The moderating role of informal economy on financial development induced ekc hypothesis in turkey

Abstract

This study conducts an empirical investigation about the moderating role of the informal economy on Turkey's environmental performance by employing advanced econometric techniques that account numerous structural breaks in series. In this extent, we created three interaction variables by captivating the impact of informal economic activities on CO₂ emissions through income, energy use, and financial sector development. Besides, we built a main effect model without the interaction variables to assess the direct effects of our variables on global environmental degradation. The outcomes of the carried analyses produced supporting evidence toward the confirmation of the Environmental Kuznets Curve (EKC) assumption. Obtained findings shown that energy use, financial development and the informal economy in Turkey transmit a deteriorating impact on environmental well-being. Furthermore, the moderating role of the informal economy was found to be statistically significant factor in terms of both economic and environmental efficiency.

Keywords

environmental performance informal economy financial development moderating effect

Introduction

Deterioration of the environment has been a central issue since the industrial era of economic activities. The atmosphere's inability to maintain an excessive amount of greenhouse gas (GHG) emissions has already produced unwanted climate changes and is expected to cause further ecological damages soon. According to the World Meteorological Organization (WMO),¹ total GHG emissions levels have increased by 175% compared to pre-industrial economic era. Heat-trapping nature of these gasses has led to rapid rises in the global temperature that used to be constantly stable for the last couple thousands of years. The records tracked by National Aeronautics and Space Administration (NASA) revealed that the mean surface heat of Earth has augmented by 0.9 degrees Celsius since the late 19^th century.² Reported figures further showed that the hottest surface temperature levels since the mid-1880s have been recorded in the last 35 years. Extreme weather changes, water scarcity, ocean/sea level rises, and lower human mortality rate are some of the permanent damages that ‘global warming’ will cause unless it can be mitigated soon³ The substantial evidence regarding the catastrophic consequences of global warming has produced international concerns and fighting against climate change has become an ambitious goal for many nations. As of January 2019, 194 parties (states and countries) who are responsible from 87% of the global GHG emissions signed the Paris Agreement within the United Nations Framework Convention on Climate Change (UNFCCC) to stop further global temperature rises.⁴

Although global warming is a natural phenomenon, GHG emissions are primarily produced by human activities which accelerated the process to alerting levels.³ Hence, understanding the dynamics between economic activities and environmental quality became a vital study field among academics and practitioners. Empirical literature of the environmental economics has generated a vast amount of studies that are focusing on the economy-energy-pollutants linkages.^5–12 Among these, the Environmental Kuznets Curve (EKC) assumption has received significant attention. Inspired from the pioneering study of Kuznets¹³ which examined the long-run shifts between unequal distribution of personal income and countries’ income growth, EKC assumption argues that environmental deterioration rises as a result of high income growth up to a point where higher economic activities starts to improve environmental quality by reducing the degradation of the environment.¹⁴ Such a relationship between income growth and environmental deterioration is referring to an inverted U-shaped curve in which the variables of environmental degradation can be modeled as functions of income growth.^15,16 Theoretically, the EKC assumption can be described by the economic stage shifts that countries experience over their growth processes. According to that, during their rapid income growth periods, nations initially move from agriculture to the industrial stage in which they generate a high amount of pollution.¹⁷ Subsequently, the pollutant-induced industrial stage is expected to be followed by the state of technologically advanced and efficient industrial production which economic activities start to improve the quality of environment.^17,18

The EKC hypothesis was initially confirmed by the work of Grossman and Krueger¹⁹ and it became a well-studied empirical phenomenon for the income growth-pollution nexus. The studies of Panayotou,²⁰ Selden and Song,²¹ Moomaw and Unruh,²² Lindmark,²³ Managi and Jena,²⁴ Ozturk and Acaravci,²⁵ Pao and Tsai,²⁶ Aslan, Destek and Okumus,²⁷ Gokmenoglu and Taspinar²⁸ and Katircioglu, Gokmenoglu and Eren²⁹ are among the ones that have delivered empirical proof for the EKC assumption by revealing the inverted U-shaped connection among income and pollution. Many other studies, on the other hand, failed to demonstrate such a relationship and argued that the hypothesis is not valid.^30–36

