Energy-growth-globalization (EGG) nexus in N-11 countries.

This study investigates the correlations of economic growth, energy consumption, and globalization in selected countries. For this purpose, the study gatheres data between 1984 and 2013 from 11 countries, where a panel causality test was employed using Kónya's (2006) approach based on the Seemingly Unrelated Regressions (SUR) system. Subsequently, the Wald test is conducted with a critical bootstrap value per country, in which the calculations obtained significant statistical figures per country. The results of the two-way causality test between economic growth and energy consumption were found in Egypt, Indonesia, Iran, South Korea, Nigeria, and Turkey. The results of the two-way causality test between globalization and economic growth are known in Bangladesh, Egypt, Indonesia, Iran, South Korea, Mexico, Nigeria, Philippines. Meanwhile, two-way causality between energy consumption and globalization exists in Bangladesh, Egypt, Indonesia, Iran, South Korea, Nigeria, Pakistan, Philippines, and Turkey. Based on these findings, strengthening cooperation is needed so that economic growth, energy consumption, and globalization continue to grow.

Introduction

Economic growth is an essential part of macroeconomics. Each country continues to try to find ways to promote higher economic growth yearly. In the case of developing countries, they try their best to become strong economies like developed countries. This effort is carried out, and it is hoped that an effect will increase welfare for the country (Ikhsan et al., 2022). According to classical theory, economic growth is determined by capital and labor. When both inputs increase, economic growth also increases. The development of the latest theory by neo-classical incorporates elements of population and technology (Romer, 2012). Another factor that is considered an input for economic growth is energy (Destek and Aslan, 2017). The discussion about energy is an exciting thing among researchers. Each country is competing to explore the natural resources they have. Up to now, at least two energies have been identified: renewable energy (water, wind, geothermal, and biomass) and non-renewable energy (crude oil, gas, and coal). One of the reasons for the drive to diversify energy is the high demand for goods. Producers catch the signal as an opportunity to produce more goods by expanding new industries. Therefore, energy and economic growth are interconnected (Saidi et al., 2017). Many previous researchers continue to confirm these two elements and some of them find consensus results (Mirza and Kanwal, 2017; Saidi et al., 2017; Khan et al., 2021; Rahman, 2021), there is only one direction (Appiah, 2018; Pinzón, 2018), and it is not proven to exist (Chontanawat, 2020).

In today's modern era, demand for goods can occur for domestic and international needs. A country generally tries to produce its goods without interference from other countries. However, the existence of resources is a problem because every country has abundant natural resources but is limited in human resources, or vice versa. Therefore, opening the door to other countries through cooperation provides economic benefits, a concept known as globalization. This concept was introduced in the early 2000s. Globalization describes a shift in the economy, becoming integrated into a more comprehensive economy and society (Hill and Hult, 2019). Shahbaz et al. (2021) explain that globalization has an important role and is believed to be an object in encouraging energy consumption and stimulating economic growth.

Regarding globalization, economic observers group countries based on specific criteria to compete with developed countries. One is N-11 Country, Next Eleven, which Goldman Sach identified. These countries have lower-middle-income statuses, such as Bangladesh, Egypt, Indonesia, Iran, Mexico, Nigeria, Pakistan, Philippines, Turkey, South Korea, and Vietnam. These 11 countries are predicted to have the same economic opportunities as group countries such as Brazil, Russia, India, China, and South Africa (BRICS). In addition, the country is also estimated to have a financial advantage because of its stable macroeconomic and political potential (Chang and Fang, 2022).

The World Development Indicator (WDI) data by the World Bank released data on the economic growth of these 11 countries, which each year continues to experience a significant increase. Figure 1 shows that in 1984 the economic growth per capita was 1.8 trillion dollars, and 6.1 trillion in 2013. In 30 years, the increase in per capita economic growth reached 227 percent, with an average economic growth rate of 4.9 per year. Of course, developing countries are not the same as developed countries. The use of this country's energy is still massive with fossil energy. Figure 2 shows that energy consumption has increased from 29 million tons of oil in 1994 to its peak in 2005 of 58 million tons of oil. However, there was a decrease until 2013 to 27 million tons of oil. This decline was due to unstable oil prices and the shift from fossil to renewable energy. At the same time, the globalization index of 11 countries experienced a reasonably high jump from 37.72 points in 1983 to 62.20 points in 2013 (an increase of 64 percent) with an average of 1.74 percent per year (Figure 3). The higher globalization shows, the more significant the role of cooperation between countries towards other countries. Indirectly, every country began to take advantage of globalization as the main gate to encourage the economy. Based on the development of the three data, there are exciting things where the patterns of movement of economic growth, energy consumption, and globalization tend to be the same. When economic growth increases, energy consumption, and globalization also increase.

