ENERGY CONSUMPTION-ECONOMIC COMPLEXITY RELATIONSHIP: AN APPLICATION ON EU COUNTRIES AND TÜRKİYE

Objective: This study aims to estimate the relationship between energy consumption and economic complexity in 25 European Union member states and Türkiye from 1995 to 2020. Theoretical Framework: The causality relationship between these concepts can be discussed in terms of four theorems (neutrality hypothesis, growth hypothesis, conservation hypothesis, and feedback hypothesis). The lack of a causal relationship between economic complexity and energy consumption is explained by the neutrality hypothesis. The growth hypothesis points to a unidirectional correlation from energy consumption to economic complexity. In the contrary case, a unidirectional causality from economic complexity to energy consumption is the subject of the conservative hypothesis. Finally, the feedback hypothesis reveals the existence of a two-way causal relationship between economic complexity and energy consumption (Gorus ve Aydin, 2019: 816; Dogan: 2015: 535; Liu vd., 2021: 3). Method: First, different approaches explaining the relationship between economic complexity and energy consumption in the relevant literature were examined to create the empirical analysis framework. After the specification of the appropriate model, pooled ordinary least square model, fixed effects model and random effects model were used to estimate the relationship between energy consumption and economic complexity within the country sample. The study investigated the relationship between fossil energy consumption and economic complexity using Emirmahmutoğlu & Köse (2011) panel Granger causality test. Results and Discussion: According to the analysis results, the growth hypothesis appears valid for EU Member States and Türkiye. Research Implications: This result reveals that fossil energy consumption plays an important role in increasing economic complexity, energy conservation policies negatively impact economic complexity, and countries in this situation are highly dependent on energy. Originality/Value: The empirical literature on the relationship between energy consumption and economic complexity is limited. In this study, where the subject of the research is discussed from an empirical perspective, the fact that 25 member countries of the European Union and Turkey were selected as samples increases the originality value of our study.

Resultados e Discussão: De acordo com os resultados da análise, a hipótese de crescimento parece válida para os Estados-Membros da UE e para a Turquia.

INTRODUCTION
Energy has been an indispensable element for the functioning of human activities throughout history.In the early periods, while this need was met by primitive means; The increase in population, the acceleration of migration from rural areas to cities, the development of societies, the industrial revolution, and changes in the production structure have made it inevitable to use different resources for energy supply.Today, fossil fuels seem to dominate the energy sector.Although energy is one of the basic needs for the economic development of countries, these fossil fuels constitute 34 percent of greenhouse gas emissions.(UN, 2024).This situation supports the idea that urgent action should be taken for a more livable world.In this context, especially since non-renewable energy consumption causes serious environmental concerns, they must be substituted with other sources.Thus, renewable energy sources such as solar energy, wind energy, and hydroelectricity constitute a significant alternative.1 1 1 1 1 1 1 1 1  1 1 1 1 1 1 1 1 1  The economic complexity index is one of the recent criteria used today to explain income and growth differences between states.The index in question particularly shows the diversity of exported products.
As seen in Table 1, the countries that rank first in terms of economic complexity index are Japan, Switzerland, Germany, and Singapore.America, which ranks first in terms of GDP variable, ranked 14th as of 2021.During the period under consideration, while the US economy appeared to become less complex, this result is associated with the US's not diversifying its exports.Türkiye is among the 40th most complex countries in the economic complexity index ranking.The Turkish economy has become more complex compared to 2000.

5
When the theoretical and empirical literature was scrutinized, the idea that there may be a relationship between the concepts of energy consumption and economic complexity emerged.
The causality relationship between these concepts can be discussed in terms of four theorems (neutrality hypothesis, growth hypothesis, conservation hypothesis, and feedback hypothesis).
The lack of a causal relationship between economic complexity and energy consumption is explained by the neutrality hypothesis.The growth hypothesis points to a unidirectional correlation from energy consumption to economic complexity.In the contrary case, a unidirectional causality from economic complexity to energy consumption is the subject of the conservative hypothesis.Finally, the feedback hypothesis reveals the existence of a two-way causal relationship between economic complexity and energy consumption (Gorus ve Aydin, 2019: 816;Dogan: 2015: 535;Liu vd., 2021: 3).
This study investigates the relationship between energy consumption and economic complexity for 25 European Union member states and Türkiye.Following the introduction section where theoretical information about the concepts of energy consumption and economic complexity is presented, the article consists of 4 chapters.Chapter 2 is titled the literature review.Chapter 3 includes the data description of the study.Chapter 4 presents the methodology and empirical evidence.Chapter 5 is titled conclusions and policy recommendations.

