Electricity Prices in the European Union Region: The Role of Renewable Energy Sources, Main Economic Factors and Market Liberalization. (2023)

Author(s): George E. Halkos (corresponding author) [*]; Apostle of Saint Tsirivis

1. Introduction

Electricity is an indispensable energy resource for both households and industry. The quality of life and economic prosperity of millions of people depend in large part on uninterrupted access to a high-quality, affordable electricity grid. The impact of the recent military conflict and the escalation of geopolitical tensions in Eastern Europe highlighted the high dependence of the European Union (EU) on imported fossil fuels to ensure security of electricity supply [1]. The entire European economy, governments and citizens were suddenly exposed to potential energy shortages and extreme price risks. This development highlighted more than ever how important it is to improve the autonomy of energy resources, removing existing market distortions that delay the transition to a deregulated electricity market.

The EU has been a key force in implementing the United Nations (UN) Sustainable Development Goals for the production of clean and affordable energy (SDG 7), which should mobilize countries across the world to take precautionary measures against the phenomenon. of climate change. change and energy poverty. Since 2015, when the UN first approved the relevant directive [2], the EU has developed an innovative energy policy that focuses on environmental protection and market opening [3]. This ambitious strategy was based on two main pillars. The first is to try to reduce CO2 emissions while increasing the EU's energy self-sufficiency, including strict environmental regulations for electricity generation and numerous incentives for the rapid deployment of renewable energy [4]. Likewise, the second intensified efforts to liberalize the electricity market and pointed to monopoly and oligopoly relationships in the generation, distribution and retail sale of energy [5]. The EU's strategic energy planning aimed to encourage the entry of new market participants by gradually removing all bureaucratic obstacles and potentially dissuasive legal frameworks in all member countries [6]. In this way, existing market barriers would finally be removed, competition would be promoted and electricity prices could be freely determined. However, according to [7, 8, 9], achieving the desired level of competition and market openness in the consolidating European energy market remains a very long and laborious process, which requires specific policy measures to reform the energy structure. existing market, making it more attractive and accessible to new entrants.

Numerous studies, including [10,11], have praised the role of electricity as a valuable social good, emphasizing the importance of electricity as the main energy product for European households. However, combining environmental protection with affordable energy and the sustainable development of the electricity sector is a great challenge and an extremely complicated task for national regulators and the EC. Taking into account the complexity of electricity markets, this study seeks to show how the percentage of renewables in the fuel mix for electricity generation, important economic factors and market liberalization affect electricity prices for consumers in many countries. Europeans.

The vast majority of scientific work focuses on the impact of renewable energy generation from specific sources, most commonly solar and wind, on electricity prices. More recent work by [12,13,14] analyzes the various consequences for spot and futures markets of the accelerated transition to green electricity generation in the EU area. A more cohesive group of research works, such as [15,16,17], examines the impact of the structure of the European electricity market on prices, examining certain indices of market liberalization, such as, for example, B. the number of retailers and the concentration of market power in the flow. Generation.

The present study is quite innovative in that it offers a more spherical view of the subject of European electricity prices under study, using two separate models. The basic model examines the impact of all RES production, including solar, wind, hydro and biofuels, and certain economic parameters on domestic electricity prices. Additionally, a second model is specified that aims to show the real effects of competition in the generation and final consumer markets on the electricity price level. For this purpose, the econometric analysis of the two models was performed using the Static Panel of Fixed Effects and Dynamic Panel methods of the GMM System and a single panel dataset that has not been used by any other relevant scientific work to date. In particular, to the authors' knowledge, this is the first time that an index of economic globalization has been used to represent the degree of a country's external economic orientation. Furthermore, the research provides energy policy makers with a comprehensive assessment of the role of market liberalization, considering different aspects of the European electricity market. In addition to production concentration and total number of retailers, two additional and quite interesting variables are included. These variables refer to the amount of electricity produced by producers of non-primary activities and the number of suppliers with a market share greater than 5%. In detail, the base model is composed of variables that represent the percentage of participation in renewable energy, the GFCF related only to the electricity sector as proposed by [18], the total level of environmental taxes and the value of the economic globalization index. . Likewise, the competition model includes the percentage market share of the largest producer, the amount of electricity from auxiliary producers, the number of large electricity suppliers and the total number of suppliers.

In addition to examining the relationships and possible interactions between the proposed model variables and domestic electricity prices, this research will further examine whether the impact of involving renewable energy in the mix of production fuels changes after a certain point. Based on the results of the econometric analysis, a series of fundamental policy adjustments are proposed to try to mitigate the recent high levels of electricity prices. Additionally, some key initiatives and incentive measures are recommended for the European authorities to successfully intervene in the process of deregulation of the electricity market and ensure the intended results for the benefit of both electricity sector investors and domestic customers. Finally, the necessary conditions are approached until electricity prices have fully converged with regard to possible political adjustments in the parameters of the two processed models.

The document is structured as follows: Section 1 and Section 2 each consist of an introduction to the research topic and the state of the art. Section 3 presents the panel data set analyzed, the specifications of the two models, the statistical diagnostic tests and the implemented methodology. Section 4 contains detailed comments on the econometric analysis and results. Section 5 provides a comprehensive summary of the most valuable economic and environmental insights and a range of potential policy implications. Finally, Section 6 summarizes the main conclusions and contributions of the document.

2. State of the art

The critical role of electricity in the quality of life, technological advancement and economic prosperity of modern societies has led several researchers to try to identify the main determinants of electricity prices. The effects and potential costs of expanding renewable energy are currently one of the most popular topics in science, as renewable electricity manages to combine security of supply and autonomy of energy resources with CO reduction [sub.2]. Due to the EU's global leadership in the deployment of renewable energy sources, the vast majority of relative studies focus on wholesale and retail price adjustments as a result of the penetration of renewable energy sources in the European electricity market. As the largest national electricity market, the epicenter is mainly in the German market, along with the general EU market. References [19,20,21] support that renewable energy generation was a positive factor in the drop in electricity prices in Germany. Likewise, in Ref. [22] examines the impact of increasing the share of renewable energies in Spain's power system, the main electricity market in the Iberian Peninsula. The survey showed a net reduction in retail electricity prices despite the cost of RES feed-in tariffs. The studies of [23,24] confirmed this result for the wholesale price of electricity in Spain, and [25] argued that electricity from renewable sources not only lowered prices, but also mitigated the likelihood of possible price spikes. . Judge. [26] indicate the beneficial effect of penetration of renewable energy on the Danish electricity market, while [27,28] further confirm this result for the entire EU region. The latter also explained that the decrease in the average price of electricity is more pronounced as the RES installed capacity increases. In contrast, examining country-specific cases, Ref. [29,30,31] will argue that increased adoption of renewable energy sources has increased end-user electricity prices in Denmark, Israel, and Germany, respectively. Judge. [32], based on a dynamic panel analysis of a dataset containing information for the seven largest OECD economies, including France, Germany, Italy and the United Kingdom, found that the share of renewable energy in the electricity generation system does not affect the current price in a statistically significant way. sensitive. However, Ref. [15,33,34], which examined the impact of RES on the electricity markets of almost the entire group of EU Member States, found that the deployment of RES generally has a direct and increasing effect on domestic electricity prices. .

Little work has been done on the impact of equity investments on electricity prices in Europe compared to research effort on renewable energy. Relative studies such as [35,36] show that electricity market reform is associated with a decline in private investment in the electricity sector, while high electricity prices reduce the ability of EU countries to attract foreign investment. However, inverse relationships were not analyzed.

However, due to increased EU efforts to accelerate the integration of European electricity markets and create a single internal market, a large group of academics has carried out research on the impact of market liberalization and the level of competition in electricity generation and distribution. electricity. References [32,34,37] concluded that specific EU policies to deregulate markets and stimulate competition benefited European citizens by lowering consumer prices. In contrast, no statistically significant effect was found between market reform progress and electricity prices at work in [33,38]. Also ref. [39,40] point out that, contrary to what was expected regarding the opening of the electricity market, in several cases there was a tendency to increase prices for the final consumer.

