Does Renewable Energy Matter for Economic Growth and Happiness?

Abstract

This paper investigates whether renewable energy influences economic growth and happiness. Using panel data from 25 European Union countries for the period 2012–2022, this study employs a panel model for estimation with fixed and random effects, and robust HAC standard errors. According to the research results, in general, the growing share of renewable energy in the energy mix has a positive impact on economic growth and the happiness of citizens. However, detailed research has shown that this effect depends on the type of energy; a significant positive impact was recorded only in solar share energy, wind share energy and economic growth. However, almost all types of renewable energy were included, i.e., biofuel, hydro, solar and other renewable share energy, and all had a significantly positive impact on the level of happiness. The exception was wind share energy, which showed a significant negative impact. The research findings of this paper provide empirical support for promoting renewable energy, which is positive both for economies and the happiness of citizens. It is one of the main aspects of sustainable economic growth.

1. Introduction

A high share of fossil fuels in the energy mix of European Union (EU) countries is unsustainable for climatic, geopolitical, economic and social reasons. Keeping fossil fuels as the most important energy source causes a number of negative effects on the environment, such as air pollution and global warming [1]. Non-renewable energy sources generate many externalities related to environmental pollution, which negatively contribute to the well-being of the population as well. Due to the fact that European countries are poor in fossil fuel resources and are largely dependent on imports, the supply system can be used as a tool of political pressure. This system has been destabilized by Russia’s actions after Western sanctions were imposed on it. Russia, seeking leverage, wanted to expose consumers to a significant increase in energy bills and supply shortages [2,3,4]. The continued use of fossil fuels is also becoming economically unjustifiable. The price of these types of energy sources are usually characterized by an increasing trend, which results from the depletion of resources and the subsequent increase in extraction costs. Therefore, it has a negative impact on the macroeconomic variables in the economy [4]. The occurrence of energy shocks in European countries also increases the dependence on the global prices of raw materials. The 2022 energy crisis was a shock of enormous scope and extraordinary complexity, which was also influenced by the overlapping effect of the rebound in demand after the pandemic. The research indicates that the contribution of the energy shock to overall inflation was approximately 60% (Q4 2022) and the contribution to core inflation was between 20 and 50% (depending on the model) [5].

The European Union has been trying to meet these threats for many years. Energy transformation is one of the political instruments at the EU level. As indicated above, this is important from the point of view of the current geopolitical situation and the increase in energy prices, which caused the energy crisis [6]. On the other hand, the EU is implementing a plan aimed at climate neutrality, which is expected to lead to zero emissions by 2050. One of the elements of this is to increase the share of renewable sources in energy production, which corresponds to the goals of sustainable development. Renewable energy sources should, by definition, limit the negative impact and reduce not only the consumption itself, but also the costs of energy consumption, which is important for achieving energy security from both a micro- and macroeconomic perspective. The EU target assumes the use of energy from renewable sources at the level of 32% in 2030 [7,8,9]. The need to increase the renewable energy capacity and phase out fossil fuels was also highlighted at the recent COP28 climate summit. At this event, the EU strongly encouraged agreement on global energy goals. As a result, the parties agreed to triple the global capacity of renewable energy sources by 2030 and to double the rate of improvement in energy efficiency [10].

The scale of the energy transformation, understood as an increase in the share of renewable energy sources (RES), is of great importance for the future of employment and quality of life and, therefore, for the level of happiness. On the one hand, the enormous potential of this revolution has been pointed out, resulting from the need for all industries to undertake new production models and investments, which in the long term should generate new jobs and contribute to economic growth. The research suggests that renewable energy consumption generally has a positive impact on economic growth [11,12,13,14,15,16,17]. Moreover, economic growth increases as the use of renewable energy increases and carbon emissions are reduced. The research also suggests that decision-making and policy bodies should work together to increase investment in renewable energy for low-carbon growth [11]. On the other hand, investments in renewable energy sources are expensive and the initial costs of their use are high. This may generate cost-related inflation, especially when the technologies used to produce renewable energy have not yet reached the “maturity level” that allows for cost reduction [3,18,19]. Therefore, the costs of energy transformation are also borne by consumers and end users of electricity. As a consequence, this may lead to a reduction in the economic growth rate. There are studies in the literature that indicate that the use of renewable energy sources is unfavorable for economic growth or even has a negative relationship. Moreover, the results are ambiguous in the short and long term and the relationships are often statistically insignificant [20,21,22,23,24,25]. Also of interest are studies showing that renewable energy consumption is more strongly correlated with economic growth in countries with a higher GDP than in countries with a lower GDP [24]. Counter-arguments to the energy transformation include the threat of the bankruptcy of many traditional industries based on fossil fuels, which will not meet the restructuring requirements. The above literature review shows that the issue of the impact of renewable energy sources on economic growth is not clear. This is a particularly important issue in the face of the European Union’s energy transformation and the implementation of the assumptions of the Green Deal. The above literature review has made it possible to specify the first research goal of this study, which is to assess the relationship between the change in the share of renewable energy in the energy mix and the economic growth rate in European Union countries.

However, research based solely on the assessment of the impact of renewable energy sources on the GDP may be assessed as insufficient. Although GDP is the most common indicator assessing economic activity, it is not the only and indisputable indicator that allows assessing the well-being of society. It ignores many important aspects such as quality and satisfaction with life and level of happiness. In the context of analyzing the energy sector, this is particularly important. Air pollution causes many health problems, causing negative social and economic impacts. The increased share of “clean” energy production determines a better quality of life by reducing environmental pollution and increasing ecological awareness. For this reason, it can be assumed that increasing renewable energy sources will also increase the level of broadly understood happiness in society. An interesting concept is an attempt to link the increase in the share of renewable energy in the energy mix with the level of happiness in societies. There are studies that confirm that such a positive relationship exists [26,27,28,29]. On the other hand, investments in renewable energy are often met with local resistance due to acoustic, odor or visual nuisance. Taking into account the type of renewable energy sources, the spatial approach and the time scope, it should be stated that this issue has not been definitively resolved [30]. Moreover, research shows that the awareness of households in EU countries regarding renewable energy sources and their requirements and expectations for improving the quality of life is low [31]. Other studies have indicated that the use of renewable energy sources is considered too expensive for low-income families to use this type of energy [32]. A number of barriers to the final acceptance of renewable energy sources are also pointed out by [33]. Also interesting is research on the emotional evaluation of renewable energy technologies and to what extent the emotional reactions of communities are correlated with the acceptance of these technologies. Such studies show that renewable energy technologies differ in terms of the emotional reactions they evoke, which may significantly affect their overall acceptance [34]. The above-mentioned issues mean that the relationship between the increase in the share of renewable energy in the energy mix and the level of happiness may not be clear. There are relatively fewer studies of this type conducted than those concerning the GDP-RES relationship. This issue, therefore, allows for the identification of a certain research gap. The above considerations allowed the formulation of the second research goal in this study, which is to assess the relationship between the change in the share of renewable energy in the energy mix and the level of happiness in European Union countries.