LITERATURE
Research on energy consumption and economic complexity is limited.In some studies, the concept of energy consumption is discussed separately as non-renewable energy consumption, renewable energy consumption, and energy efficiency.Azizi (2019) observed that the increasing complexity of the economy, specifically in the industrial sector, in the Iranian economy between 1971 and 2016 led to an increase in energy consumption.Using FMOLS and DOLS panel estimation from 1995 to 2016, Neagu and Teodoru (2019) show that there is a long-term equilibrium relationship between economic complexity, energy consumption structure, and greenhouse gas emission in European countries.This is interpreted that the more the risk of pollution increases the more economic complexity rises.Shahzad et al. (2020)

DATA
This study analyzes the relationship between energy consumption and economic complexity for the sample of 25 European Union member states2 and Turkey for the period 1995-2020.The reason for choosing the years 1995-2020 in the study is that the economic complexity index data starts from 1995 and fossil energy consumption data is available for the considered country holistically till 2020.
The variables used in the analysis are seen in Table 2.The dependent variable is the share of primary energy consumption from fossil fuels (LNFEC), economic complexity index (LNECI), income per capita (LNGDPP), urbanization rate (LNURB), and foreign direct capital inflows (LNFDI) were determined as explanatory variables.

Economic complexity index LNECI
Countries are ranked with an economic complexity index depending on how diverse and complex export baskets are.This study used the index calculated using the harmonized system product classification.

Income per capita LNGDPP
It is the ratio of a country's GDP to its mid-year population for a given year.Worldbank

Urbanization rate LNURB
It is the proportion of people living in urban areas.The natural logarithm of the variable was used.Worldbank

LNFDI
Foreign direct investment shows direct investment capital flows into the reporting economy.Worldbank

METHOD AND FINDINGS
Table 3 presents   Table 4 shows the bilateral correlations between the variables.While there is a positive and statistically significant relationship between LNFEC and LNECI, LNGDPP and LNURB; there is a negative and statistically insignificant relationship between LNFEC and LNFDI.Pooled Least Squares, Fixed Effects Model, and Random Effects Model were used to test the relationship between energy consumption and economic complexity.

Without artificial variables showing the effects of each cross-sectional unit, In Pooled
Least Squares, where the data of the cross-section group is combined/pooled, the coefficients of the fixed parameters and independent variables do not change according to units and time.
In Fixed Effects Model, slope coefficients are the same for time and cross-section units, whereas the constant coefficient varies according to units.In other words, in this model, differences 9 between units are captured by differences in the constant term.In Random Effects Model, differences occurring in cross-sectional units or units and time are included in the model as a component of the error term (Özer and Özer, 2014: 132).
To first determine which of pooled least squares, fixed effects model, and random effects model would produce more consistent results in the analyses, basic diagnostic tests were examined and the findings were summarized in   After fixed effect model was selected, to ensure the reliability of the findings, it was investigated whether there were heteroskedasticity and autocorrelation problems and the results were given in Table 6.Since the statistical values of the autocorrelation tests are less than 2, there is an autocorrelation problem in the model findings; since the modified Wald test statistic value was found to be statistically significant (the null hypothesis indicating constant variance was rejected), it was determined that there was also a heteroscedasticity problem in the model findings.10 complexity index and urbanization rate variables have a statistically significant and positive effect on fossil energy consumption, and foreign direct capital investments have a statistically significant and negative effect.Accordingly, as economic complexity and urbanization rate increase in EU countries and Türkiye, fossil energy consumption increases, and as foreign direct capital investments increase, fossil energy consumption decreases.(Ağır andTıraş, 2018: 1563).There are three stages in applying the test.First of all, the appropriate lag length (p) is determined by estimating a standard Panel Vector Autoregression (PVA) model; then, the integration level (dmax) of the unit with the highest integration degree is added to this delay level; finally, for the (p+dmax) delay level, the Panel VAR model is estimated with the level values of the variables (Şahin and Durmuş, 2019: 195).

