Financial Development, Financial Inclusion and Primary Energy Use: Evidence from the European Union Transition Economies

Abstract

The main objective of the research is to analyze the impact of financial sector development indicators and financial institutions access on primary energy use based on a sample of European Union transition members over 20 years period (1996–2017) through panel cointegration and causality tests that allow for cross-section dependence. The causality analysis revealed that the direction of the causality among financial development indicators, financial institutions access, and primary energy use varied among the countries. On the other side, panel cointegration coefficients disclosed that the financial development index positively affected the primary energy use, but private credit did not have a significant effect on the primary energy use. Furthermore, financial institutions’ access had a significant negative impact on primary energy use. However, country-level cointegration coefficients indicated that the financial development index positively affected the primary energy use in Bulgaria, Croatia, Czechia, Hungary, and Slovenia, and private credit also had a positive impact on primary energy use in Bulgaria, Czechia, Estonia, Hungary, Lithuania, Poland, and Slovakia, but the effect of financial development index on primary energy use was found to be very higher than that of private credit. Moreover, financial institutions’ access negatively affected the primary energy use in Croatia, Estonia, Hungary, Poland, and Romania.

1. Introduction

Energy is one of the main inputs for all the production processes, considered one of the crucial components underlying economic growth and development but at the same time, we confront a lack of energy access for many people especially in African and South Asian countries [1]. Global energy use has accompanied the considerable increases in global production and consumption with the contribution of the accelerating globalization and liberalization trend in the world becoming concomitantly a key contributor to global warming and climate change.

Furthermore, global energy use is expected to increase about 50% during 2018–2050 [2] but still, the primary energy sources are fossil fuels, CO₂ emissions from their combustions being the major source of greenhouse gas emissions [3].

In this context, investigation of the main factors for energy demand can help us to understand patterns of energy production and consumption and to develop energy consumption, forecasting models. Actual context imposes substantial transformation of the global energy system: rapid transition from a system based on fossil fuels to another based on renewable sources, adoption of innovative solutions and technologies for energy production, transmission and consumption, energy efficiency improvements in all economic sectors, etc.

The connection between economic growth and energy consumption has been intensely debated and indicated different results: some studies found unidirectional causality running from energy consumption to economic growth [4,5], others consider that a unidirectional causality exists between economic growth and energy consumption [6], and the rest found either a bidirectional relationship [7,8], either asserted that no connection exists between these variables [8]. Having in mind one of the main challenges that humanity is confronting nowadays: climate change and greenhouse gas emissions, the current studies focus more on issues like energy efficiency, solutions to decrease energy demand maintaining at the same time economic development. Therefore, it is essential to realize that the dynamics between the energy sector and economic growth are still need to be further investigated, more attention is necessary on issues like energy protection policies, patterns of energy supply and consumption, energy consumption by sector and CO₂ emissions, etc. In this context, we consider it important to analyze the impact of financial sector development on energy consumption.

Financial sector development can influence energy use through diverse direct and indirect channels. First, financial development can positively affect energy use by enhancing economic growth. Secondly, financial institutions can provide funds to the economic units at reasonable rates and maturity for purchases durable products such as refrigerators, cars, and houses etc. and, in turn, may raise the energy use [9,10]. Thirdly, a well-functioning financial system can enable the economic units to hedge in case of volatility in energy prices. On the other side, financial development can enhance the reduction of energy use by providing funds for businesses to develop energy-efficient technologies and products [11]. Therefore, the net influence of financial sector development on energy use may be changed by imposing the right policies, regulations, initiatives, by finding the proper solutions to sustain large-scale industries and companies to adopt energy efficiency measures.

Most of the scholars have discovered a positive influence of financial sector development on energy use for different countries and country groups [12,13,14,15]. Furthermore, energy use can affect the development of the financial sector, because the acquisition and use of houses, buses, cars, refrigerators, and dishwashers can influence the money needs of individuals [16]. Therefore, energy use can also affect the financial sector’s development through the aforementioned indirect channels.

Financial institution access can also positively affect energy use by facilitating access to finance for both people and companies that will support investments and will increase demand for energy-intensive commodities [14]. However, financial inclusion enables the economic units to make more energy-efficient investments and to buy more energy-efficient commodities, having a direct influence on global energy consumption. Furthermore, increases in energy consumption can lead to an increased amount of funding and resources required by the economic units that in turn will affect the financial institution’s access. The target of the research is to explore the impact of financial sector development indicators and financial institutions’ access on primary energy use in a sample of European Union (EU) transition states. Our selected countries have undergone a remarkable institutional and economic transformation of the late 1980s with the collapse of the Soviet Block and with the contribution of EU membership negotiations and in turn, made significant improvements in both financial sector development and financial institutions access. The research was conducted through cointegration and causality tests that allowed cross-sectional dependence and heterogeneity, considering the results of pretests. However, the study duration was limited to 1996–2017, because the data for the variable of private credit by deposit money and other financial institutions were only available for 1996–2017. The findings of the study were evaluated to be useful to offer support to policymakers to develop efficient tools that can support financial development while contributing to sustainable development goals. Our main research hypotheses based on the above theoretical arguments are as following:

Hypotheses1(H1).