Several factors can be put forward to explain the mixed empirical outcome regarding the EKC hypothesis. Potential inaccuracies in the estimation process of the EKC can be given as one of these reasons for the contradictory outcome. Perman and Stern³⁷ criticized that most of the earlier empirical studies suffer from having incorrect statistical properties. For instance, only a few studies paid attention to diagnostic statistics such as serial correlation and heteroscedasticity. It is also stated that for the constructed EKC models, a potential cointegration relationship between variables is neglected most of the time. The issue of omitted variable bias constitutes an important share in differing conclusions about the validation of the hypothesis as well. In response, researchers have shifted away from conventional EKC models in which global pollutants used to be regressed on real income and energy consumption. To prevent a possible omitted variable bias, particular segments of the economy have been included in the EKC modelling. Frankel and Romer³⁸ argued that as the countries reach into higher levels of financial development, economic growth escalates through higher local and international investment activities which in turn affect the environmental quality. In this extent, financial development has become increasingly examined factor in terms of its constructive or destructive environmental impacts. According to Lundgren³⁹ and Ma and Stern,⁴⁰ financial development can stimulate technological enhancements in industrial goods and services production of the companies. Such improvements in technology help enterprises to achieve economies of scale and as a consequence, environmental degradation decreases by reduced waste and resource pollution. Although financial sector development can generate an expansion in industrial production level and conveys economic growth, failing to adopt technological innovations might leave a deteriorating impact on the environment. Sadorsky⁴¹ indicated that financial development enables easy access to capital for businesses to expand their operations and help individuals to afford big ticket items such as cars and houses. As a result, energy consumption is expected to rise to the degree that energy conservation policies might fail to reach their targets if they do not include the impact of financial development. Katircioglu and Taspinar⁴² highlighted that increased financial sector growth might lead to environmental pollution through higher energy consumption unless sustainable growth transition can be achieved.

By examining the linkage between financial sector development and environmental degradation, Tamazian and Rao⁴³ found empirical evidence showing that financial sector development promotes capital mobility toward green investment projects and contributes to CO₂ emission reductions. Jalil and Feridun⁴⁴ suggested an inverse linkage between financial development and pollution for China. For the case of Indonesia, Shahbaz, Hye, Tiwari, and Leitao⁴⁵ determined a one-way causality from financial sector development to carbon dioxide emissions. They emphasized the supportive role of financial sector improvement for businesses to adopt advanced green energy technologies. Similarly, Shahbaz, Tiwari, and Nasir⁴⁶ revealed the mitigating role of banking sector development in reducing CO₂ emissions in South Africa as financial reforms are introduced in the country. Bekhet, Matar, and Yasmin⁴⁷ identified the financial development as an essential determinant of energy emissions reduction for Gulf Cooperation Council countries. Katircioglu and Taspinar⁴² tested the direct and moderating effects of the financial sector development. The test results confirmed the short and long-term effects of the financial development on the EKC framework through the channels of energy use and economic growth. Moreover, Tamazian and Rao⁴³ and Saidi and Mbarek⁴⁸ presented empirical outcome about financial sectors’ mitigating role of GHG emissions. On the other hand, Abbasi and Riaz⁴⁹ and Shahzad, Kumar, Zakaria, and Hurr⁵⁰ showed for the case of Pakistan that financial development increases the level of global environmental degradation due to its pre-mature structural transformation in the economy. He, Sheng and Vochozka⁵¹ demonstrated that 1% rise in financial development leads to pollution to increase by 4.29%. Similarly, Cetin and Bakirtas⁵² provided empirical evidence from emerging markets and showed financial development increases CO₂ emissions. However, Ozturk and Acaravci⁵³ documented that financial development is an insignificant variable for the changes in CO₂ emissions in the long-run.