CONCLUSION AND EVALUATION
Climate change, environmental degradation, and global warming have been one of the most discussed topics by researchers in recent years.The rapid increase in industrial production after the Industrial Revolution brought about the spread of greenhouse gases released by humans into nature and the atmosphere due to the increase in energy consumption.Then emissions from greenhouse gases have made the exacerbation of environmental problems inevitable.On the other hand, in today's international competitive environment, producing highvalue-added products and exporting these products have become the main goals for countries.
For this purpose, the development of an information and technology-centered production technique will give countries an advantage.
The economic complexity index which is an indicator of the diversity of knowledge with skills of countries and the level of knowledge in addition to the technology of exported products was developed by Hidalgo and Hausmann (2009).Thus, it provides information about a country's industrial capabilities, knowledge, diversity in production techniques, and how growth occurs.Japan, Switzerland, South Korea, Germany, Singapore, Czech Republic, Austria, England, Slovenia, and Sweden are at the top of the economic complexity index classification as of 2021.With diversified export baskets, these countries' export products are not exported much by other countries.Thus, while countries aiming for competitiveness and high economic growth in international trade try to bring their economies to a more complex level, they also take action on energy consumption and climate change.When the theoretical and empirical literature on economic complexity is examined, it is viewed that the issue is discussed with different dimensions such as ecological footprint, technological innovation, This result reveals that fossil energy consumption plays an important role in increasing the level of economic complexity; energy conservation policies have a negative impact on economic complexity, and countries in this situation are highly dependent on energy (Gorus ve Aydin, 2019: 816).
, based on data covering the period 1965Q1-2017Q4 in the United States investigated the relationship between economic complexity, fossil fuels, energy, and ecological footprint.Evidence from the framework of quantile autoregressive distributed lag (QARDL) analysis shows that economic complexity increases the ecological footprint of fossil fuel energy consumption.The impact of Energy Consumption-Economic Complexity Relationship: an Application on EU Countries and Türkiye ___________________________________________________________________________ Rev. Gest.Soc.Ambient.| Miami | v.18.n.1 | p.1-15 | e07739 | 2024.6 economic complexity on energy consumption in Iran was investigated by Azizi (2020).According to the regression analysis findings for the period 1971-2013, increasing economic complexity leads to a decrease in energy consumption.Liu et al. (2021) examined the relationship between energy consumption and economic complexity in Lancang-Mekong Cooperation countries from 1991 to 2017 using the panel vector autoregression (PVAR) model.This study has revealed a one-way relationship between energy consumption and economic complexity index.Fang et al. (2021) used Augmented Average Group (AMG) forecast management to explain the relationship between economic complexity and energy demand.The study, which used the data set of 25 OECD countries from 1978 to 2016, revealed the inverse relationship between the variablesKazemzadeh et al. (2023)found the existence of an inverted U-shaped relationship between economic complexity and energy intensity.62 countries were selected as the country sample for the period 1995-2019, and panel quantile regression (PQR) models were preferred in the empirical part of the study.Türk (2023) mentions the existence of a strong causality relationship from economic complexity to energy consumption for the period 2000-2020, using theEmirmahmutoglu and Kose (2011) causality test in the BRICS-T group countries.In another study,Shahzad et al. (2023) empirically examine the impact of biomass energy consumption and economic complexity on environmental sustainability, specifically for G7 economies.Salimi et al. (2023) examined the effect of economic complexity on renewable energy consumption in Iran and stated that, according to the results of the ARDL test for the period 1369-1400, economic complexity increased renewable energy consumption.
descriptive statistics of the variables used in testing the energy consumption-economic complexity relationship.When the skewness values of the variables are examined, it is understood that the LNFEC and LNURB variables are positively skewed to the right, while the LNECI, LNGDPP and LNFDI variables are negatively skewed to the left.According to the kurtosis values, since the kurtosis values of the LNGDPP and LNECI variables are greater than 3, their distributions are sharp compared to the normal distribution; Since the kurtosis values of LNFEC, LNFDI and LNURB variables are less than 3, their distributions are flat compared to the normal distribution.The sample of the study consists of 676 observations in total, with the number of units N = 26 and the time dimension T = 26.
Energy Consumption-Economic Complexity Relationship: an Application on EU Countries and Türkiye ___________________________________________________________________________ Rev. Gest.Soc.Ambient.| Miami | v.18.n.1 | p.1-15 | e07739 | 2024.13 income, and economic growth.On the other hand, studies on determining whether there is a relationship between economic complexity and energy consumption are limited in the literature.In this context, the relationship between economic complexity and energy consumption was chosen as the subject of the study.The causality relationship between economic complexity and energy consumption is explained by four different theories.Contrary to the neutrality hypothesis, which suggests that there is no causal relationship between economic complexity and energy consumption, the Growth Hypothesis shows unidirectional causality from energy consumption to economic complexity and the Conservation Hypothesis indicates unidirectional causality from economic complexity to energy consumption.The feedback hypothesis reveals a two-way causality relationship between economic complexity and energy consumption(Gorus ve Aydin, 2019:   816; Dogan: 2015, 535; Liu vd., 2021).The sample of this study consists of 25 European Union member states and Türkiye.The reason for choosing the years 1995-2020 in the study is that the economic complexity index data starts from 1995 and fossil energy consumption data is accessible for the considered country holistically till 2020.The relationship between fossil energy consumption and economic complexity was investigated using Emirmahmutoğlu & Köse (2011) panel Granger causality test.Findings from Emirmahmutoğlu & Köse (2011) panel Granger causality test indicate that the Growth hypothesis is valid for EU states and Türkiye.
When a country's exports are more diverse and there are fewer competitors for those export products, the economic complexity index gets a higher value (The Atlas of Economic Complexity, 2024).The change in country rankings in terms of economic complexity index for the period 2000-2021 is given in Table1.

