Coal Consumption Efficiency in the European Union—Trends and Challenges

Abstract

Coal plays a significant role in the economies of many countries and serves as an energy source for numerous societies. However, its combustion causes various environmental problems and contributes to climate change. This article examines the efficiency of coal consumption in 26 European Union countries and its changes from 2014 to 2022. Data Envelopment Analysis (DEA) methodology was applied to measure the extent of overall technical, pure technical, and scale technical efficiency, based on data concerning three production factors (labour, fixed assets, and energy), with GDP as a desirable output and CO₂ emissions as an undesirable output. The empirical findings revealed that Cyprus, Denmark, Luxembourg, and Poland were efficiency leaders throughout the entire study period. France, Germany, Italy, and the Netherlands managed energy and non-energy resources efficiently but were found inefficient in terms of operational scale. Countries that do not use their resources at optimal levels in the production of goods and services should provide greater technical and financial support to their production processes and improve the organisation and structure of labour.

1. Introduction

Fossil fuels remain a crucial energy resource, significantly contributing to the current economic and social development [1]. Among them, coal continues to play a particularly important role in the energy sector, as well as in many high-temperature industrial processes, such as cement and steel production [2]. It is also used to produce chemical substances that serve as material bases in the manufacturing of various products (fertilisers, plastics, pharmaceuticals, etc.). However, coal combustion is one of the main sources of carbon dioxide emissions, contributing to air pollution and climate change [3]. Furthermore, coal mining causes numerous environmental problems, such as contamination of surface and groundwater, destruction of natural habitats, and depletion of currently exploited deposits [4,5,6].

In order to reduce the negative environmental impact, the European Union (EU)’s climate and energy policy is gradually shifting away from coal in favour of more sustainable energy sources, which constitutes one of the greatest challenges of the current energy transition. A central element of the strategies being implemented is also energy efficiency, which may have an immediate and cost-effective, albeit indirect, effect on reducing fossil fuel consumption [7]. Energy efficiency means ‘the ratio of output of performance, services, goods or energy to input of energy’ [8], or in other words, the practice of using less energy to produce the same amount of result [9,10]. According to the above priorities, the EU’s energy targets for 2030 include increasing the share of renewable energy sources in the energy mix to at least 42.5% [11] and reducing energy consumption by 11.7% compared to the 2020 projections so that primary energy consumption does not exceed 992.5 Mtoe [12]. This should contribute to reaching climate neutrality by 2050 [13].

Achieving the goals set by the European Union is a significant challenge for its member states, particularly those heavily reliant on coal, such as Germany and Poland. The necessity to comply with new regulations requires the implementation of innovative technological solutions at the national level, including the restructuring of electricity transmission and distribution systems, the adoption of energy-efficient technologies and/or best available techniques (BATs), and the construction of energy storage facilities, as well as effective coordination and cooperation at the EU level. In light of the evident climate impacts from fossil fuel combustion, the energy transition process is, however, inevitable. Russia’s recent invasion of Ukraine and the associated sharp increase in prices for coal, gas, and other resources have further underscored the need for this transition.

Accepting the role of the coal phase-out process and the improvement of energy efficiency in ensuring both energy and economic security, as well as a high quality of life for present and future generations, it is recommended to monitor and evaluate the progress made to support decision-making processes and adapt national and regional strategies to changing conditions [14]. Accordingly, there are various reports and scientific articles in journals on energy, sustainable development, and climate policy that address the issues of energy efficiency and coal consumption reduction in EU countries. However, they mostly refer to total energy efficiency, where researchers take into account the total consumption of various energy inputs (e.g., coal, oil, and gas) [15,16,17,18,19,20], or present the current state of progress in reducing coal use based on a single-factor indicator [21,22,23]. Until now, there have been no comprehensive comparative studies on coal consumption efficiency in EU economies that focus on the analysis of multidimensional processes by simultaneously considering multiple input and output variables. Moreover, in a situation where many European countries still rely on coal as a primary energy source, determining the extent of its waste and estimating its savings potential may be of interest to policymakers, as it could help identify economies that need significant improvements, i.e., those requiring greater financial and technical investment to optimise their production processes. Consequently, improving the efficiency of existing coal resource consumption may be one way to reduce the problem of pollution [24].

To address the various aspects of energy efficiency, the non-parametric Data Envelopment Analysis (DEA) method is commonly used. It enables the transformation of multiple inputs and outputs into a single aggregated indicator, which facilitates the relatively simple interpretation of the obtained results [25]. Moreover, unlike parametric methods, it does not require any information about the functional relationships between inputs and outputs [26]. Fidanoski et al. [27] provide a comprehensive literature review on the application of DEA for estimating energy efficiency, explaining its usefulness and arguing why researchers should adopt it for this type of empirical analysis.

