The Relationship between Geothermal Energy Consumption, Foreign Direct Investment, and Economic Growth in Geothermal Consumer Countries: Evidence from Panel Fourier Causality Test

Abstract

This paper investigates the relationship between geothermal energy consumption, economic growth, and foreign direct investments in countries where geothermal energy production is possible. Panel Fourier Granger causality and panel Fourier Toda–Yamamoto causality tests (2020–2021) were applied, which can take into account smooth transitional structural breaks with trigonometric functions using quarterly data for the period 2016 Q1–2020 Q3. Data were obtained from the International Energy Agency (IEA), Federal Reserve Economic Data (FRED), and the OECD official website. According to the results obtained based on panels, there is one-way causality from economic growth to geothermal energy and one-way causality from geothermal energy consumption to foreign direct investments. The results obtained based on individual countries indicate that one-way causality from foreign direct investment to geothermal energy consumption was found for Mexico and Portugal, and one-way causality from geothermal energy consumption to economic growth was found for Italy and Mexico. On the other hand, causality from economic growth to geothermal energy consumption was observed for Germany, Japan, and USA. No significant results were found for Turkey and New Zealand, and it is understood that the macroeconomic structures of these countries are not affected by geothermal energy. The difference in the results reveals that the application recommendations on this subject should also be different.

1. Introduction

Energy has been one of the basic building blocks of economic and social development in countries, and it continues to increase its importance in our 21st century. In addition to its role as an indispensable element of production and consumption activities, energy also plays an important role as a “strategic commodity” in the context of international competition. Thus, the energy consumption and energy supply processing conditions of the countries are an indicator of economic growth [1]. Economic activities prompted the energy needs of countries that obtain momentum on a global scale with industrialization initiatives. In the process of meeting these needs, the limited life of fossil energy resources (coal, natural gas, oil, etc.) and the negative environmental effects of greenhouse gas emissions cause concerns in economies [2]). The price uncertainties caused by these problems in fossil energy resources have made it necessary for countries to turn to renewable energy sources in order to meet their energy needs.

Renewable energy is an important type of energy provided from natural resources (solar, wind, biomass, hydroelectric, geothermal, etc.) that do not have a limited life and reserves, and that help to protect the environment by reducing greenhouse gas emissions, in comparison to fossil energy sources. Renewable energy sources possess advantages such as offering the opportunity to diversify fuel mixtures, reducing environmental and air pollution, active use in ensuring sustainable economic development, and solving the problem of access to energy [3]). Reducing foreign dependency in economies, due to the supply of renewable energy from domestic sources, also reveals the importance of energy in terms of economic growth [4].

In terms of renewable energy sources, solar energy and wind energy stand out with their different advantages. The advantages of solar energy include protecting natural energy resources and increasing the quality of these resources, reducing water consumption and carbon dioxide emissions, and preventing environmental pollution [5].

Among the advantages of wind energy are its low cost, its increasing prevalence, and the fact that wind turbines will not increase acid rain or greenhouse gas.

Another type of renewable energy is biomass energy. Biomass energy is a type of carbon-neutral energy that is stored by plants through solar absorption, obtained from plants as well as animals, which also helps to reduce climate change [6,7].

Hydroelectric energy is the provision of electricity by utilizing the movement of water [8]. The advantages of this type of energy, which is accepted as a primary energy source, include the fact that it is safe and efficient, that hydroelectricity does not deplete the water supply, requires a low investment of capital compared to other energies, and that hydroelectric facilities create employment for the country.

According to the BP Statistical Review [9], the production volumes of energy types are as follows: hydroelectric energy (4297 terawatt-hours), wind energy (1591 terawatt-hours), and solar energy (856 terawatt-hours) (BP). These statistics are provided in Figure 1:

Geothermal energy, which is also the subject of our paper, is, generally, the energy obtained from underground water and steam. Geothermal energy is obtained by the slow cooling of underground magma, which can reach 6650 °C (12,000 °F), and, partly, by the decay of radioactive elements in rock minerals [10,11]. The use of geothermal energy dates back to ancient times. The ancient Romans were the first to use geothermal energy [12]. According to the BP Statistical Review of Global Energy [13], the countries that produce the most geothermal energy in the world are as follows: USA (2587.00 MW), Indonesia (2131.00 MW), Philippines (1928.00 MW), Turkey (1613.00 MW), and New Zealand (984.00 MW). A graph of this information is shown in Figure 2.