This study contributes to the current discussion on the energy transition in Europe. Research on the impact of renewable energy sources on economic growth and citizen happiness is becoming increasingly popular for various countries, which was also the reason for undertaking research in this area for European Union countries. This study is innovative because it takes into account not only the impact of the share of renewable energy in the energy mix in general, but also by individual energy sources, taking macroeconomic variables as control variables.

2. Materials and Methods

Both research objectives were verified using panel models. Twenty-five European Union countries were considered in this research for the period from 2012 to 2022. Cyprus and Malta were excluded due to a lack of statistical data on renewable energy sources. As previously explained, these variables were adopted: GDP at current prices and the happiness index. The selected happiness index is based on life evaluations from the Gallup World Poll that provides the basis for the annual happiness rankings. They are based on answers to the main life evaluation question. The Cantril Ladder asks respondents to think of a ladder, with the best possible life for them being a 10 and the worst possible life being a 0. They are then asked to rate their own current lives on that 0 to 10 scale. They include GDP per capita, social support, healthy life expectancy, freedom, generosity and corruption. The happiness rankings are based on the individuals’ own assessments of their lives, in particular, their answers to the single-item Cantril ladder life-evaluation question (https://worldhappiness.report/about/, accessed on 20 March 2024). Initially, the renewable share energy (% of total primary energy) and then the share of individual types of energy were adopted as explanatory variables. Macroeconomic factors have the greatest impact on economic growth and happiness. Renewable energy sources should, therefore, not be studied in isolation from these variables. They were entered as control variables. To check the correctness of the obtained results, the models were gradually turned on. The models included control variables such as the inflation rate, unemployment rate and current account (as % of GDP). Afia [35] also took these variables into account in his research.

The first stage of this research was to verify the statistical criteria of the collected data. Pearson’s linear correlation coefficients were used for this purpose [36]:

$${\rho }_{X,Y}=\frac{COV(X,Y)}{{\sigma }_X{\sigma }_Y} .$$

(1)

Then the stationarity of the variables was checked using the ADF test.

As a result, the following were adopted as explanatory variables:

• Renewable share energy (% of total primary energy).
• Biofuel energy (% of total primary energy).
• Hydro energy (% of total primary energy).
• Solar energy (% of total primary energy).
• Wind energy (% of total primary energy).
• Other renewable energy (% of total primary energy).
• Inflation rate HICP (%).
• Unemployment rate (%).
• Current account balance (% of GDP).

This study employed the Ordinary Least Squares method (OLS), models with fixed effects (FE) and models with random effects (RE); afterwards, the Wald, Breusch–Pagan and Hausman tests were used to select the most accurate form of the model.

The following model was adopted for the OLS approach [37]:

$${y_{it}=x_{it}\beta +v}_{it} ,$$

(2)

where $Y_{it}$ is the dependent variable, $x_{it}$ is the independent variable (in general, the vector of independent variables), β is the vector of the N dimension of the model’s structural parameters, $v_{it}$ is the total random error composed of the purely random part ${\epsilon }_{it}$ and the individual effect $u_{it}$ pertains to the specific, i-th unit of the panel.

The model with fixed effects (FE) assumed the form [37]:

$$y_{it}=x_{it}\beta +{\mathrm{u}}_i+{\epsilon }_{it}\mathrm{z},$$

(3)

and the model with random effects (RE) looked as follows [37]:

$${\widehat{\beta }}_{RE}={\left(X^T{\mathsf{\Omega }}^{-1}X\right)}^{-1}{X}^T{\mathsf{\Omega }}^{-1}y ,$$

(4)

where ${\widehat{\beta }}_{RE}$ is the generalized estimator of the least squares of structural parameters, $X$ is a matrix of independent variables, $y$ is a vector of dependent variables and $\mathsf{\Omega }$ is a reversible matrix of variance and covariance of the total random error. The validity of the models was assessed with the Wald, Breusch–Pagan and Hausman tests [38].

The Wald test serves to determine whether the inclusion of different intercept terms allows one to attain more precise estimates of the parameters of a model of panel data. This test is based on F statistics and the verified hypotheses assume the form [38]:

$$H_0:{\beta }_i={\beta }_0=const.,$$

(5)

$$H_1:\exists i,j\hspace{1em}{\beta }_i\ne {\beta }_j ale {\beta }_{it}={\beta }_{is}={\beta }_i\hspace{1em}i=\mathrm{1,2},\dots ,N\hspace{1em}t,s=\mathrm{1,2},\dots ,T,$$

(6)

The null hypothesis ($H_0$) assumes that all intercepts, regardless of the object or time, are equal, whereas the alternative hypothesis ($H_1$) assumes that intercepts are constant in the time function, but varied for the i-th objects. The choice of the null hypothesis indicates the homogeneity of the evaluated objects and provides justification for building a model with the help of the OLS method, whereas the choice of the alternative hypothesis proves the lack of homogeneity of the objects and inclines one to employ an estimator of panel data models with individual effects (FE).

The Breusch–Pagan test enables the user to verify the assertion about the constant variance of the random factor of N objects. The null and alternative hypotheses are as follows [38]:

$$H_0:{\sigma }_{{\beta }_i}^2=0,$$

(7)

$$H_1:{\sigma }_{{\beta }_i}^2\ne 0.$$

(8)

The adoption of the null hypothesis means that the variance of the random factor of individual effects is zero, hence it does not change when entering individual effects into the model. Thus, the OLS model should be applied to processing panel data. In turn, the choice of the alternative hypothesis indicates that the variance of the random factor of individual effects differs from zero and, therefore, the model should be estimated with the RE estimator.