On the other hand, informal economies have been also at the agenda of research many times where it has been shown that in developing nations the size of informal economies are considerably higher than those in developed ones.^54–57 Therefore, it is clear that informal economies are likely to have environmental concerns. Among the vast amount of studies focusing on the pollution-energy-growth nexus, the role of the informal economies has not received enough attention. Especially in emerging countries, the size of the informal economies is expected to reach into significant levels which might lead to changes in environmental quality. Although the diverse measurement approaches make the informal economy hard to define, the phenomenon can be referred as “market-based production of goods and services, whether legal or illegal, that escapes detection in the official estimates of GDP”.⁵⁸ ¹ Schneider and Enste⁵⁹ presented a detailed overview regarding the size and consequences of the unofficial economy on the official one. They stressed that despite the estimation differences and complexity of measurements, the informal economic activities had reached an enormous size in most transition and OECD countries since the 1970s. In the aspect of environmental quality, the informal economy may have hazardous influences. Unlike the firms who operate under the authority of governments, unregistered firms tend to evade environmental regulations and might cause higher pollution.⁶⁰ For Latin American countries, Loayza⁶¹ demonstrated that as the share of the underground economy rises, public services quality falls due to inefficient use of the existing facilities. In other words, some of the most essential services of general interests such as electricity distribution, environmental protection practices, public transportation, and waste management are expected to function poorly in the existence of the informal economy. Biswas, Farzanegan, and Thum⁶² investigated the linkage between corruption, informal economy and pollution in the context of both local and global pollutants. The panel estimations of the study for more than 100 countries documented the positive effect of the informal economy on SO2 and CO₂ emissions. Abid⁶³ tested the EKC hypothesis for Tunisia by taking the informal economy into the derivation of the total GDP. The study concluded an increase in total GDP stimulates global pollution by increasing the CO₂ emissions in a monotonic relationship.

As the size of informal economies is likely to impact environmental quality through income effect, it is also documented in the relevant literature that there is significant links between financial markets and informal economies.^55,64–68 Imamoglu et al.⁵⁵ find that financial markets are major contributors to the size of informal economies while Elgin and Uras⁶⁶ find that informal economies significantly impact financial markets. Therefore, a new research question is available if any significant links between financial markets and informal economies would result changes in the levels of environmental quality. To the best of our knowledge, there isn't any attempt till the moment yet to investigate the nature of the links between financial markets, informal economies, and environmental degradation.

Based on the given theoretical and empirical linkages between environment and its determinants, this study focuses on the role of informal economies on the nexus of financial sector-pollutant emissions under the EKC framework. As an emerging economy aiming to achieve a higher level of economic growth with a sustainable environment, Turkey is a compelling case for such a study. According to the Climate Change Performance Index (CCPI)⁶⁹ Turkey is ranked as the 50^th out of 61 countries in managing the GHG emissions. World Bank⁷⁰ reported that the CO₂ emissions of the country increased at a fast pace in the last three decades. Also, the figures showed that the total energy use of the country which mostly contains fossil-based energy, increased by 50% in the last decade. That being said, Turkey introduced a number of policies towards achieving sustainable economic growth in the process of its European Union membership accession. Throughout its economic past, Turkey went through economic liberalization policies, experienced severe economic crises and performed successful recoveries. During this volatile process, Turkey abounded the fixed exchange rate regime in 2001 and liberalized its financial market by letting free capital movements.⁷¹ Throughout its transition period of being a market economy, Turkey introduced numerous economic reforms and carried out institutional structural changes to promote the private business sector. The economic survey of the OECD⁷² stated that these promotions helped Turkey in accelerating the production at official level. However, it is showed that the country continues to struggle with the informality as the size of its unrecorded economy makes up 27% of the total GDP.^73,74

The linkage between income growth and global environmental pollution in Turkey has been subject to a lot of empirical studies. Mixed evidence about the validity of the EKC assumption has made Turkey one of the most examined case studies in the empirical literature. Table 1 presents and summarizes the findings from the studies that investigated the EKC for Turkey. The role of the informal economy has only covered by limited studies in the relevant literature. Elgin and Oztunali⁷⁵ examined the link between the informal sector and environmental degradation by using both local and global pollutants for the case of Turkey. Conducted time series procedures demonstrated an inverse U-shaped association between the pollutants and the extent of the informal sector. Recently, Imamoglu⁷⁶ studied the impacts of informal economy and development of financial sector on environmental well-being. The test results of the ARDL methodology showed a positive effect of the informal economy on environmental pollution for Turkey. Besides, the study pointed out that financial development has a contributing influence on the environmental performance, and it decreases the pollution in the country.