Geothermal energy was first used in USA for residential heating in 1891. In study conducted in Italy in 1904, electrical energy was produced from geothermal steam. In 1969, a large city in France started to be heated entirely with geothermal energy [14]. This energy stored underground is clean, emission-free, recyclable, sustainable, and economical compared to fossil energy sources [15]. It creates “base-load generation” because geothermal energy is not affected by weather conditions. Due to this feature, the continuity of production is ensured, and it differs from other technologies that provide power or heat [10]. Geothermal energy and activities involved in supporting the industry increase the level of employment in economies more than other renewable energy sources [16].

There are many papers dealing with the relationship between renewable energy and economic growth. The first papers to address the relationship between renewable energy consumption and economic growth determined that per capita GDP increased renewable energy consumption in developing countries for the 1994–2003 term and in G7 countries for the 1980–2005 term, particularly in the USA [17,18]. In the following processes, the relationship between energy consumption and economic growth was investigated by application of the neutrality hypothesis (no causality), conservation hypothesis, growth hypothesis, and feedback hypothesis [19].

Another factor related to renewable energy is the increasing population and industrialization levels of countries [20,21,22,23].

Many motivational factors are revealed throughout the course of this paper. The issues that can be evaluated in this context can be grouped under three headings: subject, method, and data frequency. Compared to fossil energy sources, renewable energy is important in terms of both creating energy security and protecting the environment. Due to intense interest in recent years, the relationship between geothermal energy consumption, foreign direct investments, and economic growth have been investigated in this paper. An important contribution of the paper is to expand the literature on foreign direct investment and economic growth. In the studies in the literature, the effect of renewable energy, in general, or its types (solar, wind, biomass, and hydroelectric) on economic growth has been discussed, but the absence of any study considering geothermal energy among renewable energy types provided one of the motivations of this paper. In the economic balance, the significance of energy is great, and geothermal energy provides renewable, low-cost, reliable, zero-emission energy which can be utilized over a wide area, among other positive benefits. The effect on foreign direct investments and economic growth due to the shift towards geothermal energy consumption by certain countries has been a matter of curiosity.

On the other hand, another motivating factor for the paper is that the results were obtained by using Fourier-based panel causality methods, which are quite new in the literature. These methods take into account smooth transitional structural breaks. The results obtained in this manner will add a new dimension to the literature in terms of the relationships between energy consumption and macroeconomic indicators. Another important aspect of this paper is that the data used are quarterly and by this means, short-term changes will be seen in more detail.

Based on the increasing demand and importance of geothermal energy, the main research goal of this study is to determine whether there is a relationship between geothermal energy consumption, foreign direct investments, and economic growth. For this reason, the panel Fourier Granger causality and panel Fourier Toda–Yamamoto causality tests were applied. The advantage of these tests is that they can take into account smooth transitional structural breaks using trigonometric functions. We considered the time period of 2016 Q1–2020 Q3.

Following this introductory section, in which the terminology of the paper is explained, the literature related to the subject is reviewed. In the methodology, the data included in the research and the economic method are specified. Following the analysis, results are presented in the Findings section and a general evaluation is presented in the Conclusion section.

2. Literature Review

There are many papers in the literature investigating the relationship between renewable energy consumption and economic indicators. These include papers that focus on economic growth, foreign direct investments, sustainability, and foreign trade indicators of total renewable energy consumption. A number of these researchers only examined a particular type of renewable energy. In addition, it was observed that most of the studies were prepared empirically, and time series and panel data analysis were used extensively. The general findings concern the relationship between renewable energy consumption and economic indicators.