In theory, a competitive electricity market with high supplier diversification and a regulatory framework that separates generation and distribution services is better able to offer lower prices to consumers. Judge. [37] argue that, under conditions of perfect competition, energy supply markets lead to significant price reductions. According to this opinion Refs. [34,41] report that moderate generation concentration leads to more affordable electricity prices. On the contrary, a wide range of research finds strong evidence against all of the above notions. According to [42,43], unbundling in power generation has led to higher average electricity prices for a large number of OECD countries. With an explicit focus on market conditions and the level of competition for energy production within the EU, Refs. [15,33] concluded that, contrary to the theory of perfect competition, decreasing market share of the largest generator can adversely affect domestic electricity prices. In line with the previous result, Ref. [16] also states that the entry of more individual electricity generators into the supply market has not reduced prices for the final consumer.

By giving consumers the possibility to choose between several electricity suppliers, it aims to avoid potential market abuse by one or a small group of large retailers, thus avoiding any price implications under monopoly and oligopoly market conditions. In line with this view, Ref. [44] argues that, regardless of whether it is a public or private retail establishment, acting as a regional monopoly causes significant deadweight losses and increases prices for the consumer. References [19,42] suggest stepping up liberalization efforts, as expanded retail access is likely to drive down electricity prices, but [43] argue that retail competition cannot significantly affect electricity prices.

3. Methodology and Data

3.1. data summary

The paper focuses on the main drivers of domestic electricity prices in Europe. Therefore, the following econometric analysis is based on a panel dataset consisting of multiple annual observations for a group of 26 European countries. In detail, the dataset contains information on the annual average price of electricity (per MWh) [45], the percentage of total renewable energy sources in the total fuel generation mix, including solar, wind, hydropower generation and biofuels [46], the sector's GFCF energy (% GDP) [47], the value of the economic globalization index [48.49], the total amount paid in environmental taxes [50], the market share percentage of the largest producer [51], the value of electricity (GWh) produced by auxiliary producers [52], the number of retailers that cover at least 5% of the total national electricity consumption [53] and the total number of electricity retailers [54] for a time horizon ending in 2003-2019 (the competitive model dataset includes observations for 24 countries, as annual values ​​of the market share percentage of the largest generators for Austria and the Netherlands were not (confidential) available in the Eurostat dataset). The data sample comes from Eurostat, World Bank, OECD Statistical Library, US Energy Information Administration and Quality of Government databases.

Table 1 presents the descriptive statistics of the entire dataset, including the price of electricity and the other 8 explanatory variables. The statistical values ​​in Table 1 provide a quick but very clear picture of the structural conditions of the electricity market in Europe. Even before the recent energy crisis, the price range for domestic electricity suffered from great instability and fluctuations, with the cost of using domestic electricity oscillating between €50.5 and a staggering €213 per MWh. Undoubtedly, the majority of energy generation in Europe comes from conventional fossil fuel power plants. Average and average FER values ​​are around 30%, indicating a heavy dependence on traditional energy commodities such as natural gas, coal and oil. However, the statistical results for the countries studied show a great imbalance in terms of dependence on renewable energies, as in the same sample there are countries that cover 99.47% of their total demand with renewable electricity with countries where the contribution of renewable energies ​is practically irrelevant. Likewise, capital investments in the electricity sector and the environmental taxes levied differ significantly across the countries included in the study. GFCF values ​​from taxes on electricity and the environment range from almost zero to several billion euros. However, average values ​​of 0.85% of GDP and €5.79 billion mean that protecting the environment and improving energy system infrastructure is becoming a priority for most European governments. Statistics from the Economic Globalization Index also confirm the above results, that most European countries are characterized by an external economic orientation and a high level of investment and financial freedom.

However, the statistical results of the electricity market competition model variables show that the objective of a common and liberalized European electricity market is far from being achieved in the near future. In particular, the mean and median values ​​of the largest producer's market share, the contribution of non-core manufacturers and the number of large retailers show that there is a low level of market competition in a large group of countries. The zero additional generation of electricity, the presence of a single large retailer and the concentration of market power by the largest producer reaching 97% are strong indications of a lack of competition. However, the median and average values ​​of the above variables mostly describe oligopolistic market conditions, with a large public company likely playing the main role in electricity generation and only a few large electricity traders. In addition, the low market concentration of the largest producer at 3% together with the maximum values ​​for electricity generation from secondary producers and for the number of main retailers point to a closed group of countries with high liberalization. This group of countries most likely refers to the Nordic market, where market barriers related to electricity generation and transmission networks have been removed, giving consumers easy access to various producers and retailers residing in their own country or in the wider Scandinavian region. Finally, consistent with the asymmetry and kurtosis values, the null hypothesis of the Jarque-Bera normality test is rejected both for the dependent variable and for all independent variables, verifying their non-normal unconditional distributions.

3.2. causality analysis

3.2.1. Pearson correlation test

Table 2 and Table 3 present the Pearson correlation test coefficients for the variables of the two processed models. For the baseline model, there is a low but statistically significant correlation between all explanatory variables and the price of electricity except for renewables, with the highest correlation being the relationship between the price of electricity and the economic globalization index. Likewise, for the competition model, there is a statistically significant low correlation between the price of electricity and three of the four independent variables. Furthermore, the largest producer's market share appears to be negatively related to non-primary producers, the number of large retailers and total retailers, although the strongest statistical association is related to secondary electricity generation and total retailers, with a correlation coefficient of 0.8083.

3.2.2. Causality test according to Dimitrescu-Hurlin (2012)

The specification of suitable econometric models requires a prior investigation of possible causal relationships between the variables of the examined sample. For this reason, a causality analysis was performed using the Dumitrescu and Hurlin (2012) test [55]. Table 4 and Table 5 highlight the statistically significant causal relationships between the variables of the proposed models for two lagged periods. In the first table containing the base model results, the price of electricity seems to affect both renewable energy and environmental taxes in one direction, while economic globalization seems to affect only the price of electricity. Likewise, renewable energy affects the GFCF of electricity or economic globalization. Interestingly, the test shows a mutual relationship between renewable energy and environmental taxes.

With regard to the competition model in the electricity market, there are indications of a bidirectional connection between the concentration of production and the price of electricity and suppliers in general. Likewise, there appears to be a unidirectional causal effect of non-primary activity producers on the electricity price, the largest producer's market share, and the total number of retailers. Finally, the price of electricity alone affects the total number of retailers, and the total number of retailers affects the number of prime retailers.

3.3. model specification

The latest causality analysis results confirmed several interactions between the variables included in the panel data sample. Based on these findings, two proposed model specifications are formed, with the base model examining the impact of renewable energy and the four economic variables. Instead, the competition model will attempt to determine the actual influence of retailers and electricity generation concentration on residential electricity prices.

Modelo básico: (1)Precio de la electricidad=ß[sub.0]+ß[sub.1] R e n e w a b l e s[sub.i, t]+ß[sub.2]E l e c t i c i d a G F C F[sub.i, t] +ß[ sub.3]T a x i o m e d i a d a l [sub.i, t]+G l o b a l i z a c i ó n e c o n ó m i c a [sub.i, t]+e[sub.i, t]

Competition model:(2)Electricity price=ß[sub.0]+ß[sub.1]Market share_largest_generator[sub.i,t]+ß[sub.2]Secondary_Autoproducers[sub.i,t]+ ß [sub . 3 ]Main Retailers[sub.i,t]+ß[sub.4]Total_Retailers[sub.i,t]+e[sub.i,t] where ß[sub.1], ß[sub.2 ], ß [sub.3] and ß[sub.4] denote the coefficients of the independent variables of the regression, which represent the price elasticity of electricity with respect to changes in the independent variables, ceteris paribus. ß[sub.0] and e[sub.i,t] symbolize the constant and error terms of relative regressions.

