels.

4.1. Sample Characteristics and Variation of Key Variables

The analysis covers the following countries: Germany (DE), France (FR), Spain (ES), Italy (IT), and Poland (PL) in the years 2012–2023. The dependent variable is RE (in p.p.). The key variable is the EBI (EUR million), which is estimated asinh(EBI/λ) transformation due to the skewness of the distribution and the presence of zeros. The set of controls includes $\ln \left(\mathrm{G}\mathrm{D}\mathrm{P}\mathrm{p}{\mathrm{c}}_{\mathrm{P}\mathrm{P}\mathrm{P}}\right)$, final energy prices (EP, EUR/kWh), and the Regulatory Policy Intensity Index (PI). All specifications use bidirectional fixed effects (country and year), which neutralizes the constant differences between countries over time and common time shocks. The panel includes a total of 60 observations.

Table 1. Descriptive statistics (2012–2023; DE/FR/ES/IT/PL).

Variable	N	Average	SD	Min	P25	P50	P75	Max
RE (p.p.)	60.00	16.73	3.00	10.96	14.90	16.64	18.93	24.85
EIB (EUR million)	60.00	511.78	500.41	0.00	129.83	405.44	748.40	2516.08
EP (EUR/kWh)	60.00	0.15	0.05	0.09	0.12	0.14	0.17	0.32
GDP per capita (PPP, USD)	60.00	47,176.53	9063.48	28,371.92	41,284.04	47,280.62	52,976.34	62,579.07
PI (index)	60.00	3.27	0.70	2.22	2.83	3.09	3.47	4.89

Source: own study based on the calculations carried out.

presents the basic characteristics of variables for the entire panel.

Due to the high volatility of EIB financing and the presence of zeros, we use the inverse hyperbolic sine transformation asinh(EBI/λ), which behaves similarly to the natural logarithm for large values and is defined at zero. Logarithmic transformations of income ln($\mathrm{G}\mathrm{D}\mathrm{P}\mathrm{p}{\mathrm{c}}_{\mathrm{P}\mathrm{P}\mathrm{P}}$) are applied to reduce the impact of extreme observations. The “two-way within” transformation (demeaning by country and year) is used only for visualisation purposes, to illustrate relationships after removing fixed effects. Due to the high variability of the EBI over time and across countries, the inverse hyperbolic sine (IHS) transformation was applied to the EBI. The utilization of the asinh(EBI/λ) and ln($\mathrm{G}\mathrm{D}\mathrm{P}\mathrm{p}{\mathrm{c}}_{\mathrm{P}\mathrm{P}\mathrm{P}}$) functions serves to mitigate the influence of extreme observations; the IHS transformation is defined at zero and approximates the logarithmic function for large values. The “two-way within” transformation (demeaning by country and year) was employed solely for visualization purposes, specifically to illustrate relationships after the removal of fixed effects.

Over the period 2012–2023 (Figure 1), the share of renewable energy (RES) has been growing in all five countries surveyed, with the largest increases in Spain (+10.6 p.p., 14.24% → 24.85%) and France (+9.0 p.p., 13.24% → 22.28%), stable growth in Germany (+8.0 p.p., 13.55% → 21.55%), and moderate changes in Italy (+4.2 p.p., 15.44% → 19.59%) and Poland (+5.6 p.p., 10.96% → 16.56%). Short-term fluctuations are visible (among others declines in 2021 in Spain, Italy, and Poland), but the overall trend remains upwards. This confirms the progressive diffusion of RES in the economies of the EU countries analysed.

It is worth noting that the EIB’s financing of RES projects (Figure 2) is characterized by high variability between countries and over time, with clear peaks in the last two years.

Spain reached a record high in 2023 (EUR 2516.08 million), France recorded a very high level in 2021 (EUR 1582.34 million), and Poland reached EUR 745.64 million in 2023, while Italy maintained moderate levels with fluctuations (2013 peak: EUR 932.15 million, EUR 594.82 million in 2023).

4.2. Model Specifications

The estimations employing two-way fixed effects (country and year) for models M1–M5 (

Table 2. Estimated effects of asinh(EBI/λ) on RE (p.p.); two-way fixed effects (country and year).