Financial sector development indicators and financial institutions’ access have a significant impact on primary energy use.

Hypotheses2(H2).

Primary energy use has a significant impact on financial sector development and financial institutions’ access.

This research aims to contribute to the relevant literature in three ways. Firstly, this study will be one of the pioneering studies examining the impact of financial institutions’ access on primary energy use. Secondly, the study employed the financial development index of the International Monetary Fund (IMF) [17], which is calculated based on the depth, access, and efficiency of financial markets and institutions, together with private credit by deposit money bank and other financial institutions, unlike the related literature. Thirdly, the use of cointegration and causality tests that allow for heterogeneity and cross-sectional dependence lead to reach more robust findings. The forthcoming section summarized the empirical studies about the energy–finance nexus and then data and methods were briefly described; Section 4 included the empirical part of the study and the research was discussed in Section 5, followed by conclusions in Section 6.

2. Literature Review

Energy is one of the main inputs for all activities and can be classified as primary energy and secondary energy. The primary energy is extracted or captured from energy sources without its chemical and physical characteristics unchanged, but secondary energy includes the energy transformation of primary energy sources. In other words, the primary energy of fossil fuels, nuclear fuels and solar radiation may be converted to a secondary energy source such as electricity [18]. Primary energy sources are important for all sectors of the economy and since the global production and population are rapidly growing, this, in turn, will increase the energy demand, especially in emerging and developing economies [19]. The increased global energy demand is already raising the energy costs and is expected that these costs would keep their uptrend in the future. Therefore, identification of the critical factors associated with the fluctuation of energy consumption (especially it such as: population growth, industrialization, trade openness, capital stock, urbanization, financial development, and energy prices are important [20] if we want to set up strategies to foster sustainable energy consumption.

In this research, we have analyzed the effect of financial development and financial institutions’ access on primary energy use. The related literature about the finance-energy nexus has reached mixed findings based on the development level of the analyzed countries, methods and the periods examined [12,14,16,21,22,23,24]. Most of the studies such as Kakar et al. [12], Bekhet et al. [13], Ma and Fu [14], Mukhtarov et al. [15], Kahouli [25] have reached the conclusion that financial sector development exerts a positive influence over energy use. Using time-series data, Kakar et al. [12] examined the influence of the financial sector development on energy use in Pakistan over the 1980–2009 period and revealed a positive influence of the sector development on energy use.

Sadorsky [16] also conducted research on the interaction between energy use and financial sector development in Bulgaria, Croatia, Estonia, Kazakhstan, Lithuania, Romania, Serbia, Slovenia and Ukraine through dynamic regression analysis and reached the same conclusion, that financial sector development raised the energy use. The findings of Xu [21] that analyzed China from 1999 to 2009 arrived at the same results, that financial development affects the demand for energy.

On the other side, Shahbaz and Lean [22] revealed a positive long-run influence of financial sector development over energy use and bidirectional causality between these variables in the case of Tunisia.

Islam et al. [26] explored the interaction among energy use, population, and production in Malaysia over the 1971–2009 period through a vector error correction model and revealed a positive influence of the development of the financial sector on energy consumption. Al-mulali and Lee [27] explored the financial development and energy consumption nexus in the Gulf Cooperation Council (GCC) economies over the period of 1980–2019 through the Pedroni cointegration test and discovered a positive influence of financial development on energy use. Bekhet et al. [13] also conducted the same research for the GCC and reached similar findings. Using data from EU members over the 1990–2011 period, Çoban and Topcu [28] discovered a strong positive effect of financial sector development on energy consumption in EU-15 economies and an inverted U-shaped interaction between financial development proxied by bank index and energy consumption in the last twelve EU members. Investigating a sample of Organization for Economic Co-operation and Development (OECD) countries through cointegration test Başarır ve Erçakar [29] also revealed a positive influence of the development of the financial sector on energy use.

