Investigating Divergent Energy Policy Fundamentals: Warfare Assessment of Past Dependence on Russian Energy Raw Materials in Europe

Abstract

The Russian invasion of Ukraine and the resulting warfare within direct proximity to the European Union (EU) have significantly changed the conditions of Europe’s energy policy. Consequently, the energy security agencies and governments have been deliberating over an energy trilemma: the problem of balancing affordability, security, and environmental considerations, with much greater ramifications than before the breakout of the current military conflict. The past structural similarities and differences of the EU member states’ dependence on Russian energy raw materials, and the factors that might have determined those differences, may be among the key drivers of future energy policy choices in Europe and the prospective paths of the European Green Deal (EGD). Historical convergence might encourage an EGD consensus, while divergent past patterns might not allow for such a single direction throughout the EU. The multifaceted research proposed in this paper, from descriptive statistics to multiple linear regression model analyses, identifies structural dissimilarities among the EU economies relating to their dependence on Russia. Key findings suggest that the past dependencies on Russian energy raw materials have not been similar across the EU countries, and have been associated with the economic size of the EU member states and their proximity to or remoteness from Russia. In this regard, economic size and distance from Russia may be significant determinants of certain dissimilarities. Some countries, especially Estonia, Lithuania, and Poland, have demonstrated a structurally different energy mix dependence on Russian fossil energy raw materials than countries more remote from Russia that boast a higher GDP per capita.

1. Introduction

Amid the escalation of the Russia–Ukraine warfare taking place within direct proximity to the EU, the conditions affecting Europe’s energy policies have changed significantly. The standard policy preconditions defining the three layers of affordability, security and environment—the so-called “energy trilemma”—have gained new dynamics. The energy security layer, threatened by the instability of supplies and the hostility of Russia, who have been accused of weaponizing energy exports [1], has been struggling with that trilemma more strongly than before the breakout of the current military conflict. Russia, which had supplied one third of European gas consumption, cut its gas flows to Europe just before the start of the current invasion of Ukraine, causing a negative supply shock on the gas markets and unprecedented price increases [2], reminiscent of the first oil crisis in 1973–1974. Nevertheless, the situation does not imply that the two remaining layers of energy considerations should be forfeited. Looking for an optimal trade-off between economic affordability and environmental sustainability is still a valid approach to policy-making, but the reoriented energy security preconditions have been altered. The question arises as to what will be required for countries and governments to agree on how to reconcile all these considerations amidst the current geopolitical outlook. The International Energy Agency (IEA) argued that, rather than a setback of the efforts to tackle climate change, today’s energy crisis should be a historic turning point towards a cleaner and more secure energy system by leveraging an unprecedented response from governments around the world [3]. On the other hand, some voices claim that early evidence reveals a prioritization of short-term, seemingly quicker policy solutions focusing on new fossil fuel supply routes in order to address the immediate energy security concerns in this time of war. They claim that the fossil fuel industry may leverage the current energy crisis for its own benefit and create new lock-ins [4]. Therefore, the prospective paths for energy transition may vary.

In this regard, the initiatives under an umbrella term of the EGD, such as Fit for 55 and REPowerEU, may be the EU’s response to the crisis. They encourage collective action to meet the objectives of the international climate agenda and to rapidly reduce the dependence on Russian fossil fuels. While actions such as the diversification of energy supplies and scaling-up of renewables appear relevant, some, mainly energy efficiency targets, may remain too ambiguous to deliver. Furthermore, the execution timeline and track for particular actions may be prone to local (national) variations. Therefore, the question again arises as to whether the EU member states can find a common path towards a homogenous EGD, and what it will take to achieve that goal. The EU Energy Council that convened on 24 November 2022 in Brussels demonstrates that EGD legislation will take time to negotiate, since even the Fit for 55 package of acts was not proposed to the EU ministers of energy in one attempt [5]. The past structural similarities and differences of the EU member states’ dependence on Russian energy raw materials, as well as the factors that might have determined them in recent years, may be among the key drivers of current negotiations and future energy policy choices in Europe. Past structural convergence might foster reaching an EGD consensus, while critically divergent past conditions might trigger withdrawals from a single EU energy agenda.