Table 1.

EKC studies carried Out for Turkey.

Studies	Independent Variables	Methodology	Conclusion
Lise, W.⁷⁷	GDP, energy supply per technology, energy consumption per technology.	Decomposition analysis.	EKC is not confirmed
Akbostancı, E., Türüt-Aşık, S., and Tunç, G. İ. (2009)<	GDP per capita, population density.	VAR analysis (first model), pooled EGLS regression (second model).	EKC is not confirmed
Ozturk, I., and Acaravci, A.²⁵	GDP per capita, energy consumption, employment.	ARDL, Granger causality.	EKC is not confirmed

Shahbaz, M., Ozturk, I., Afza, T., and Ali, A.⁷⁸	GDP per capita, GDP² per capita, energy intensity per capita, globalization index.	Zivot Andrews (1992) unit root test, bounds test, Johansen cointegration test, Gregory-Hansen⁷⁹ cointegration test,VECM Granger causality test	EKC is confirmed
Yavuz, N. C.⁸⁰	GDP, GDP², energy consumption.	ADF, PP and Zivot Andrews (1992) unit root tests, Johansen and Gregory-Hansen (1996) cointegration test	EKC is confirmed
Bölük, G., and Mert, M.⁸¹	GDP, GDP², energy consumption.	Bounds test, ARDL error correction model.	EKC is not confirmed (U-shaped)
de Vita, G., Katircioglu, S., Altinay, L., Fethi, S., and Mercan, M.⁸²	GDP, energy consumption, tourism development.	Carrion-i-Silvestre et al.⁸³ unit root test, Maki⁸⁴ cointegration test, DOLS regression.	EKC is confirmed

Dogan, N.⁸⁵	GDP, GDP², energy consumption, agriculture growth.	Bounds test, ARDL.	EKC is confirmed
Gökmenoğlu, K., and Taspinar, N.⁸⁶	GDP, energy consumption, foreign direct investment.	Zivot Andrews (1992) unit root test, Bounds test, ARDL error correction model, Toda-Yamamoto (1995) causality test .	EKC is confirmed
Gozgor, G., and Can, M.⁸⁷	GDP, GDP² per capita, energy consumption, export product diversification.	Strazicich et al. (2004) unit root test, Maki⁸⁴ cointegration test, DOLS regression.	EKC is confirmed
Katircioglu, S.¹⁵	GDP, GDP², energy consumption, oil price.	Carrion-i-Silvestre et al.⁸³ unit root test, Maki⁸⁴ cointegration test, DOLS regression.	EKC is confirmed
Katircioglu, S. T., and Taspinar, N.⁴²	GDP, GDP², financial development.	Maki⁸⁴ cointegration test, DOLS regression Granger causality/block exogeneity Wald test	EKC is confirmed
Katircioglu, S., and Katircioglu, S.⁸⁸	GDP, urbanization, energy consumption	Carrion-i-Silvestre et al.⁸³ unit root test, Maki⁸⁴ cointegration test, ARDL.	EKC is not confirmed
Ozcan, B., Apergis, N., and Shahbaz, M.⁸⁹	GDP, ecological footprint.	Hacker-Hatemi-J (2006) bootstrap Granger causality.	EKC is not confirmed
Pata, U. K.⁹⁰	GDP, financial development, trade openness, industrialization, urbanization, coal and noncarbohydrate energy consumption.	Bounds test, ARDL.	EKC is confirmed