Figure 1

Real GDP development in N-11 countries 1984–2013. Source: World development indicator (WDI).

Figure 2

Development of energy consumption in N-11 1984–2013. Source: World development indicator (WDI).

Figure 3

Development of globalization in next eleven countries 1984–2013. Source: KOF swiss economic institute.

Most of the literature is very consistent in finding that economic growth and energy consumption are interrelated, so it is natural for this pattern to form. However, the assumptions used are only two variables. The era of globalization provides opportunities for every country to open trade gates to other countries in the form of cooperation. The goal is to get access to resources (i.e, oil, technology, capital) from a country intended to increase economic growth. Based on this situation, the problem of this research is whether economic growth, energy consumption, and globalization have a causal relationship?

This research can be considered a development of previous literature in the field of energy economics. However, in this study, some things are different from the existing literature, among others. First, to our knowledge, most of the research literature presents the relationship between economic growth and energy consumption in particular countries but has little application in panel data, especially N-11. Anticipate that our study will significantly contribute to the literature in filling this gap. Second, the combination of variable relationships that we use, including elements of globalization, is scarce, especially the relationship between globalization and energy consumption and globalization on economic growth. So this finding is the first introduction to the energy-growth-globalization (EGG) nexus. Third, this study uses Konya's (2006) bootstrap panel Granger causality method in the trivariate system of equations. Fourth, unlike the previous method, it only answers one answer so that its implementation applies to all so that the presentation of results is more specific for each country. Fifth, we employed the Cross-sectional Dependence (CSD) method to determine the presence or absence of linkages between countries. Additionally, we used Pesaran and Yamagata's (2008) study to test for heterogeneity between the countries.

The last part of this introduction is the structure of the writing, with Section 2 describes literature of variables. Meanwhile, Section 3 discusses the data and strategies used, and Section 4 describes the test results, ending with the analysis results and policy implications in Section 5.

Literature review

Linkage between energy consumption and economic growth

Studies on the relationship between energy consumption and economic growth have had their charm in recent years. According to the literature of Ozturk (2010), four hypotheses test the relationship between the two variables. The first is the growth hypothesis, in which energy consumption presents a one-way relationship with economic growth. An indication of this hypothesis explains that an increase in energy consumption will increase economic growth. Second is the conservative hypothesis that economic growth is related to energy consumption. This idea explains that increasing energy consumption is highly dependent on economic growth. The last element is the feedback hypothesis, which describes the effect of energy consumption on economic growth. The final hypothesis is neutral, indicating no relationship between the two factors. The four hypotheses are strengthened by previous studies (Table 1), where various methods and different samples obtained the results of these hypotheses. Apergis and Tang (2013), Destek and Aslan (2017), Ozcan and Ari (2015) and Appiah (2018) found the growth hypothesis findings. Ozcan and Ari (2015), Destek and Aslan (2017),Pinzón (2018) and Munir et al. (2020) found the conservative hypothesis findings. While the feedback findings were found in the study of Apergis and Payne (2010), Destek and Aslan (2017), Gorus and Aydin (2019), Khan et al. (2021), Rahman (2021), and Ikhsan et al. (2022). Meanwhile, the neutral hypothesis was found by Sari and Soytas (2009), Ozturk and Acaravci (2011).

Table 1

Selected literature energy consumption (EC) and economic growth (GDP).