Therefore, the aim of this paper is to investigate, economy by economy, the energy mix dependence rates on Russian fossil energy raw materials (the country-level dependence of the EU economies on Russia in terms of coal, oil, and gas imports combined), throughout various layers and dimensions. DR_i,t stands for the dependence rate of an energy mix of a country, “i”, on Russian fossil energy raw materials within a period, “t.” The scope of the analysis will comprise the EU member states and the EU itself throughout the period after its 2004 enlargement. This analysis will relate to four features of the DRs: (a) magnitude; (b) dynamics; (c) consistency; and (d) idiosyncrasy. The following research questions are posed in this paper:

• Research question 1: have the past dependencies on Russian energy raw materials been similar across the EU countries?
• Research question 2: have the past dependencies on Russian energy raw materials been associated with the economic size of the EU member states and their proximity to or remoteness from Russia?

Finding the answers to these research questions is the objective of this paper. In order to achieve that goal, quantitative research on secondary data was conducted using a multifaceted approach, from descriptive statistics to multiple linear regression model analyses (the conditions for using a multiple linear regression model analysis are satisfied). The study claims that investigating the dissimilarities among the EU member states, with respect to their dependence on Russian energy raw materials, and proving (or disproving) a linkage of the potentially identified dissimilarities to the geographical proximity to Russia, offers a contribution to the existing literature by explicitly focusing on the perspectives of Poland, the Baltic States, and the wider Central and Eastern Europe (CEE) region. Understanding this perspective may be helpful, if not essential, to designing the optimal EGD operationalization. Additionally, investigating the potential disparities, with a focus on the economic size of the EU member states, may enhance the discussion of energy policy fundamentals.

Common sense suggests that a transition away from fossil fuels might be a remedy for the destabilized economic and geopolitical foundations of the EU, as well as a panacea for the climate concerns that have evolved within the global agenda throughout the past decades. It has become apparent that the climate and energy security strategies adopted throughout the EU over the past decades have resulted in an increased vulnerability to external shocks from Russia, and requires the rethinking of priorities and vulnerabilities [6]. Such rethinking may, however, be prone to the path dependency phenomenon, as well as the different preconditions across the EU member states. In this regard, some researchers have explicitly underlined that the large European countries do have very different energy situations [7]. It has also been noted that the EU member states such as France, formerly the UK, and especially Germany, have been pushing the climate agenda, while the Eastern member states, particularly Poland, have prioritized security considerations [8]. It implies that the EU energy policy outlook is far from homogenous. Focusing specifically on the dependencies on Russian energy raw materials, rather than the generic energy mix and energy policy circumstances, leads to the formulation of the first research question of whether the past dependencies on Russian energy raw materials were similar across the EU countries throughout the period 2004–2020.

Regardless of whether the past dependencies on Russian energy raw materials were similar across the EU member states, it is clear that an extensive dependency persists. Accordingly, some claim that—as for the dependency on natural gas supplies—time, money, and sustained political effort could reduce Europe’s dependence on Russia, but it would not be an easy task [9,10]. All these prerequisites seem challenging in the context of long-term energy policy formulations—time and money are generally scarce, and political effort consensus is not a given, especially considering that the gas market is where Russia has been seeking to leverage its warfare actions by exposing consumers to higher energy bills and supply shortages (yet the prices of all energy raw materials increased considerably in 2022). In general, the war-driven energy crisis has resulted in inflationary pressures and the risk of recession [3]. However, potential recession scenarios are not homogenous, since the economic prospects resulting from Russia’s aggression vary greatly across countries—from catastrophic economic losses due to the war, to spillover effects from the warfare through commodity, trade, and financial channels [11]. The consequences of the crisis may be heterogeneous, as are the paths for recovery. The prospects of economies in terms of diversifying from Russia may vary and depend on country-by-country specificities. Potential specificities on the causal side lead to the formulation of the second research question of whether the past dependencies on Russian energy raw materials have been associated with the economic size of the EU member states and their proximity to (or remoteness from) Russia.