Various articles investigated the relationship between renewable energy consumption and FDI. The importance of the subject stems from FDI’s relationship with efficiency sequences, management skills, new technology transfers, the introduction of the new process, management costs, technical know-how, and international production networks [24]. [25] for the 1995–2010 term in 22 Central and South American countries, and [26], for the 1980–2015 term in Bangladesh, found bidirectional causality between FDI and renewable energy consumption. The same conclusion—that one-way causality exists between the FDI and renewable energy consumption—was reached by [27] for Denmark, Finland, France, India, Italy, Morocco, Norway, Portugal, and Sweden, by [28] for Bangladesh in the periods 1980–2016, and by [29] for 21 countries in the periods 2002–2010. The following researchers demonstrated a positive effect of FDI on renewable energy consumption. [30] in the 2004–2019 term, for Indonesia; [31], in the 1990–2015 term, for 28 European Union member countries; [32], in the 1992–2016 term, for the Next Eleven countries; [33] in the 1990–2012 term, for Brazil and China; and [34] in the 1985–2012 term, for 74 countries. On the other hand, the following researchers showed a negative effect on renewable energy consumption by FDI: [35] in the 1990–2019 term, for South Asian countries [36] in the 1981–2014 term, for Nigeria. In addition, [37] found FDI to have both positive and negative effects on renewable energy consumption in the 1995–2014 term for 32 European countries. Finally, the following researchers reached the conclusion that FDI had no effect on renewable energy consumption: [38] in the 1989–2019 term, for the UAE; [39] in the 2002–2014 term, for 85 developing and developed countries; [32] in the 1992–2016 term, for the BRICS countries; and [40] in the 1971–2009 term, for G20 countries. It is possible to conclude that results of these papers were mixed.

Authors investigating the relationship between renewable energy consumption and economic growth who found evidence to support the growth hypothesis include the following: [41] for the 1990–2015 term in OECD and non-OECD countries; [42] for the 1990–2010 term in OIC countries; [43] for the 1991–2012 term of the Renewable Energy Country Attractiveness Index [44] for the 1990–2018 term in renewable energy-consuming countries; [45] for the 1995–2015 term in the European Union countries; [46] for the 1970–2007 in Italy; [47] (2013), for the 1990–2010 term in the OECD countries; and [48] for the 1996–2012 term, regarding the different income levels of various countries. The following authors found evidence to support the conservation hypothesis: [49] for the 1955–1956 and 1995–1996 terms in Pakistan; [50] for the 1990–2010 term in Turkey; and [51] for the 1990–2011 term in sub-Saharan African countries. The following researchers confirmed the feedback hypothesis: [52] for the 1971–2011 term in African countries that belong to OPEC; [53] for the 1985–2005 term in the OECD countries; [54] for the 1990–2014 term in the European Union member states in Central and Eastern Europe; [55] for the 1995–2012 term in the OECD countries; [56] for the United States (1965–2019), Germany (1971–2019), Japan (1970–2019), and Brazil (1970–2019); and [57] for the 1971–2002 term, for net energy importers and exporters. No study in the literature argues that there is no relationship between economic growth and renewable energy consumption.

Regarding studies examining the relationship between FDI and renewable energy types, Ref. [58] evaluated the relationship between solar energy consumption and FDI, for the 1990–2018 term, in the top 10 solar energy-consuming countries. The study concluded that increasing the inflow of FDI will provide incentives for solar energy technology. Ref. [59] investigated the determinants of FDI. At the end of the study, they determined that solar and wind energy investments play an important role. Ref. [60] examined the role of attracting FDI in wind energy for the 2004–2010 term in India. The results remarked that wind energy is significant in attracting FDI. Ref. [61] demonstrated that FDI initiatives increased the use of wind energy in developing countries [62] found that FDI initiatives increased the use of renewable electricity, using data for the 1972–2015 term in Bangladesh. Ref. [63] found a causality between biomass energy consumption and FDI, for the 1970–2017 term, in selected high-income, upper-middle-income, and middle-income countries. Ref. [64] investigated the relationship between FDI and electric power consumption for the 1980–2011 term in India and Pakistan. The results revealed long-run causality between FDI and consumption. In the literature, no studies were found researching geothermal energy consumption and FDI.