The Hausman test validates the assumption on the unbiasedness of the FE and RE estimators. The hypotheses are as follows [38]:

$$H_0:E\left({\beta }_{it},x_{itk}\right)=0,$$

(9)

$$H_1:E({\beta }_{it},x_{itk})\ne 0.$$

(10)

When the null hypothesis is accepted, individual effects are independent from explanatory variables, hence both estimators are unbiased. According to this hypothesis, the RE estimator is more efficient. In turn, according to the alternative hypothesis, the FE estimator is unbiased and the RE estimator is biased, which suggests selecting the former for the estimation of the model.

3. Results

3.1. Renewable Energy Sources in EU-25 Countries

European Union countries constitute a diverse group in terms of the share of renewable energy in the energy mix, as shown in Figure 1. The share of renewable energy in the energy mix of all EU countries increased between 2012 and 2022. In 2022, the largest share was recorded in the Scandinavian countries, such as Sweden (53.31%), Denmark (43.04%) and Finland (38.50%), while the smallest share was found in the countries of Central and Eastern Europe, which include the Czech Republic (7.40%), Hungary (9.12%) and Poland (9.37%). At the same time, the Scandinavian countries increased their share of renewable energy sources at the fastest rate compared to other countries.

There is also diversity in the European Union countries in terms of the structure of renewable energy sources, as shown in Figure 2. The structure of renewable energy in the EU-25 countries in 2022 was diverse. In countries such as Austria, Croatia, Latvia, Romania, Slovakia and Sweden, hydropower was the dominant renewable energy source. Its share ranged from 0.43% in Poland to 28.78% in Sweden. The second source of renewable energy that dominated in the EU-25 economies was wind energy. These countries included Denmark, Greece, Spain, The Netherlands, Ireland, Lithuania, Germany, Poland and Portugal. This share ranged from 0.001% in Slovakia to 26.16% in Denmark. Other renewable sources dominated in the Czech Republic and Estonia.

3.2. Econometric Modeling

Calculations were made in Statistica 13.3 and Gretl ver. 2024b. The descriptive statistics of each variable are shown in

Table 1. Data description and sources.

Variable	Data Description and Unit	Sources
Gross Domestic Product (GDP)	GDP at current prices; million EUR	Eurostat
Happiness index (HI)	The country scores are based on a survey in which respondents evaluated the quality of their current lives on a scale of 0 to 10	https://worldhappiness.report/data/, accessed on 20 March 2024
Renewable share energy (RSE)	% of total primary energy	https://ourworldindata.org/energy, accessed on 20 March 2024
Biofuel share energy (BSE)	% of total primary energy	https://ourworldindata.org/energy, accessed on 20 March 2024
Hydro share energy (HSE)	% of total primary energy	https://ourworldindata.org/energy, accessed on 20 March 2024
Solar share energy (SSE)	% of total primary energy	https://ourworldindata.org/energy, accessed on 20 March 2024
Wind share energy (WSE)	% of total primary energy	https://ourworldindata.org/energy, accessed on 20 March 2024
Other renewable share energy (ORSE)	% of total primary energy	https://ourworldindata.org/energy, accessed on 20 March 2024
Inflation rate (IR)	Harmonized Indices of Consumer Prices (HICP); %	Eurostat
Unemployment rate (UR)	Unemployed persons aged 15 to 74 as a percentage of the labor force; %	Eurostat
Current account balance (CAB)	All transactions (other than those in financial items) in goods, services, primary income and secondary income which occur between resident and non-resident units; % of GDP	Eurostat

.

The descriptive characteristics of the variables are presented in

Table 2. Descriptive statistics.

Variable	Mean	Min	Max	Standard Deviation	Coefficient of Variation
GDP	522,750	17,917	3,876,810	7.82	1.50
HI	6.43	3.99	7.89	0.78	0.12
RSE	17.13	2.99	53.31	10.81	0.63
BSE	1.05	0.01	4.18	0.64	0.61
HSE	7.10	0.01	33.18	8.29	1.17
SSE	1.18	0.00	5.91	1.21	1.02
WSE	4.50	0.00	26.16	4.74	1.05
ORSE	3.30	0.10	14.78	2.87	0.87
IR	2.20	−1.60	19.40	3.25	1.48
UR	8.49	2.00	27.80	4.67	0.55
CAB	1.30	−19.90	13.7	4.24	3.25

.

Table 3. Correlation matrix.

	GDP	HI	RSE	BSE	HSE	SSE	WSE	ORSE	IR	UR	CAB
GDP	1.00	0.23 *	−0.03	−0.03	−0.17 *	0.47 *	0.11	−0.01	−0.02	−0.05	0.23 *
HI	0.23 *	1.00	0.32 *	0.46 *	0.06	−0.07	0.31 *	0.47 *	0.07	−0.44 *	0.38 *
RSE	−0.03	0.32 *	1.00	0.39 *	0.78 *	−0.03	0.52 *	0.55 *	0.01	0.02	0.12 *
BSE	−0.03	0.46 *	0.39 *	1.00	0.20 *	−0.05	0.25 *	0.26 *	0.17 *	−0.23 *	0.21 *
HSE	−0.17 *	0.06 *	0.78 *	0.20 *	1.00	−0.23 *	−0.06	0.16 *	−0.05	0.09	−0.06
SSE	0.47 *	−0.07	−0.03	−0.05	−0.23 *	1.00	0.17 *	−0.14 *	0.19 *	0.19 *	0.02
WSE	0.11	0.31 *	0.52 *	0.25 *	−0.06	0.17 *	1.00	0.35 *	−0.01	0.04	0.25 *
ORSE	−0.01	0.47 *	0.55 *	0.26 *	0.16 *	−0.14 *	0.35 *	1.00	0.08	−0.28 *	0.17 *
IR	−0.02	0.07 *	0.01	0.17 *	−0,05	0.19 *	−0.01	0.08	1.00	−0.27 *	−0.22 *
UR	−0.04	−0.44 *	0.02	−0.23 *	0.09	0.19 *	0.04	−0.28 *	−0.27 *	1.00	−0.20 *
CAB	0.23 *	0.38 *	0.12 *	0.21 *	−0.06	0.02	0.25 *	0.17 *	−0.22 *	−0.20 *	1.00

Notes: * denotes significance levels of 10%.

presents Pearson’s linear correlation coefficients for the analyzed variables.