Mahalik et al. [30] researched the linkage between energy use and the development of the financial sector in Saudi Arabia over the duration of 1971–2011 through the panel ARDL approach and found that the development of the financial sector raised the energy use in the long run. Lebe and Akbaş [31] analyzed the influence of development in the financial sector on energy use for the 1960–2012 duration through the Maki cointegration test and disclosed that financial development promotes energy consumption. Kahouli [25] explored the interaction among energy use, financial development, and economic growth in Algeria, Egypt, Israel, Lebanon, Morocco, and Tunisia over the 1995–2015 period and reported a positive influence of financial development over the energy use in Morocco, Lebanon, Israel, and Algeria. Same results have also been reported by Mukhtarov et al. [32] in the case of Azerbaijan, based on a cointegration analysis conducted over the 1992–2015 period and by Mukhtarov et al. [15] that carried out an analysis through a vector error correction model based on the data from Kazakhstan that covers 1993–2014 period.

Using data for 1991–2014 collected from 120 nations, Ma and Fu [14] pointed out that financial sector development promotes energy consumption in the developing countries, but could not find a significant influence for the developed countries. Based on cointegration and causality analyses, Janpolat et al. [33] investigated 32 Belt and Road countries over the 2000–2015 period and reported the same findings, that the development of the financial sector exerts a positive influence on energy consumption.

In the relevant literature, relatively few studies have reached the conclusion that financial development contributes to the reduction of energy consumption. In this sense, Chtioui [23] employed a panel cointegration analysis to explored the influence of financial development on energy use in Tunisia over the 1972–2010 period and disclosed a negative influence of financial sector development on overall energy consumption by improving energy efficiency. On the other side, Chang [24] conducted similar research for a panel of 53 countries through threshold regression analysis and reached similar findings with Chtioui [23].

Furthermore, Altay and Topçu [34] explored the finance-energy nexus in Turkey over the 1980–2011 period and revealed no significant interaction between the two variables. The same result has been reached by Topcu and Payne [35] that examined the same relationship using 32 high-income economies during the period 1990–2014 and by Denisova [36] for the case of Germany. Moreover, in Nigeria, Ozdeser et al. [37] found a negative influence of the development of the financial sector on fossil fuel use based using the ARDL method.

Aslan et al. [38] explored the interaction among financial development, economic growth, and energy use in G7 economies and emerging economies over the 1990–2015 duration through the VAR approach and reached the conclusion that stock market development positively affected the energy use in both group countries, but banking sector development positively affected the energy use in emerging economies and decreased the energy use in G7 economies. Godil et al. [39] analyzed the factors behind energy use in India over the 1995–2018 duration through Quantile Autoregressive Distributed Lag and disclosed a positive influence of financial development over energy use.

Some scholars have explored the causal interaction between energy use and the development of the financial sector but reached different causality directions. In this context, Furuoka [40] explored the causal interaction between the development of the financial sector and energy use in Asia countries during 1980–2012 and discovered a unilateral causality from energy use to the development of the financial sector, but Çağlar and Kubar [41] disclosed a unilateral causality from financial sector development to fossil fuel use for Turkey. Çetin [42] explored the relationship between financial sector development and energy use in Turkey for the duration of 190–2015 through the ARDL approach and disclosed that financial sector development raised the energy use and revealed a unilateral causality from financial development to energy use. Ayaydın et al. [43] and Kurt [44] also reached a unilateral causal interaction from the development of the financial sector to energy use for Turkey. Fernandes and Reddy [45] explored the determinants of energy consumption in China, India, Indonesia, Malaysia, Philippines and Thailand and disclosed a unilateral causality from financial development to energy consumption in China. Shahbaz et al. [46] discovered a bilateral causality between energy use and financial development in Pakistan. Gungor and Uzoamaka Simon [47] also disclosed a bilateral causality between financial development and energy consumption for South Africa.

3. Data and Econometric Methodology

The research investigated the impact of financial development indicators and financial institutions’ access on primary energy use. In this context, the dependent variable of primary energy use was proxied by primary energy use as million tonnes of oil equivalent and was collected from BP [48]. The independent variable of financial sector development was proxied by the financial development index of IMF [17] and private credit by deposit money banks and other financial institutions as a percent of gross domestic product (GDP) by World Bank [49]. The financial development index is calculated based on the access, depth, and efficiency of financial markets and institutions. On the other side, financial institutions access was represented by the financial institutions access the index of IMF [17] and the index is calculated based on bank branches and ATMs per 100,000 adults, bank accounts per 1000 adults, the percent of firms with a line of credit, and usage of mobile phones to send and receive money considering the lack of similar data for insurance companies, mutual funds and pension funds (see Svirydzenka [50] for detailed information about the methodology of financial development index and financial institutions index.). All the variables in

Table 1. Data description.