3.4. Cross-sectional dependency testing

After specifying the model, it is important to carry out a series of statistical tests that will guide researchers in selecting the most appropriate econometric methods to successfully process the two proposed models and obtain reliable and robust results. Testing for cross-correlation within and between country panels is a crucial step, as failing to take this aspect into account when working with panel datasets can lead to ill-founded conclusions and policy implications. The presence of a cross-dependency in the error terms can introduce serious biases in the approximation of the coefficients of the variables and the standard errors, unless an appropriate estimator is used.

To detect possible signs of cross-sectional dependence, the analysis uses Pesaran's (2015) CD test [56] for weak cross-sectional dependence and Pesaran's (2004) CD test [57], which is based on mean and paired correlation of the residuals OLS, derived from separable panel regressions. Table 6 strongly demonstrates the existence of an interdependence between the model variables. The null hypotheses (H[sub.0]) of both tests are rejected with 1% significance, which implies that the examined variables, which concern several EU member states, are influenced by the common energy strategy and for the fiscal and environmental measures implemented. guidelines

After discarding the cross-section independence assumption for the panel time series, the analysis examines the probability that different panel units in the dataset are cross-section dependent. Table 7 shows the relative results of nonparametric Q distribution tests by Friedman (1937) [58] and Frees (1995) [59] and Pesaran (2004). The results of the test statistics in Table 7 indicate a strong cross-sectional dependence between the different country panels, as the null hypothesis (H[sub.0]) of the three tests with a significance level of 1% is rejected for the baseline. The same applies to the Pesaran (2004) and Frees (1995) tests for the competition model. This is a perfectly reasonable conclusion, as all countries included in the study belong to the 'core' EU or have very close political ties and special trade agreements with the EU, indicating a high level of integration. This will, to some extent, propagate the consequences of a potential economic and environmental impact in one country to all other country panels in the dataset.

3.5. Dashboard Unit Root Tests

Standard econometric panel data techniques, commonly used in academia to process similar data sets, assume stationary model variables. Given the importance of this aspect, the present investigation uses two first-generation panel unit root tests and a second-generation unit root test to examine the stationarity of all variables included in the dataset, both in terms of levels how much of first differences. The LLC test is a basic unit root test that can be reasonably reliable for data sets with observations over a wide range of time periods. However, the ADF-Fisher test relaxes the limitations of the LLC test and allows the delay lengths to be different between panels, making it superior. However, a weakness of the ADF-Fisher test is that it relies on Monte Carlo simulations to estimate the corresponding p-values. Therefore, the CIPS unit root test introduced by Pesaran (2007) is additionally included in the analysis as a basis for its results in a non-standard distribution and, at the same time, manages to take into account cross-sectional dependence.

The statistical results for the three unit root tests are presented in Table 8. What can be seen in the table is that in the levels, the null hypothesis (H[sub.0]) for the unit roots in the panels of the sentence data is not rejected in most cases. independent variables of the trend option. On the contrary, applying the first differences to the data set, the three tests confirm the stationarity of the variables at a significance level of 1% for trend and non-trend, which implies that the variables of the models are stationary and integrated of order an I(1).

3.6. Panel-Kointegrationstests

After verifying the absence of unity roots, the two proposed models need to be checked again to verify cointegration. Table 9 shows the statistical results of the panel cointegration tests by Kao (1999) [60] and Pedroni (1999, 2004) [61,62]. In Table 9, the p-values ​​for the test statistics suggest panel cointegration for the baseline models and performance on three of Kao's (1999) five tests. To confirm panel cointegration in both models, the analysis also uses Pedroni's test (1999, 2004) with the adjustment recommended by Levin, Lin and Chu (2002) for the test to account for cross-sectional dependence. The statistical results in Table 10 categorically reject the null hypothesis of no joint cointegration at a significance level of 1%.

3.7. Tests for heteroscedasticity, serial correlation, and omitted variables

Before proceeding with the main econometric analysis and being able to properly orient yourself in choosing the most suitable panel methods and estimators, it is necessary to carry out a series of diagnostic tests. The two proposed models are tested for heteroscedasticity and serial correlation using a set of relative tests generally recognized in the scientific community as the most statistically robust. Table 11 illustrates the test statistics P-values ​​for the Breusch-Pagan (1979) [63], Glejser (1969) [64], Harvey (1976) [65] and White (1980) [66] tests of heteroscedasticity. . . ]. and the Breusch-Godfrey/Wooldridge (2010) serial correlation test [67]. It can be seen in the table that four of the five heteroscedasticity tests confirm the presence of heteroscedasticity in the models, while the Breusch-Godfrey/Wooldridge (2010) test strongly rejects the null hypothesis of lack of serial correlation at a significance level of 1 ,% for both models.

3.8. econometric methodology

Considering the verification of cross-sectional dependence, heteroscedasticity and serial correlation through the results of the statistical tests presented in the previous sections, it is imperative to implement appropriate panel econometric techniques that guarantee the accuracy and robustness of the generated results. Table A1 and Table A2 in Appendix A consist of a series of econometric tests on the static econometric model best suited for the current analysis, showing a clear indication of the use of the fixed effects methodology. Taking into account the results of previous diagnostic tests, to improve the predictability of the standard fixed effects model, the study will apply a Driscoll-Kraay standard error correction as suggested by Driscoll and Kraay (1998) [68] and Hoechle (2007 ) [ 69].

Despite the wealth of merits of the qualified static panel method, it can lead to biased estimates and underestimate or overestimate the real effect of explanatory variables in the case of an underlying long-term relationship. Addressing this eventuality requires a more in-depth dynamic analysis of the two proposed models. Blundell and Bond's (1998) [70] model of the GMM system is used for this. The GMM system approach is very effective in dealing with possible unobserved effects related to dependent and explanatory variables, among other advantages. By developing and incorporating a set of tools based on differentiated and lagged levels of the variables under study and integrating them into the baseline model, the System GMM method manages to compensate for the lack of external tools. However, to ensure the accuracy and robustness of the generated results, strict orthogonality is required. Furthermore, the validity of the GMM model of the system also depends on the lack of a second-order serial correlation and on non-overidentified instrumental variables. These conditions are checked by Arellano and Bond's (1991) AR(1) and AR(2) serial correlation tests [71] and by the Hansen-J and Sargan-J overidentification tests. The dynamic analysis in the present study is mainly based on Windmeijer's (2005) two-stage system GMM model with robust standard error correction (WCSE) and orthogonal deviations [72], an econometric approach that is widely accepted in science because of its precision and endogeneity. , concordant results for autocorrelation and heteroscedasticity.

4. Empirical analysis and results

4.1. static analysis

Table 12 summarizes the final results obtained from static and dynamic panel regression analysis. The P-values ​​for the coefficients of all baseline model variables are statistically significant at 1%, except for economic globalization, which is significant at 5%. Likewise, all independent variables of the competition model are significant at 1%, except the number of top dealers, which is considered statistically insignificant. According to the static modeling results of [15,32,33], renewable energies have a positive estimated coefficient; Therefore, it is implied that increasing the share of electricity from renewable sources increases domestic prices in all 26 EU Member States examined. The fact that the renewable energy coefficient[sup.2] is also statistically significant but negative shows the complexity of the impact of RES on the price of electricity. Based on this result, the initial increasing effect of renewable energy reverts beyond the ownership of 49.27% ​​of RES in the generation scheme, as shown in Figure 1. With regard to the GFCF effect of electricity, it should be noted that the high investment costs in modernizing the electricity network are largely reflected in the price of the final consumer, since a 1% increase in the GFCF of electricity ceteris paribus results in an increase of 8.2735 in the short term % effected. The same applies to environmental taxes, with a 1% increase in relative taxation resulting in 0.6855% higher electricity prices. Furthermore, economic globalization is having a similar impact on electricity prices.