Author	Period	Country	Method	Findings
Sari and Soytas (2009)	1971–2002	OPEC Countries	Granger causality	EC = GDP
Apergis and Payne (2010)	1992–2004	11 Common welath	Panel causality	EC↔GDP
Ozturk and Acaravci (2011)	1971–2006	11 Mena	ARDL-ECM	EC = GDP
Apergis and Tang (2013)	1975–2007	85 countries	Toda-Yamamoto	EC→GDP
Ozcan and Ari (2015)	1980–2012	15 OECD	Bootstrap causality	EC→GDP (UK, Finland, MexicoEC←GDP (US, Czech)
Destek and Aslan (2017)	1980–2012	17 emerging ecnomies	Panel bootstrap causality	EC↔GDP (Turkey)EC→GDP (China, Colombia, Mexico, Philippines)EC←GDP (Egypt, Peru, Portugal)
Mirza and Kanwal (2017)	1971–2009	Pakistan	VECM Granger	EC↔GDP
Saidi et al. (2017)	1990–2014	53 countries	VECM Granger	EC↔GDP
Appiah (2018)	1960–2015	Ghana	Toda-Yamamoto	EC→GDP
Pinzón (2018)	1970–2015	Ecuador	VAR	EC←GDP
Gorus and Aydin (2019)	1975–2014	MENA	Frequency Domains	EC↔GDP (FD)
Ozcan and Ozturk (2019)	1990–2016	Emerging	Bootstrap panel causality	EC→GDP (Poland)
Chontanawat (2020)	1971–2015	ASEAN	VECM Granger	EC = GDP
Munir et al. (2020)	1980–2016	ASEAN-5	Dumitrescu-Hurlin	EC←GDP
Khan et al. (2021)	1972–2017	South Asian	Dumitrescu-Hurlin	EC↔GDP
Rahman (2021)	1990–2017	BRICS & ASEAN	Dumitrescu-Hurlin	EC↔GDP
Ikhsan et al. (2022)	1971–2014	Indonesia	Asymmetric Bootstrap causality	EC↔GDP (negative direct)

Note: → unidirectional ( ), ← unidirectional ( ), ↔ bidirectional, = neutral.

Linkage between globalization and economic growth

Globalization has a critical role in economic growth besides energy consumption, albeit the literature on these studies is limited (Table 2). Globalization provides more space for countries to open cooperation with other countries. Of course, globalization is not the same as open trade which is limited to the definition of commodity value only. More than that, the role of globalization is very important in the economy such as cooperation in improving technology, capital, or human resources so that countries bind each other in the long term. If this limits globalization, the country will only rely on local resources and the economy will stagnate. Kilic (2015), using the Dumitrescu-Hurlin method in developing countries, reveals two-way results vis--vis globalization and economic growth. Meanwhile, Olimpia & Stela (2017) found a one-way relationship with the paired Granger method in Romania. This finding is in line with the study of Yang et al. (2021), focusing on the countries with the most significant health expenditures. Etokakpan et al. (2020) also found a neutral relationship in Malaysia.

Table 2

Selected literature globalization (GLO) and economic growth (GDP).

Author	Period	Country	Method	Finding
Kilic (2015)	1981–2011	Developing countries	Dumitrescu-Hurlin	GLO↔GDP
Olimpia and Stela (2017)	1990–2013	Romania	Pairwise Granger	GLO→GDP
Etokakpan et al. (2020)	1980–2014	Malaysia	Granger causality	GLO = GDP
Yang et al. (2021)	1995–2018	High healthcare expenditure countries	Dumitrescu-Hurlin	GLO→GDP

Note: → unidirectional ( ), ↔ bidirectional, = neutral.

Linkage between globalization and energy consumption

Globalization and energy consumption have contributed to developing energy usage increasing domestic oil imports and exports to foreign countries. Consequently, countries are less concerned about their natural resource limitations or the status of net oil-importing countries. For countries producing net oil exports, globalization is an advantage for trade. Therefore, in terms of energy consumption, the two statuses of these countries are mutually beneficial. Just like the literature on globalization and economic growth, this literature is also still rare. In Table 3, the first study of this study was conducted by Shahbaz et al. (2018) using the Dumitrescu-Hurlin method and found a two-way relationship in 25 developed countries, and this finding is reinforced by Fahimi et al. (2021). Most of the findings on globalization have a one-way effect on energy consumption (Etokakpan et al., 2020; Usman et al., 2021; and Shahbaz et al., 2021). The relationship between the direction of energy consumption and globalization was found by Shahbaz et al. (2021) through the Time-varying Granger method.

Table 3

Selected literature globalization (GLO) and energy consumption (EC).

Author	Period	Country	Method	Finding
Shahbaz et al. (2018)	1970–2014	25 developed economies	Dumitrescu-Hurlin	GLO↔EC
Etokakpan et al. (2020)	1980–2014	Malaysia	Granger causality	GLO→EC
Fahimi et al. (2021)	1970–2015	MINT	Bootstrap panel Granger	GLO↔EC
Usman et al. (2021)	1980–2018	Resource-rich Countries	Dumitrescu-Hurlin	GLO→EC
Shahbaz et al. (2021)	1970Q1-2017Q4	Ten top global and ten bottoms global	Time-varying Granger causality	GLO→EC (Austria, Singapore, Kazakhstan)GLO←EC (Namibia)

Note: → unidirectional ( ), ← unidirectional ( ), ↔ bidirectional.