The varying energy interests across the EU have existed since long before the current warfare and energy crisis. To a certain extent, they have determined the energy mix of the respective EU economies and have been constraining policy coordination and the creation of a fully integrated European energy market. Experts claimed that two major clusters of countries emerge with respect to energy priorities: those countries seeking a higher security of supply and those countries seeking a stronger position in the energy market [12]. Having said that, most of the countries in the CEE have articulated fears about politically rooted gas disruptions from Russia. Given that, and the great dependence of the CEE economies on the energy production from fossil fuels [13], energy policy and the target energy mix should be analyzed with due consideration for geopolitics and their legacy.

2. Materials and Methods

2.1. Literature Review

The literature review indicated that there is no specific consideration for a perspective of the CEE region on the issue of Europe’s dependence on Russian energy raw materials. The occurrence of the “dependence” and “dependency” terms, combined with the terms “Russia” or “Russian”, in the abstracts and keywords of the research papers across leading journals related to energy and energy economics, is limited (28 for Energies, ISSN 1996-1073; 12 for Energy, ISSN 0360-5442; 12 for Energy Economics, ISSN 0140-9883; and 3 for Renewable & Sustainable Energy Reviews, ISSN 1364-0321). The occurrence of the “CEE” (and “Central and Eastern Europe”) term alongside these is non-existent. However, there is a voice recognizing that the EU’s relative disregard for the economic, geopolitical, and climatic concerns of its peripheral Eastern countries towards Russia has existed [6].

2.2. Data

The research relies on energy mix data sourced from Our World in Data (OWD), from where they were retrieved [14,15,16]. Such data have been gathered by OWD from various primary sources and converted into ready-to-analyze datasets, using appropriate inherent conversion methods. Given the overall high quality of OWD projects, such reliance is assumed to be justified. The energy mix was calculated by taking the normalized energy volumes of all the energy sources as a denominator, where the primary energy in exajoules was converted to terawatt hours by a conversion factor of 278, in order to get an approximation of the final energy consumption that takes into account the inefficiencies of the primary energy transformation into final energy. Then, the final energy consumption of the given source (energy raw material, or fuel) is a numerator taken for the energy mix calculation. The research is also based on the Eurostat data, relating to the import rates for coal, oil, and gas [17,18,19], used for calculating the shares of the imports from Russia for the respective energy raw materials.

The research also explores the data relating to: (1) the values of countries’ gross domestic product per capita, measured at the purchasing power parity (sourced from the World Bank database [20]), where these values are money amounts in current international dollars; and (2) the distance of a given country’s capital from Moscow, measured in kilometers (sourced from Open Street Map [21]).

Malta is not considered in the research due to data unavailability, which means that the analysis covers 26 of the 27 EU member states: Austria (AT), Belgium (BE), Bulgaria (BG), Croatia (HR), Cyprus (CY), Czechia (CZ), Denmark (DK), Estonia (EE), Finland (FI), France (FR), Germany (DE), Greece (EL), Hungary (HU), Ireland (IE), Italy (IT), Latvia (LV), Lithuania (LT), Luxembourg (LU), Netherlands (NL), Poland (PL), Portugal (PT), Romania (RO), Slovakia (SK), Slovenia (SI), Spain (ES), and Sweden (SE).

2.3. Variables

The dependence rates of a given country’s energy mix on Russian fossil energy raw materials in a given period (DRs) are the key variables used throughout this paper. They are calculated as follows:

$${\mathrm{DR}}_{\mathrm{i},\mathrm{t}}= \mathrm{DR}\_{\mathrm{C}}_{\mathrm{i},\mathrm{t} \mathrm{weighted}}+ \mathrm{DR}\_{\mathrm{O}}_{\mathrm{i},\mathrm{t} \mathrm{weighted}}+ \mathrm{DR}\_{\mathrm{G}}_{\mathrm{i},\mathrm{t} \mathrm{weighted}}$$