As for the geographical scope of existing studies on coal consumption efficiency, they focus exclusively on the area of China, due to the heavy reliance of its economy on this energy source. For instance, Cui et al. [28] assessed the efficiency of coal resource utilisation in 29 Chinese provinces in 2012. Long et al. [29] conducted a similar study in 30 Chinese provinces for the period 2000–2012, while Guo et al. [30] examined the years 2003–2014. Subsequently, Guo et al. [24] estimated coal consumption efficiency in six energy-intensive sub-industries in China for the year 2015, and Qi et al. [31] analysed 14 major coal-intensive Chinese industries from 2006 to 2015. Finally, Xue et al. [32] measured coal resource efficiency in 30 Chinese provinces between 2000 and 2015. All of the aforementioned studies evaluated efficiency by jointly considering coal consumption and economic indicators using various DEA models. However, since energy alone cannot produce an economic output [33], it is combined with additional inputs, such as capital and human capital. Furthermore, due to the adverse environmental impact of fossil fuels, environmental variables—such as CO₂ emissions—are also considered, representing undesirable outputs in the DEA model [34].

The purpose of this paper is to analyse the efficiency of coal consumption in the European Union during the period 2014–2022, which is the first attempt of this kind. The study continues to apply the DEA method, as it demonstrates great potential in energy efficiency research [27,29]. It fills a gap in the literature related to the lack of comprehensive analyses on coal consumption efficiency at the European level. Monitoring changes in efficiency over time is important, as it may assist policymakers in formulating better-targeted policies aimed at the rational management of existing energy and non-energy resources, thereby accelerating the achievement of the EU’s energy and climate goals.

The remainder of the article is structured as follows. Section 2 presents the empirical methodology and the data used in the study. Section 3 describes the results obtained and outlines the policy implications. The final section concludes with the limitations of the conducted research and suggestions for future studies.

2. Materials and Methods

The DEA method was first introduced in 1978 by Charnes et al. [35] as an approach to identify efficient units among a set of observations. The objects of analysis in this method, the so-called decision-making units (DMUs), can be any homogeneous entities capable of transforming inputs into outputs. They are considered efficient if they minimise inputs at a given level of outputs (in input-oriented models) or maximise outputs at a given level of inputs (in output-oriented models). Otherwise, they are inefficient and should, therefore, take examples from benchmark units that achieve a better ratio of weighted outputs to inputs [36]. Since the DEA method allows for the assessment of activities through the simultaneous application of multiple variables expressed in different units and also avoids subjective determination of their weights [37,38], it is widely used in efficiency studies at the micro (enterprise), meso (regional), and macro (national) levels [39,40,41].

Among the many available DEA models, the most widely known are the CCR (Charnes, Cooper, and Rhodes) model, which measures Overall Technical Efficiency (OTE), and the BCC (Banker, Charnes, and Cooper) model, which measures pure technical efficiency (PTE). Unlike the CCR model, which assumes linear scaling of inputs and outputs (the so-called constant returns to scale—CRS), the BCC model assumes that changes in inputs do not necessarily lead to proportional changes in outputs (the so-called Variable Returns to Scale—VRS) [42]. The use of the CCR model is most appropriate when DMUs operate in a competitive market, and therefore it can be assumed that they operate at their Most Productive Scale Size (MPSS) [43]. Although the OTE indicator on its own allows for determining the overall inefficiency of a given unit, by encompassing both managerial inefficiency and inefficiency resulting from operating at a non-optimal scale, it does not allow for distinguishing between these two sources of inefficiency. Meanwhile, the PTE indicator on its own does not consider Scale Efficiency and reflects only input management efficiency. Therefore, both indicators should be applied if one seeks to obtain information on both managerial improvement and scale optimisation.

In this study, overall and pure technical efficiency is calculated for 26 EU countries (i.e., Austria, Belgium, Bulgaria, Croatia, Cyprus, the Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Ireland, Italy, Latvia, Lithuania, Luxembourg, the Netherlands, Poland, Portugal, Romania, Slovakia, Slovenia, Spain, and Sweden). Malta is excluded, as it phased out coal use entirely, including the discontinuation of hard coal in 1996 [44]. It should be noted that this is methodologically correct, as the DEA method is sensitive to variables with zero values [45]. Since reducing energy consumption while maintaining the level of produced goods and services should be a priority at the national level, an input-oriented approach was applied. The dual form of the model used to calculate OTE is as follows:

$$\mathit{min}{\theta }_o-\epsilon ({\sum }_{n=1}^N{s}_{on}^{-}+{\sum }_{r=1}^R{s}_{or}^{+}),$$

(1)

$${\sum }_{j=1}^J{x}_{nj}{\lambda }_{oj}+s_{on}^{-}\le x_{no}{\theta }_o,$$

(2)

$${\sum }_{j=1}^J{y}_{rj}{\lambda }_{oj}-s_{or}^{+}\ge y_{ro},$$

(3)

$${\lambda }_{oj},s_{on}^{-},s_{or}^{+}\ge 0,$$

(4)

where

• ${\theta }_{\mathit{o}}$—efficiency indicator for the tested DMU;
• $x_{nj}$—the n-th input for the j-th DMU;
• $x_{no}$—the n-th input for the tested DMU;
• $y_{rj}$—the r-th output for the j-th DMU;
• $y_{ro}$—the r-th output for the tested DMU;
• ${\lambda }_{oj}$—weight coefficients;
• $s_{on}^{-}$—slack variables for inputs;
• $s_{or}^{+}$—slack variables for outputs;
• $\epsilon$—non-Archimedean value;
• j = 1, …, J; n = 1, …, N; r = 1, …, R.