Data and methodology

One of the significant challenges this study faces is the minimal data accessibility. Hence, this study utilized the annual data of 11 countries between 1984 and 2013 with 330 observations. In this study, the variables are measured as follows: economic growth (GDP) using a constant 2010 GDP in dollars and Energy Consumption (EC) in kg of oil equivalent. Meanwhile, Globalization (GLO) is measured by the economic, social, political, and social globalization index (Dreher, 2006; Gygli et al., 2019). The economic growth and energy consumption data are sourced from the World Development Indicators (WDI), while the globalization data is sourced from the KOF Swiss Economic Institute.

This study uses Granger Causality to answer the causality relationship. The causal concept is found in the VAR model, assuming that the variables x and y (if using two) are stationary and cointegrated. In addition, the VAR model has one answer for all samples. However, problems arise when using panel data in heterogeneous conditions. As a result, the resulting analysis does not provide a unique solution. Kónya (2006) introduced the bootstrap panel causality method with the Seemingly Unrelated Regression (SUR) estimator to solve this problem. This method has several advantages; among others, first, it is very robust against stationarity and cointegration, so these two conditions do not need to be tested again for panel data (Destek and Aslan, 2017). Second, each country's statistical results are based on bootstrapping (Zhang et al., 2016). Third, causality equations can be formed as bivariate and trivariate SUR. Fourth, SUR is more efficient in estimating country by country than Pooled Least Square (PLS) (Zhang et al., 2016). We construct the equations of this system with the trivariate SUR (Tekin, 2012) with the following equation:

andWhere Y is GDP, X is EC, and Z is GLO. The three elements are formed as endogenous variables in the configuration of a triaviate, and thus causality testing is conducted between GDP and EC, GLO and GDP, and GLO and EC. Meanwhile, N is the area, t is the time, i is the amount of lag, and ε is the white noise. We form these three variables in the natural logarithm.

For the Granger causality test in the system, the causal relationship for each tested country will be found as follows: (i) there is a one-way Granger causality relationship from X to Y if not all are zero, but all is zero; (ii) there is a one-way Granger causality relationship from Y to X if all are zero, but not all are zero; (iii) there is a bidirectional Granger causality relationship between X and Y if and are not zero; (iv) there is no Granger causality relationship between X and Y if and are zero. The results of causality testing are very sensitive to lag, so optimal lag testing is necessary. However, lag testing in the case of large panel systems is very computationally taxing. Therefore, we assume a lag between 1 and 4; our best lag measurement uses the Schwartz Criterion (SC) information. The procedure was strongly suggested by Kónya (2006) and has been implemented by Zhang et al. (2016). In addition, we limit the use of variables to three to avoid multicollinearity problems in the system. Kónya (2006) also explains that the SUR estimator applied in panel Granger causality because there are contemporaneous correlations (correlated between residuals). If no contemporaneous correlation then the OLS estimator is more efficient. Detection can be tested through the latest methods, namely the Cross-Sectional Dependence (CSD) test, and the Slope Homogeneity Test.

The CSD tests is used on panel data that has specific conditions. In particular, the relationship in panel data caused by globalization that the economy between countries affects the economies of other countries (Destek and Aslan, 2017; Ozcan and Ozturk, 2019). Therefore, it is necessary to test the existence of the CSD. There are three forms of tests to observe this idea, namely the Lagrange Multiplier (LM), otherwise known as Cross-sectional Breusch-Pagan (CSBP) (Breusch and Pagan, 1980) and the two cross-sectional tests (Pesaran, 2021). The latter is a derivative of LM in the form of one, and the final test is based on the pairwise correlation.

The CDBP test for CSD is valid when T→∞ where N is constant and the assumptions used in the test are in the form of unbalanced panel data. In other words, this test is feasible for a few N and a sufficient T. The formulation of this test is as follows:

However, the test is not valid when N and T exhibit large sizes, so Pesaran (2021) reformulates this test with a scaled version, namely:

The CDLM test is distributed with N (0, 1) with the null hypothesis, in which there is no CSD. Furthermore, Pesaran (2021) performs the CSD calculations based on the average with the equation:

Pesaran and Yamagata (2008) discovered the slope homogeneity test with two types of shapes, which determine the existence of heterogeneous parameters in the panel data (Guzel and Okumus, 2020). Generally, countries have a distinct method of developing their economy with various policies, thus making the test necessary (Ozcan and Ozturk, 2019). This test is applied to the method in Eqs. (1), (2), and (3), presented in the form as follows:Where k, S, and N indicate the number of exogenous regressors, Swamy's (1970) statistics, and the number of cross-sections. The test can be increased for a small sample by assuming that the residuals are normally distributed. In this case, the mean and variance are adjusted as follows:where E () is k, and Var () is .