When the studies are evaluated in terms of the particular renewable energy types studied, Ref. [65] examined the relationship between economic growth and solar energy consumption, analyzing the 1990–2019 period in 19 countries. As a result of the analysis, a positive relationship was found between economic growth and solar energy consumption. Ref. [66] investigated the relationship between wind energy consumption and economic growth. They examined developing countries in the period of 2004–2016, concluding that increased use of wind energy has a positive and significant effect on countries’ economic growth. Considering biomass energy consumption, Ref. [67] analyzed the 1990–2011 terms in Estonia, Belarus, Latvia, Georgia, Lithuania, and Moldova. According to these findings, biomass energy consumption has a positive effect on economic growth. Ref. [68] found that the conservation hypothesis is valid for Austria, Germany, Finland, and Portugal and the feedback hypothesis is valid for USA from 1970 to 2013. Ref. [69] determined that the conservation hypothesis is valid for Canada, France, the UK, and USA in terms of short-run causality. The author reached the same conclusions regarding the growth hypothesis for Australia, Belgium, Finland, and Japan, and regarding the feedback hypothesis for USA, UK, and France in the 1980–2010 period. Ref. [70] attempted an examination of the BRICS countries for the 1991–2015 period, and concluded that a bidirectional relation exists between biomass energy consumption and economic growth. Ref. [71] revealed that feedback effects exist between biomass energy consumption and economic growth in Mali, Togo, Nigeria, The Gambia, and Burkina Faso. Ref. [72] studied the relationship between economic growth and biomass energy consumption for the 1980–2013 term in OECD countries. The results revealed bidirectional causality between economic growth and biomass energy consumption. Ref. [73] investigated the relationship between hydroelectricity consumption and economic growth in the 1985–2014 period in Indonesia, concluding that two-way causality occurs between the two. Ref. [74] studied the economic growth and hydroelectricity consumption nexus in the 1965–2012 period in Canada, Brazil, China, France, Japan, India, Sweden, Norway, the USA, and Turkey. The results supported the feedback hypothesis. Ref. [75] investigated the nexus between hydroelectricity consumption and economic growth in the 1966–2015 period in Latin American countries. The results suggested the existence of a feedback hypothesis. Ref. [76] examined the relationship between hydroelectricity consumption and real GDP. In the study, which analyzed the 1965–2013 period data of China and India, a long-run bidirectional relationship was found in both countries between hydroelectricity consumption and real GDP. Ref. [77] examined the relationship between hydroelectricity and economic growth in the 1980–2009 period in African countries. The results of the study supported the neutrality hypothesis for Egypt, the feedback hypothesis for Algeria, and the conservation hypothesis for South Africa. The relationship between economic growth and geothermal energy consumption has not been studied. As opposed to other researchers, Ref. [78] adopted a descriptive and inferential survey design. They surveyed 60 senior executives in Kenya. The study revealed that there is a relationship between FDI variables such as infrastructure, technology diffusion, and trade.

A general evaluation of the studies that consider the relationship between economic growth and renewable energy consumption shows that while the effect of renewable energy consumption on economic growth is limited in the short term, it has an intense effect in the long term. On the other hand, no study has been found that deals with the relationship of geothermal energy with either FDI or GDP, which indicates the originality of this paper. Studies show that the importance of renewable energy types in macroeconomic balances is increasing. In this context, studies on renewable energy types such as solar, hydropower, and wind reveal the necessity of investigating the relationship between geothermal energy and macroeconomic factors. The other original characteristic of this paper is the sample that it used. As far as we were able to determine, no other study has used this sample.

3. Data and Econometric Method

Quarterly data for the period 2016 Q1–2020 Q3 were used in the paper. The dataset covers Germany, Iceland, Italy, Japan Mexico, New Zealand, Portugal, Turkey, and the USA. The reason for choosing these countries is related to the fact that not all countries possess the same characteristics, in terms of natural resource infrastructure, that would permit the use of geothermal energy. In this respect, even if geothermal energy is desired to be produced, in some countries this is not possible due to a lack of infrastructure. Accordingly, the dataset constitutes a certain constraint. Geothermal energy consumption data were obtained from the IEA’s official website. Economic growth data were obtained from the FRED, and foreign direct investment data were obtained from the OECD official website. The fact that no previous studies using quarterly data have been found emphasizes the originality of this paper. The dataset and descriptive statistics used in the paper are detailed in

Table 1. Information on the dataset of the paper.

Variables	Unit of Measurement	References
Geothermal Energy Consumption	Megawatt (MW)	-
FDI	FDI, net (BoP, current USD)	[25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,58,59,60,62,63].
Economic Growth	Normalized (GDP) per capita	[41,43,44,45,46,47,48,49,50,51,53,54,55,56,57,65,66,67,68,69,70,71,72,74,75,76,77].

and

Table 2. Descriptive statistics.