To calculate PTE, an additional restrictive condition was introduced into the above model in the following form:

$${\sum }_{J=1}^J{\lambda }_{oj}=1$$

(5)

In the coal consumption efficiency model, variables representing energy (coal), capital, and human capital are used as inputs, while variables representing desirable and undesirable results are treated as outputs [24]. Therefore, considering previous studies [15,16,19,20,28,29,30,31,32], national coal consumption was introduced as the main input variable in the DEA models, while Gross Domestic Product (GDP) at current prices was chosen as the main output variable (

Table 1. Variables in the DEA model.

Type	Name	Unit
Main input	Coal consumption	TJ
Additional inputs	Employment	Thousand persons
	Fixed assets (net)	Million Euro
Output	GDP	Million Euro
Undesirable output	CO2 emission	Mt

). Adopting the approaches of Guo et al. [24], the total number of employed persons by main industry was used as the indicator representing human capital (the first additional input), whereas net total fixed assets at the industry level, valued at current replacement costs, were chosen as the indicator of capital input (the second additional input). Furthermore, in this study, as in other studies [17,20,30,31], CO₂ emissions from coal combustion were considered an undesirable output. It is important to note that CO₂ emissions have a different preference direction in the constructed DEA-CCR and DEA-BCC models, meaning that they are not an output that should be maximised, and countries with the lowest values in this regard are more preferred. However, in the case of an opposite preference direction for a given variable Z in the classical DEA models, it can be replaced with its inverse (1/Z) or the difference from a certain constant, i.e., a − Z, where a − Z > 0 [46,47]. Another possibility is to reclassify the undesirable output as an additional input variable, as it represents a form of environmental cost that should be minimised. Since treating undesirable outputs as inputs does not reflect real processes [47], in this study the transformation based on a constant a = 318 was applied. Moreover, this is a linear transformation that does not cause any loss of information [46]. It should also be noted that the selected number of variables provided adequate discriminatory power for the DEA method, as the condition was met that the number of DMUs is at least three times greater than the sum of input and output variables [48].

Since researchers have already assessed energy and/or coal consumption efficiency using the variables and DEA models adopted in this study at the national [25], regional [29] and industry [24] levels, the research method has therefore been validated.

Data for the analysis were obtained from the International Energy Agency (IEA) [49] and the European Statistical Office (EUROSTAT) [50]. Considering the delay in publishing data on net fixed assets, the analysis period covered the years 2014–2022. The lack of data prevented the inclusion in the analysis of additional emissions resulting from coal combustion, such as nitrogen oxides (NO_x), sulphur oxides (SO₂), or other harmful substances. The values of the variables used in the study are presented in Appendix A.

3. Results

The estimation results for overall technical efficiency (OTE) and pure technical efficiency (PTE), presented in

Table 2. Estimation of coal consumption efficiency in 2014–2022.

Country	2014		2015		2016		2017		2018		2019		2020		2021		2022
	OTE	PTE	OTE	PTE	OTE	PTE	OTE	PTE	OTE	PTE	OTE	PTE	OTE	PTE	OTE	PTE	OTE	PTE
Austria	0.64	0.85	0.63	0.84	0.64	0.85	0.62	0.82	0.63	0.82	0.64	0.79	0.60	0.75	0.57	0.67	0.61	0.62
Belgium	0.69	0.99	0.70	0.97	0.70	0.99	0.71	0.99	0.71	0.97	0.73	0.94	0.70	0.91	0.68	0.83	0.74	0.76
Bulgaria	0.51	0.52	0.51	0.52	0.54	0.55	0.56	0.56	0.57	0.58	0.60	0.60	0.61	0.61	0.61	0.61	0.64	0.64
Croatia	0.49	0.49	0.50	0.50	0.52	0.52	0.50	0.51	0.50	0.50	0.51	0.51	0.47	0.47	0.48	0.48	0.55	0.55
Cyprus	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1
Czechia	0.60	0.69	0.61	0.70	0.61	0.71	0.62	0.70	0.61	0.69	0.62	0.70	0.58	0.67	0.54	0.59	0.57	0.57
Denmark	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1
Estonia	0.91	0.92	0.88	0.89	0.88	0.89	0.86	0.87	0.83	0.84	0.81	0.81	0.85	0.85	0.83	0.83	0.81	0.81
Finland	0.64	0.82	0.65	0.86	0.65	0.87	0.66	0.84	0.66	0.81	0.66	0.80	0.66	0.79	0.62	0.70	0.66	0.66
France	0.73	1	0.73	1	0.70	1	0.75	1	0.81	1	0.87	1	0.85	1	0.60	1	0.65	1
Germany	0.66	1	0.67	1	0.68	1	0.69	1	0.68	1	0.68	1	0.66	1	0.63	1	0.65	1
Greece	0.68	0.84	0.69	0.86	0.72	0.87	0.73	0.89	0.76	0.91	0.82	0.97	0.77	0.93	0.80	0.91	1	1
Hungary	0.55	0.66	0.56	0.67	0.57	0.67	0.58	0.67	0.58	0.68	0.59	0.68	0.56	0.66	0.53	0.60	0.56	0.56
Ireland	0.81	1	0.99	1	0.96	1	1	1	1	1	1	1	1	1	1	1	1	1
Italy	0.75	1	0.81	1	0.83	1	0.91	1	0.84	1	0.85	1	0.84	1	0.79	1	0.80	1
Latvia	0.86	0.86	0.86	0.86	0.86	0.86	0.88	0.88	0.87	0.87	0.85	0.86	0.86	0.86	1	1	1	1
Lithuania	0.82	0.82	0.81	0.82	0.80	0.81	0.83	0.84	0.83	0.85	0.83	0.86	0.83	0.84	0.77	0.78	0.83	0.87
Luxembourg	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1
The Netherlands	0.89	1	0.86	1	0.85	1	0.81	1	0.81	1	0.84	1	0.87	1	0.91	1	0.91	1
Poland	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1
Portugal	0.96	1	0.96	1	1	1	1	1	0.92	1	1	1	1	1	1	1	1	1
Romania	0.61	0.74	0.62	0.76	0.60	0.73	0.63	0.76	0.64	0.77	0.66	0.78	0.65	0.80	0.62	0.71	0.69	0.69
Slovakia	0.52	0.54	0.53	0.56	0.53	0.56	0.53	0.55	0.53	0.55	0.54	0.55	0.54	0.56	0.52	0.53	0.52	0.52
Slovenia	0.67	0.67	0.68	0.68	0.70	0.71	0.73	0.74	0.75	0.75	0.78	0.78	0.78	0.78	0.85	0.85	0.77	0.77
Spain	0.77	1	0.79	1	0.80	1	0.72	0.88	0.72	0.90	0.77	0.92	0.70	1	0.63	1	0.67	1
Sweden	0.69	1	0.71	1	0.70	1	0.69	1	0.69	0.99	0.71	0.97	0.70	0.93	0.68	0.85	0.71	0.73
Average UE-26	0.75	0.86	0.76	0.87	0.76	0.87	0.77	0.87	0.77	0.86	0.78	0.87	0.77	0.86	0.76	0.84	0.78	0.84