This study presents a novel approach to the existing EKC literature by testing the moderating role of the informal economies on the effects of financial sector on environmental performance of Turkey. In the statistical context, the moderation effect enables a third variable to affect the strength of the linkage between dependent and independent variables.⁹¹ In this study, the informal economy variable is introduced to the constructed model as the moderator in order to observe its channeling impact on CO₂ emissions through financial development. The intuition behind this methodology is based on the possibility of having an indirect impact from informal economic activities to environmental performance. According to that, the informal economy first leads to a change in income growth, energy use and financial development and subsequently, these variables lead to a change on the environment. In the context of EKC hypothesis, Katircioglu and Taspinar⁴² followed a similar approach and they introduced financial development as the moderating variable. It should be noted that such an empirical practice is quite rare in the relevant literature and to the authors’ best knowledge, moderating effect of informal economy on global environmental pollution has not been researched before. The present study aims to fulfill this gap.

The rest of the study is planned as follows: Section 2 highlights the sample data and methodology, section 3 demonstrates the empirical findings and finally, section 4 provides the conclusion and policy prescriptions.

Data and methodology

Data

The present study covers the yearly sample time span of 1960–2013 to construct a multivariate EKC model which captures the effects of income growth, energy usage, financial development, and informal economic activities on CO₂ emissions². The variables that are used to model the given relationship include; carbon dioxide emissions (CO₂) (total amount of kiloton), total energy consumption (EN) (kilogram of oil equivalent), economic growth proxy (GDP per capita) (constant 2010 US$), financial development index (FD) and the volume of the informal economy (INF). Except the INF variable, we collected all other variables from the World Bank Development Database⁷⁰ The INF variable was collected from the study of Elgin and Oztunali⁹² in which a two-sector dynamic general equilibrium model is used to assess the scope of the informal economy for 161 economies around the world³. Appendix section of the present study provides glossary of terms where the variables are described.

Creating the compound financial development Index

Studies in the empirical literature propose several alternatives for the proxy preference of financial development. Based on the discussions of Beck, Demiguc-Kunt and Levine (2000), Levine, Loayza and Beck (2000) and Ang⁹³ three main determining factors of financial development are identified to quantify the financial sector development. These determinants are size of money in financial markets measured by the broad money in the system (M2), financial intermediation measured by the domestic credits delivered by the financial sector (DC), and commercial bank effectiveness derived by the ratio of deposit money bank assets to central bank assets (DBC). Many studies have employed only one of these given indicators which might result in not capturing the full aspect of financial development. Hence, the present study constructs a compound financial development index (FD) which aims to reflect the development of the financial sector completely.

The functional representation of the financial development proxy which is used in this study can be identified as follows:

FD = f (M 2, DC, DBC)

(1)We collected the components of the index from the database of World Bank⁷⁰ All of the variables are represented in percentage numbers by taking the GDP% of the M2 and DC variables. We implemented the principal component analysis to calculate the index whose formula can be written as follows:

FD Index = \sum_{k = 1}^{k} w_{k} \times S_{k}

(2)The FD index is calculated by the ratio of variance by individual financial component and the equivalent factor scores of each component (

S_{k}

w_{k}

designates the weight of separate k^th element for the financial development factor and it can be calculated as;

w_{k} = (\frac{{var}_{k}}{\sum_{k = 1}^{n} v_{k}}) \times 100

(3)where

{var}_{k}

is the explained variance by every k^th factor and n represents the amount of factors.⁹⁴

Theoretical setting and specification of the models

Empirical investigation of the EKC hypothesis starts with examining the impacts of the initial and later phases of income growth variables (GDP and GDP²) on the global emissions level. Including the GDP² variable sets up a quadratic equation where a parabolic curve can be generated. This empirical practice is also referred as the conventional EKC framework in the literature. Estimation of such quadratic function should produce an inverted U-shaped curve where an increase in GDP leads to an increase in CO₂ emissions initially and then later, further GDP growth decreases the CO₂ emissions. Therefore, estimated coefficient signs of GDP and GDP² must be positive and negative, respectively.