(1)

where DR_C_{i,t weighted} stands for the weighted dependence rate of a country, “i”, on Russian coal in a period, “t”, DR_O_{i,t weighted} stands for the weighted dependence rate of a country, “i”, on Russian oil in a period, “t”, and DR_G_{i,t weighted} stands for the weighted dependence rate of a country, “i”, on Russian gas in a period, “t”, and are calculated from:

$${\mathrm{DR}}_{\mathrm{i},\mathrm{t}} =\frac{\mathrm{IMP}\_\mathrm{RU}\_{\mathrm{C}}_{\mathrm{i},\mathrm{t}}}{\mathrm{IMP}\_\mathrm{TOT}\_{\mathrm{C}}_{\mathrm{i},\mathrm{t}}} \times \frac{\mathrm{PE}\_{\mathrm{C}}_{\mathrm{i},\mathrm{t}}}{\mathrm{PE}\_{\mathrm{TOT}}_{\mathrm{i},\mathrm{t}}}+\frac{\mathrm{IMP}\_\mathrm{RU}\_{\mathrm{O}}_{\mathrm{i},\mathrm{t}}}{\mathrm{IMP}\_\mathrm{TOT}\_{\mathrm{O}}_{\mathrm{i},\mathrm{t}}} \times \frac{\mathrm{PE}\_{\mathrm{O}}_{\mathrm{i},\mathrm{t}}}{\mathrm{PE}\_{\mathrm{TOT}}_{\mathrm{i},\mathrm{t}}}+\frac{\mathrm{IMP}\_\mathrm{RU}\_{\mathrm{G}}_{\mathrm{i},\mathrm{t}}}{\mathrm{IMP}\_\mathrm{TOT}\_{\mathrm{G}}_{\mathrm{i},\mathrm{t}}} \times \frac{\mathrm{PE}\_{\mathrm{G}}_{\mathrm{i},\mathrm{t}}}{\mathrm{PE}\_{\mathrm{TOT}}_{\mathrm{i},\mathrm{t}}}$$

(2)

where IMP_RU_C_i,t, IMP_RU_O_i,t, and IMP_RU_G_i,t stand for the imports of Russian coal, oil, and gas, respectively, in a country, “i”, in a period, “t”; IMP_TOT_C_i,t, IMP_TOT_O_i,t, and IMP_TOT_G_i,t stand for the total (global) imports of coal, oil, and gas, respectively, in a country, “i”, in a period, “t”; PE_C_i,t, PE_O_i,t, and PE_G_i,t stand for the primary energy use in a country, “i”, sourced from coal, oil, and gas, respectively, in a period, “t”; and PE_TOT_i,t stands for the entire primary energy use in a country, “i”, in a period, “t”. The use of a primary energy notion in this analysis is justified, as primary energy is a form of energy accounted for in statistical energy balances, representing the energy volumes prior to any transformation to secondary or final energy. The rationale behind the proposed composition of the DRs is that oil, gas, and coal represent the three largest energy raw materials (commodities) markets globally [22].

The dependence rate trend (DRT or r_i,*), defined as an average yearly rate of dependence of given country, is derived from the following:

$${\mathrm{DR}}_{\mathrm{i},\mathrm{t}}={\mathrm{DR}}_{\mathrm{i},0}{\left(1+{\mathrm{r}}_{\mathrm{i},*}\right)}^{\mathrm{t}}$$

(3)

for i = 1, 2 … m, and t = 1, 2 … n; where m = 26 (number of countries at scope), and n = 17 (annual periods between 2004 and 2020).

2.4. Research Process

The research is twofold. The first step of the research is based on a descriptive statistics analysis relating to four features of the DRs: (a) magnitude, (b) dynamics, (c) consistency, and (d) idiosyncrasy. The former (a–b) are analyzed to investigate the characteristics and distribution of the DR values. In this respect, the simple concepts of the descriptive statistics are claimed to be important and useful methods in today’s era of big data, where effectiveness matters [23]. The magnitude of the DRs is investigated by analyzing the distribution of the DRs. The dynamics of the DRs are investigated by analyzing the sign and magnitude of the DRT of a given country. The latter features (c–d) are investigated with novel indicators, proposed to capture, respectively, the consistency (stability) of the DRs dynamics, and their idiosyncrasy, respective to the wider EU patterns. The consistency of the DRs is investigated by analyzing the occurrence of periods when the DRs had a similar sign to the DRT (an occurrence investigation based on the following conditions: if r_i,* > 0, consider a period, “i”, in the count when DR_i;t—DR_i;t-1 > 0; if r_i,* < 0, consider a period, “i”, in the count when DR_i;t—DR_i;t-1 < 0). The idiosyncrasy of DRs is investigated by analyzing the country-specific deviations of the DRs from its DRT. Such a deviation measure, used in various macroeconomic research [24], is defined as a portion of the whole DR deviation, that for a given country is not explained by the deviations of the EU DR from the EU DRT, and will be called idiosyncratic deviation rates (IDR or x_i,t):