Additionally, the result of the concentration of electricity production again corresponds to that of [15.33], since the increase in the market share of the largest producer seems to be able to lower the price of electricity. On the other hand, contrary to the assumption of [16], the influence of the production of secondary activity producers is positive and statistically significant. Finally, allowing additional electricity retailers to enter the market has been shown to have a slightly dampening effect on home prices.

4.2. dynamic analysis

The econometric results for the dynamic panel analysis show that none of the processed two-stage system GMM models fail the required specification tests. Both models reject the null hypothesis (H[sub.0]) of the AR(1) test without autocorrelation, indicating that the correct specification is indeed a dynamic model. Then, the results of the AR(2) tests confirm that no additional autocorrelation is observed in the models after the inclusion of the lagged dependent variable. Likewise, the null hypotheses (H[sub.0]) of the Sargan and Hansen-J overidentification tests cannot be rejected by any of the models, verifying the validity of the implemented instrumental variables.

Looking at the estimates of the base variables and the competitive model, the statistical significance of the lagged electricity price shows a strong path dependence, with the last value of the variable correlated with the previous ones. In contrast to the results of [32,34] for the overall dynamic effect of RE generation, the coefficient for the renewable energy indicator is negative and statistically significant at the 1% level. Specifically, an additional 1% increase in the contribution of renewable energy sources to the fuel mix for power generation, ceteris paribus, is expected to reduce domestic electricity prices by -0.1957%. This finding is consistent with the work of [28], in which the author states that the increasing use of renewable energy can lead to an increasingly rapid drop in electricity prices. Regarding the base model, it should be noted that the impact of the GFCF of electricity, although still statistically significant, is lower than that estimated in the static model with a 1% increase in expenditure (% of GDP) of the electrical system. update increases the consumer price by 6.7119%. Likewise, a 1% improvement in the Economic Globalization Index affects a price increase of the order of 1.25%, while only 0.1379% of a potential 1% increase in environmental taxes appears to have a longer-term impact. in the final price of the house.

In line with the findings of [15.33] on the role of production concentration, the largest producer's market share percentage coefficient is negative and statistically significant at 10%. However, the magnitude of this effect is significantly smaller than that estimated by the static model. Contrary to what is expected from the perfect competition theory, a 1% increase in market concentration, keeping all other explanatory variables constant, leads to a -0.1612% reduction in domestic electricity prices. Furthermore, a 1% increase in complementary generation is expected to lead to an increase in consumer electricity prices by around 1.95%. However, according to [19,42] the progressive opening of the final consumer market is an essential step towards reducing electricity prices. Both the number of large retailers and the total number of retailers have high statistical significance, but the influence of retailers with more than 5% market power is many times greater. In particular, it is estimated that adding 1% more resellers to the market will reduce electricity prices by -0.0207%, while increasing the number of large resellers will cause a notable drop of -0.7942%. Finally, the statistically significant coefficient for the lagged price of electricity in both studied dynamic models allows approximating the rate of convergence of the variable in relation to changes in the explanatory variables of the two models. Therefore, the electricity price adjustment occurs at a rate of just under 14% (1-0.86) per year for the basic model and 17% (1-0.82) per year for the competitive model. These results essentially mean that it takes almost 7 years for electricity prices to fully converge in terms of changes in RES production and the other three economic variables, while it would take almost 6 years for the level of competition in the electricity market. stabilizing completely affects house prices. .

5. Discussion and political implications

In an effort to examine the actual impact of renewable energy sources on domestic electricity prices, the present study used advanced econometric panel methods that revealed the complexity of the link between the two factors. Empirical analysis showed a quadratic relationship between renewable energy and electricity price, best described by an inverted U-shaped curve. This result clearly shows that the initial phase of positive influence of renewable energies on the electricity price is reversed after the inflection point of 49.27% ​​renewable generation and followed by a gradual increase in price pressure as suggested by [28 ]. However, the overall effect of renewable energy estimated by the dynamic model is negative and significant for the 26 EU countries included in the dataset. Considering that the marginal cost of generating electricity from renewable energy sources tends to be zero, the above results are quite reasonable and provide important answers to the debate and the body of conflicting scientific work on the real impact of the sources. of renewable energy in electricity prices.

Combining the results of the econometric analysis with the knowledge from [18] that excessive use of renewable energy sources can have adverse environmental impacts in terms of CO production [sub.2], results in an optimal range for the percentage share of renewable energy sources in the mix of fuels for power generation. by the national authorities of each country. Within this optimal range of use of renewable energy sources, CO2 reduction could be combined with lower consumer prices, allowing a specific country or region to reach the UN target for the production of clean and affordable energy from from renewable sources (SDG 7). However, the widespread development of the RES market and the transition to a “green” and sustainable energy future require governments and organizations such as the EU to establish strict environmental regulations and provide financial incentives to encourage strong green and sustainable innovations. According to [73,74,75,76] sustainable corporate innovation and environmental performance are based on corporate responsibility and employees with high environmental awareness; Therefore, specific policies and information and education campaigns must be created to improve these essential characteristics of private companies.

As for the role of GFCF for electricity, it is clear that these costs need to be mitigated, at least in part, through the implementation of subsidy policies to avoid possible price increases due to high investment costs related to the modernization of the electrical system. Furthermore, the modest effect that green taxes seem to have on electricity prices in the long run allows the European Commission to encourage its member countries to raise relative taxes and mobilize significant funds essential to promote environmental protection and other green policies. Instead, the potential improvement in the Economic Globalization Index score comes at the expense of higher electricity prices for households. A stronger external economic orientation brings with it a series of advantages, such as greater openness to trade and freedom of financing and investment. Broad interaction with other economies promotes economic growth and attracts foreign investment; However, if electricity prices are higher in neighboring countries, energy suppliers may choose to export electricity through a unified transmission and distribution network. In this scenario, domestic electricity prices would be pulled towards the export price.

The econometric analysis showed that the liberalization of the electricity market is of great importance for the level of domestic electricity prices. In contrast to the conclusion of [34] for the negligible dynamic effect of generation concentration, it was found that the market power of the main power generator is statistically significant and negatively associated with the price of electricity. This result implies that the presence of a large supplier is beneficial to maintain a lower price level. Although it may seem paradoxical, this finding is quite reasonable, given that most European countries have a state-owned power generation company, where profitability is not the main objective and therefore provides the market with low-cost electricity, usually derived from carbon-intensive conventional sources. . production units. In such market conditions, newly introduced power generators are forced to operate with a very limited profit margin, which discourages new potential investments in the electricity sector, especially new investments in new environmentally friendly production lines that require significant amounts of capital. Likewise, despite the theoretical expectation of additional electricity production by generators outside the main activity, the econometric results in Section 4 show that this additional supply of electricity is a positive factor in increasing domestic prices. it is simply based on sophisticated and relatively expensive small FER units with limited capacity. In order to prevent these initial investment costs from being passed on to end consumers, the EU's energy strategy must therefore provide for affordable finance for such investment projects, as well as subsidies for 'green' electricity generation.

Finally, the deregulation of the retail electricity market and the gradual abolition of the entire legal framework, as well as bureaucratic problems related to market entry barriers may lead to lower domestic prices; therefore, it is recommended that the European Commission act accordingly. It has been shown that the ability of retailers to maintain a market share greater than 5% in reducing electricity prices is much greater than that of retailers with limited market power. Therefore, it is reasonable for the EU to invest more in creating the necessary conditions in the retail electricity market that stimulate and trigger a process of mergers and acquisitions, while at the same time attracting new investment plans for the creation of retail companies.