GEOTHERMAL ENERGY CONSUMPTION
	Germany	Italy	Japan	Mexico	New Zealand	Portugal	Turkey	USA
Mean	39.72	1437.6	585.5	1344.7	1896.35	46.82	1476.7	3939.2
Maximum	54.19	1488.5	670.6	1472.6	1986.08	52.79	2461.9	4197.9
Minimum	29.74	1380	513.2	1155.2	1731.99	26.84	949.5	3498.6
Std. dev.	6.98	41.89	41.89	97.69	73.71	6.30	389.63	161.28
ECONOMIC GROWTH
	Germany	Italy	Japan	Mexico	New Zealand	Portugal	Turkey	USA
Mean	298.41	297.69	299.77	298.43	298.37	297.81	298.94	298.47
Maximum	303.43	302.98	303.75	302.9	301.96	301.81	304.81	302.06
Minimum	273.19	260.89	281.32	270.90	282	267.52	285	277.69
Std. dev.	7.25	10.41	5.41	8.08	5.61	8.76	4.35	5.81
FOREIGN DIRECT INVESTMENT
	Germany	Italy	Japan	Mexico	New Zealand	Portugal	Turkey	USA
Mean	20,265.3	5141	40,460	1515	43.91	392.29	747.62	42,729
Maximum	62,163	10,640	90,740	4887	645.31	3490.4	1109.1	147,026
Minimum	−23,476	−4257	12,259	−2055	−837.9	−1373	103.7	−80,083
Std. dev.	25,260	4172.6	16,132	1866.34	368.87	1103.73	246.36	64,222

respectively; the section on research on geothermal energy consumption in could not be filled because there are no studies on this sample in the literature.

Before performing the cointegration and causality tests, which present the main results of the panel data analysis, the following diagnostic tests are required: homogeneity, unit root, and cross-section dependency. Cross-sectional dependence reveals cross-sectional relationships. In the case where cross-sectional dependence is found in the panels, second-generation unit root, cointegration, and causality tests should be performed in the following stages; in the absence of cross-sectional dependence in the panels, first-generation methods should be used. Therefore, four horizontal fraction tests measuring cross-sectional dependence were found in the literature. This was carried out due to the case in which the time dimension and the cross-section size were larger than each other or if they were equal in number. In this paper, the CDLM test improved by [79] and the CDLM_adj test improved by [80] were used because the time dimension is larger than the cross-section dimension. The method developed by [80] is an improved version of the method improved by [79] to eliminate possible error term deviations and is measured as follows:

$$L{M}_{adj}=\sqrt{\frac{2}{N\left(N-1\right)}}\left({\displaystyle \sum }_{i=1}^{N-1}{\displaystyle \sum }_{j=i+1}^{N}T{\widehat{P}}_{ij}\frac{\left(T-k\right){\widehat{P}}_{ij}^2-{\mu }_{Tij}}{\sqrt{{\nu }_{Tij}^2}}\right)$$

(1)

$$CDL{M}_{adj}={\left(\frac{2}{N\left(N-1\right)}\right)}^{1/2}\left({\displaystyle \sum }_{i=1}^{N-1}{\displaystyle \sum }_{j=i+1}^N{\overline{\rho }}_{i{j}^2}\frac{\left(T-k-1\right)Pi{j}^2-{\widehat{\mu }}_{Tij}}{V{T}_{ij}}N\left(0,1\right)\right)$$

(2)

In the equation, ρ is the binary sample estimate of the residues. The assumption states that the sections possess a correlation with each other and have a chi-square asymptotic distribution with 𝑇 → ∞ 𝑖𝑘𝑒𝑛 𝑁 if constant (𝑁−1) possesses two degrees of freedom. It also states that the time dimension will be used because the cross-section of T data is larger than N.

There are some second-generation panel unit root tests, such as SURADF, CADF, and Hadri–Kurozumi, that can be used in a case of cross-sectional dependence in the panels. One of these tests, the CADF panel unit root test, can be used in both T > N and N > T cases. The CADF test statistics obtained as a result of the analysis are compared with the critical values and the decision is made. The null hypothesis of the method is that the panels contain a unit root, and the alternative hypothesis is that the panels possess a stationary structure. The CADF test statistic is calculated as follows:

$$Y_{it}=\left(1-{\varnothing }_i\right){\mu }_i+{\varnothing }_i{Y}_{i,t-1}+u_{i,t}$$

$$i=1,2,\dots \dots ,N$$

$$t=1,2,\dots \dots ,T$$

where $u_{\mathit{it}}$ is standard error:

$$u_{it}=y_i{f}_t+{\epsilon }_{it}$$

and ${\epsilon }_{\mathit{it}}$ determines the individual error.