$${\mathrm{x}}_{\mathrm{i},\mathrm{t}}=\left({\mathrm{DR}}_{\mathrm{i},\mathrm{t}}-{\mathrm{r}}_{\mathrm{i},*}\right)-\left({\mathrm{DR}}_{\mathrm{EU},\mathrm{t}}-{\mathrm{r}}_{\mathrm{EU},*}\right)$$

(4)

derived from Equations (5)–(7)

$${\mathrm{DR}}_{\mathrm{i},\mathrm{t}}-{\mathrm{r}}_{\mathrm{i},*}={\mathrm{x}}_{\mathrm{EU},\mathrm{t}}+{\mathrm{x}}_{\mathrm{i},\mathrm{t}}$$

(5)

where x_EU,t is a deviation of the EU DR from the EU DRT

$${\mathrm{x}}_{\mathrm{i},\mathrm{t}}={\mathrm{DR}}_{\mathrm{i},\mathrm{t}}-{\mathrm{r}}_{\mathrm{i},*}-{\mathrm{x}}_{\mathrm{EU},\mathrm{t}}$$

(6)

$${\mathrm{x}}_{\mathrm{EU},\mathrm{t}}={\mathrm{DR}}_{\mathrm{EU},\mathrm{t}}-{\mathrm{r}}_{\mathrm{EU},*}$$

(7)

The second step of the research is based on a multiple linear regression model analysis. The dynamics, consistency, and idiosyncrasy of the DRs are fitted with: (i) the gross domestic product per capita at the purchasing power parity (GDP PC), and (ii) the distance from Russia (DFR) as the independent variables, in order to investigate the relationship between the DRs and the basic economic and spatial independent variables.

It should, however, be recognized that the scope of this paper is limited, and that a more granular analysis on per-raw-material basis could enhance the conclusions. The enhancement potential could be also provided by data relating to non-fossil fuels and energy carriers, such as uranium [25], if included in the analysis. This is, however, out of the scope of this paper, and will remain a scope for further development.

3. Results

3.1. Magnitude

The average value of the DRs in 2020 stood at 24% (likewise its median), with a standard error of 3.9 percentage points. The kurtosis of about −0.8 indicates that the DR distribution has no extreme deviations. The skewness of less than 0.6 indicates that the DR distribution is not substantially skewed. More characteristics of the DR data set for 2020 are presented in

Table 1. Descriptive statistics of DRs in 2020 (26 EU member states).

Measure	Value
Mean	0.24247
Standard error	0.03924
Median	0.24225
Standard deviation	0.20012
Sample variance	0.04005
Kurtosis	−0.82844
Skewness	0.57776
Range	0.63285
Minimum	0.01651
Maximum	0.64935
Sum	6.30416
Count	26

.

Investigating the distribution of the DRs for the analyzed EU countries in 2020 indicates that most of the countries had DRs of less than 27%, with 12 countries having DRs of no more than 14%, and 2 countries having DRs between 14% and 27%. On the other hand, five countries had DRs of more than 40%, with two countries (Estonia and Poland) having DRs of over 62%. Consequently, seven countries demonstrated DRs between 27% and 40%.

3.2. Dynamics

Figure 1 presents the DRs for the analyzed EU countries and the whole EU. The tones of red represent a given country’s upper DRs (historically), and the tones of blue represent a given country’s lower DRs (historically). The DRs of the whole EU have been oscillating between 20% and 24%, with 2012 showing the lowest DR, and 2017 demonstrating the highest DR. The highest DR could be observed in Lithuania (92% in 2010), while the lowest DR (nil dependence) could be observed in several countries in various periods (mainly Ireland between 2004 and 2012).