Model	asinh(EBI/λ)	$\mathbf{ln}\left(\mathbf{G}\mathbf{D}\mathbf{P}\mathbf{p}{\mathbf{c}}_{\mathbf{P}\mathbf{P}\mathbf{P}}\right)$	EP (EUR/kWh)	PI (Indeks)
M1	0.0031 (SE_DK 0.0028) [p = 0.333]	−0.0489 (SE_DK 0.0261) [p = 0.135]	−0.1247 (SE_DK 0.0627) [p = 0.118]
M2	0.0033 (SE_DK 0.0027) [p = 0.284]	−0.0446 (SE_DK 0.0259) [p = 0.161]	−0.1131 (SE_DK 0.0607) [p = 0.136]	−0.0095 (SE_DK 0.0058) [p = 0.176]
M3 (t − 1)	0.0014 (SE_DK 0.0030) [p = 0.672]
M4 (t)	0.0030 (SE_DK 0.0033) [p = 0.415]	−0.0455 (SE_DK 0.0277) [p = 0.176]	−0.1186 (SE_DK 0.0622) [p = 0.129]	−0.0064 (SE_DK 0.0079) [p = 0.467]
M4 (t − 1)	0.0000 (SE_DK 0.0033) [p = 0.991]
M5 (t)	0.0018 (SE_DK 0.0038) [p = 0.660]	−0.0454 (SE_DK 0.0315) [p = 0.222]	−0.1144 (SE_DK 0.0611) [p = 0.134]	−0.0089 (SE_DK 0.0084) [p = 0.349]
M5 (t − 1)	−0.0011 (SE_DK 0.0034) [p = 0.755]
M5 (t − 2)	0.0020 (SE_DK 0.0035) [p = 0.597]

Source: own study based on the calculations carried out.

) elucidate the relationship between the dependent variable RE and the transformed variable asinh(EBI/λ), as well as a set of control variables: ln($\mathrm{G}\mathrm{D}\mathrm{P}\mathrm{p}{\mathrm{c}}_{\mathrm{P}\mathrm{P}\mathrm{P}}),$ EP, and PI. The specifications encompass both contemporaneous and lagged components (up to two years). The coefficient β associated with asinh(EBI/λ) captures the marginal effect of EBI financing; the IHS transformation is defined at zero and approximates the logarithm for large values. This effect is contingent upon the regulatory environment (PI) and network capacity, thereby justifying the prioritization of network investments and regulatory reforms.

In the current effect variants (M1–M2) presented in Table 2, the coefficient associated with asinh(EBI/λ) is positive, small in magnitude, and statistically insignificant. Specifically, in M1, the estimate β = 0.0031 with a standard error ${SE}_{DK}$ = 0.0028 and $p=0.333$. Upon inclusion of the policy/regulation indicator (PI) in M2, the parameter remains similarly estimated (β = 0.0033, ${SE}_{DK}$= 0.0027, p = 0.284), indicating that the addition of PI does not materially alter the conclusions regarding the magnitude and significance of the current funding effect of the EBI. In the dynamic variants (M3–M5), the lagged effects do not attain statistical significance. In M3, the coefficient t − 1 is β = 0.0014 with ${SE}_{DK}$ = 0.0030 and p = 0.672. In M4, the current period (t) coefficient is β = 0.0030 (${SE}_{DK}$= 0.0033, p = 0.415), while the lagged component at t − 1 approaches zero ($\beta \approx 0.0000$, ${SE}_{DK}$ = 0.0033, p = 0.991). In M5, the coefficients are ${\beta }_t=0.0018$ (${SE}_{DK}$ = 0.0038, $p=0.660$), ${\beta }_{t-1}=-0.0011$ (${SE}_{DK}$ = 0.0034, $p=0.755$) and ${\beta }_{t-2}=0.0020$ (${SE}_{DK}$ = 0.0035, $p=0.597$). In all variants, the distribution of lagged effects does not reach statistical significance at the 10% level.