6. conclusions

As electricity is the most valuable energy asset for households, it is of the utmost importance that governments and international organizations such as the EU ensure that their citizens have affordable and unimpeded access to this vital social asset. In line with the UN Sustainable Development Goals and in particular with SDG 7 on the production of clean and affordable energy, this study analyzes how domestic electricity prices are influenced by renewable energies, the parameters specific economic conditions and the liberalization of generation and supply of electricity market. The research is based on a static and dynamic panel analysis of a dataset comprising 26 EU countries for a time horizon from 2003 to 2019.

Given the importance of electricity prices for European consumers, the study also analyses, for the first time, the impact of a set of variables that explain the role of the national economy. and the structure of the electricity market. This set of variables includes the GFCF of the electricity sector, the total amount of environmental taxes paid, the value of the economic globalization index, the energy generation of producers from non-primary activities and the number of retailers with market power greater than 5%. . . Other innovations concern the use of linear and quadratic forms of the renewable energy variables, as well as the approximation of the time required for electricity price convergence to policy changes in relation to the parameters of the studied models.

The econometric results obtained showed the complexity of the relationship between renewable energies and electricity prices. The results showed that a low share of renewable energy sources in the fuel mix for power generation tends to result in higher electricity prices for the consumer, but this effect is gradually reversed with greater shares of renewable energy sources. The increasing use of renewable energy sources is expected to drive down electricity prices at an accelerated rate, as suggested by [28]. However, such a development may mask adverse environmental and market impacts. Therefore, before taking any action to encourage the deployment of renewable energy, national energy policymakers must first assess whether the expected level of reliance on renewable energy will deliver the expected economic and environmental benefits in that particular country. On the contrary, it is recommended that the EU as a whole encourage investment in the improvement and sustainable development of the electrical system, as these high costs may have an impact on the final price for the consumer. The fact that electricity producers and traders are gradually absorbing possible increases in environmental taxes allows EU countries to increase relative taxes and accumulate important funds for green measures. Finally, an improvement in the Economic Globalization Index score implies greater financial and investment freedom and economic growth, although greater market openness allows for the export of electricity, likely leading to an increase in domestic electricity prices.

Likewise, electricity market liberalization is a complicated process in which several aspects must be taken into account. The econometric analysis showed that the decrease in the market share of the largest generator can become an aggravating factor for residential electricity prices. Although this contradicts the expectations of perfect competition theory, this result is consistent with previous work by [15,33] on the European utility market. This negative correlation between market concentration and electricity prices can be explained by the presence of a public primary producer in several European countries, which serves as a regulatory pillar. These public companies are not for profit. Instead, its main objective is to keep consumer prices low. Consequently, to reduce domestic electricity prices, European governments can either rely on an oligopoly built around a leading utility or proceed with full liberalization of the generation market with privatization of public generators and total reliance on the terms of free competition. on the open market. economy. As for complementary producers, the EU must subsidize their activity if it involves the use of renewable energy sources. Additionally, the extension of deregulation policies to the retail market may benefit electricity prices in European homes.

Finally, the low convergence of electricity prices in terms of policy adjustments related to the parameters studied shows that both the European Commission and individual state authorities need to develop long-term strategic energy planning. However, all proposed measures must be modified taking into account the specificities of each country. This limitation of the present article, which provides an overview of the examined parameters for the entire EU area, may become the main topic of future research involving econometric methods of the FMOLS and PDOLS panels.

Author Contributions

G.E.H. and A.S.T. designed and engineered the analysis. Both G.E.H. and A.S.T. wrote the manuscript and contributed to the final version of the manuscript. G.E.H. He oversaw the article, provided critical feedback, and helped shape the research. All authors read and accepted the published version of the manuscript.

Institutional Review Board Statement

Not applicable.

declaration of consent

Not applicable.

Declaration of Data Availability

Dataset derived from public domain resources. The complete dataset is available from the authors upon request.

interest conflicts

The authors declare no conflict of interest.

Disclaimer/Editor's Note: The statements, opinions and data contained in all publications are the sole responsibility of the individual authors and contributors and not of MDPI or the publishers. MDPI and/or the publisher(s) disclaim all liability for any damage to persons or property resulting from any ideas, methods, instructions or products mentioned in the content.

Anhang A

Energias-16-02540-t0A1_Table A1 Table A1 Static Econometric Analysis Tests. Basic model competency model Statistics Statistics P-value statistics test 50.02 0.0000 48.82 0.0000 Note: The pool test from the PLM software package "R" was used for the poolability test. The null hypothesis (H[sub.0]) of the pool test assumes the stability and robustness of the POLS model compared to a fixed effects model. The Breusch-Pagan LM test (1980) [77] from the plm “R” software package was used to examine the presence of a panel effect in the data. The null hypothesis (H[sub.0]) of the Breusch-Pagan (1980) LM test does not assume panel effects, which implies that the POLS model is more effective than a random effects model. The joint significance F-test (u_i = 0) in Table A1 is the one specified in the statistics table of a fixed effects model using the xtreg command of the "STATA" software. To investigate whether it is necessary to implement the time-fixed-effects model in relation to the basic single-effects fixed-effects model, pFtest from the plm package of the "R" software was used. The null hypothesis (H[sub.0]) of the pFtest assumes the presence of significant temporal effects, implying the superiority of the temporal fixed effects model over the baseline fixed effects model. Finally, the Hausman test (1978) [78] of the "STATA" software with the Sigmamore option and annual variables was applied to choose between the notional control of time-invariant heterogeneity and its possible random variance in the error term, as implied by the Fixed Effects or Random Effects methodology.

energies-16-02540-t0A2_Table A2 Table A2 Variance Inflation Factor Test. Basic model Competitive model Variable VIF 1/VIF Variable VIF 1/VIF Renewable energies 1.12 0.8896 Market share 1.48 0.6753 Electricity GFCF 1.08 0.9272 Secondary self-producers 3.00 0.3332 Tax environmental 1.06 0.9400 Large retailers 1.37 0.7276 Total Flows10 Total Retailers 0.91 , the stat vif command of the "STATA" software was used, which calculates the centers of the variance inflation factors (VIF) for the independent variables specified in a linear regression model.

Research Highlights

The impact of renewable energies and external economic guidance on electricity prices is evaluated; The optimal share of renewable energy sources in the fuel mix for power generation is approaching; The role of market liberalization on electricity prices in Europe is examined; It evaluates the importance of the deregulation process in the formulation of energy policies; The time horizon until the convergence of electricity prices is estimated in the event of political changes.

glossary

Abbreviations and Acronyms

RES Renewable Energy Sources EU European Union CE European Commission GDP Gross Domestic Product GFCF Gross Fixed Capital Formation GWh Gigawatt hour MWh Megawatt hour UN UN SDG for sustainable development

references

1. European Commission REPowerEU: Joint European Action for More Affordable, Secure and Sustainable Energy, 8 March 2022 COM 108 Final. 2022,. Available online: https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=COM:2022:108:FIN <date-in-citation content-type="access-date" iso - 8601-date="2023-01-24">(accessed January 24, 2023)</date-in-citation>.

2nd UN General Assembly: Transforming Our World: The 2030 Agenda for Sustainable Development. October 21, 2015. A/RES/70/1.. Available online: https://www.refworld.org/docid/57b6e3e44.html <date-in-citation content-type="access-date" iso - 8601 - date="2023-01-24">(accessed January 24, 2023)</date-in-citation>.

3. European Commission Directorate-General for Communication, Decarbonising our energy system to meet our climate targets, Publications Office of the European Union. 2021,. Available online: https://data.europa.eu/doi/10.2775/582723 <date-in-citation content-type="access-date" iso-8601-date="2023-01-24">(accessed at January 24, 2023)</date-in-citation>.

4. European Parliament Amendments adopted by the European Parliament on 8 October 2020 on the proposal for a Regulation of the European Parliament and of the Council establishing a framework for achieving climate neutrality and amending Regulation (EU) 2018/1999. 2020,. Available online: https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:52020AP0253 <date-in-citation content-type="access-date" iso-8601-date= "2022-12-31">(Accessed December 31, 2022)</Citation Date>.