Variables	Description	Source
ENERGY	Primary energy consumption (million tonnes oil equivalent)	BP [48]
FDI	Financial development index	IMF [17]
PCREDIT	Private credit by deposit money banks and other financial institutions to GDP (%)	World Bank [49]
FIA	Index of financial institutions access	IMF [17]

were annual, and the study period was specified as 1996–2017, because the private credit by deposit money banks and other financial institutions existed between the 1996–2017 period.

The eleven EU transition states constituted the sample of the research. The software programs of Gauss 10.0 (APTECH Systems, Higley, AZ, USA), EViews 11.0 (HIS Global, Irvine, CA, USA), and Stata 15.0 (StataCorp LLC, TX, USA) have been used in the econometric analyses. The main characteristics of the series were denoted in

Table 2. Main characteristics of the series.

Characteristics	ENERGY	FDI	PCREDIT	FIA
Mean	24.22727	0.341942	42.55798	0.525661
Median	17.45000	0.320000	42.75970	0.500000
Maximum	103.4000	0.580000	102.5380	0.930000
Minimum	3.200000	0.110000	6.392100	0.150000
Std.Dev.	25.63619	0.111437	19.85068	0.196679
Skewness	1.827448	0.076805	0.302910	0.363396
Kurtosis	5.582011	2.103280	2.772085	2.485036

. The average primary energy consumption was 24.34 million tonnes of oil-equivalent, but considerably varied among the countries. On the other hand, the mean of financial development index was 0.33, the mean of private credits by deposit money banks and other financial institutions was 42.55% of GDP and the average financial institutions access index was 0.502 during the study duration, but the variable of private credits by deposit money banks and other financial institutions exhibited considerable variation among the countries.

In making a choice among the econometric tests, the heterogeneity and cross-sectional dependence firstly should be checked. In this sense, ignoring the existence of cross-sectional dependency would probably lead to bias and size distortions in the econometric analysis [51,52]. Furthermore, the existence of heterogeneity is fundamental for econometric analyses. The homogeneity presupposition of panel parameters cannot regard the heterogeneity between the cross-sections resulting from country-specific characteristics [53].

The econometric tests of heterogeneity and cross-sectional dependence denoted the existence of heterogeneity and cross-sectional dependence among the series therefore the cointegration and causality interactions among financial development indicators, financial institutions access and primary energy use were examined by employing Westerlund and Edgerton [54] Lagrange multiplier (LM) bootstrap cointegration test and Konya [55] bootstrap Granger causality test which takes into account the cross-sectional dependence and heterogeneity. Westerlund and Edgerton [54] cointegration test counts on the cross-sectional dependence and allows heteroskedasticity and autocorrelation in the cointegration equation and also regards the possible endogeneity problem through the use of bootstrap fully modified ordinary least squares. Therefore, the cointegration produces relatively more robust results when compared with the first-generation cointegration tests.

The Konya [55] bootstrap Granger causality test is based on seemingly unrelated regression (SUR) and the critical values of the test are derived for each country through the bootstrap process. For this reason, integration levels of variables are not necessary for the test application. The causality test relies on the SUR estimations below:

$$y_{i,t}={\alpha }_{1,1}+{\displaystyle \sum }_{i=1}^{t{y}_1}{\beta }_{1,1,i}{y}_{1,t-i}+{\displaystyle \sum }_{i=1}^{l{x}_1}{\gamma }_{1,1,i}{\chi }_{1,t-i}+{\epsilon }_{1,1,t }$$

$$y_{2,t}={\alpha }_{1,2}+{\displaystyle \sum }_{i=1}^{l{y}_1}{\beta }_{1,2,i}{y}_{2,t-i}+{\displaystyle \sum }_{i=1}^{l{x}_1}{\gamma }_{1,2,i}{\chi }_{2,t-i}+{\epsilon }_{1,2,t}$$

(1)

$$y_{N,t}={\alpha }_{1,N}+{\displaystyle \sum }_{i=1}^{l{y}_1}{\beta }_{1,N,i}{y}_{N,t-i}+{\displaystyle \sum }_{i=1}^{l{x}_1}{\gamma }_{1,N,i}{\chi }_{N,t-i}+{\epsilon }_{1,N,t}$$

and:

$${\chi }_{1,t}={\alpha }_{2,1}+{\displaystyle \sum }_{i=1}^{l{y}_2}{\beta }_{2,1,i}{y}_{1,t-i}+{\displaystyle \sum }_{i=1}^{l{x}_2}{\gamma }_{2,1,i}{\chi }_{1,t-i}+{\epsilon }_{2,1,t }$$

$${\chi }_{2,t}={\alpha }_{2,2}+{\displaystyle \sum }_{i=1}^{l{y}_2}{\beta }_{2,2,i}{y}_{2,t-i}+{\displaystyle \sum }_{i=1}^{l{x}_2}{\gamma }_{2,2,i}{\chi }_{2,t-i}+{\epsilon }_{2,2,t}$$