, indicate that throughout the analysed period of 2014–2022, only Cyprus, Denmark, Luxembourg, and Poland demonstrated full technical efficiency (OTE = 1 and PTE = 1). In 2017, Ireland joined the group of top performers, followed by Latvia in 2021 and Greece in 2022. For most of the studied period, Portugal was also among the efficiency leaders; however, in 2014, 2015, and 2018, it achieved only 100% pure technical efficiency. Countries such as France, Germany, Italy, the Netherlands, and Spain (since 2020) were efficient in resource management for GDP creation (PTE = 1), but they were inefficient in terms of operational scale (OTE < 1). The remaining countries were both inefficient in managing available resources and operated at a non-optimal scale. This group included all Central and Eastern European countries, except Poland and Latvia, namely Bulgaria, Croatia, the Czech Republic, Estonia, Hungary, Lithuania, Romania, Slovakia, and Slovenia, as well as only four countries from the so-called ‘old EU’, namely Austria, Belgium, Finland, and Sweden (from 2018 onwards). Among the inefficient countries, Croatia and Slovakia in particular recorded the weakest performance during the study period (with OTE values below 0.55 and similarly low PTE values below 0.56). Slightly higher efficiency scores were observed in Bulgaria, Hungary, and the Czech Republic. In general, the more economically developed countries of the ‘old EU’ had greater capacity and efficiency in coal management than the Central and Eastern European countries.

The average OTE values for the EU-26 obtained in subsequent years indicate a slight upward trend until 2019 (an increase from 0.75 to 0.78), followed by a stagnation. Meanwhile, the average PTE value for the EU-26 remained at 0.86–0.87 until 2020, after which it decreased to 0.84 in 2021 and 2022.

Figure 1 presents the classification of EU countries into three groups based on the arithmetic average values of the OTE and PTE indicators for the years 2014–2022, i.e., countries with high or full efficiency, having efficiency indicators of 0.91–1 (Group A); countries with medium efficiency, having efficiency indicators of 0.61–0.9 (Group B); and countries with low efficiency, having efficiency indicators of 0.31–0.6 (Group C). This concept has the advantage of providing fixed intervals for both the OTE and PTE indicators, allowing for consistent classification and observation of individual countries’ positions relative to the same efficiency thresholds. Furthermore, it is in line with approaches used in previous studies [51,52,53]. Only 23% of the EU member states achieved the highest efficiency level according to the average OTE indicator and were classified into Group A. Since PTE is always greater than or equal to OTE, 54% of countries had the highest average PTE value and therefore were included in Group A. Thus, 31% of the countries managed their resources efficiently; however, due to operating at a non-optimal scale, they were unable to fully exploit their production potential. None of the analysed countries fell into Group D, defined as the group with very low efficiency (average OTE or PTE ≤ 0.30). The lowest arithmetic average for both overall and pure technical efficiency was observed in Croatia (average OTE and PTE equal to 0.50). Thus, this country wasted the highest share of energy and non-energy resources in the production of goods and services. In other words, it used on average 0.50 too much of its resources to achieve GDP.