The interpretation of the magnitude of the coefficients accounts for the applied IHS transformation. Specifically, the coefficient associated with asinh(EBI/λ) should be understood as the marginal effect of the transformed financing variable on RE. The IHS transformation is defined at zero and approximates the logarithm for large values; consequently, the conventional semi-elasticity based on the rule “+10% → 0.1 × β p.p.” is not employed here. Inference is based on Driscoll–Kraay standard errors with a Bartlett kernel and a small bandwidth (L = 1–2), which are robust to heteroskedasticity, serial correlation, and cross-sectional dependence in short panels.

The model fit remains high and stable, with an adjusted R² in the range of approximately 0.87 to 0.90. However, the scale of the estimated effects of EBI financing is small, and their statistical significance does not manifest across the examined specifications. When combined with the influence of regulatory conditions (PI) and network constraints, this suggests that an increase in financing volume does not mechanically translate into a short-term rise in the share of renewable energy sources in final energy consumption. This finding aligns with the conclusions presented in Table 2.

The comprehensive regression results for models M1–M5, estimated within a fixed effects framework accounting for countries and years, incorporating Driscoll–Kraay standard errors (L = 2), as well as adjusted R² metrics (within), are presented in

Table 3. Complete regression results—FE (country and year); SE: Driscoll–Kraay (L = 2). The funding measure: asinh(EBI/λ), with λ = 447.38 million EUR. We report the coefficient β, SE_DK, and the p-value; at the bottom: the sample size N and the adjusted R².

Variable	M1	M2	M3 (t − 1)	M4 (t)	M4 (t − 1)	M5 (t)	M5 (t − 1)	M5 (t − 2)
asinh(EBI/λ)	0.0031 (0.0028) [p = 0.333]	0.0033 (0.0027) [p = 0.284]	0.0014 (0.0030) [p = 0.672]	0.0030 (0.0033) [p = 0.415]	0.0000 (0.0033) [p = 0.991]	0.0018 (0.0038) [p = 0.660]	−0.0011 (0.0034) [p = 0.755]	0.0020 (0.0035) [p = 0.597]
$\mathbf{ln}\left(\mathbf{G}\mathbf{D}\mathbf{P}\mathbf{p}{\mathbf{c}}_{\mathbf{P}\mathbf{P}\mathbf{P}}\right)$	−0.0489 (0.0261) [p = 0.135]	−0.0446 (0.0259) [p = 0.161]	−0.0458 (0.0274) [p = 0.169]	−0.0455 (0.0277) [p = 0.176]		−0.0454 (0.0315) [p = 0.222]
EP (EUR/kWh)	−0.1247 (0.0627) [p = 0.118]	−0.1131 (0.0607) [p = 0.136]	−0.1185 (0.0625) [p = 0.131]	−0.1186 (0.0622) [p = 0.129]		−0.1144 (0.0611) [p = 0.134]
PI (indeks)		−0.0095 (0.0058) [p = 0.176]	−0.0074 (0.0080) [p = 0.404]	−0.0064 (0.0079) [p = 0.467]		−0.0089 (0.0084) [p = 0.349]
N	60	60	55	55		50
R2 corrected	0.901	0.901	0.886	0.884		0.869

Notes: Bidirectional FE (country and year). Estimation method: Driscoll–Kraay standard errors (L = 2). Financialization measure: asinh(EBI/λ), where λ = 447.38 million EUR (median of positive EBI). In the first row: β with significance indicators; in parentheses: SE_DK and p-values. Source: own study based on the calculations carried out.

.

The high consistency of signs and the comparable magnitudes of the coefficients confirm the robustness of the estimates summarized in Table 2. Consequently, Table 3 provides a comprehensive empirical foundation for evaluating both the direction and the strength of the relationship between EBI financing (transformed via asinh(EBI/λ)) and the dynamics of renewable energy share (RE) in final energy consumption.

It therefore becomes crucial whether the additional complexity of the specification (PI and delays) translates into a measurable improvement in matching. The answer is provided by comparing model metrics, which is included in

Table 4. Model matching metrics (M1–M5).

Model	N	R2 Corrected	AIC	BIC
M1	60	0.901	−373.98	−334.19
M2	60	0.901	−373.79	−331.9
M3	55	0.886	−336.99	−298.85
M4	55	0.884	−335.64	−295.49
M5	50	0.869	−300.1	−261.86

Source: own study based on the calculations carried out.