Emperical results

Table 4 is an overview of the research data from 11 countries. Economic growth, energy consumption, and globalization are transformed into a natural logarithm. The average economic growth is 26.08, energy consumption is 15.02, and globalization is 3.89. The standard deviation describes how significant the fluctuation of the data itself is. It can be seen that the economic growth is 0.973 and the energy consumption is 0.723, which indicates quite a significant deviation compared to the globalization deviation of 0.257. The table also shows the correlation between research data. Economic growth and energy consumption have a positive relationship of 81.8 percent. Economic growth and globalization have a positive relationship of 65.7 percent. Meanwhile, the relationship between globalization and economic growth is positive at 39.7 percent. This result indicates that there is a suspicion that the three are related.

Table 4

Descriptive statistics and correlation matrix.

	Obs	Mean	Std. dev	Min.	Max.
GDP	330	26.086	0.973	23.835	27.856
EC	330	15.028	0.723	13.314	16.229
GLO	330	3.891	0.257	3.206	4.345
Correlation matrix
GDP	1
EC	0.818	1
GLO	0.657	0.397	1

Before analyzing the causal relationship of the three variables, the pre-analysis stage is required. Accordingly, critical tests are conducted: the Cross-Sectional Dependence (CSD) (Eqs. (4), (5), and (6)) and slope homogeneity (Eqs. (7) and (8)). Table 5 shows the results of the CSD test, revealing that all variables reject the null hypothesis where there is no relationship between the regions. This result implies that the change in the region's economy will impact other regions or countries. The slope homogeneity clarifies that the impact of one country presents a different effect between countries. Table 5 shows that 11 countries reject the null hypothesis at the one percent significance level or heterogeneous slope. In other words, the parameters between countries do not have the same form.

Table 5

CSD and slope homogenity results.

Testing	GDP	EC	GLO
CDBP	1577.217∗∗∗	1130.008∗∗∗	1528.902∗∗∗
CDLM	145.137∗∗∗	102.498∗∗∗	140.531∗∗∗
CD	39.709∗∗∗	33.092∗∗∗	39.093∗∗∗
Slope Test	Statistic	Prob
Δ˜	21.368∗∗∗	0.000
Δ˜_adj	22.948∗∗∗	0.000

Note: ∗∗∗ significant at 1 percent.

After testing the CSD and slope homogeneity, the next step is the panel second generation unit root using the Cross-sectionally IPS (CIPS), and Cross-sectionally ADF (CADF) approaches (Pesaran, 2007). These two tests are carried out if the CDS condition and slope homogeneity are proven to exist. According to Phillips and Sul (2003), the estimation results become inefficient, especially in panel data, if there are CSD problems and slope homogeneity. Although based on the guidelines of Kónya (2006), it is stated that this stationarity test is unnecessary, we include the test as a form of information that the use of this causality model should not be stationary at the level. The results in Table 6 show that economic growth, energy consumption, and globalization are not stationary at the levels of both the CADF and CIPS approaches. Testing at the first difference (Δ) level found that all variables were stationary in both CADF and CIPS at 1 percent.

Table 6

CADF and CIPS stationarity tests.

Variables	CADF		CIPS		Conclusion
	Level	Δ	Level	Δ
GDP	-2.133	-2.473∗∗∗	-1.426	-3.621∗∗∗	I (1)
EC	-1.566	-2.790∗∗∗	-1.455	-4.773∗∗∗	I (1)
GLO	-2.222	-2.294∗∗∗	-1.391	-5.127∗∗∗	I (1)

Note: ∗∗∗ significant at 1 percent.

Causality results of energy consumption (EC) and economic growth (GDP)

After testing the CSD and slope detection, the next step tests the Granger bootstrap panel. This method is based on the SUR estimator and tests each region from the panel data through the Wald test (Kónya, 2006). This test is similar to the previous Granger causality test, which examines the influence of the past X on the Y variable. However, it is probable for the panel data to find erroneous or misleading results due to its heterogeneous nature (Kar et al., 2011), and thus this bootstrap is appropriate (Menyah et al., 2014).