The least square method (LSM) trend lines of the DR dynamics across the analyzed countries and the EU imply that 15 countries have been trending negative in their combined dependence on Russian fossil energy raw materials, while 11 countries have been trending positive in such dependence (likewise the whole EU). The DRTs imply that 19 countries have been decreasing their combined dependence, contrary to 7 countries, and the whole EU, that have increased that dependence.

Figure 2 presents the DRTs for the analyzed EU countries and the whole EU. The DRTs across the EU countries have been oscillating between −12.1% and 39.2% (0.5% for the whole EU), with half of the countries showing DRTs between −2% and 0%. The highest decrease of the DRs could be observed in Croatia and Austria, while the highest increase of the DRs could be observed in Ireland, Netherlands, Germany and Denmark (Irish DRT should, however, be interpreted with caution due to a base effect, since its nil or quasi-nil DRs throughout the period between 2004 and 2015 increased to the level of 3.1% in 2020, which, despite such dynamics, represents the third-lowest DR across the analyzed countries). Most of the countries have experienced weak DRTs with magnitude of yearly rates of less than 2%. The green bars in Figure 2 represent the countries with a historically decreasing dependence on Russia (negative DRTs), while the red bars represent the countries with a historically increasing dependence on Russia (positive DRTs).

The average value of DRT stood at 0.7%, with a standard error of 1.7 percentage points, while the median stood at the level of −1.0%, implying that positive DRTs were less frequent but of a higher magnitude. The high kurtosis of almost 18.0 indicates that the DRT distribution includes more outliers than the normal distribution would. The skewness of 3.7 indicates that the right tail of the DRT distribution is longer, and the distribution is more concentrated to the left. More characteristics of the DRT distribution are presented in

Table 2. Descriptive statistics of DRTs (26 EU member states and the whole EU).

Measure	Value
Mean	0.00066
Standard error	0.01651
Median	−0.00968
Standard deviation	0.08577
Sample variance	0.00736
Kurtosis	17.98135
Skewness	3.71438
Range	0.51339
Minimum	−0.12146
Maximum	0.39192
Sum	0.01770
Count	27

.

3.3. Consistency

Table 3. Occurrence of periods when country’s DRs had similar sign to its DRT.

EU27	8	50%	FI	11	69%	LU	8	50%
AT	7	44%	FR	10	63%	NL	10	63%
BE	10	63%	DE	10	63%	PL	9	56%
BG	6	38%	EL	8	50%	PT	7	44%
HR	10	63%	HU	10	63%	RO	8	50%
CY	9	56%	IE	5	31%	SK	12	75%
CZ	11	69%	IT	7	44%	SI	10	63%
DK	9	56%	LV	8	50%	ES	10	63%
EE	11	69%	LT	10	63%	SE	8	50%

presents the analysis of the occurrence of periods when a country’s DRs had a similar sign to its DRT. The analysis reflects conditional per country counts, where for positive DRTs, the counted periods are the ones when year-on-year DRs were positive, and for negative DRTs, the counted periods are the ones when year-on-year DRs were negative.

Figure 3 presents the consistency of the DRs versus the DRT per country, where additionally, a direction of the DRT is distinguished (the green bars represent negative DRTs, implying a decreasing dependence, while red bars represent positive DRTs, implying an increasing dependence). In such a view, Slovakia has demonstrated the most consistent decrease dynamics of DRs, while Czechia has been most consistently increasing its DRs. The least consistent dynamics of decreasing DRs can be observed in Bulgaria, while the least consistent dynamics of increasing DRs can be observed in Ireland. Such consistency indicators of the DR dynamics for the whole EU stood at 50%, meaning that its increasing combined dependency on Russian energy raw materials has been caused by half of the periods, while during the other half of the analyzed periods, this dependence was decreasing (2006, 2008, 2012, 2014, 2018–2020).

3.4. Idiosyncrasy

Figure 4 presents the idiosyncratic deviation rates (IDR), defined per country as a portion of the whole DR deviation, that for a given country, is not explained by the deviations of the EU DR from the EU DRT. The tones of red and green represent the country’s IDRs in a way that faded red and green values imply a weak idiosyncratic deviation from the European-wide structural dynamics of the dependence rates. The IDRs of such weak deviation kinds can be observed for most of the EU member countries. However, several countries have demonstrated distinguished idiosyncratic deviations throughout the whole period (bright red and green values)—Estonia, Hungary, Ireland, Lithuania, and Poland, where Irish IDRs should be interpreted with caution, as per the above section. Some countries have demonstrated such idiosyncratic deviations throughout part of the analyzed period—mainly Croatia and Finland.