By dividing OTE by PTE, Scale Efficiency (SE) indicators were calculated for all analysed EU countries for the years 2014–2022. Since SE does not indicate whether a given country operates under increasing or decreasing returns to scale, the nature of returns to scale was determined based on the sums of the weight coefficients (λ_oj) obtained in the CCR model. The results of these calculations are presented in

Table 3. Scale Efficiency (SE) indicators and type of returns to scale (RTS) in 2014–2022.

Country	2014		2015		2016		2017		2018		2019		2020		2021		2022
	SE	RTS	SE	RTS	SE	RTS	SE	RTS	SE	RTS	SE	RTS	SE	RTS	SE	RTS	SE	RTS
Austria	0.76	DRS	0.76	DRS	0.74	DRS	0.76	DRS	0.77	DRS	0.81	DRS	0.81	DRS	0.85	DRS	0.98	DRS
Belgium	0.70	DRS	0.71	DRS	0.71	DRS	0.71	DRS	0.74	DRS	0.77	DRS	0.77	DRS	0.82	DRS	0.97	DRS
Bulgaria	0.98	IRS	0.98	IRS	0.99	IRS	0.99	IRS	0.99	IRS	0.99	IRS	0.99	IRS	0.99	IRS	≈1	IRS
Croatia	≈1	IRS	≈1	IRS	≈1	DRS	≈1	IRS	≈1	IRS	≈1	IRS	≈1	IRS	≈1	IRS	≈1	IRS
Cyprus	1	CRS	1	CRS	1	CRS	1	CRS	1	CRS	1	CRS	1	CRS	1	CRS	1	CRS
Czechia	0.87	DRS	0.87	DRS	0.87	DRS	0.88	DRS	0.88	DRS	0.88	DRS	0.86	DRS	0.92	DRS	≈1	IRS
Denmark	1	CRS	1	CRS	1	CRS	1	CRS	1	CRS	1	CRS	1	CRS	1	CRS	1	CRS
Estonia	0.98	IRS	0.99	IRS	0.99	IRS	0.99	IRS	0.99	IRS	≈1	IRS	≈1	IRS	≈1	IRS	≈1	IRS
Finland	0.78	DRS	0.75	DRS	0.75	DRS	0.78	DRS	0.81	DRS	0.83	DRS	0.83	DRS	0.89	DRS	≈1	IRS
France	0.73	DRS	0.73	DRS	0.70	DRS	0.75	DRS	0.81	DRS	0.87	DRS	0.85	DRS	0.60	DRS	0.65	DRS
Germany	0.66	DRS	0.67	DRS	0.68	DRS	0.69	DRS	0.68	DRS	0.68	DRS	0.66	DRS	0.63	DRS	0.65	DRS
Greece	0.81	DRS	0.80	DRS	0.83	DRS	0.82	DRS	0.83	DRS	0.84	DRS	0.83	DRS	0.88	DRS	1	CRS
Hungary	0.83	DRS	0.83	DRS	0.85	DRS	0.86	DRS	0.86	DRS	0.86	DRS	0.85	DRS	0.88	DRS	≈1	DRS
Ireland	0.81	DRS	0.99	DRS	0.96	DRS	1	CRS	1	CRS	1	CRS	1	CRS	1	CRS	1	CRS
Italy	0.75	DRS	0.81	DRS	0.83	DRS	0.91	DRS	0.84	DRS	0.85	DRS	0.84	DRS	0.79	DRS	0.80	DRS
Latvia	1	DRS	1	DRS	1	DRS	1	DRS	0.99	DRS	0.99	DRS	1	DRS	1	CRS	1	CRS
Lithuania	0.99	DRS	0.99	DRS	0.99	DRS	0.99	DRS	0.98	DRS	0.97	DRS	0.98	DRS	0.98	DRS	0.95	DRS
Luxembourg	1	CRS	1	CRS	1	CRS	1	CRS	1	CRS	1	CRS	1	CRS	1	CRS	1	CRS
The Netherlands	0.89	DRS	0.86	DRS	0.85	DRS	0.81	DRS	0.81	DRS	0.84	DRS	0.87	DRS	0.91	DRS	0.91	DRS
Poland	1	CRS	1	CRS	1	CRS	1	CRS	1	CRS	1	CRS	1	CRS	1	CRS	1	CRS
Portugal	0.96	DRS	0.96	DRS	1	CRS	1	CRS	0.92	DRS	1	CRS	1	CRS	1	CRS	1	CRS
Romania	0.82	DRS	0.81	DRS	0.82	DRS	0.83	DRS	0.83	DRS	0.84	DRS	0.82	DRS	0.87	DRS	≈1	IRS
Slovakia	0.95	DRS	0.95	DRS	0.95	DRS	0.96	DRS	0.96	DRS	0.97	DRS	0.96	DRS	0.99	DRS	≈1	IRS
Slovenia	≈1	IRS	≈1	IRS	≈1	IRS	≈1	IRS	≈1	IRS	≈1	IRS	≈1	IRS	≈1	IRS	≈1	IRS
Spain	0.77	DRS	0.79	DRS	0.80	DRS	0.82	DRS	0.80	DRS	0.83	DRS	0.70	DRS	0.63	DRS	0.67	DRS
Sweden	0.69	DRS	0.71	DRS	0.70	DRS	0.69	DRS	0.69	DRS	0.73	DRS	0.75	DRS	0.80	DRS	0.97	DRS

.