The first results of the Granger SUR causality test are economic growth and energy consumption. The test results are shown in Table 7. The results of this causality test we found a relationship between energy consumption and economic growth in Egypt, Indonesia, Iran, South Korea, Nigeria, and Turkey, where the Wald test was statistically significant at the 1 percent level. Meanwhile, the opposite test between economic growth and energy consumption shows that all countries are significant at the 1 percent level. Based on these findings, it can be concluded that there is a two-way relationship between Egypt, Indonesia, Iran, South Korea, Nigeria, and Turkey. This study is in line with Mirza and Kanwal (2017), Destek and Aslan (2017), Saidi et al. (2017), Gorus and Aydin (2019), Khan et al. (2021), Rahman (2021), and Ikhsan et al. (2022). Interestingly, the sample we used was found in several ASEAN regions, such as Indonesia, which found different things from research by Chontanawat (2020), where the relationship between energy and economic growth was not proven to exist. On the other hand, a one-way causality relationship is shown for Bangladesh, Mexico, Pakistan, the Philippines, and Vietnam from economic growth to energy consumption, and this finding is in line with Ozcan and Ari (2015), Destek and Aslan (2017) and Pinzón (2018). This finding indirectly explains that every country in the N-11 will increase energy if its economic growth grows.

Tabel 7

The panel causality results in energy consumption (EC) and economic growth (EG).

Country	Statistics EC→GDP	Critical Value			Statistics	Critical Value
		1%	5%	10%	GDP→EC	1%	5%	10%
Bangladesh	0.369	0.765	0.735	0.720	6.238∗∗∗	0.159	0.154	0.151
Egypt	9.256∗∗∗	0.433	0.417	0.407	10.518∗∗∗	0.011	0.010	0.009
Indonesia	1.558∗∗∗	0.037	0.034	0.033	13.493∗∗∗	0.621	0.605	0.598
Iran	6.237∗∗∗	0.169	0.165	0.163	7.697∗∗∗	0.033	0.032	0.031
South Korea	0.980∗∗∗	0.763	0.755	0.751	51.853∗∗∗	0.123	0.119	0.116
Mexico	0.001	0.412	0.401	0.395	52.823∗∗∗	0.435	0.411	0.401
Nigeria	1.695∗∗∗	0.346	0.333	0.324	16.855∗∗∗	0.003	0.003	0.002
Pakistan	0.570	0.929	0.923	0.920	16.153∗∗∗	0.072	0.070	0.069
Philippines	0.005	0.632	0.571	0.530	25.078∗∗∗	0.632	0.620	0.615
Turkey	3.428∗∗∗	0.746	0.723	0.713	41.452∗∗∗	0.024	0.023	0.022
Vietnam	1.299	1.394	1.360	1.342	5.025∗∗∗	0.259	0.250	0.246

Note: ∗∗∗ significant at 1 percent. The critical value is obtained by bootstrap 10,000.

Causality results of globalization (GLO) and economic growth (GDP)

A second Non-Granger causality test was conducted between globalization and economic growth. The results of this test are shown in Table 8. The first column is the result of the Wald Statistics test and is followed by the critical bootstrap value. Our findings show that feedback causality between economic growth and globalization occurs in 9 countries, except for Pakistan and Turkey, where the Wald statistic is less than the critical value of 10 percent. Meanwhile, the causal test of economic growth against globalization found that the Wald statistic obtained for each country was more significant than the critical value of 1 percent. This means that all countries have a significant influence. Based on these findings, it can be concluded that a two-way causal relationship occurs in Bangladesh, Egypt, Indonesia, Iran, South Korea, Mexico, Nigeria, and the Philippines. This finding is in line with Kilic (2015). Meanwhile, the hypothesis of a one-way relationship between economic growth and globalization only occurs in Pakistan and Turkey. This finding supports the findings by Olimpia and Stela (2017) and Yang et al. (2021); however, it differs from the study of Etokakpan et al. (2020).

Table 8

Panel causality results from globalization (GLO) and economic growth (GDP).