—a comparison of the quality, fit, and complexity of five FE specifications (country+year) was made on a common set of countries DE/FR/ES/IT/PL in the years 2012–2023.

All models attain high adjusted R² values, ranging from 0.869 to 0.901. Such levels are characteristic of estimations incorporating extensive fixed effects and should not be interpreted as indicative of a strong influence exerted by the individual predictor asinh(EBI/λ) on RE. The appropriateness of extending the specifications is primarily informed by the AIC and BIC criteria, which integrate measures of model fit with penalties for model complexity. It is important to note that the absolute values of AIC and BIC are dependent on the scale of the dependent variable; therefore, model comparisons are most reliable when the estimations are performed on the same sample of observations [83,84]. Among the models estimated on a sample size N equal to 60, the M1 and M2 models yield remarkably similar results concerning the criterion values. In the sample with N equal to 55, the M3 model demonstrates a marginal advantage, achieving an AIC of approximately −336.99 and a BIC of approximately −298.85, thereby slightly improving the fit relative to M4, which has an AIC of around −335.64 and a BIC of approximately −295.49. The M5 model, estimated on the smallest sample size N equal to 50, does not exhibit a substantial enhancement in fit compared to M3 and M4, as evidenced by an AIC of roughly −300.10 and a BIC of approximately −261.86. The varying sample sizes result from the inclusion of delays and underscore the necessity for caution when comparing criterion values across different model groups. Consequently, we recommend Model M3 as the reference specification for economic interpretation. This model aligns with the logic of delayed realization of financial effects and ensures comparability of samples within variants featuring delays. We regard Model M4 as a supplementary framework that distinguishes between the current and delayed components. Model M5 remains a control extension applied to a shorter sample, serving to evaluate the robustness of the conclusions drawn.

Following the selection of the reference specification, the subsequent step involves evaluating the cumulative impact of financing, understood as the sum of both contemporaneous and lagged effects. In model M5, we estimate the long-term effect, denoted as β_LR, which is formulated as the sum of the coefficients β_t, β_(t − 1), and β_(t − 2) for the measure asinh(EBI/λ).

Table 5. The long-term effect in Model 5 (sum of coefficients). The long-term effect is calculated as β_LR = β_t + β_{t − 1} + β_{t − 2}; the confidence intervals and hypothesis tests are determined using the delta method applied to the Driscoll–Kraay covariance matrix (L = 2).

Measure	Value
Sum β (t + t − 1 + t − 2)	0.0027
SE (DK, L = 2)	0.0044
p (DK)	0.539
95% CI (DK)	[−0.0059; 0.0113]
N	50
R2 corrected	0.869

Source: Own calculations based on data from 2012 to 2023; the measure of financing: asinh(EBI/λ), where λ represents the median of positive EBI values within the sample; bidirectional FE (country and year).

reports the value of this sum of coefficients, along with the associated standard errors and tests derived using the delta method based on the Driscoll–Kraay (L = 2) covariance matrix.

We obtained ${\beta }_{LR}=\mathrm{0,0027}$ with ${SE}_{DK}$ = 0.0044 I $p_{DK}=0.539$, resulting in a 95% confidence interval of [−0.0059; 0.0113]. This conclusion is consistent across both covariance matrix specifications: the cumulative effect of EBI funding on RE is minimal and statistically non-significant in the short-term horizon. These findings align with the estimates obtained in models M1 through M4, wherein both current and lagged effects are small and not statistically significant, with their signs and magnitudes remaining coherent. The interpretation of the coefficient magnitude considers the inverse hyperbolic sine (IHS) transformation. Specifically, the coefficient associated with asinh(EBI/λ) has a marginal interpretation in the context of the transformed variable and does not correspond to a logarithmic “semi-elasticity” based on the rule that “a ten percent increase results in a 0.1 times beta change,” as discussed in reference [79]. The IHS transformation is defined at zero and approximates the logarithm for large values, thereby justifying its application in the presence of zeros and skewness. Inference is primarily conducted based on Driscoll–Kraay standard errors (L = 2), which are robust to heteroskedasticity, autocorrelation, and cross-sectional dependence in short panels. Overall, the M5 estimator is derived from a sample of 50 observations, and the R-squared value, Rskoryg.2, is approximately 0.869, remaining consistent with the fit levels observed in other specifications.