3.5. Multiple Linear Regression Model Analysis

A multiple linear regression model analysis was performed for the features of the DRs that were analyzed in the previous sections (from Section 3.2 to Section 3.4)—the dynamics, consistency, and idiosyncrasy. The dynamics of the DRs were operationalized by the DRTs, while the idiosyncrasy of the DRs was operationalized by the IDRs. The consistency was analyzed through a notion of the percentage occurrence of the periods, in respect to all the periods, when a country’s DRs had similar sign to its DRT (indicated in Table 3 above). The GDP PC and DFR were tested as independent variables separately for dynamics, consistency, and idiosyncrasy. The results are set out in

Table 4. Multiple linear regression model analysis in respect of DRs dynamics.

Regression Statistics—Dynamics
Multiple R	0.35615
R Square	0.12684
Adjusted R Square	0.04747
Standard Error	0.08383
Observations	25
ANOVA	df	SS	MS	F	Significance F
Regression	2	0.02246	0.01123	1.59796	0.22491
Residual	22	0.15462	0.00703
Total	24	0.17708
	Coeff	SE	t Stat	p-value	Lower 95%	Upper 95%
Intercept	−0.0825	0.0532	−1.5511	0.1352	−0.1928	0.0278
GDP PC	0.0000	0.0000	1.3331	0.1961	−0.0000	0.0000
DFR	0.0000	0.0000	0.8634	0.3972	−0.0000	0.0001

,

Table 5. Multiple linear regression model analysis in respect of DRs consistency.

Regression Statistics—Consistency
Multiple R	0.36050
R Square	0.12996
Adjusted R Square	0.05087
Standard Error	0.10237
Observations	25
ANOVA	df	SS	MS	F	Significance F
Regression	2	0.03444	0.01722	1.64311	0.21624
Residual	22	0.23056	0.01048
Total	24	0.26500
	Coeff	SE	t Stat	p-value	Lower 95%	Upper 95%
Intercept	0.6720	0.0649	10.3475	* 0.0000	0.5373	0.8067
GDP PC	−0.0000	0.0000	−0.4320	0.6699	−0.0000	0.0000
DFR	−0.0000	0.0000	−1.6197	0.1195	−0.0001	0.0000

* Statistically significant value at α = 0.05.

and

Table 6. Multiple linear regression model analysis in respect of DRs idiosyncrasy.

Regression Statistics—Idiosyncrasy
Multiple R	0.81884
R Square	0.67050
Adjusted R Square	0.64054
Standard Error	0.15565
Observations	25
ANOVA	df	SS	MS	F	Significance F
Regression	2	1.08459	0.54229	22.38386	* 0.00000
Residual	22	0.53299	0.02423
Total	24	1.61758
	Coeff	SE	t Stat	p-value	Lower 95%	Upper 95%
Intercept	0.6979	0.0987	7.0678	* 0.0000	0.4931	0.9027
GDP PC	−0.0000	0.0000	−3.5442	* 0.0018	−0.0000	−0.0000
DFR	−0.0002	0.0000	−4.7406	* 0.0001	−0.0003	−0.0001

* Statistically significant value at α = 0.05.

(a least square estimation with significant values at α = 0.05 [26] marked with *).

Neither the dynamics nor consistency of the DRs can be explained with the GDP PC and DFR based on the performed analyses. On the contrary, the modeled relationship between the idiosyncrasy of the DRs and the GDP PC and DFR demonstrates statistical significance and a sound goodness of fit (adjusted R² at 0.64). Therefore, it is right to claim that, while the dependent variables of the dynamics and consistency of the DRs are not related to the GDP PC and DFR, the idiosyncrasy of the DRs is related to these independent variables. The results do not materially change when the outlying Irish data is omitted [27], meaning that the model is stable in this regard.