Country	Statistics GLO→GDP	Critical Value			Statistics	Critical Value
		1%	5%	10%	GDP→GLO	1%	5%	10%
Bangladesh	3.912∗∗∗	0.603	0.584	0.573	16.375∗∗∗	0.059	0.048	0.043
Egypt	27.194∗∗∗	0.080	0.074	0.070	3.146∗∗∗	0.017	0.011	0.008
Indonesia	4.330∗∗∗	0.267	0.259	0.255	9.421∗∗∗	0.718	0.675	0.646
Iran	5.001∗∗∗	0.003	0.002	0.002	7.804∗∗∗	0.071	0.054	0.043
South Korea	2.669∗∗∗	0.469	0.463	0.460	11.237∗∗∗	0.228	0.170	0.143
Mexico	4.734∗∗∗	0.524	0.514	0.508	35.006∗∗∗	0.625	0.554	0.517
Nigeria	2.645∗∗∗	0.037	0.031	0.028	10.430∗∗∗	0.043	0.029	0.024
Pakistan	0.064	0.184	0.181	0.180	11.383∗∗∗	0.116	0.098	0.085
Philippines	0.960∗∗∗	0.073	0.049	0.038	2.590∗∗∗	0.986	0.872	0.830
Turkey	0.007	0.121	0.116	0.113	22.964∗∗∗	0.112	0.087	0.075
Vietnam	2.435∗∗∗	0.188	0.175	0.169	3.050∗∗∗	0.388	0.325	0.291

Note: ∗∗∗ significant at 1 percent. The critical value is obtained by bootstrap 10,000..

Causality results of globalization (GLO) and energy consumption (EC)

Our third test of the Granger causality relationship is between globalization and economic growth. The results of the tests with Wald statistics and critical bootstrap values are presented in Table 9. We find the Wald statistics on energy consumption against globalization rejecting H0 at a significance level of 1 percent for each country. However, the opposite test of globalization on economic growth is only nine countries that reject H0 while two countries, namely Mexico and Vietnam, show the statistical Wald value obtained is smaller than the 10 percent significance level. From these findings, it can be concluded that a two-way relationship occurs in Bangladesh, Egypt, Indonesia, Iran, South Korea, Nigeria, Pakistan, Philippines, and Turkey. This finding is in line with the study of Fahimi et al. (2021). For the countries of Mexico and Vietnam, it is concluded that there is only a one-way relationship between globalization to energy consumption. These findings show similarities to Shahbaz et al. (2018) and Shahbaz et al. (2021) because countries with high globalization play an essential role in energy availability. This opportunity is not found for low globalized countries to obtain energy resources from outside countries.

Table 9

Panel causality results from globalization (GLO) and energy consumption (EC).

Country	Statistics EC→GLO	Critical Value			Statistics GLO→EC	Critical Value
		1%	5%	10%		1%	5%	10%
Bangladesh	6.123∗∗∗	0.531	0.520	0.514	1.202∗∗∗	0.419	0.388	0.370
Egypt	4.955∗∗∗	0.065	0.062	0.061	15.033∗∗∗	0.491	0.445	0.415
Indonesia	9.225∗∗∗	0.282	0.275	0.270	1.968∗∗∗	0.147	0.133	0.122
Iran	9.327∗∗∗	0.002	0.001	0.001	10.192∗∗∗	0.359	0.306	0.274
South Korea	30.160∗∗∗	0.465	0.454	0.449	7.211∗∗∗	1.055	0.875	0.826
Mexico	45.933∗∗∗	0.554	0.535	0.522	0.306	0.424	0.375	0.345
Nigeria	15.061∗∗∗	0.031	0.029	0.028	2.576∗∗∗	0.288	0.259	0.245
Pakistan	16.573∗∗∗	0.176	0.173	0.171	2.447∗∗∗	1.411	1.279	1.207
Philippines	20.471∗∗∗	0.022	0.019	0.018	4.301∗∗∗	0.520	0.448	0.409
Turkey	17.404∗∗∗	0.126	0.123	0.122	2.645∗∗∗	1.151	1.083	1.037
Vietnam	1.484∗∗∗	0.171	0.164	0.161	0.728	1.742	1.518	1.397

Note: ∗∗∗ significant at 1 percent. The critical value is obtained by bootstrap 10,000.