To complement the findings presented in Table 3, Figure 3 synthesizes the impact of financing, represented as asinh(EBI/λ), in both contemporaneous and lagged specifications, while also illustrating the uncertainty associated with the estimates. The points depicted correspond to the β coefficient values derived from FE models incorporating country and year fixed effects, and the vertical lines denote the 95 percent confidence intervals calculated using the Driscoll–Kraay method with a memory length of two. The horizontal line at zero serves as a reference to facilitate the assessment of statistical significance.

In Figure 3, all confidence intervals encompass zero, and the estimates are relatively small, which aligns with the findings presented in Table 2 and Table 3. In the variants incorporating delays M3–M5, statistically significant effects are not observed, and both the signs and magnitudes of the estimates remain stable and modest. The primary inference is based on Driscoll–Kraay standard errors. In all specifications, the qualitative conclusion of no sizeable short-term or lagged effects remains unchanged.

After subtracting the constant effects of the country and the year (see Figure 4) in the form of a ‘cloud’, the points were almost flat with a slightly positive slope.

The estimated slope of the regression line is approximately 0.0038, and the correlation coefficient r ≈ 0.154 with N = 60, indicating a very weak positive association between RE I asinh(EBI/λ) and its fluctuations within countries over time. The data presented in Figure 4 are of a diagnostic nature: constant differences across countries and common temporal shocks have been accounted for through the ‘within’ estimator; however, this does not establish causality nor elucidate the dynamics of lagged effects. We consider it as a supplementary element to the conclusions derived from the fixed effects models across variants M1–M5.

Visual diagnostics were supplemented with formal residual tests.

Table 6. Cross-sectional dependence on Pesaran CD residues.

Model	CD	p	N	T
M1	−1.8844	0.0595	5	12
M2	−1.9207	0.0548	5	12
M3	−1.7751	0.0759	5	11
M4	−1.8664	0.062	5	11
M5	−1.9808	0.0476	5	10

Source: own study based on the calculations carried out.

presents the Pesarana CD statistics for models M1–M5, which serve to evaluate the cross-sectional dependence scale, thereby justifying the selection of the error estimators. The null hypothesis of the test assumes the absence of cross-sectional dependence. The columns labelled “CD” and “p” report, respectively, the standardized value of the statistic—approaching a standard normal distribution, N(0,1), under the null hypothesis—and the corresponding significance level. N denotes the number of cross-sectional units, while T represents the length of the effective time series in each specification, with shorter durations observed in models incorporating lag structures.

The rejection of the null hypothesis is indicated by the absolute values of the test statistic (CD) exceeding approximately 1.96 at a significance level of p < 0.05. The highest degree of comparability among the results is observed when the models are estimated using the same sample, which is particularly relevant when differing sample lengths (T) arise due to the incorporation of lag structures.

In the examined sample, the values of the statistics fluctuate near the threshold of significance. Specifically, for M1 CD is −1.884 with p = 0.060, for M2 −1.921 with p = 0.055, for M3 −1.775 with p = 0.076, for M4 −1.866 with p = 0.062, and for M5 −1.981 with p = 0.048. The negative sign bears no bearing on the test decision and merely indicates that the average residual correlation across countries is negative. The absolute value of the statistic and the p-value are of primary importance. Variations in the T statistic among the models stem from the incorporation of lagged variables, which may influence the test’s power; therefore, direct comparisons between model groups should be approached with caution. In light of these findings, the primary conclusions of the article are based on Driscoll–Kraay standard errors, which are robust to cross-sectional dependence and autocorrelation in short panels.

As a result of the studies carried out, there was no evidence of a positive short-term effect of the scale of EIB financing on the RE. At most, a small, delayed effect is visible. The estimates remain statistically insignificant in inference based on Driscoll–Kraay standard errors, which can be interpreted through the prism of selection and a long path of materialization. This does not mean that public funding is ineffective, but points to the need for better targeting and coordination—especially network investments. In this sense, the EIB and the European Commission (EC), in cooperation with governments, operators, and the banking sector, have tools that can increase the efficiency of the conversion of euros into RE percentage points, if projects with a clear, technically documented translation into RES capacity, connections, and integration become priority.