The Pearson correlation coefficient between the IDRs versus the GDP PC and IDRs versus DFR are −0.58 and −0.69, respectively. This indicates that the higher the GDP PC and DFR, the lower the IDRs.

4. Discussion and Conclusions

The study proposes an assessment of the past dependence of EU economies on Russian energy raw materials (or DRs), in which a self-designed analysis framework is populated with secondary historical data. This assessment covers dependence rates defined as the combined dependence of a given country’s energy mix on Russian coal, oil, and gas, analyzed annually throughout the four dimensions of magnitude, dynamics, consistency, and idiosyncrasy.

The pre-warfare outlook of 2020 suggests that the mean and median value of the DRs across the EU member states stood at 24%, ranging between 2% and 65%. While most of the countries had DRs of less than 27%, and 12 countries had DRs of no more than 14%, two countries, Estonia and Poland, had DRs of over 62%. This implies heterogeneous magnitudes of the DRs in 2020. The longer time horizon of 2004–2020 reveals that the highest DRs were observed in Lithuania (92% in 2010), while the lowest DRs (nil dependence) were observed in several countries throughout various periods. Thus, it also demonstrates heterogeneity even for longer-term magnitudes. Most of the EU economies have been decreasing their combined dependence on Russian fossil energy raw materials since 2004, while the whole EU economy has increased such dependence. Most of the EU countries have experienced weak DRTs, with an average yearly rate (compound rate) of less than 2%, where positive DRTs were less frequent than the negative ones, but of a higher magnitude. This implies divergent dynamics of the heterogeneous DRs throughout the period between 2004 and 2020. Investigating the consistency through a self-designed indicator of the DR dynamics demonstrates that the increasing combined dependency of the EU on Russian energy raw materials occurred during half of the periods, while during the other half of the analyzed periods, this dependence was decreasing (the value of such an indicator for the whole EU at 50%). To a certain extent, it infers a volatility of the divergent dynamics of the DRs. Several countries demonstrated distinguished idiosyncratic deviations of the DRs from the EU-wide patterns throughout the whole period, especially Estonia, Lithuania, and Poland. Some other countries demonstrated such idiosyncratic deviations throughout part of the analyzed period, mainly Croatia and Finland. This analysis demonstrated that the past dependency on Russian energy raw materials was not similar across the EU countries (answer to research question 1). This entails that some countries demonstrate inherently divergent past features that may cause the differences in the energy consensus now and in the future.

Since the modeled relationship between the idiosyncrasy of the DRs and GDP per capita, as well as between the idiosyncrasy of the DRs and the distance from Russia, have demonstrated statistical significance, it is possible to claim that the idiosyncrasy of the DRs is related to those independent variables. In this regard, strongly negative correlation coefficients have been obtained, suggesting that the EU countries neighboring Russia and having a lower economic output per capita than their more developed peers (especially in the cases of Estonia, Lithuania, and Poland) have demonstrated, in general, structurally different dependence rates of energy mix on Russian fossil energy raw materials than the EU countries more remote from Russia that boast a higher GDP per capita. This also suggests that the past dependencies on Russian energy raw materials are, to a certain extent, associated with the economic size of the EU member states and their proximity to, or remoteness from, Russia (answer to research question 2).

Such findings are consistent with the known problem of the so-called East–West divergences [28] and, to a certain extent, reflect observations that the energy-related divergence between Eastern European and Western European economies began as soon as central planning was adopted [29]. However, the findings presented in this paper constitute a value added, in the sense that the perspectives of Poland, the Baltic States, and the wider CEE region are advocated and backed by the latest empirical data. While divergent energy transition interests within the EU have been known and articulated in the past [30], they matter most now, in the context of the EGD decision-making. This perspective was indeed discussed before, especially for Poland and Lithuania [31], but prior to the current warfare and without the contextual concept of the idiosyncratic analysis proposed in this paper.

The paper is limited by the high-level granularity of its analysis, and therefore a more granular analysis on a per-raw-material basis could enhance the conclusions. A dataset could also be extended to cover more energy-related commodities such as uranium (Russia is an important supplier in the nuclear energy sector [31]). Another area for further improvement could be to analyze such enriched data using structural equation modeling.