Conclusion and policy implication

This study examined the causal relationships of energy consumption, economic growth, and globalization in 11 countries between 1984 and 2013. In this context, we introduce a new branch of research in the field of economics, namely energy-growth-globalization (EGG). To our knowledge, the study of the combination of the three variables has never been carried out. On this basis, we are motivated to fill the void in the literature, which has been discussing energy consumption and economic growth, especially in panel data. We use the bootstrap causality testing method by Kónya (2006), which produces more country-specific Wald statistics and tests better with the SUR estimator. The procedure for using this method was carried out using the Cross-sectional Dependence (CSD) approach by Pesaran (2021) and the slope homogeneity by Pesaran and Yamagata (2008), as well as the Cross-sectional ADF (CADF) and Cross-sectional IPS (CIPS) stationarity testing by Pesaran (2007). Our findings in these pre-estimates explain that statistical results in CSD, across three types of tests, are significant at the 1 percent level, meaning that each country in the study has a relationship with the other countries. The results of the slope homogeneity test explain that the two tests get a statistically significant value at the 1 percent level. This result indicates that each country has different parameters, so the country in the sample is heterogeneous. The last test was stationary, showing that all variables, based on the CADF and CIPS approaches, were stationary at the first difference level.

Based on the emperical results, we found that there findings, First, the results of the Granger non-causality test between energy consumption and economic growth found two conclusions: a two-way relationship and a one-way one. Two-way relations occur in Egypt, Indonesia, Iran, South Korea, Nigeria, and Turkey. Meanwhile, the one-way relationship between economic growth to energy consumption applies in Bangladesh, Mexico, Pakistan, the Philippines, and Vietnam. This reciprocal relationship signals that the country needs energy as the initial foundation for promoting economic growth (Saidi et al., 2017). If there is an increase in economic growth, energy demand will appear to meet all economic sectors (Destek and Aslan, 2017; Khan et al., 2021). However, the ability of the state to buy oil becomes a challenge if the price of crude oil continues to increase and its volatility is high, as was the case in 2012 when the price of oil touched 100 dollars per barrel. This result has caused lower-middle-income countries such as Bangladesh, Mexico, Pakistan, the Philippines, and Vietnam to rethink and take steps to reduce the amount of oil consumption. As a result, energy does not have a tangible impact on economic growth.

Second, the results of the Granger non-causality test between globalization and economic growth found two forms of causality, namely reciprocity and one direction from economic growth to globalization. Two-way relationships were confirmed for Bangladesh, Egypt, Indonesia, Iran, South Korea, Mexico, Nigeria, Philippines. At the same time, the one-way relationship applies in Pakistan and Turkey. These findings explain that efforts to increase economic growth can be made by opening up workspaces with other countries. Indirectly, globalization acts as a space for exchanging technology and knowledge (Fahimi et al., 2021). On the other hand, the state has the opportunity to produce goods massively if the conditions for domestic demand have been met. The steps taken by the state are to export these goods to the destination country. This result is very beneficial for developing countries (Kilic, 2015).

Third, the non-causality test results between globalization and economic growth found a two-way and a direct relationship. Bangladesh, Egypt, Indonesia, Iran, South Korea, Nigeria, Pakistan, Philippines, and Turkey have two-way relations. Meanwhile, the one-way relationship between globalization to energy consumption is only found in Mexico and Vietnam. The linkage of globalization and energy consumption reflects how a country obtains sufficient resources to revive its economy. With globalization, oil-importing countries such as Indonesia and oil-exporting like Nigeria have an advantage over the two (Fahimi et al., 2021). However, increasing globalization also creates problems if one country closes access to needed goods (Shahbaz et al., 2018). For example, a decline in oil prices is so unfavorable to oil-exporting countries that they reduce their oil exports. This effect creates problems for oil-importing countries, namely supply shortages, resulting in stalled economic growth.

Based on these findings, important policies need to be observed. The existence of cross-sectional dependence (CSD) in N-11 indicates that if one country has problems with globalization, such as the economic crisis, other countries will also feel the impact. This evidence shows the existence of a significant relationship between energy-growth-globalization (EGG). In particular, the role of globalization in increasing economic growth and encouraging energy consumption. Therefore, N-11 needs to strengthen cooperation. Not only to strengthen economic growth but also to strengthen and share environmentally friendly-based technologies and divert fossil energy resources to renewable energy in order to future economic and sustainability.

Declarations

Author contribution statement

Kamal Fachrurrozi: Conceived and designed the experiments; Analyzed and interpreted the data; Wrote the paper.

Raja Masbar: Analyzed and interpreted the data; Wrote the paper.

Aliasuddin: Conceived and designed the experiments; Performed the experiments; Analyzed and interpreted the data; Wrote the paper.

Chenny Seftarita: Contributed reagents, materials, analysis tools or data; Wrote the paper.

Funding statement

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability statement

Data will be made available on request.

Declaration of interest's statement

The authors declare no conflict of interest.

Additional information

No additional information is available for this paper.