None of the pairs of variables showed a strong correlation. Variables positively correlated with GDP included the happiness index (0.23), solar share energy (0.47), wind share energy (0.11) and the current account balance (0.23). Variables that were negatively correlated with GDP were renewable sources of energy (−0.03), biofuel source energy (−0.03), hydro source energy (−0.17), other renewable source energy (−0.01), the inflation rate (−0.02) and the unemployment rate (−0.04). Variables positively correlated with the happiness index included GDP (0.23), renewable source energy (0.32), biofuel source energy (0.46), hydro source energy (0.06), wind source energy (0.31), other renewable source energy (0.47), the inflation rate (0.07) and current account balance (0.38). Variables that were negatively correlated with the happiness index were solar share energy (−0.07) and the unemployment rate (−0.44). In most cases, the correlations were relatively weak, which did not raise concerns about the occurrence of collinearity.

Additionally, the variance inflation factor (VIF) was calculated for each model. Its value below 10 indicates that the model has acceptable levels of multicollinearity [39]. According to

Table 4. VIF coefficients of independent variables.

Variable	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6	Model 7	Model 8
RSE	1.02	1.00	1.00	1.02	1.02	1.00	1.02
HICP	1.19	1.08	1.00	1.05
UR	1.18	1.08			1.05	1.00
CAB	1.17			1.07	1.06		1.02

and

Table 5. VIF coefficients of independent variables.

Variable	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6	Model 7	Model 8
BSE	1.28	1.25	1.19	1.25	1.24	1.23	1.18	1.15
HSE	1.19	1.18	1.12	1.14	1.19	1.18	1.13	1.12
SSE	1.24	1.23	1.17	1.17	1.14	1.15	1.12	1.12
WSE	1.36	1.32	1.28	1.31	1.36	1.31	1.30	1.26
ORSE	1.36	1.36	1.26	1.28	1.36	1.36	1.25	1.25
HICP	1.33	1.18	1.10	1.19
UR	1.43	1.34			1.28	1.25
CAB	1.29			1.21	1.14		1.11

there is no multicollinearity problem.

Then, the ADF tests for the stationarity of the level variables were performed. Their results are presented in

Table 6. ADF test for stationarity.

Variable	At Level	p Value
GDP	−3.5258	0.04
HI	−4.4204	0.01
RSE	−5.2766	0.01
BSE	−5.2631	0.01
HSE	−4.4598	0.01
SSE	−4.4752	0.01
WSE	−5.0284	0.01
ORSE	−3.5679	0.04
IR	−10.4060	0.01
UR	−4.9409	0.01
CAB	−5.1466	0.01

.

The ADF test results show the stationarity of all of the variables. Therefore, the variables were adopted for one model. Then, the Wald, Breusch–Pagan and Hausman tests were carried out, the results of which indicate the most optimal estimator. OLS, models with fixed effects and random effects were taken into account.

Before modeling, the variables were presented in a scatterplot form, which is shown in Figure 3.

First, the models containing the total renewable energy sources as explanatory variables were estimated. Then, the models with explanatory variables divided into individual types of energy sources were estimated. In addition, control variables, namely the inflation rate, unemployment rate and current account balance, were included in the models to check the robustness of the results.

Table 7. Results for the estimation of parameters in panel models. Dependent variables: GDP (robust HAC).

Variables	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6	Model 7	Model 8
Const.	429,535 *** (0.001)	427,682 *** (0.001)	337,316 *** (0.003)	336,498 *** (0.003)	437,226 *** (0.001)	436,312 *** (0.001)	319,259 *** (0.004)	318,324 * (0.052)
RSE	7940.99 * (0.074)	8240.71 * (0.060)	9988.94 ** (0.020)	9783.45 ** (0.024)	9205.02 ** (0.046)	9249.87 ** (0.043)	11,926.3 *** (0.009)	11,935.4 *** (0.000)
HICP	5928.74 *** (0.010)	5179.82 *** (0.002)	6524.98 *** (0.001)	7178.51 *** (0.000)
UR	−6989.55 ** (0.018)	−6766.38 ** (0.027)			−8585.93 *** (0.005)	−8476.89 *** (0.006)
CAB	2707.17 ** (0.026)			2225.58 ** (0.043)	598.496 (0.730)		−598.163 (0.769)
Obs.	275	275	275	275	275	275	275	275
F-Value	4.144 (0.003)	0.387 (0.763)	0.192 (0.825)	5.522 (0.001)	5.368 (0.001)	0.466 (0.628)	8.082 (0.003)	0.322 (0.571)
Wald test	F(24.246) = 1003.2 (0.000)	F(24.247) = 1057.76 (0.000)	F(24.248) = 1024.46 (0.000)	F(24.247) = 966.572 (0.000)	F(24.247) = 964.357 (0.000)	F(24.248) = 1022.61 (0.000)	F(24.248) = 1259.63 (0.000)	F(24.249) = 966.833 (0.000)
Breusch–Pagan test	LM = 1267.28 (0.000)	LM = 1330.68 (0.000)	LM = 1328.98 (0.000)	LM = 1271.59 (0.000)	LM = 1259.16 (0.000)	LM = 1338.24 (0.000)	LM = 1163.96 (0.000)	LM = 1333.24 (0.000)
Hausman test	H = 4.337 (0.362)	H = 3.470 (0.321)	H = 2.788 (0.248)	H = 4.705 (0.175)	H = 1.194 (0.363)	H = 0.651 (0.722)	H = 3.443 (0.179)	H = 1.019 (0.313)
CD Pesaran	z = 0.483 (0.629)	z = 0.422 (0.673)	z = 5.237 (0.000)	z = 5.409 (0.000)	z = 4.800 (0.000)	z = 4.338 (0.000)	z = 7.270 (0.000)	z = 7.888 (0.000)
Woolridge test	F = 18.191 (0.000)	F = 18.441 (0.000)	F = 23.403 (0.000)	F = 23.139 (0.000)	F = 24.040 (0.000)	F = 18.441 (0.000)	F = 33.439 (0.000)	F = 33.822 (0.000)
Breusch–Pagan test	BP = 1302.3 (0.000)	BP = 1272.5 (0.000)	BP = 1137.7 (0.000)	BP = 1154.0 (0.000)	BP = 1183.5 (0.000)	BP = 1167.1 (0.000)	BP = 1005.4 (0.000)	BP = 999.9 (0.000)
Estimator	RE	RE	RE	RE	RE	RE	RE	RE