5. Regulation of the European Parliament (EU) 2019/943 of the European Parliament and of the Council of June 5, 2019 on the internal electricity market. 2019,. Available online: https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:32019R0943 <date-in-citation content-type="access-date" iso-8601-date= "2023-01-24">(accessed January 24, 2023)</date-in-citation>.

6. B. Felsman; A. Vékony, Publications Office of the European Union: Luxembourg, 2021, ISBN: 978-92-76-30267-4.

7. T. Schittekatte; V.Reif; L. Meeus Welcome New Entries to the European Electricity Markets., 2021, 14, 4051. DOI: https://doi.org/10.3390/en14134051.

8. N. Wasilewska; M. Wasilewski; S. Sabolotnyj; D. Osiichuk The Impact of M&A on the Competitive Positioning of European Energy Companies and Market Power Concentration on EU National Energy Markets., 2022, 15, 8798. DOI: https://doi.org/10.3390/en15238798.

9. L. de Almeida; F. Esposito; J. van Zeben When indicators fail in electricity policy: pitfalls of the EU retail energy market barrier index, 2022, 165, p. 112892. DOI: https://doi.org/10.1016/j.enpol.2022.112892.

10.G.G. Álvarez; RS Tol The impact of the Social Electricity Title on the energy poor in Spain., 2021, 103, p. 105554. DOI: https://doi.org/10.1016/j.eneco.2021.105554.

11. J. Priesmann; S. Spiegelburg; R. Madlener; A. Praktiknjo Does renewable electricity hurt the poor? Research on Tax Programs to Reduce Income Inequality and Feed Poverty in German Households, 2022, 93, page 102812. DOI: https://doi.org/10.1016/j.erss.2022.102812.

12. K. Tselika The influence of variable renewable energies on the distribution of hourly electricity prices and their variability: a panel approach., 2022, 113, p. 106194. DOI: https://doi.org/10.1016/j.eneco.2022.106194.

13. E.O. Jastad; IM Trotter; T. F. Bolkesjø Long-term energy prices and renewable energy market values ​​in Norway: a probabilistic approach., 2022, 112, p. 106182. DOI: https://doi.org/10.1016/j.eneco.2022.106182.

14. F. Schoniger; UB Morawetz What goes down must go up: why fluctuating renewables doesn't necessarily increase fluctuations in the spot price of electricity in Europe., 2022, 111, p. 106069. DOI: https://doi.org/10.1016/j.eneco.2022.106069.

3 pm Oosthuizen; R. Inglesi-Lotz; GEORGIA. Thopil The relationship between renewable energy and retail electricity prices: Evidence Panel from OECD Countries, 2021, 238, p. 121790. DOI: https://doi.org/10.1016/j.energy.2021.121790.

16. T.C. lung; K.P. Whistle; K.K. Tsui What can deregulators deregulate? The Strom case, 2019, 56, pp. 1-32. DOI: https://doi.org/10.1007/s11149-019-09386-9.

17. TSA Loi; G. Deregulation of Jindal's Electricity Market in Singapore: Initial Wholesale Price Assessment, 2018, 127, S. 1-10. DOI: https://doi.org/10.1016/j.enpol.2018.11.043.

18. G. E. halcos; COM. Tsirivis Electricity Production and Sustainable Development: The Role of Renewable Energy Sources and Specific Socioeconomic Factors., 2023, 16, 721. DOI: https://doi.org/10.3390/en16020721.

19. J. Cludio; H. Hermann; FC Matthäus; V. Graichen The Merit Order Effect of Wind and Photovoltaic Electricity Generation in Germany 2008–2016: Estimation and Distribution Implications., 2014, 44, pp. 302-313. DOI: https://doi.org/10.1016/j.eneco.2014.04.020.

20. M. Dillig; M. Jung; J. Karl The influence of renewable energies on electricity prices in Germany: an estimate based on historical spot prices in the years 2011–2013, 2016, 57, pp. 7-15. DOI: https://doi.org/10.1016/j.rser.2015.12.003.

21. L. Hirth What caused electricity prices to fall in Europe? A factorial analysis., 2018, 39, pp. 143-156. DOI: https://doi.org/10.5547/01956574.39.1.lhir.

22. G. Saenz de Miera; P. del Rio González; I. Vizcaíno Analysis of the impact of renewable electricity support schemes on energy prices: The case of wind power in Spain., 2008, 36, pp. 3345-3359. DOI: https://doi.org/10.1016/j.enpol.2008.04.022.

23. L. Gelabert; X. Labandeira; P. Linares An ex post analysis of the impact of renewable energy and cogeneration on electricity prices in Spain., 2011, 33, pp. S59-S65. DOI: https://doi.org/10.1016/j.eneco.2011.07.027.

24. HA Gil; C. Gomez-Quiles; J. Riquelme Large-scale integration of wind power and benefits of wholesale electricity trade: Estimation by an ex-post approach., 2012, 41, pp. 849-859. DOI: https://doi.org/10.1016/j.enpol.2011.11.067.

25. C. Angry; D. Ballester Effects of renewables on stylized electricity price facts., 2015, 52,pp. 1596-1609. DOI: https://doi.org/10.1016/j.rser.2015.07.168.

26.W.Chen; Z. Bak-Jensen; B. Hu The relationship between electricity prices and wind power generation in Danish electricity markets. DOI: https://doi.org/10.1109/appeec.2010.5448739.

27. M. Rathmann Do E-RES support systems reduce EU ETS-driven electricity prices?, 2007, 35, pp. 342-349. DOI: https://doi.org/10.1016/j.enpol.2005.11.029.

28. L. Hirth The Market Value of Variable Renewable Energy: The Effect of Solar Wind Energy Variability on Its Relative Price., 2013, 38, pp. 218-236. DOI: https://doi.org/10.1016/j.eneco.2013.02.004.

29.J.Munksgaard; PHYSICAL EDUCATION. Morthorst wind energy in the liberalized Danish energy market: policy measures, price implications and incentives for investors, 2008, 36, pp. 3940-3947. DOI: https://doi.org/10.1016/j.enpol.2008.07.024.

30. I. carpenter; A. Milstein Intermittent renewable energy, optimal mix of capacity and prices in a deregulated electricity market., 2011, 39, pp. 3922-3927. DOI: https://doi.org/10.1016/j.enpol.2010.11.008.

31.F.Erni; D Pietsch; R. Paraschiv The influence of renewable energies on daily electricity prices EEX., 2014, 73, pp. 196-210. DOI: https://doi.org/10.1016/j.enpol.2014.05.004.

32. A. Cross; JS Iimura The impact of renewable energy on domestic electricity prices in liberalized electricity markets: an analysis of cross-country panel data, 2018, 54, S. 96-106. DOI: https://doi.org/10.1016/j.jup.2018.08.003.

33. B. Moreno; AJ Lopez; MOUNTAIN. García-Álvarez Electricity prices in the European Union. The role of renewable energies and regulatory reforms in the electricity market., 2012, 48, pp. 307-313. DOI: https://doi.org/10.1016/j.energy.2012.06.059.

34. P.P. Cerqueira; PENNSYLVANIA. Da Silva Evaluating the determinants of domestic electricity prices in the EU: a panel data approach of the GMM system, 2017, 73, pp. 1131-1137. DOI: https://doi.org/10.1016/j.rser.2017.02.016.

35. E. Erdogdu Investment, security of supply and sustainability after three decades of energy sector reform., 2014, 31, pp. 1-8. DOI: https://doi.org/10.1016/j.rser.2013.11.014.

36. E. Ziesemer; T.H.W. Barteková The impact of electricity prices on foreign direct investment: evidence from the European Union., 2018, 51, pp. 1183-1198. DOI: https://doi.org/10.1080/00036846.2018.1524983.