(2)

$${\chi }_{N,t}={\alpha }_{2,N}+{\displaystyle \sum }_{i=1}^{l{y}_2}{\beta }_{2,N,i}{y}_{N,t-i}+{\displaystyle \sum }_{i=1}^{l{x}_2}{\gamma }_{2,N,i}{\chi }_{N,t-i}+{\epsilon }_{2,N,t}$$

In the above equations, primary energy use is represented by y, the financial development index is represented by x in the first system; y indicates the primary energy use and x indicates the financial institution access index in the second system. In this sense, a significant one-way causality from x to y is accepted if not all the ${\gamma }_{1,j,i}$s are 0, but all ${\beta }_{2,j,i}$s are 0. On the other hand, a significant one-way causality from y to x is accepted if all ${\gamma }_{1,j,i}$s are 0, but not all ${\beta }_{2,j,i}$s are 0. Lastly, a mutual causality between x and y is accepted if neither ${\gamma }_{1,j,i}$s nor ${\beta }_{2,j,i}$s are zero.

4. Empirical Analysis

In the applied part, heterogeneity and cross-sectional dependence were firstly checked to make a choice among the causality tests. Therefore, cross-sectional dependence was checked by using LM, LM CD (cross-section dependence), and LM_adj. tests, respectively developed by Breusch and Pagan [56], Pesaran [57], and Pesaran et al. [58], and the tests’ findings were denoted in

Table 3. Cross-sectional dependence tests’ findings.

Test	Test Statistic	Prob. (Critical Values at 5% Significance Level)
LM	144	0.0000 (11.15)
LM adj *	17.95	0.0000 (9.50)
LM CD *	8.512	0.0000 (5.75)

* two-sided test.

. The probability values of the three tests were lower than 5%, therefore, the null hypothesis of cross-sectional independence was declined and the existence of cross-sectional dependence among the countries was reached.

The presence of heterogeneity is also important for causality test selection. Therefore, heterogeneity presence was checked with delta tilde tests of Pesaran and Yamagata [59], and the results of the two tests were revealed in

Table 4. Homogeneity tests’ findings.

Test	Test Statistic	Prob. (Critical Values at 5% Significance Level)
$\tilde{\Delta }$	7.598	0.000 (5.80)
${\tilde{\Delta }}_{\mathrm{adj}.}$	8.643	0.000 (4.20)

. The probability values of the tests were lower than 5%, therefore, the null hypothesis of homogeneity was declined and heterogeneity presence was reached. In the light of the findings, employment of causality test-taking notices of heterogeneity and cross-sectional dependence help us to obtain more robust results.

The presence of unit root in the series was checked by Pesaran [60] CIPS (Cross-sectionally augmented IPS (Im, Pesaran and Shin [61])) test given the presence of cross-sectional dependence and the findings were presented in

Table 5. Unit root test.

Variables	Constant		Constant + Trend
	Zt-Bar	Probability Value	Zt-Bar	Probability Value
ENERGY	3.266	0.999	1.193	0.883
d(ENERG)	−6.228	0.000	−5.099	0.000
FDI	0.905	0.817	−0.578	0.282
d(FDI)	−2.929	0.002	−6.482	0.282
PCREDIT	−0.224	0.411	2.551	0.995
d(PCREDIT)	−2.764	0.003	−2.506	0.006
FIA	−1.121	0.131	−1.104	0.135
d(FIA)	−2.774	0.003	−2.414	0.008

. The unit root test results indicated that all series were I(1).

The cointegrating relationship among primary energy use, indicators of financial sector development, and financial institutions access was checked by employing the LM bootstrap cointegration test of Westerlund and Edgerton [54] and the findings were exposed in

Table 6. Panel cointegration test.

Constant			Constant and Trend
Test Statistic	Asymptotic Probability Value	Bootstrap Probability Value	Test Statistic	Asymptotic Probability Value	Bootstrap Probability Value
4.509	0.000	0.890	6.422	0.000	0.970

Note: Bootstrap critical values were derived from 10.000 simulations and lag and lead values were taken as 1.

. The null hypothesis of significant cointegration interaction among the series was accepted given the bootstrap probability values and the significant cointegration relationship among the variables was concluded.