Notes: *, ** and *** denote significance levels of 10%, 5% and 1%, respectively.

and

Table 8. Results for the estimation of parameters in panel models. Dependent variables: GDP (robust HAC).

Variables	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6	Model 7	Model 8
Const.	475,114 *** (0.001)	473,042 *** (0.001)	443,768 *** (0.001)	443,155 *** (0.002)	478,453 *** (0.001)	476,449 *** (0.001)	442,595 *** (0.002)	443,102 *** (0.002)
BSE	−17,295.3 (0.476)	−17,283.4 (0.476)	−17,037.4 (0.485)	−17,036.9 (0.488)	−10,849.8 (0.650)	−11,827.4 (0.617)	−10,216.6 (0.675)	−11,113.5 (0.645)
HSE	−2806.49 (0.184)	−2575,35 (0.217)	−3323.52 (0.110)	−3635.23 * (0.085)	−3296.49 (0.138)	−3056.91 (0.161)	−4265.41 * (0.056)	−3969.62 * (0.071)
SSE	40,472.7 ** (0.032)	39,064.5 ** (0.048)	41,494.2 ** (0.024)	43,148.9 ** (0.014)	47,226.3 ** (0.014)	45,184.6 ** (0.024)	50,643.7 *** (0.004)	48,524.5 *** (0.009)
WSE	20,154.5 * (0.087)	20,911.7 * (0.073)	21,881.5 ** (0.047)	21,242.2 * (0.054)	19,718.9 * (0.090)	20,336.6 * (0.078)	20,916.4 * (0.056)	21,398.5 * (0.051)
ORSE	−12,203.8 ** (0.031)	−11,466.7 * (0.058)	−10,401.2 * (0.092)	−11,007.2 * (0.064)	−12,121.5 ** (0.038)	−11,595.9 * (0.057)	−10,768.5 * (0.083)	−10,383.7 (0.101)
HICP	3857.37 *** (0.010)	3054.14 ** (0.026)	3269.36 *** (0.016)	4040.82 *** (0.005)
UR	−2989.07 (0.462)	−2719.90 (0.516)			−3355.50 (0.042)	−3102.10 (0.448)
CAB	3276.44 ** (0.030)			3101.42 ** (0.040)	2236.25 (0.121)		1987.19 (0.181)
Obs.	275	275	275	275	275	275	275	275
F-Value	13.540 (0.000)	13.719 (0.000)	14.059 (0.000)	14.594 (0.000)	14.958 (0.000)	14.237 (0.000)	16.843 (0.000)	15.755 (0.000)
Wald test	F(24.242) = 880.558 (0.000)	F(24.243) = 904.583 (0.000)	F(24.244) = 933.530 (0.000)	F(24.243) = 893.034 (0.000)	F(24.243) = 875.382 (0.000)	F(24.244) = 923.4 (0.000)	F(24.244) = 880.080 (0.000)	F(24.245) = 937.943 (0.000)
Breusch–Pagan test	LM = 111.24 (0.000)	LM = 1086.84 (0.000)	LM = 1190.77 (0.000)	LM = 1164.47 (0.000)	LM = 1128.96 (0.000)	LM = 1111.53 (0.000)	LM = 1163.96 (0.000)	LM = 1186.4 (0.000)
Hausman test	H = 9.289 (0.318)	H = 10.029 (0.187)	H = 6.055 (0.418)	H = 7.030 (0.428)	H = 8.661 (0.278)	H = 9.230 (0.161)	H = 7.048 (0.316)	H = 6.009 (0.305)
CD Pesaran	z = 3.347 (0.001)	z = 3.188 (0.001)	z = 4.888 (0.000)	z = 4.893 (0.000)	z = 2.602 (0.009)	z = 1.842 (0.065)	z = 3.524 (0.000)	z = 3.342 (0.001)
Woolridge test	F = 16.112 (0.000)	F = 16.401 (0.000)	F = 19.078 (0.000)	F = 18.674 (0.000)	F = 18.881 (0.000)	F = 18.801 (0.000)	F = 22.543 (0.000)	F = 22.540 (0.000)
Breusch–Pagan test	BP = 1128.0 (0.000)	BP = 1094.8 (0.000)	BP = 1034.6 (0.000)	BP = 1058.4 (0.000)	BP = 1065.4 (0.000)	BP = 1039.5 (0.000)	BP = 988.56 (0.000)	BP = 971.34 (0.000)
Estimator	RE	RE	RE	RE	RE	RE	RE	RE

Notes: *, ** and *** denote significance levels of 10%, 5% and 1%, respectively.

present the results of econometric modeling using selected estimators with GDP as the dependent variable. Additionally, the models included robust standard errors (robust HAC), which made it possible to reduce the problems of autocorrelation, cross-sectional dependence and heteroscedasticity [40].

The results of the diagnostic tests show that the optimal estimator for all models was the random effects estimator. In all estimated models, the renewable share energy had a positive, significant impact on GDP, which means that their increase in the energy mix of EU economies resulted in an increase in GDP. This significant and positive effect was confirmed regardless of the combination of control variables used. A clear positive impact was recorded in the case of the inflation rate, while a negative impact was recorded in the case of the unemployment rate. All mentioned variables were significant.