The effectiveness of EIB financing is determined primarily by regulatory and institutional conditions (PI; pace and certainty of permits), the system’s ability to absorb new capacity (grid capacity, storage, demand flexibility), and the quality and maturity of projects, together with financial assembly and implementation risk mitigation instruments. The empirical results are consistent with this. When PI is considered, the current effect weakens (M2), and the signal materializes with a delay (M3–M5). Simply increasing the volume of funds for RES investments is not enough without meeting the above conditions. Effectiveness only increases when funding goes hand in hand with regulatory reforms and investments into the grids.

5. Discussion

The estimation results of models M1–M5, incorporating country and year fixed effects, alongside Driscoll–Kraay inference, indicate that the scale of EIB financing—measured via the transformed asinh(EBI/λ)—does not exert a statistically significant short-term influence on the share of RES in final energy consumption. From an interpretative standpoint, the efficacy of financing primarily depends on the regulatory and institutional environment (PI), network capacity, as well as the quality and maturity of project portfolios. Once the regulatory environment is accounted for, the current impact component diminishes, and the lagged effects stabilize, reflecting the temporal requirements for permitting, implementation, and grid connection processes. Practically, this suggests that the magnitude of funding flows alone is insufficient; effectiveness is enhanced when financing is linked to regulatory reforms and investments in grid infrastructure, storage, and demand flexibility. Moreover, project selection tends to favour initiatives with a direct, technically substantiated impact on capacities, interconnections, and integration. The limitations of this study include its focus on five countries and the moderate length of the time series, which warrants cautious interpretation. The RE indicator also responds to fluctuations in the denominator. The long-term effect in model M5 remains small and statistically insignificant, consistent with the findings from models M1–M4. Consequently, to contextualize these results within a broader framework, cross-country comparisons within the EU are employed. Such analysis enables the alignment of short-term estimates with the variability in demand-side and supply-side indicators and facilitates an examination of their co-occurrence with energy performance outcomes.

These findings align with an academic research trajectory focused on evaluating the role of RES in gross final energy consumption. However, these results cannot be directly extrapolated from the existing body of research, as the authors concentrated on different research questions concerning RE. Specifically, the analyses primarily addressed the following:

• The disparities among EU member states concerning two categories of indicators are examined: (1) those reflecting changes in end-user energy demand, including considerations of energy derived from renewable sources; and (2) those associated with the security of raw material supplies, specifically, energy dependency indicators segmented by primary energy sources and the overall energy efficiency metric. The analysis revealed no statistically significant correlation between the level of sustainable energy consumption and either energy efficiency or reliance on energy imports. Conversely, a statistically significant correlation was identified between the dependency of crude oil and petroleum products on energy imports and primary energy consumption. Additionally, a notable positive correlation was observed between the share of renewable energy in gross final energy consumption and the overall dependency on energy imports [86].
• Identification of the determinants that significantly influence the share of total RES in the individual voivodeships of Poland within the context of the country’s overall electric energy consumption. Through qualitative analysis covering the years 2005–2019, a negative correlation was established between energy consumption levels and the proportion of renewable energy sources in the total energy production. Conversely, regarding the expenditures allocated to research and development activities as well as the total investments in environmental protection and water management, no statistically significant impact on the development of RES was observed [87].
• The examination of the relationship between GDP and the share of RES in final energy consumption was a central focus of the analysis. Certain segments of the study concentrated on elucidating this dependence while incorporating GDP per capita as an additional variable. Utilizing an analytical framework applied to the EU-28 member states over the period from 2007 to 2017, it was demonstrated that a low but positive correlation exists between these variables. This finding suggests that more developed nations possess a greater potential for increased RES consumption. In essence, countries with higher GDP per capita tend to exhibit a larger share of renewable energy in their final energy use. Nonetheless, other latent factors may also significantly influence the intensive adoption of RES in energy production and consumption [88]. Furthermore, some researchers incorporated additional variables—such as investment rates and energy intensity—into their analyses involving GDP per capita. Such studies, conducted on EU member states between 2015 and 2023, established that the influence of GDP per capita on RES share is structurally positive but lacks statistical significance regarding temporal changes within individual countries. Conversely, the investment rate was found to be positively and significantly correlated with the share of renewable energy. Moreover, during periods in which investment rates increased within a given member state, the share of renewable sources in gross final energy consumption also generally rose. No significant relationship was identified between energy intensity and RES share [89].
• The analysis examined the correlations between the share of renewable energy in final energy consumption and the economic strength of a country, assessed through seven indicators: real GDP growth, unemployment rate, inflation rate, exports of goods and services, public debt, foreign direct investment, and the labour cost index. The findings indicated that the selected macroeconomic indicators do not exert a statistically significant influence on the development of the renewable energy sector within the European Union. Only in certain countries (notably Denmark, Ireland, Greece, France, and Cyprus) was a statistically significant relationship established between several discrete indicators, such as the unemployment rate, export levels, and the labour cost index. More substantial correlations were identified between renewable energy deployment and the volume of foreign direct investment across numerous EU member states. The conclusions emphasized that, although specific macroeconomic factors do not directly drive the growth of renewable energy, its expansion is predominantly determined by broader national conditions and the overarching economic, political, and institutional environment [90].
• Determination of the impact of sectoral economy on the share of renewable energy consumption (case study: Ethiopia). It has been established that sustainable financing programs are indispensable for the development and support of renewable energy projects in both the short- and long-term perspectives [91].
• Assessment of the implementation of models and financing mechanisms within the energy efficiency sector (exemplified by Bulgaria, Croatia, Greece, Romania, and Slovenia) across various domains of economic activity. The analysis emphasizes that the development of sustainable financial instruments is contingent upon robust political support and the acknowledgment of their significance by all relevant stakeholders.