37. W. Kruseman; G. Lise Long-term price and environmental effects in a liberalized electricity market., 2008, 30, pp. 230-248. DOI: https://doi.org/10.1016/j.eneco.2006.06.005.

38. M. Hyland Restructuring European Electricity Markets: An Analysis of Panel Data., 2016, 38, pp. 33-42. DOI: https://doi.org/10.1016/j.jup.2015.11.004.

39. H. Nagayama Patterns of energy sector reform liberalization and energy prices in developing countries: an empirical analysis using international panel data, 2009, 31, S. 463-472. DOI: https://doi.org/10.1016/j.eneco.2008.12.004.

40. D. Bruneckiene; J. Cibinskiene; A. Streimikiene A Review of the Impacts of Electricity Market Liberalization on Electricity Prices., 2013, 12, S. 40-60.

41. C. The Girl of the Day; H. Patokos; T. Amountzias Pricing decisions and market power in the UK electricity market: a VECM approach., 2017, 108, pp. 467-4 DOI: https://doi.org/10.1016/j.enpol.2017.06.016.

42.T.tsutsui; M. Hattori Economic implications of regulatory reforms in the electricity supply sector: an analysis of panel data for OECD countries., 2004, 32, pp. 823-832. DOI: https://doi.org/10.1016/s0301-4215(03)00004-1.

43. H. Nagayama Impact of regulatory reforms in the electricity industry on electricity prices in developing countries., 2007, 35, p. 3440-3462. DOI: https://doi.org/10.1016/j.enpol.2006.12.018.

44. S.Taymaz; E. Bahçe The Effects of Electricity Market Liberalization in Turkey., 2008, 30, pp. 1603-1624. DOI: https://doi.org/10.1016/j.eneco.2007.03.003.

45. Electricity prices for domestic customers from European Commission, Eurostat, (nrg_pc_204). 2022,. Available online: https://ec.europa.eu/eurostat/databrowser/view/NRG_PC_204/default/table?lang=en <date-in-citation content-type="access-date" iso-8601-date=" 2022-07-26">(accessed July 26, 2022)</date-in-citation>.

46. ​​​​​​Renewable Electricity Generation (GWh) from US Energy Information Administration, International/Electricity. 81.2022,. Available online: https://www.eia.gov/international/data/world/electricity/electricity-generation <date-in-citation content-type="access-date" iso-8601-date="2022-07 - 26">(Accessed 26 July 2022)</date-in-citation>.

47. OECD Gross Fixed Capital Formation/Power Sector, Capital Formation by Activity ISIC rev4. 2020,. Available online: https://stats.oecd.org/Index.aspx?DataSetCode=SNA_TABLE8A <date-in-citation content-type="access-date" iso-8601-date="2022-07-26"> ( accessed July 26, 2022)</date-in-citation>.

48. S.Gygli; F. Halg; N. Potrafke; JE Storm The KOF Globalization Index – Revisited. 543-5 DOI: https://doi.org/10.1007/s11558-019-09344-2.

49. A. Dreher Does globalization affect growth? Evidence for a new globalization index., 2006, 38, pp. 1091-1110. DOI: https://doi.org/10.1080/00036840500392078.

50. European Commission environmental taxes by economic activity, Eurostat ((NACE Rev. 2), (env_ac_taxind2)) 2022c. Available online: https://ec.europa.eu/eurostat/databrowser/view/ENV_AC_TAXIND2/default/table ?lang=en <date-in-citation content-type="access-date" iso-8601-date=" 2022-07-26">(accessed July 26, 2022)</date-in-citation>.

51. European Commission market share of the largest producer in the electricity market, Eurostat, (ten00119), 2022d. Available online: http://data.europa.eu/88u/dataset/vk1YAqU00TW2OBXK8KqR0A type="access-date" iso-8601-date="2022-07-26">(Accessed 26 July 2022)</ /date- in quotes>.

52. European Commission electricity generation capacities by main fuel groups and operators (autogenerators), Eurostat, (nrg_inf_epc, [PRR_AUTO]), 2022e. Available online: https://ec.europa.eu/eurostat/databrowser/product/page /NRG_IND_PEH <date-in-citation content-type="access-date" iso-8601-date="2022-07-26 " >(accessed July 26, 2022)</date-in-citation>.

53. European Commission electricity retailers selling at least 5% of total electricity consumption, Eurostat, (nrg_ind_market, [ERTL_SELL_EC5]), 2022f. Available online: https://ec.europa.eu/eurostat/databrowser/view/NRG_IND_MARKET__custom_5201392/default/table?lang=en <date-in-citation content-type="access-date" iso-8601-date=" 2022-07-26">(accessed July 26, 2022)</date-in-citation>.

54. European Commission Electricity Retailers, Eurostat, (nrg_ind_market, [ERTL]), 2022g. Disponível online: https://ec.europa.eu/eurostat/databrowser/view/NRG_IND_MARKET__custom_5201420/default/table?lang=en <date -in-citation content-type="access-date" iso-8601-date=" 2022-07-26">(acessado em 26 de julho de 2022)</date-in-citation>.

55. E.-I. Dumitrescu; C. Hurlin Tests for Granger Non-Causality in Heterogeneous Panels., 2012, 29, pp. 1450-1460. DOI: https://doi.org/10.1016/j.econmod.2012.02.014.

56. M. H. Pesaran proves weak cross-sectional dependence on large panels., 2014, 34, pp. 1089-1117. DOI: https://doi.org/10.1080/07474938.2014.956623.

57. M. H. Pesaran, Faculty of Economics, University of Cambridge: Cambridge, UK, 2004. Cambridge Working Papers No: 0435 in Economics

58. M. Friedman The use of intervals to avoid implicit acceptance of normality in analysis of variance., 1937, 32, S. 675-701. DOI: https://doi.org/10.1080/01621459.1937.10503522.

59. E.W. Frees Evaluation of Cross-Sectional Correlation in Panel Data., 1995, 69, pp. DOI: https://doi.org/10.1016/0304-4076(94)01658-M.

60. C. Kao Spurious Regression e Residual-Based Tests for Cointegration in Panel Data, 1999, 90, S. 1-44. DOI: https://doi.org/10.1016/S0304-4076(98)00023-2.

61. P. Pedroni's Critical Values ​​for Cointegration Tests in Heterogeneous Panels with Multiple Regressors., 1999, 61, pp. 653-670. DOI: https://doi.org/10.1111/1468-0084.61.s1.14.

62. Panel cointegration P. Pedroni: finite and asymptotic sample properties of clustered time series tests with an application to the PPP hypothesis., 2004, 20, pp. 597-625. DOI: https://doi.org/10.1017/S0266466604203073.

63.S.T. Breusch; ARKANSAS. Pagan A simple test for heteroscedasticity and random coefficient variation., 1979, 47, p. 1287. DOI: https://doi.org/10.2307/1911963.

64. H. Glejser A new proof of heteroscedasticity., 1969, 64, p. 316. DOI: https://doi.org/10.1080/01621459.1969.10500976.

65. A. C. Harvey Estimating Regression Models with Multiplicative Heteroscedasticity., 1976, 44, p. 461. DOI: https://doi.org/10.2307/1913974.

66. H. White A covariance matrix estimator consistent with heteroscedasticity and a direct test of heteroscedasticity., 1980, 48, pp. 817-838. DOI: https://doi.org/10.2307/1912934.

67. J.M. Wooldridge, MIT Press: Cambridge, MA, EE. UU., 2010,

68. J.C. Driscoll; Consistent estimate of the covariance matrix of A.C. Kraay with spatially dependent panel data., 1998, 80, pp. 549-560. DOI: https://doi.org/10.1162/003465398557825.

69. D. Hoechle Robust Standard Errors for Cross-Dependent Panel Regressions, 2007, 7, S. 281-312. DOI: https://doi.org/10.1177/1536867X0700700301.