The results of the diagnostic tests showed that the optimal estimator for all models was the random effects estimator. The results of the estimated models were slightly different after taking into account the individual types of renewable energy in the models. In all models, a clear significant positive impact on GDP was recorded for solar share energy and wind share energy. In the case of the other renewable share energy variable, its impact was negative and significant in most models. The exception was model 8, in which this variable was not significant. Biofuel share energy was a variable with a negative, but insignificant influence, as well as hydro share energy, which was negative but significant only in models 4, 7 and 8. The inflation rate had a clear positive, significant impact, while the impact and significance of the remaining control variables was unclear.

Table 9. Results for the estimation of parameters in panel models. Dependent variables: happiness index (robust HAC).

Variables	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6	Model 7	Model 8
Const.	6.759 *** (0.000)	6.765 *** (0.000)	5.895 *** (0.000)	5.883 *** (0.000)	6.789 *** (0.000)	6.767 *** (0.000)	5.860 *** (0.000)	5.864 *** (0.000)
RSE	0.015 ** (0.039)	0.014 * (0.058)	0.029 *** (0.001)	0.032 *** (0.024)	0.013 (0.129)	0.014 * (0.065)	0.035 *** (0.001)	0.033 *** (0.000)
HICP	−0.003 *** (0.717)	0.0002 (0.968)	0.015 ** (0.020)	0.009 (0.234)
UR	−0.068 *** (0.000)	−0.069 *** (0.000)			−0.067 *** (0.000)	−0.069 *** (0.000)
CAB	−0.010 (0.126)			−0.017 ** (0.047)	−0.012 * (0.062)		−0.021 *** (0.010)
R2				0.87	0.91		0.87
Adjusted R2				0.16	0.41		0.15
Obs.	275	275	275	275	275	275	275	275
F-Value
Wald test	F(24.246) = 62.372 (0.000)	F(24.247) = 68.290 (0.000)	F(24.248) = 60.176 (0.000)	F(24.247) = 50.895 (0.000)	F(24.247) = 62.606 (0.000)	F(24.248) = 68.916 (0.000)	F(24.248) = 52.305 (0.000)	F(24.249) = 59.110 (0.000)
Breusch–Pagan test	LM = 822.881 (0.000)	LM = 989.379 (0.000)	LM = 965.541 (0.000)	LM = 687.096 (0.000)	LM = 823.48 (0.000)	LM = 997.47 (0.000)	LM = 679.754 (0.000)	LM = 955.024 (0.000)
Hausman test	H = 8.754 (0.068)	H = 3.056 (0.383)	H = 0.237 (0.888)	H = 13.382 (0.004)	H = 8.658 (0.034)	H = 0.986 (0.611)	H = 14.764 (0.001)	H = 0.658 (0.417)
CD Pesaran	z = 0.986 (0.324)	z = 1.022 (0.307)	z = 8.900 (0.000)	z = 8.121 (0.000)	z = 1.547 (0.122)	z = 0.983 (0.325)	z = 6.676 (0.000)	z = 6.684 (0.000)
Woolridge test	F = 37.683 (0.000)	F = 37.287 (0.000)	F = 69.600 (0.000)	F = 68.456 (0.000)	F = 38.626 (0.000)	F = 39.302 (0.000)	F = 62.957 (0.000)	F = 66.256 (0.000)
Breusch–Pagan test	BP = 97.721 (0.000)	BP = 102.93 (0.000)	BP = 91.086 (0.000)	BP = 87.803 (0.000)	BP = 89.546 (0.000)	BP = 91.728 (0.000)	BP = 88.694 (0.000)	BP = 93.683 (0.000)
Estimator	RE	RE	RE	FE	FE	RE	FE	RE

Notes: *, ** and *** denote significance levels of 10%, 5% and 1%, respectively.

and

Table 10. Results for the estimation of parameters in panel models. Dependent variables: happiness index (robust HAC).

Variables	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6	Model 7	Model 8
Const.	6.641 *** (0.000)	6.624 *** (0.000)	6.877 *** (0.000)	5.908 *** (0.000)	6.635 *** (0.000)	6.644 *** (0.000)	5.909 *** (0.000)	5.875 *** (0.000)
BSE	0.159 ** (0.032)	0.151 ** (0.027)	0.170 * (0.089)	0.164 (0.130)	0.147 ** (0.046)	0.151 ** (0.047)	0.159 (0.140)	0.172 * (0.086)
HSE	0.022 *** (0.008)	0.015 ** (0.046)	0.0004 (0.953)	0.002 (0.807)	0.023 *** (0.008)	0.021 *** (0.009)	0.002 (0.755)	0.0003 (0.968)
SSE	0.042 (0.385)	0.041 (0.385)	0.096 * (0.080)	0.105 * (0.065)	0.029 (0.474)	0.038 (0.364)	0.099 ** (0.042)	0.098 ** (0.032)
WSE	−0.043 ** (0.016)	−0.033 ** (0.012)	−0.014 (0.402)	−0.017 (0.395)	−0.042 ** (0.016)	−0.045 ** (0.012)	−0.017 (0.398)	−0.014 (0.386)
ORSE	0.069 ** (0.026)	0.066 *** (0.009)	0.096 *** (0.005)	0.096 ** (0.024)	0.069 ** (0.025)	0.066 ** (0.035)	0.096 ** (0.023)	0.096 *** (0.004)
HICP	−0.001 (0.425)	−0.005 (0.566)	0.001 (0.895)	−0.003 (0.765)
UR	−0.069 *** (0.000)	−0.067 *** (0.000)			−0.068 *** (0.000)	−0.069 *** (0.000)
CAB	−0.012 ** (0.041)			−0.016 ** (0.029)	−0.010 * (0.079)		−0.015 ** (0.018)
R2 Adjusted R2	0.93 0.49			0.89 0.28	0.92 0.48	0.92 0.48	0.89 0.27
Obs.	275	275	275	275	275	275	275	275
F-Value	28.782 (0.000)	29.679 (0.000)	24.133 (0.000)	25.970 (0.000)	32.909 (0.000)	33.630 (0.000)	30.051 (0.000)	29.046 (0.000)
Wald test	F(24.242) = 58.958 (0.000)	F(24.243) = 61.421 (0.000)	F(24.244) = 48.375 (0.000)	F(24.243) = 44.437 (0.000)	F(24.243) = 58.834 (0.000)	F(24.244) = 62.517 (0.000)	F(24.244) = 44.837 (0.000)	F(24.245) = 48.584 (0.000)
Breusch–Pagan test	LM = 742.691 (0.000)	LM = 821.884 (0.000)	LM = 795.571 (0.000)	LM = 636.894 (0.000)	LM = 739.748 (0.000)	LM = 829.57 (0.000)	LM = 640.882 (0.000)	LM = 796.408 (0.000)
Hausman test	H = 16.475 (0.036)	H = 12.587 (0.083)	H = 6.574 (0.362)	H = 14.688 (0.040)	H = 17.000 (0.017)	H = 12.707 (0.048)	H = 15.174 (0.019)	H = 6.876 (0.230)
CD Pesaran	z = 0.941 (0.347)	z = 0.854 (0.393)	z = 5.961 (0.000)	z = 6.352 (0.000)	z = 1.977 (0.048)	z = 1.645 (0.100)	z = 7.016 (0.000)	z = 5.706 (0.000)
Woolridge test	F = 38.833 (0.000)	F = 39.098 (0.000)	F = 82.516 (0.000)	F = 80.222 (0.000)	F = 42.724 (0.000)	F = 43.382 (0.000)	F = 79.141 (0.000)	F = 82.624 (0.000)
Breusch–Pagan test	BP = 109.53 (0.000)	BP = 110.66 (0.000)	BP = 82.743 (0.000)	BP = 78.759 (0.000)	BP = 99.218 (0.000)	BP = 98.864 (0.000)	BP = 73.982 (0.000)	BP = 75.565 (0.000)
Estimator	FE	RE	RE	FE	FE	FE	FE	RE