The cointegration coefficients were estimated through the Augmented Mean Group (AMG) estimator of Eberhardt and Teal [62] and were presented in

Table 7. Estimation of cointegrating coefficients.

Country	FDI	PCREDIT	FIA
Bulgaria	9.3329 ***	0.0432 *	−0.3885
Croatia	11.7707 ***	0.0511	−9.0606 ***
Czechia	12.4426 **	0.0987 *	−11.4918
Estonia	14.6545	0.0215 **	−6.4027 **
Hungary	29.4532 ***	0.0804 **	−17.4174 ***
Latvia	4.3393	0.00071	−0.8668
Lithuania	8.3183	0.1246***	9.107 **
Poland	31.8007	0.6293 *	−46.6360 *
Romania	20.6391	0.0296	−18.1561 *
Slovakia	1.3160	0.0682 ***	−7.7112
Slovenia	2.0616 **	0.0002	1.4355
Panel	13.2845 ***	0.0234	−9.7807 **

***, **, and * indicates it is respectively significant at 1%, 5%, and 10%; Source: own processing.

. The panel-level cointegration analysis indicated that financial development proxied by the financial development index had a significant positive impact on primary energy use, but financial institutions access had a significant negative impact on primary energy use. Furthermore, the country level coefficients revealed that financial development proxied by the financial development index positively affected the primary energy use in Bulgaria, Croatia, Czechia, Hungary, and Slovenia and private credits by banks and other financial institutions also had a positive impact on primary energy use in Bulgaria, Czechia, Estonia, Hungary, Lithuania, Poland, and Slovakia, but the effect of financial development index on primary energy use was found to be very higher than that of private credits by banks and other financial institutions. On the other side, the effect of financial institutions’ access on primary energy use was negative in Croatia, Estonia, Hungary, Poland, and Romania, and positive only in the case of Lithuania.

The causal interaction among financial development indicators, financial institutions access, and primary energy use in a sample of EU transition economies was explored through bootstrap Granger causality test and test findings were denoted in

Table 8. Causality analysis between the financial development index and primary energy consumption.

Countries	H0: FDI Is Not the Cause of Energy				H0: Energy Is Not the Cause of FDI
	Wald St.	Bootstrap Critic Value			Wald St.	Bootstrap Critic Values
		1%	5%	10%		1%	5%	10%
Bulgaria	0.4006	26.67847	13.55287	9.00102	2.2946	39.06666	19.72467	12.79664
Croatia	0.7018	34.51029	16.45071	10.77189	2.7336	26.86199	14.97572	10.21100
Czechia	5.3645	31.95752	16.96166	11.44204	0.2666	33.10591	17.81830	12.22203
Estonia	9.6069	37.75689	19.86430	13.40915	1.9819	36.81016	20.08998	13.78954
Hungary	5.3296	28.74907	15.98021	10.85757	6.3971	35.82092	18.03127	12.48281
Latvia	2.8344	33.85149	17.77009	11.58308	9.0389 *	24.96835	12.90118	8.64302
Lithuania	13.162 *	33.28920	17.28678	12.13703	1.8850	32.76813	17.96951	12.17321
Poland	13.635 *	35.83556	18.66600	12.72647	3.4616	36.02957	18.54708	12.68043
Romania	0.1885	35.88219	17.99552	12.24344	25.584 **	37.26994	20.63097	14.30034
Slovakia	0.3006	28.65764	13.88025	9.35830	21.065 **	28.85599	16.30899	11.08176
Slovenia	2.8032	36.13179	18.04382	11.82997	13.805 **	29.10432	15.54230	11.07975

**, * respectively indicates that it is significant at 5%, 10%.

,

Table 9. Causality analysis between private credits and primary energy consumption.

Countries	H0: Pcredit Is Not the Cause of Energy				H0: Energy Is Not the Cause of Pcredit
	Wald St.	Bootstrap Critic Value			Wald St.	Bootstrap Critic Values
		1%	5%	10%		1%	5%	10%
Bulgaria	0.2176	27.63072	14.68623	10.07791	1.5707	32.94590	16.63359	11.08531
Croatia	2.1805	28.26325	15.22205	10.49066	12.724 *	32.86288	16.63140	11.18377
Czechia	46.236 ***	40.19300	20.69382	14.23874	53.862 ***	35.79637	18.11481	12.63455
Estonia	5.6867	37.08983	18.90815	13.20339	1.1105	37.03863	19.38042	13.50000
Hungary	2.3175	36.17370	19.13005	13.30780	76.384 ***	38.53119	20.26179	14.08773
Latvia	9.2645 *	24.30420	12.64270	8.66469	0.1669	37.87162	20.44536	14.46659
Lithuania	4.7854	31.62632	17.16496	11.69833	11.560	37.72914	20.39363	14.23688
Poland	8.1733	37.59888	21.50852	15.01293	8.9713 *	23.85979	12.53312	8.39434
Romania	8.7325	32.94564	18.07043	12.29582	12.881	41.45920	22.49262	15.55226
Slovakia	1.4713	36.94018	18.53431	12.56840	8.8233	35.80409	19.78699	13.60470
Slovenia	2.7451	28.14703	15.29530	10.66410	29.586 **	41.49527	21.86419	14.85829