Therefore, the demonstrated absence of a statistically significant relationship between the scale of EIB funding and short-term RE growth within the examined cohort of EU countries may serve as a valuable reference point for future researchers interested in this particular issue. This is especially pertinent given that, consistent with prior studies, the analyses suggest that the increase in the share of RES within gross energy consumption can be better understood within a broader political and institutional context, which encompasses factors such as the enhancement of regulatory frameworks governing energy services. Consequently, further investigations into RE development that incorporate the role of the EIB as the principal climate-oriented financial institution should be pursued to deepen our understanding of the multifaceted influences shaping renewable energy trajectories.

6. Conclusions

As a result of the research, the relationship between the EIB’s financing of RES projects and the level of energy transition, measured by the share of RES in gross final energy consumption (RE), was determined. In the course of the analyses, the following were determined:

(1)

the level and dynamics of EIB financing in the DE/FR/ES/IT/PL sample in the period 2012–2023;

(2)

factors determining the effectiveness of this financing, identified as an increase in the share of RES (RES) in final energy consumption.

On the basis of the analysis, our results indicate that that the scale of EIB financing alone does not translate into short-term growth of the RE in the studied sample (EU countries) (positive validation of hypothesis 1). The estimates for asinh(EBI/λ) are statistically insignificant in both current and lagged variants (M3–M5). Research shows that the volume of funds alone is not enough to quickly increase the RE ratio. Delays resulting from the investment and connection cycle create time gaps between the financial decision and the effect in the statistical data. Additionally, design selection and potential endogenicity (including inverse causality) weaken the short-term relationship.

According to the results of the research, it should be stated that the effectiveness of financing depends on the regulatory and institutional context (partially “captured” by PI in models (Regulatory Policy Intensity Index), the capacity of the grid to absorb new capacities (capacity, storage, demand flexibility) and the time horizon—delayed materialisation of effects (positive validation of hypothesis 2). The amount of expenditure alone does not generate a significant increase in RE in the short term (i.e., up to two years).

Based on the empirical findings, the results point to several priority directions for improving the effectiveness of EIB and public financing of RES. Acceleration of permitting and connection procedures is essential. Simplifying, standardising and digitalising permitting and grid-connection processes can shorten the time between financial close and commissioning, thereby reducing the investment lag and strengthening the short-term translation of financing into additional RE [92].

Investments in grid integration need to be further prioritised. Shifting the emphasis from standalone generation projects to integrated “power + grid” packages—including reinforcement and expansion of transmission and distribution networks, interconnectors, storage, and demand-side flexibility—is a precondition for absorbing new RES capacities and avoiding curtailment or long connection queues [93].