70. R. Blundell; S. Bond Initial Conditions and Moment Constraints in Dynamic Panel Data Models., 1998, 87, pp. 115-143. DOI: https://doi.org/10.1016/S0304-4076(98)00009-8.

71.M. Arellano; S. Bond Some specification tests for panel data: evidence from Monte Carlo and an application to employment equations., 1991, 58, pp. 277-297. DOI: https://doi.org/10.2307/2297968.

72. F. Windmeijer A finite sample correction for the variance of two-stage efficient linear GMM estimators., 2005, 126, pp. 25-51. DOI: https://doi.org/10.1016/j.jeconom.2004.02.005.

73.S.K. sing; MD Giudice; R. Chierici; D. Graziano Green Innovation and Environmental Performance: The Role of Green Transformation Leadership and Green Human Resource Management., 2019, 150, p. 119762. DOI: https://doi.org/10.1016/j.techfore.2019.119762.

74. F. Khatami; V. Scuotto; N. Krueger; V. Cantino The impact of the entrepreneurial ecosystem model on sustainable innovation at the macro level., 2021, 18, pp. 1419-1451. DOI: https://doi.org/10.1007/s11365-021-00788-w.

75. See Cillus; I AM Petruzzelli; L. Ardito; M. Del Giudice Understanding Sustainable Innovation: A 121st Systematic Literature Review., 2019, 26, pp. 1012-1 DOI: https://doi.org/10.1002/csr.1783.

76. M. Del Giudice; A. Di Vaio; R Hassan; R. Palladino Digitalization and new technologies for sustainable business models at the ship-port interface: a bibliometric analysis., 2021, 49, pp. 410-446. DOI: https://doi.org/10.1080/03088839.2021.1903600.

77. T. S. Breusch; ARKANSAS. Pagan The Lagrange Multiplier Test and its Applications to the Specification of Models in Econometrics., 1980, 47, S. 239-253. DOI: https://doi.org/10.2307/2297111.

78. J. A. Hausman's Specification Tests in Econometrics., 1978, 46, pgs. 1251-1271. DOI: https://doi.org/10.2307/1913827.

figure and tables

Figure 1: Domestic electricity price (Euro €/MWh) vs. Polynomial function of the percentage of renewable energy use (authors' elaboration). [Download the PDF to view the image]

Table 1: Descriptive statistics from the full panel dataset (years: 2003-2019).

Note: *** indicates significance at 1% and ** at 5%.

Table 2: Pearson correlation coefficients.

Note: *** indicates significance at 1% or * at 10%. Renewables [sup.2] is the variable Renewables raised to the second power.

Table 3: Pearson correlation coefficients.

Note: *** Indicates significance at the 1% level. The numbers in parentheses show the p-values ​​corresponding to the test.

Table 4: Dimitrescu-Hurlin's base model (2012) – causality test (lag order: 2).

Note: To estimate Dimitrescu-Hurlin's (2012) causality test, the analysis used the xtgcause command with two "STATA" software delays.

Table 5: Dimitrescu-Hurlin (2012) Competence Model – Causality Test (Lag Order: 2).

Note: To estimate Dimitrescu-Hurlin's (2012) causality test, the analysis used the xtgcause command with two "STATA" software delays.

Table 6: Cross-sectional dependence of panel time series.

Note: *** Significance at 1%. The numbers in parentheses show the p-values ​​corresponding to the test. The null hypothesis (H[sub.0]) of Pesaran's DC test (2004) assumes strict cross-sectional independence. The null hypothesis (H[sub.0]) of Pesaran's DC test (2015) assumes poor cross-sectional independence. The xtcd and xtcd2 commands of the "STATA" software were used for the Pesaran CD (2004) and Pesaran CD (2015) tests. Correlation and Absolute (correlation) are the mean (absolute) value of the off-diagonal elements of the cross-sectional residual correlation matrix. Renewables [sup.2] is the variable Renewables raised to the second power.

Table 7: Cross-sectional dependency between groups.

Note: *** Significance at 1%. The numbers in parentheses show the P values ​​corresponding to the test. For the cross-dependency tests of the Pesaran, Friedman and Frees groups, the commands xtcsd Pesaran abs, Friedman xtcsd, Frees xtcsd post were used after the xtreg POLS regression in the "STATA" software.

Table 8: Unit root tests.

Note: *** means significance at 1% and ** at 5%. The numbers in parentheses show the p-values ​​corresponding to the test. The null hypotheses (H[sub.0]) of the tests assume non-stationary variables. The ADF–Fisher, LLC, and CIPS unit root tests used the STATA software xtunitroot and xtcips commands. The critical values ​​for Pesaran's (2007) CIPS test are -2.07 (10%), -2.15 (5%) and -2.32 (1%) for constant and -2.58 (10%) , -2.67 (5%). and -2.83 (1%) for the trend. Optimal delay selection was performed based on Akaike's information criterion while selecting the Bartlett kernel with the maximum number of delays determined by Newey and West's bandwidth selection algorithm. Renewables [sup.2] is the variable Renewables raised to the second power.

Table 9: Kao-Panel-Cointegration test.

Note: Kao's panel cointegration test (1999) used the STATA software xtcointtest kao command with the kernel option (bartlett). The optimal delay length was automatically selected based on Akaike information criteria. All other bandwidth orders are established according to the rule 4(T/100)[sup.2/9] ∼ 3. However, the null hypothesis of the test (H[sub.0]) does not assume a cointegration in the examined models, the alternative hypothesis (H[sub.a]) assumes that all panels are cointegrated.

Table 10: Pedroni-Panel-Kointegrationstest.

Note: For Pedroni's panel cointegration test (1999, 2004), the "STATA" software command xtcointtest pedroni was used with the options Kernel (Bartlett), Trend and Demean. The optimal delay length was automatically selected based on the Akaike Information Criterion (AIC). All other bandwidth orders are configured according to rule 4(T/100)[sup.2/9] ˜ 3. The null hypothesis of the test (H0) assumes that there is no cointegration in the studied models, while the alternative hypothesis (Ha) assumes that all panels are cointegrated. The null hypothesis of the test (H0) assumes that there is no cointegration in the studied models, while the alternative hypothesis (Ha) assumes that all panels are cointegrated. Decrement option: Stata calculates the line mean across the panels and subtracts that line mean. Levin, Lin and Chu (2002) propose this procedure to mitigate the effects of interdependence.

Table 11: Model of diagnostic tests.

Note: The null hypothesis (H[sub.0]) of the tests by Breusch-Pagan (1979), Glejser (1969), Harvey (1976) and White (1980) does not assume heteroscedasticity in the models. Likewise, the null hypothesis (H[sub.0]) of the Breusch-Godfrey/Wooldridge (2010) test assumes a non-serial correlation (pbgtest{plm} from "R" software).

Table 12: Empirical results under different specifications.

Note: *** represents significance at 1%, ** at 5%, and * at 10%, respectively, with numbers in parentheses indicating corresponding p-values. The fixed effects function of the xtscc command by Hoechle (2007) for the "STATA" software was used for the static analysis. The corresponding statistical functions of the xtbond2 command by Roodman (2009) for the "STATA" software were used for the dynamic analysis and the GMM models of the system. In particular, the two-level system GMM models were estimated applying the relative options of robust estimators (robust) and strict orthogonality (orthogonal). The two-level robust option in the STATA xtbond2 command calls for the Windmeijer finite-sample correction for the two-level covariance matrix. AR(1) and AR(2) are serial first- and second-order autocorrelation tests. Sargan and Hansen-J denote tests for overidentification of instrument limitations in GMM system models. Renewable Energy [sup.2] is the second power of the renewable variable.

links do autor):

Faculty of Economics, University of Thessaly, 38333 Volos, Greece

Note(s) of the author(s):

[*] Correspondence: halkos@econ.uth.gr; Phone: +30-2421074920

DOI: 10.3390/en16062540