***,**, * indicates that it is significant at 1%, 5% and 10%.

and

Table 10. Causality analysis between financial institutions accesses and primary energy consumption.

Countries	H0: FIA Is Not the Cause of Energy				H0: Energy Is Not the Cause of FIA
	Wald St.	Bootstrap Critic Value			Wald St.	Bootstrap Critic Value
		1%	5%	10%		1%	5%	10%
Bulgaria	3.0605	25.96186	13.90103	9.67348	2.1632	30.27421	16.07647	10.61107
Croatia	2.8364	30.03195	15.07134	10.43451	15.729 **	29.03331	15.05772	10.22306
Czechia	14.343 *	36.28749	19.47985	13.28411	11.582 *	28.38119	14.71876	9.84669
Estonia	0.3065	40.99779	21.76701	14.64907	8.2643	36.04422	19.21063	13.36456
Hungary	8.3000	35.28400	18.85081	13.08202	2.2852	26.80771	14.16399	9.55937
Latvia	4.8194	24.07212	13.34364	8.95185	2.4580	31.98911	16.12541	11.22346
Lithuania	3.0451	32.82661	16.87194	11.75151	11.1823 *	29.59585	15.52793	10.64909
Poland	5.4631	41.41375	21.77966	14.83314	0.1232	26.53811	13.94197	9.21189
Romania	15.844 *	35.76161	18.13337	12.42248	0.9986	38.38235	19.02732	13.09244
Slovakia	21.285 **	29.48135	16.11610	11.38036	3.6843	28.59840	14.69981	9.57808
Slovenia	3.1894	27.99279	15.33151	10.70887	0.8259	36.28547	17.78769	12.03944

**, * indicates that it is respectively significant at 5% and 10%.

. The causality analysis between financial development index and primary energy use disclosed a unilateral causality from financial development index to primary energy use in Lithuania and Poland, and a unilateral causality from primary energy use to financial development index in Latvia, Romania, Slovakia, and Slovenia.

On the other side, the causality analysis between financial development proxied by private credits by deposit money banks and other financial institutions and primary energy use indicated a unilateral causality from private credits by banks and other financial institutions to primary energy in Latvia, a unilateral causality from primary energy use to private credits by deposit money banks and other financial institutions in Croatia, Hungary, Poland, and Slovenia, and a bilateral causality between the series in the case of Czechia.

The results of causality analysis between financial institutions access and primary energy use revealed a unilateral causality from financial institutions access to primary energy use in Romania and Slovakia, a unilateral causality from primary energy use to financial institutions access in Croatia and Lithuania and a bilateral causality in Czechia.

5. Discussion

The panel-level cointegration coefficients revealed that financial development proxied by the financial development index had a positive impact on primary energy use, but private credit by deposit money banks and other financial institutions did not have a significant effect on primary energy use. The financial development index is calculated based on the access, depth, and efficiency of financial markets and institutions, but private credit by deposit money banks and other financial institutions is a bank-based indicator. Therefore, the financial development index reflects all the aspects of the financial system. On the other side, the country level coefficients revealed that financial development proxied by the financial development index positively affected the primary energy use in Bulgaria, Croatia, Czechia, Hungary, and Slovenia and financial development proxied by private credits by banks and other financial institutions also had a positive impact on primary energy use in Bulgaria, Czechia, Estonia, Hungary, Lithuania, Poland, and Slovakia, but the effect of financial development index on primary energy use was found to be higher than that of private credits by banks and other financial institutions. So, the country-level cointegration coefficients also verify that the financial development index is more inclusive than the variable of private credit.

A well-functioning financial system can provide funds to the economic units at reasonable rates and maturity and this can contribute to production expansion that in turn will raise energy consumption. On the other side, a well-developed financial system may increase the access of the consumers to the credit markets to buy more energy consumer products that obviously will influence energy consumption. Furthermore, the findings of cointegration analysis were found to be compatible with the empirical findings of Kakar et al. [12], Bekhet et al. [13], Ma and Fu [14] Mukhtarov et al. [15], and Kahouli [25]. However, our findings were not consistent with Çoban and Topcu [28] that analyzed the same issue in a sample of EU states during 1990–2011 by dynamic regression analysis. We evaluated that the contradiction can result from different financial development indicators, study duration, and methods.