Throughout the entire study period, Cyprus, Denmark, Luxembourg, and Poland had Scale Efficiency (SE) indicators equal to 1 and operated under constant returns to scale (CRS). In their economies, an n-fold increase in energy and/or non-energy resources resulted in an n-fold increase in GDP. In Bulgaria, Croatia, Estonia, Lithuania, Slovakia, and Slovenia, the SE indicator exceeded 0.9 or was even close to 1, indicating that the scale of activity in these countries was near or almost at the optimal level. Thus, their inefficiency was mainly due to improper resource management practices and incorrect combinations of resources in generating GDP. Additionally, Bulgaria, Croatia (except in 2016), Estonia, and Slovenia operated under increasing returns to scale (IRS), meaning that their GDP grew at a faster rate than the value of resources used, including coal consumption. On the other hand, as previously mentioned, France, Germany, Italy, the Netherlands, and occasionally Spain and Sweden managed resources efficiently but operated under a non-optimal scale throughout the study period. Moreover, they experienced decreasing returns to scale (DRS), which is unfavourable from a sustainable development perspective, as it implies a faster increase in resource consumption compared to GDP growth, as well as an excessive rise in CO₂ emissions. In addition, the countries that operated under decreasing returns to scale (DRS) from 2014 to 2022 include Austria, Belgium, Hungary, and Lithuania, as well as the Czech Republic, Finland, Romania, and Slovakia, which, however, exhibited increasing returns to scale in 2022.

Inefficient EU countries should reduce their use of both energy and non-energy resources to achieve OTE and PTE indicators of 1. Figure 2 presents the excessive use of resources by EU countries in the years 2014–2022, which contributed to their inefficiency in terms of OTE, while Figure 3 shows the excessive resource use associated with inefficiency in terms of PTE. Countries are denoted by symbols according to the international ISO 3166 standard.

As shown in the figures, the potential reductions in the consumption of energy and non-energy resources from 2014 to 2022 vary depending on the country and its inefficiency. Generally, during the analysis period, most countries exceeded optimal employment levels to the greatest extent, indicating significant potential to improve the structure and organisation of labour resources. Efforts to improve labour utilisation in GDP generation should be a priority, especially in Bulgaria, Croatia, Hungary, Romania, and Slovenia. Until 2021, the overall technical inefficiency of countries such as Belgium, Finland, and Sweden can be largely attributed to excessive use of coal resources. Ireland’s overall technical inefficiency in the years 2014–2016 can also be linked to overconsumption of this energy resource. In the case of pure technical inefficiency, coal consumption was the main contributing factor in Belgium. It should be noted that in 2022, the share of countries for which suboptimal coal use during GDP generation was the main cause of pure technical inefficiency increased to as much as 46% (compared to 21% in 2021).

4. Discussion

Research conducted in 26 EU countries using the DEA method indicates a varied level of coal consumption efficiency. This may result from differences in economic structure, the level of workforce specialisation, and access to technical knowledge, as well as disparities in the quality and level of fixed assets in the industrial sector, including different technologies used in the energy and mining industries. Local climatic and energy conditions may also play a significant role, as they influence energy demand and the extent of coal use as an energy resource [54,55].

The consistent presence of countries such as Cyprus, Denmark, Luxembourg, and Poland among those fully efficient, in terms of both OTE and PTE, indicates exceptionally effective strategies for managing available energy and non-energy resources while maintaining an optimal scale of operation. The inclusion of Portugal, Ireland, Latvia, and Greece in subsequent years suggests a gradual improvement in efficiency in some countries, possibly resulting from structural reforms, investments in energy efficiency, and/or changes in resource management policy.

High efficiency may, among other things, stem from the low level of domestic coal consumption, which is observed in Cyprus, Portugal, Latvia, Luxembourg, Denmark, and Greece. The average coal consumption for the period 2014–2022 in these countries was as follows: Cyprus—250.56 TJ, Portugal—355.44 TJ, Latvia—1389.33 TJ, Luxembourg—1886.11 TJ, Denmark—5136.56 TJ, and Greece—8153.89 TJ, compared to the European average of 47,098.32 TJ. Low coal use can be attributed to various factors specific to each country, such as the implementation of advanced energy-saving technologies, effective policy frameworks promoting energy efficiency, a high share of renewable energy sources (RES) and other sources, or the dominance of services in the economic structure.

According to the conducted research, countries with medium or even very high levels of coal consumption, such as Ireland and Poland, also achieved efficiency. In the examined period, average coal consumption amounted to 19,540.11 TJ in Ireland and 436,601.20 TJ in Poland. Interestingly, the Polish economy is highly dependent on coal, and the country has the highest coal consumption in the European Union. However, full technical efficiency in the DEA model does not imply that Poland’s energy system is optimal and sustainable, but rather that there is no waste of resources in relation to the GDP generated, and countries considered inefficient exhibit a less favourable input-to-output ratio. Moreover, despite its high-emission model, Poland has a long-standing tradition and good skills in coal management, and the large scale and centralisation of its energy sector may positively influence technical efficiency. Nevertheless, this country still requires the implementation of further strategies and programmes to support energy efficiency and the transition away from coal-based energy, although many have already been introduced. As a result, domestic coal consumption is on a downward trend; namely, in 2021 it was about 22% lower compared to 2014, and in 2022 there was already a decline of over 35% [49]. Furthermore, although coal remains significant for electricity generation, the solar photovoltaic market is expanding here at the fastest rate in the EU [56]. In Ireland, the ratio of coal and other resource consumption to the country’s increasingly better economic performance also enabled it to attain fully efficient status according to the DEA method. The high share of the services sector and advanced technologies in the structure of the Irish economy, which generate substantial added value, may have significantly improved technical efficiency.