Notes: *, ** and *** denote significance levels of 10%, 5% and 1%, respectively.

present the results of econometric modeling using selected estimators with the happiness index as the dependent variable and robust HAC.

The results of the diagnostic tests show that the optimal estimator for models 1–3, 6 and 8 was the random effects estimator and for models 4–5 and 7, it was the fixed-effect estimator. In most of the estimated models, renewable share energy had a positive, significant impact on the happiness index, which means that its increase in the energy mix of EU economies results in an increase in the happiness index. Only in model 5 was this variable not significant. A positive effect was confirmed regardless of the combination of control variables used. A clear negative significant impact was recorded in the case of the unemployment rate. The impact and significance of the remaining control variables was unclear.

The results of the diagnostic tests show that the optimal estimator for models 1 and 4–6 was the fixed effects estimator and for models 2–3 and 7–8, it was the random effects estimator. The results of the estimated models were slightly different after taking into account individual types of renewable energy in the models. In all models, a clear positive impact on the happiness index was recorded for solar biofuel share energy, hydro share energy, solar share energy and other renewable share energy. In most of the estimated models, this impact was significant. The only negative impact was recorded in the case of wind share energy. The inflation rate was an insignificant variable in all models, with a negative impact in most models. However, the unemployment rate and current account balance were significant and had a negative impact in all models.

4. Discussion

In general, the conducted research showed that renewable energy sources have a positive impact on both economic growth and citizens’ sense of happiness. The research results are consistent with those previously obtained by other researchers. The research results of Pirlogea and Cicea [41] showed a positive impact of renewable energy sources on GDP per capita in the case of Romania, Spain and the EU-27 countries. Moreover, the authors indicated that these countries should replace traditional energy sources such as gas, oil and coal with renewable sources in order to reduce their negative impact on the natural environment. Pao and Fu [42] have investigated relationships between the real GDP and types of energy consumption in Brazil from 1908—2010. Their study has shown that the influence of renewable energy consumption on real output is positive and significant. Similar results were also seen for 10 Asian countries in the years 1990–2020 (Bangladesh, China, India, Indonesia, Japan, Malaysia, Maldives, Nepal, Pakistan and Vietnam) [43], for OECD countries in the years 1990–2015 [44] and for 38 countries with the highest consumption of renewable energy in 1991–2012 [11].

This study also confirmed the positive impact of renewable energy sources on happiness. The conclusions are consistent with the results obtained by other authors. According to Karthik et al. [45], the use of renewable energy sources improves the quality of the natural environment and thus, the sense of subjective satisfaction, while fossil fuels have the opposite effect. The positive impact of renewable energy and the negative impact of environmental degradation on life satisfaction was demonstrated by Omri et al. [46] in a sample of 36 emerging countries from 2005 to 2014. In turn, Aldieri et al. [47] examined the impact of ecoefficiency, measured by the percentage of the total energy derived from renewable fuels and waste. The research results showed a positive impact of ecoinnovation on happiness. Moreover, happiness is related with environmental conditions, the pace of introducing ecoinnovations and their effectiveness. The research covered 10 European Union countries in the years 1981–2011. According to Wang et al. [48], the ecological footprint and globalization had a negative impact on subjective wellbeing in OECD countries in the years 2008–2020. According to that study, policymakers should be promoting sustainable economic growth while enhancing subjective wellbeing with investment in innovation and sustainable development. The links between renewable energy sources, economic growth and happiness were included in research conducted by Hafez and Adris [49], showing positive correlations between them. The findings of Kumari et al. [27] have revealed that renewable energy consumption and environmental quality enhanced subjective wellbeing in G20 countries in the years 2006–2019 and non-renewable energy consumption degraded subjective wellbeing. Moreover, the study also found bidirectional causality between renewable energy consumption, non-renewable energy consumption and economic growth. The positive impact of renewable energy sources on economic growth and happiness was also confirmed by Afia [35]. Moreover, these studies have also confirmed the significant impact of macroeconomic variables, such as the inflation rate.