$$\Delta y_{it}=a_i+{\beta }_i{y}_{i,t-1}+y_i{f}_i+{\epsilon }_{it}$$

$$H_0:{\beta }_i=0 for all i$$

$$H_1:{\beta }_i<0$$

$$i=1,2,\dots \dots ,N_1$$

$${\beta }_i=0,$$

$$i=N_1+1.N_1+2\dots \dots N$$

CIPS (cross-sectionally augmented IPS) statistics calculated according to this method are obtained by averaging the unit root test statistics of each cross-section [81]:

$$CIPS=N^{-1}{\displaystyle \sum }_{t=1}^{N}CAD{F}_t$$

After determining the cross-section dependency and stationarity levels in the panels, the next step is to determine cointegration and causality. In this paper, direct causality methods were not used because of the presence of stationarity at different levels, and because the panel ARDL test in Fourier form suitable for this situation has not yet been derived in the literature. The use of Fourier forms for causality coincides with the Fourier causality test of [82] for time series. On the other hand, Fourier forms for panel cointegration and causality were applied in 2020 and 2021 in studies by the following authors: [83,84,85]. According to the aforementioned methods, the Fourier form of the traditional [86], panel causality test is used when the panels are stationary at the same level, and the Fourier form of the [87] panel causality test is used when the panels are stationary at different levels. These methods are based on the [88,89] tests used for time series.

As in all Fourier form methods, soft transition structural breaks are detected in these causality tests, and the results obtained will be more reliable. The following mathematical modeling is used to calculate test statistics for these methods:

$${\tilde{v}}_{i, t}={\widehat{v}}_{i,t}-{\widehat{\alpha }}_i-{\widehat{\omega }}_i\sin (\frac{2\pi kt}{T})+{\widehat{\phi }}_i\cos (\frac{2\pi kt}{T})+v_t$$

4. Empirical Findings

The first stage of panel data analysis consists of investigating whether there is a relationship between the horizontal and vertical sections of the panels. With this goal, the cross-section dependency test investigates the relationship between horizontal sections, while the homogeneity/heterogeneity tests investigate the relationship between vertical sections. It is possible to benefit from four different test statistics, depending on whether the time dimension is larger than the cross-section dimension or not, regarding the determination of the cross-sectional dependence. If the time dimension is larger than the slice size, it is possible to use the CDLM test improved by [79] and the corrected CDLM test improved by [81]. In this context, the results of each method are presented in

Table 3. Cross-section dependency test results.

	CDLM	CDLMADJ
Geothermal Energy Consumption	42.254 (0.041) **	4.826 (0.000) ***
Foreign Direct Investment	58.580 (0.001) ***	5.926 (0.000) ***
Economic Growth	91.881 (0.000) ***	7.601 (0.000) ***

Note: ** and *** are the representation of p values less than 0.05 and 0.01 respectively.

. Although the mathematical backgrounds of the methods are different from each other, the results obtained support each other because, while the CDLM test does not take into account the probability of the errors that are not distributed normally, the CDLM_adj test obtains test statistics by taking this possible deviation into account.

At this stage of the paper, it was determined that use of second-generation unit root tests would be appropriate because of the cross-section dependence of the panels. In this context, the CADF panel unit root test improved [80] was applied to the panels examined within the scope of the research. When the results obtained were examined, it was determined that the geothermal energy consumption and economic growth panels have a unit root at the level, and the foreign direct investments panel is stationary at its level. The first differences in the economic growth and geothermal energy consumption panels were subjected to the unit root test again, and we concluded that they became stationary in the first difference values (

Table 4. CADF panel unit root test results.

	Level	First Diff.
Geothermal Energy Consumption	−2.71	−3.129 ***
Foreign Direct Investment	−3.797 ***	-
Economic Growth	−1.796	−3.377 ***

Note: *** Critical values are at 1% confidence levels. These values were obtained [81].

).