It is worth noting that among the leaders demonstrating full technical efficiency in the DEA model were all EU-26 countries with a zero share of coal in electricity generation, namely Cyprus, Latvia, and Luxembourg. Meanwhile, in Denmark, Greece, Ireland and Portugal, the share of coal steadily declined between 2014 and 2022, reaching, respectively, 12.6%, 10.9%, 7.7% and 0% in 2022 [57,58,59,60]. Therefore, energy demand in these countries is not associated with an increase in coal consumption, which is favourable from the perspective of energy transition.

The conducted research suggests possible improvements for all inefficient countries, which could have achieved their economic growth with lower consumption of energy and non-energy resources. For example, Germany, France, Italy, and the Netherlands, although characterised by exceptionally high capabilities and efficiency in managing energy (coal) and non-energy resources, should adjust their scale of operations, as increasing resource consumption in their economies does not bring benefits, i.e., it does not translate into a corresponding increase in GDP. These countries are among the most economically and technologically advanced EU member states, having invested for many years in technological development, human capital, and management systems, as well as consistently accumulating knowledge and experience. This undoubtedly translates into a high quality of production and organisational infrastructure, a high level of professional specialisation, and effective performance control mechanisms, which in turn contribute to a high level of efficiency in managing available resources. Between 2014 and 2022, all of the countries under discussion experienced growth in the value of fixed assets and employment levels, while GDP growth was very low or even negative, except for 2022, when a more noticeable economic recovery occurred [61]. At the same time, Germany, France, and Italy rank high among the EU-26 in terms of coal consumption, occupying the second, fourth, and fifth positions, respectively, according to the arithmetic average for 2014–2022, while the Netherlands ranked twelfth. The development and energy policies of these countries should therefore focus on better matching resources to the actual needs of the economy, rather than merely increasing the amount of resources used. This means, in practice, the need to move away from a strategy based solely on expanding production resources and employment in favour of actions aimed at improving their efficiency and utilisation, as well as ensuring better allocation to higher-productivity sectors. It is also recommended to review existing support instruments for material- and energy-intensive sectors and provide greater support for investments in clean technologies and resource-efficient solutions. Similar challenges should also concern Spain, which shows decreasing returns to scale, although it has been efficiently managing energy and non-energy resources since 2020.

Recommendations concerning the need to broaden the impact of energy and development policy tools can also be formulated for those countries where, according to the conducted research, coal was the main source of inefficiency. In the case of Belgium, for example, excessive coal consumption prevented the achievement of both overall and pure efficiency. The Belgian economy remains dependent on coal, which indicates the need for more assertive policies aimed at reducing the use of this resource and accelerating emissions reductions, especially in the industrial sector [62]. Sweden and Finland should also pay attention to optimising coal usage if they are to improve their overall and/or pure efficiency. Although in Sweden most electricity supply comes from hydropower, nuclear and wind energy [63], coal consumption, mainly in industry, has for years placed the country among the top ten coal consumers in the EU-26. In Finland, coal use in industry, agriculture, and forestry remains significant; between 2014 and 2022, the country ranked between 10th and 12th among EU-26 countries. It is therefore essential to strengthen support for investments in low- and zero-carbon innovations, which will help reduce the emissions intensity of economic sectors.

Finally, in the context of the EU’s efforts to phase out coal, particularly from electricity generation, attention should be paid to inefficient countries such as Bulgaria, Czechia, Estonia, Germany, and Slovenia. They are characterised by the highest shares of coal in electricity production (e.g., in 2022 ranging from approximately 20% to 50%). However, the pace of coal phase-out in these countries may be constrained by high energy demand, which may be one of the causes of inefficiency, albeit an indirect one. It is also worth noting that in 2022, the share of countries exhibiting PTE inefficiency due primarily to excessive coal consumption increased to as much as 46%. This may suggest increasing difficulties in managing this resource effectively. On the other hand, the level of wastefulness related to coal consumption only slightly exceeded that associated with other non-energy resources. Continued efficiency analysis in the coming years is therefore recommended in order to monitor the situation and respond appropriately through targeted policy measures and investment supporting sustainable resource management.