The cointegration analysis revealed that financial institutions’ access negatively affected the primary energy use at panel level and in a series of countries like Croatia, Estonia, Hungary, Poland, and Romania, but positively affected only Lithuania. So, the findings indicated that financial institutions access enabled the economic units to make more energy-efficient investments and buy more energy-efficient products.

The bootstrap Granger causality analysis disclosed a unilateral causality from financial development index to primary energy use in Lithuania and Poland, and a unilateral causality from primary energy use to financial development index in Latvia, Romania, Slovakia, and Slovenia. On the other side, a unilateral causality running from private credits by banks and other financial institutions to primary energy was noticed in Latvia, and a unilateral causality from primary energy use to private credits by deposit money banks and other financial institutions was found in Croatia, Hungary, Poland, and Slovenia. A bilateral causality between series was reported in Czechia. Furthermore, the causality analysis between financial institutions access and primary energy use revealed a unilateral causality from financial institutions access to primary energy use in Romania and Slovakia, a unilateral causality from primary energy use to financial institutions access in Croatia and Lithuania and a bilateral causality in Czechia. The causality direction varied from country to country depending on country-specific characteristics, the results being similar with other empirical findings: Furuoka [40], Çağlar and Kubar [41], Çetin [42], Fernandes and Reddy [45], Shahbaz et al. [46], and Gungor and Uzoamaka Simon [47].

6. Conclusions

Energy is a vital resource for all economic sectors and for all the economic units. Therefore, identifying the major factors which affect energy consumption is important for designing and implementing the right policies for the public and private sector. Improving the standard of living is the key to sustainable development but this can result in an increased demand for goods and services that in turn will increase energy consumption. Investments in energy-efficient technologies, development of energy-efficient products must be the solution for sustainable development, but this implies that all involved stakeholders understand and sustain the transition to a sustainable future. Right policies and regulations imposed by governments, combined with subsidies to promote renewable energy production and consumption, incentives for investments in renewable energy technologies, low-interest loans for renewable energy projects can have a major impact on a future country’s development.

In this sense, this research analyzed the causality among financial development indicators, financial institutions access, and primary energy use in EU transition states during 1996–2017 through LM bootstrap cointegration and bootstrap Granger causality tests. The causality analysis between financial development indicators and primary energy use revealed a unilateral causality from financial development index to primary energy use in Lithuania and Poland, a unilateral causality from private credits by banks and other financial institutions to primary energy in Latvia, a unilateral causality from primary energy use to financial development index in Latvia, Romania, Slovakia, and Slovenia, a unilateral causality from primary energy use to private credits by deposit money banks and other financial institutions in Croatia, Hungary, Poland, and Slovenia, and a bilateral causality between private credits by deposit money banks and other financial institutions and primary energy use in Czechia.

The panel-level cointegration analysis disclosed that the financial development index had a significant positive impact on primary energy use, but private credits by banks and other financial institutions did not have a significant effect on primary energy use and financial institutions’ access had a significant negative impact on primary energy use. Furthermore, the country-level cointegration coefficients revealed that the financial development index positively affected the primary energy use in Bulgaria, Croatia, Czechia, Hungary, and Slovenia and the variable of private credits by banks and other financial institutions also had a positive impact on primary energy use in Bulgaria, Czechia, Estonia, Hungary, Lithuania, Poland, and Slovakia, but the effect of financial development index on primary energy use was found to be higher than that of private credits by banks and other financial institutions. Lastly, the effect of financial institutions’ access on primary energy use was found to be negative in Croatia, Estonia, Hungary, Poland, and Romania, but positive in Lithuania.

The positive effect of two financial development indicators on primary energy use verified that a well-functioning financial sector is an important factor for the economies given the finance-energy nexus. Furthermore, the financial sector also enables the economic units to hedge against the volatility in energy prices in the globalized world. However, the negative effect of financial institutions’ access on primary energy use indicated that the increasing financial institution’s access fosters investments in energy-efficient technologies and influences purchases of energy-efficient products. Primary energy use affects the financial development in some countries from our sample by increasing the amount of funds required for energy efficiency investments, for purchasing energy-efficient products, etc.

The policy-makers can develop an optimal environment that helps the financial system to operate efficiently and encourage the economic units to use different financial services and, in turn, raise the access, depth, and efficiency of this sector. Furthermore, a developed financial sector can also contribute to economic growth through the energy–finance nexus.