It should be emphasized that the results were different when divided into individual types of renewable energy. In the case of economic growth, a significant positive impact was demonstrated for solar and wind energy. In relation to other sources, this impact was negative, but insignificant. A positive impact of solar and wind energy on economic growth for 28 European Union countries in the years 2004–2013 was also obtained by Armeanu et al. [50]. The results of the research conducted by Li et al. [14] showed a positive impact of geothermal, hydro and wind energy on the economic growth of the South Asian Association for regional cooperation (SAARC) countries in the years 1995–2018. Moreover, hydropower had the highest impact compared to the others. The research results in the case of the hydro energy source are different from those obtained by other authors. For example, Apergis et al. [51] showed bidirectional positive relationships between hydropower and economic patterns in the years 1965–2012 for countries such as Brazil, Canada, China, France, India, Japan, Norway, Sweden, Turkey and the U.S.A. The positive impact of wind energy on GDP in 23 developing countries in 2004–2016 was demonstrated by Doğan and Doğan [52]. Similar results were also obtained by Koç and Apaydın [53] for the G-20 group of countries for the years 1991–2017.

The only type of renewable energy that showed a negative impact on happiness was wind energy. The reasons for this may be various. It generates many external effects, including the destruction of soil, birds and the landscape. Windmills are often built near homes, which negatively affects life satisfaction [54]. Onakpoya et al. [55] reported negative health effects due to noise from wind turbines. Krekel and Zerrahn [56] reached less radical conclusions: the construction of wind turbines close to households exerts significant negative external effects on residential well-being, although they seem both spatially and temporally limited, being restricted to about 4000 m around households and decaying after five years at the latest.

A certain limitation in this research is the fact that the level of happiness, i.e., satisfaction, is a subjective measure. The influence of interest groups and lobbies cannot be ruled out. Li et al. [57] showed that Japan has been almost one-sidedly leaning towards the more expensive solar PV due to the solar lobby that comprises bureaucracies, politicians, solar PV manufacturers and independent power producers. Li et al. [57] pointed out that due to the higher costs of solar PV than wind electricity, Japan has been spending more economic resources than foreign countries in utilizing renewable energy. This conclusion has shown that the activities of interest groups may result in an economically suboptimal energy mix. On the other hand, the development of renewable energy sources in many European countries has encountered barriers [58]. The following barriers and shortcomings have been identified [58,59]: social, cultural, political and organizational; a lack of experience setting up communities and low trust in the community model; legal, administrative and bureaucratic; complicated legal frameworks, bureaucratic barriers to grid connections and microgrids operations; and a lack of authority support and a lack of RES support schemes. Furthermore, there may be technical barriers such as a lack of expert knowledge about energy communities. Financial barriers may exist as well, such as access to finance, unfair payments for the energy produced, weak incentives to use renewable energy in heating and unjust taxes. Challenges may also be found in existing communities, including the membership share value and investments into existing installations. These barriers particularly concern energy communities, which may be directly related to happiness levels.

Another issue that should be paid attention to is the regulations related to the implementation of renewable energy sources. The share of hydropower varied in the surveyed countries (see Figure 2) and depended primarily on geographical conditions. It is a reliable source of renewable energy; however, the legislation is very rigorous. For example, ensuring the protection and safe passage of migratory fish is required. The costs associated with environmental mitigations represent a large proportion of the total costs required for the licensing of hydropower facilities [60]. Hydropower development should not endanger the social environment, cultural heritage or local communities and, therefore, the principles of sustainable development should be observed and the environmental protection and social development interests should be harmonized by ensuring a clean and healthy environment, the effective use of natural resources as well as social security [61]. The impact of dams on the environment, society and economy is multi-faceted. Reservoirs help to provide water not only for hydropower, but also for agriculture and enabling flood management [62].

5. Conclusions

This study analyzed data from 2012 to 2022 for 25 European Union countries to investigate the impact of renewable energy shares on economic growth and the happiness index using panel data models. In general, the results of the estimated models indicate that there is a significant and positive impact of the renewable energy share on both direct variables. This means that an increase in the share of renewable energy in the energy mix by 1%, on average, contributes to economic growth from 7940.99 in model 1 to 11,935.4 mln EUR in model 8 and to the happiness index from 0.014 in model 2 to 0.035 in model 7. Various relationships were noted by specific types of renewable energy sources. In the case of GDP, a significant positive impact was recorded for solar and wind energy, while the remaining types were insignificant and negative. This means that an increase in the share of solar energy by 1% in the energy mix, on average, will contribute to economic growth from EUR 39,064.5 mln in model 2 to EUR 50,643.7 mln in model 7. An increase in the share of wind energy by 1% in the energy mix, on average, will contribute to economic growth from EUR 19,718.9 mln in model 5 to EUR 21,881.5 mln in model 3. An increase in hydro energy by 1%, on average, will result in decreased GDP from EUR 3635.23 mln in model 4 to EUR 4265.41 mln in model 7. An increase in other energy source productions will decrease GDP from EUR 10,401.2 mln in model 3 to EUR 12,203.8 mln in model 2. In relation to the happiness index, all renewable energy sources had a significant positive impact, except wind energy. An increase in biofuel energy in the mix of total primary energy by 1%, on average, will result in an increase in the happiness index from 0.147 in model 5 to 0.170 in model 3, hydro energy from 0.015 in model 2 to 0.023 in model 5, solar energy from 0.096 in model 3 to 0.105 in model 4 and other renewable energy from 0.066 in model 2 and 6 to 0.096 in model 1, 3, 4, 7 and 8. An increase in wind energy production will result in a reduction in the happiness index from 0.033 in model 2 to 0.045 in model 6. However, this is justified by many external effects generated by wind turbines, which adversely affect, for example, the landscape.

Based on the research conducted, several recommendations can be made. European Union countries should increase renewable energy resources in the energy mix, which will increase the chance of accelerating economic growth and increasing the sense of happiness among citizens. Actions should be taken at both the macroeconomic and microeconomic levels. From the EU perspective, it is important to implement transparent policies appropriate to a given country. In this respect, a coherent strategy should be developed, taking into account the conditions of individual countries, including their geographical location and the costs they are able to incur. It should be emphasized that the implementation of renewable energy sources depends not only on the willingness of the authorities of a given country, but also on access to financing. In this respect, the financial sector plays an important role, including primarily banking institutions, whose main activity is to provide loans. Banks are obliged to implement the principles of green banking, the main goal of which is to grant loans to entities that demonstrate the application of the principles of green governance in their operation, as well as to expand the banking offer with green banking products. Education about renewable energy sources and the benefits they can bring is also important for bank clients and citizens.