After determining the stationarity levels of the panels, it is necessary to investigate the cointegration and causality relations between these panels. However, the panel cointegration test is not included in this paper because the FDI panel is stationary at its level and a second-generation panel cointegration test in Fourier form for different levels of stationarity has not yet been included in the literature.

In the panel causality stage, the panel causality test in the Fourier form was used, which shows the relationship between economic growth and geothermal energy consumption as the same (stationary level). The relationship between geothermal energy consumption and foreign direct investment panels, which are stationary at different levels, was examined by the panel Fourier Toda–Yamamoto causality test. The results obtained are presented in

Table 5. Panel Fourier causality and panel Fourier Toda–Yamamoto causality test results.

	Panel Fourier Toda–Yamamoto		Panel Fourier Granger
	GEO → FDI	FDI → GEO	GEO → EG	EG → GEO
Germany	0.05 (0.70)	0.18 (0.50)	0.005 (0.93)	4.59 (0.03) **
Italy	0.02 (0.85)	0.13 (0.70)	5.15 (0.02) **	0.06 (0.81)
Japan	0.47 (0.55)	1.38 (0.70)	0.77 (0.37)	7.26 (0.00) ***
Mexico	8.70 (0.00) ***	0.51 (0.55)	4.07 (0.03) **	0.43 (0.50)
New Zealand	1.37 (0.20)	0.12 (0.90)	0.61 (0.43)	0.19 (0.65)
Portugal	7.17 (0.04) **	0.01 (0.95)	0.00 (0.99)	0.36 (0.54)
Turkey	0.85 (0.35)	0.07 (0.95)	0.01 (0.95)	1.78 (0.18)
USA	0.17 (0.70)	0.01 (0.99)	0.83 (0.36)	4.35 (0.03) **
Panel Fisher	2.42 (0.09) *	6.48 (0.98)	1.43 (0.38)	2.38 (0.08) *

Note: *, ** and *** are the representation of p values less than 0.1, 0.05 and 0.01 respectively.

. When the results were examined, it was determined that there is no effect from foreign direct investments on geothermal energy consumption and that there is causality between geothermal consumption and foreign direct investments. The causality relationship in question was found to be valid for Mexico and Portugal. On the other hand, we concluded that unidirectional causality occurs between geothermal and economic growth. It was shown that the growth hypothesis is valid for Italy and Mexico; a causal relationship from geothermal to economic growth has been demonstrated for these countries. Unidirectional causality from economic growth to geothermal was demonstrated for Germany, Japan, and USA; thus, the conservation hypothesis is valid for these countries. For the remaining countries—New Zealand, Portugal, and Turkey—there was no causality found and it was determined that the neutrality hypothesis was valid for these countries. The results did not demonstrate the validity of the feedback hypothesis for any country.

5. Discussion

Ensuring continuity in the economy is directly proportional to the availability of energy resources, the basic inputs of production. Since non-renewable energy resources are not evenly distributed in the world, the economic growth potentials and rates of countries vary. Supply problems, price and quota changes, emission emissions, and significant reduction of resources from non-renewable energy sources have initiated increased research on new energy sources.

However, renewable energy sources that are constantly present in nature and available to be converted into energy have been discovered. Each economy possesses resources according to its renewable energy potential. As it has been determined that the use of renewable energy has positive effects on the economy, investments and incentives have started to shift in this direction. As renewable energy consumption increases, positive effects are seen in investments made in this field. Implementing and increasing incentives for the use of renewable energy will help to produce macroeconomic implications for the implementation of future policies. Properly directed incentives will succeed in attracting foreign direct investments and contribute to economic growth.

Among renewable energy sources, geothermal energy sources have succeeded in attracting foreign direct investments due to their initial cost, reliability, diversity of usage areas, and because they do not involve increased emissions. The entry of foreign direct investments into the economy provides a positive effect on the economy and produces economic growth. When the main input of production is the use of geothermal energy, problems such as dependency, access, security, and price are minimized and improvements in the GDP ratio occur; as a result, the residual energy item cost is significantly reduced.