Table 3
Descriptive Statistics

Table 4
Correlation Matrix ** : It is statistically significant at the 5% level.

Table 5 .
Since the coefficients of the F test and the Breusch and Pagan test are rejected, pooled least squares is not valid.Therefore, to determine which of fixed effects model and random effects model will give consistent results, the results of the Hausman test were examined.Since the Hausman test statistic value was found to be statistically significant, it was determined that fixed effects model would produce more consistent results.

Table 5 F
, Breusch and Pagan and Hausman Test Results

Table 6
Tests for Deviation from Basic Assumptions Since there were autocorrelation and heteroscedasticity in the model findings, Driscoll-Kraay standard errors were used and presented in Table 7.As in the table, the economic ___________________________________________________________________________ Rev. Gest.Soc.Ambient.| Miami | v.18.n.1 | p.1-15 | e07739 | 2024.

Table 7
Fixed Effects Model Findings Obtained Using Driscoll-Kraay Standard Errors

Table 8 including
Emirmahmutoğlu & Köse (2011)panel Granger causality test findings is examined, there is no causality relationship from the economic complexity index to fossil energy consumption in terms of the panel overall (since Panel Fisher test statistic value = 99.045 is statistically insignificant); on the other hand, it has been demonstrated that there is causality from fossil energy consumption to the economic complexity index (as the Panel Fisher test statistic value = 237.076 is statistically insignificant).When the results are examined on a country basis, the countries where there is causality from the economic complexity index to fossil energy consumption are; France, Greece, Hungary, Ireland, Italy, and Latvia.From fossil energy consumption to economic complexity, it is seen that there is a causality relationship in all countries considered.