The research findings indicated that a significant issue among inefficient countries was the exceeding of optimal employment levels. This may suggest inefficient organisation of labour resources and/or a mismatch between the employment structure and the actual needs of the economy. Development policy should therefore support actions to modernise the labour market, including the digitalisation of processes, the development of employee skills, and the elimination of unproductive employment, particularly in Central and Eastern European countries such as Bulgaria, Croatia, Hungary, Romania, and Slovenia, where the potential for improvement is greatest. Although the role of human capital in enhancing coal consumption efficiency has not yet been fully explained, it is known that employee education helps to reduce the use of dirty energy, understood as traditional fossil fuels [64], and to decrease CO₂ emissions, including those from fossil fuel combustion [65]. High-quality human capital contributes to the evolution of knowledge and the implementation of new technologies that reduce negative environmental impacts [66].

Lastly, the scope of potential resource reductions identified in this study should serve as a basis for making more informed management decisions aimed at approaching the efficiency frontier and for designing better-targeted policy strategies, particularly in those countries that exhibit poor resource management and operate at a non-optimal scale associated with decreasing returns to scale, such as Austria, Belgium, Hungary, Lithuania, and Sweden, as well as the Czech Republic, Finland, Romania, and Slovakia (although the latter succeeded in operating under increasing returns to scale in 2022). Since each of these countries has unique characteristics, such as specific economic structures, energy and development policies need to be adapted to local conditions. In general, in Central and Eastern European countries such as the Czech Republic, Hungary, Lithuania and Romania, it is recommended to eliminate outdated industrial capacities in order to improve production efficiency and conserve coal resources, as well as to replace coal-fired heating with clean energy sources. For example, policymakers could strengthen regulations that limit industrial and domestic coal-fired emissions. The use of coal stoves in households as the main source of heating remains common in the Czech Republic, Romania, Hungary, and Lithuania. Although the issue of coal heating is less significant in Slovakia, industrial coal consumption places the country in a high seventh position in the EU based on the average for 2014–2022 and is undoubtedly one of the reasons for its very low efficiency. For more economically and technologically developed ‘old EU’ countries such as Austria, Belgium, Finland, and Sweden, further organisational improvements and the implementation of resource-saving technologies are recommended, as well as limiting the excessive concentration of activity in energy- and resource-intensive sectors. Moreover, in the context of decreasing returns to scale, all inefficient countries need to use both energy and non-energy resources more sparingly and effectively. This can be achieved by providing greater technical and financial support to existing production processes to optimise them, as well as through flexible human resource management to better match workforce allocation to changing market needs.

As a final point, it should be added that although the potential reductions identified in this study should not be taken literally, they highlight the essence of the problem, namely the excessive consumption of coal and non-energy resources in relation to the achieved economic growth, as well as the disparities between efficient and inefficient EU member states. Therefore, they should serve as an impetus to find ways to conserve resources and improve energy efficiency without harming economic growth.

5. Conclusions

In the face of global environmental problems caused by fossil fuel combustion, as well as the energy crisis of recent years, the phase-out of coal and improvements in energy efficiency are two key issues that shape the direction of the European Union’s energy and climate policies and form the foundation of the energy transition. However, due to the continuous increase in energy demand, many countries still need to generate energy from coal, and certain industrial processes require the presence of coal’s carbon [52]. To avoid worsening the negative environmental impact, it is therefore essential to manage coal resources rationally, taking into account environmental, social, economic and security aspects. Moreover, in the process of decision-making and defining policy objectives aimed at reducing coal-related overconsumption, it is advisable to rely on appropriate analytical tools that provide essential information and performance indicators. One such tool is the DEA method.

This study employs the DEA method to measure various indicators of relative coal consumption efficiency, namely overall technical efficiency (OTE), pure technical efficiency (PTE), and Scale Efficiency (SE), in both temporal and spatial dimensions. Additionally, it identifies EU countries operating under decreasing or increasing returns to scale, as well as the extent of resource wastage. Therefore, this study not only fills a gap in the existing literature, given the lack of previous analyses focused on coal consumption efficiency at the EU level, but also offers insights that may support the design of more optimal policies to improve energy efficiency. This is possible because it reveals whether countries are using their coal resources at an optimal level or whether they should provide greater technical and financial support for their production processes and pay more attention to policies regarding organisation and employment structure.

Although this research is a valuable source of information on coal consumption efficiency within EU economies, it is not without limitations. Firstly, the efficiency calculated using the DEA method is a relative measure. This means that even if all countries demonstrate low efficiency, some will be relatively efficient (with an efficiency score equal to 1). Similarly, even if all countries were highly efficient, some might still achieve a score below 1. However, efficiency itself is a relative concept, measured by comparing achieved productivity against established standards, norms, or targets. Secondly, the efficiency scores were calculated based on selected traditional DEA models; thus, using other computational models might lead to different results. Thirdly, although the classical DEA method allows for the identification of efficient and inefficient units, it does not, in itself, enable statistical verification of the reliability of the obtained results, so they should be approached with caution. The use of the bootstrap DEA model could enhance the robustness and granularity of the analysis, thereby supporting the interpretation of the final findings. Finally, a significant challenge in this study remains the lack of detailed data, such as separate CO2 emissions data for industry and households.

Since many challenges related to energy efficiency have a strong regional dimension, measuring and assessing efficiency at this level could have important implications for regional policies regarding the implementation and distribution of measures aimed at reducing coal consumption, limiting emissions, and utilising non-energy resources more efficiently. Therefore, further research should expand the cross-sectional framework of this article and conduct comparative analyses at the level of EU regions.