This paper studies the relationship between geothermal energy use, foreign direct investments, and economic growth in Italy, Germany, Japan, Turkey, Mexico, Portugal, USA, and New Zealand. The results showed variations based on countries; this is thought to be related to the economic priorities of the countries. Germany, Japan, and USA currently hold the status of developed countries. Economic growth has encouraged the tendency to use existing energy. Here, economic growth has affected the use of geothermal energy and it has been seen to support the “Conservation Hypothesis”. Compared to the results obtained for other countries, this shows that developed and economically strong countries have a strong hand in providing high-cost renewable energy investments, such as investments in geothermal energy. It has been seen that Italy and Mexico support the “Growth Hypothesis” by providing a positive effect on economic growth by using rich geothermal energy resources. On the other hand, the “Neutrality Hypothesis” has been observed in New Zealand, Turkey, and Portugal, where economic growth and geothermal energy consumption do not affect each other. A country in which the “Feedback Hypothesis” is demonstrated could not be identified. Regarding Mexico and Portugal, it has been seen that foreign direct investments in these two countries affect geothermal energy consumption. The geothermal energy potential and the “Growth Hypothesis” for economic growth has been demonstrated in Mexico by directing foreign direct investment to this field. Therefore, the relationship between these three variables, which is the research subject of this study, has been partially confirmed. The results are in the line with the findings of previous studies such as those by [27,28,29,41,42,43,44,45,46,47,48,49,50,59,60,63,64,66,67,68,69,77].

6. Conclusions and Policy Recommendations

In the paper, the relationship between the geothermal energy consumption of geothermal energy-producing countries, economic growth, and foreign direct investments was studied. Of the world’s geothermal-producing countries, Italy, Germany, Mexico, Japan, New Zealand, Turkey, USA, and Portugal were included in the sample. The data for the period 2016 Q1–2020 Q3 were analyzed with the panel Fourier Toda–Yamamoto causality test.

In line with the findings obtained in the paper, it has been determined that geothermal energy consumption has an important role in attracting foreign direct investments in Mexico and Portugal and in promoting economic growth in Italy and Mexico. The countries’ orientation to geothermal energy sources assists their economic development by reducing their foreign dependency, thus reducing their deficits; in Germany, Japan, and USA it was seen that the capital increase resulting from the real income-increasing effect of economic growth also causes geothermal energy consumption.

Price uncertainties in energy and supply, and security problems in recent times have made it necessary for countries to reach renewable energy sources. In addition, many governments have declared that, within the scope of global warming and climate change action plans, turning to renewable energy sources instead of fossil fuels prevents environmental degradation. With the importance of this in mind, policymakers and governments in countries with geothermal resources should be aware of the fact that geothermal energy is an important indicator, both in the struggle against climate change and for the realization of sustainable development goals in the economy, and they should develop strategies that develop geothermal energy infrastructure. In particular, tax increases for fossil fuel use in the country, tax reductions for geothermal energy use, a reduction of procedures for the installation of geothermal energy systems, and the provision to investors of low-cost research and development loan options will be effective in intensifying the transition to geothermal energy use. As a result of these practices, foreign dependency on energy, which is one of the biggest obstacles to the development of countries, is reduced, reducing energy costs and improving other macroeconomic indicators.

In this paper, a model consisting of two basic macroeconomic indicators, FDI and economic growth, and their relationship with geothermal energy, has been established.

According to the results obtained, a one-way causality from economic growth to geothermal energy, and a one-way causality from geothermal energy consumption to foreign direct investments were observed. In the cases of Mexico and Portugal, on the other hand, this relationship is unidirectional, from foreign direct investment to geothermal energy consumption. Unidirectional causality from geothermal energy consumption to economic growth was observed for Italy and Mexico. In addition, causality from economic growth to geothermal energy consumption was found in Germany, Japan, and USA. No significant results were obtained for Turkey and New Zealand.

However, macroeconomic factors affecting the economies of countries are not limited to FDI and economic growth. In this regard, in future studies, the model used in the paper can be expanded with macroeconomic indicators such as financial development, unemployment, current account balance, exchange rate, and stock market and interest rates. In addition, the results that can be obtained with a model that includes energy variables, such as CO₂ emissions and oil prices, which have an important place in the world economy, and precious metals, such as gold, silver, and platinum, will point to important findings. However, if time series data with a longer frequency are obtained, it will be possible to reach conclusions about the individual dynamics of particular countries, in terms renewable energy. Future studies may wish to consider different country groups and add different variables to the model. As a result, future research, which will examine in detail the various types of renewable energy—whose importance is increasing day by day—will provide important information to readers, investors, and policymakers to understand the dynamics of renewable energy.