Portfolio rebalancing is required. Public and EIB support should move from dispersed, generation-only portfolios towards coherent portfolios combining generation, networks and flexibility assets. Such a rebalanced structure increases the probability that each euro of financing will result in a measurable increase in the RE indicator, rather than in delayed, underutilised, or curtailed capacity [94].

Optimisation of risk-sharing mechanisms is also important. A higher share of mixed (public–private) mechanisms and instruments that mitigate implementation risks—such as completion guarantees, CAR/EAR insurance, DSRA, LDs, and step-in rights—can reduce delays in project execution and improve the conversion of approved financing into commissioned and operating RES projects [95].

Finally, more targeted and selective support is needed. Financing should prioritise projects with direct and relatively near-term leverage on the RE indicator, including connection-ready RES projects, grid-enhancing investments, and flexibility solutions in regions with high curtailment or constrained network capacity, rather than broadly distributed support with distant or tenuous links to final energy consumption [96].

Based on the conducted analysis, conclusions can be drawn for EU energy policy and its framework governing the financing system for RE expansion (including financing provided by the EIB). From the perspective of the EIB, the EC, national governments, transmission system operators (TSOs) and distribution system operators (DSOs), the banking sector, RES developers, and local governments, the emphasis shifts from the amount of funds invested to the quality and subject of the investment. This means prioritising grid investments (bottlenecks, energy storage, demand flexibility) for the absorption of new capacity and regulatory improvements shortening permit and connection paths (which partly reflects PI). There is also a need for greater selectivity, a focus on projects with direct leverage on the RE indicator, rather than distributed support that has distant or tenuous translation into final energy consumption. For the EIB and the EC, the conclusions of the research may imply the codification of the allocation criteria in the thematic programmes (power +grid). For TSOs/DSOs, the recommendations from the research may concern the programming of investments in conjunction with connection plans and the location of sources, and for commercial banks and local governments, co-financing and closing the implementation chain. Bearing in mind that the RE index is recorded as the quotient “energy from RES/final consumption”, it also reacts to fluctuations in the value of the denominator (economic situation, temperature, consumption structure). The EIB part of the financing channelled through grids and upgrades does not increase RES production in the year of commissioning but creates an integration capacity that materialises with a delay. Negative and small short-term estimations therefore do not cancel out the usefulness of financing in the long term. The criteria of bankability, maturity, and connection capacity of projects should be explicitly taken into account in the selection process in order to increase the likelihood of translating financing into a real increase in the production and consumption of energy from RES.

EIB operations are implemented within the evolving legal framework of EU energy policy. Therefore, in the light of the results of the research, it is worth recalling that since November 2023, the RED III Directive has been in force with a binding target of 42.5% of RES by 2030 (aspirationally 45%), which increases the pressure for rapid capacity increases and system integration [97]. In turn, in May 2024, a reform of the EU electricity market (EMD) was adopted, which, among others strengthens long-term contracts (PPAs, bidirectional CfDs) to reduce the cost of capital for low-carbon investments [98]. In parallel, the EC has launched a Grid Action Plan and is preparing the European Grid Package 2025, highlighting the scale and urgency of grid inputs [99]. In light of these EU energy policy regulations, in 2024 the EIB recorded record financing of power grids, interconnectors, and storage (EUR 8.5 billion), confirming the shift in emphasis to grid capacity. At the same time, the auctions of the European Hydrogen Bank are progressing [100].

Increasing the efficiency of the conversion of the euro into RE percentage points requires better targeting (power + grids), simplification of procedures, and good financial assembly with the right allocation of risks. In these conditions, public financing, including from the EIB, has the greatest chance of translating into a sustainable growth of the RE on a system scale.

This approach is particularly significant, given that the implementation strategy of the European Union’s energy policy must be founded upon broad societal acceptance. Beyond strictly academic forums, outreach activities using accessible formats (e.g., long-form blog posts, explainer videos, and podcast series) can help build social acceptance for EU energy policy. Popularizing evidence-based findings on how banking sector participation and EIB co-financing translate into grid capacity, connections, and RES integration may support public understanding and the legitimacy of the transition.