C O_{2_{t}} = f (GDP, GD P^{2})

(4)In parallel with the given functional relationship, many studies have included the variable of energy consumption (EN) into the conventional EKC to observe the effect of energy consumption on CO₂ emissions. Energy augmented model can be written as:

C O_{2_{t}} = f (GDP, GD P^{2}, EN)

(5)Equation 5 can be written in the form of a double logarithmic regression equation which appropriately represents the impact estimations of the regressors on CO₂ emissions.

\ln C O_{2_{t}} = α_{0} + α_{1} (\ln GD P_{t}) + α_{2} (\ln GD P_{t}^{2}) + α_{3} (\ln E N_{t}) + ε_{t}

(6)Energy augmented EKC model has extended by including various economic sectors throughout the literature. By being one of the most crucial economic sector variables, financial development is included to capture the impact of financial development on CO₂ emissions.

\ln C O_{2_{t}} = α_{0} + α_{1} (\ln GD P_{t}) + α_{2} (\ln GD P_{t}^{2}) + α_{3} (\ln E N_{t}) + α_{4} (\ln F D_{t}) + ε_{t}

(7)Finally, the informal economy variable is added to the model to observe the environmental effects of underground activities.

\ln C O_{2_{t}} = α_{0} + α_{1} (\ln GD P_{t}) + α_{2} (\ln GD P_{t}^{2}) + α_{3} (\ln E N_{t}) + α_{4} (\ln F D_{t}) + α_{5} (\ln IN F_{t}) + ε_{t}

(8)Equation 8 demonstrates the main model of this study where the effects of income growth (GDP, GDP²), energy usage (EN), financial development (FD) and informal economic activities (INF) on carbon dioxide emissions (CO₂) can be estimated for the case of Turkey. In addition to the main model, this study suggests that the informal economy may have a moderating effect on CO₂ emissions through GDP, EN and FD variables. By adopting the similar works of Chen and Myagmarsuren⁹⁵ and Katircioglu and Taspinar⁴² this study constructs a multivariate model where the moderating effect of the informal economy can be investigated.

\begin{aligned} \ln C O_{2_{t}} = & α_{0} + α_{1} (\ln GD P_{t}) + α_{2} (\ln GD P_{t}^{2}) + α_{3} (\ln E N_{t}) + α_{4} (\ln F D_{t}) + α_{5} (\ln IN F_{t}) \\ + α_{6} (\ln GD P_{t} x \ln IN F_{t}) + α_{7} (\ln E N_{t} x \ln IN F_{t}) + α_{8} (\ln F D_{t} x \ln IN F_{t}) + ε_{t} \end{aligned}

(9)The INF variable in equation 9 is included as a standalone component as well as an interaction variable. The interaction variables in equation 9 are derived by multiplying the GDP, EN and FD variables with the INF variable. Hence, both direct and moderating role of the INF variable on CO₂ emissions can be observed from the output of this model.

Empirical methodology

Unit root test

The empirical investigation of this study starts with identifying the stationarity characteristics of the variables presented in equation 8. Unlike the many numbers of available tests, Carrion-i-Silvestre et al.⁸³ unit root test allows multiple structural breaks to be included. Specifically, the test tolerates at most five structural breaks by employing the quasi-GLS procedure suggested by Elliot, Rothenberg, and Stock⁹⁶ Carrion-i-Silvestre et al.⁸³ discuss that similar other unit root procedures assume the occurrence of a break point under the alternative hypothesis of stationarity only. Not allowing a break under the null hypothesis leads test statistics to diverge or not being invariant to break parameters. Also, it is quite likely that such tests suffer from having low power as a result of failing to fully utilize the information about the structural break. The methodology of Carrion-i-Silvestre et al.⁸³ overcomes the mentioned issues by allowing the structural breaks under both null (series are not stationary) and alternative (series are stationary) hypotheses. The test calculates five test statistics under multiple structural breaks; Gaussian point optimal statistics (P_t), modified feasible point optimal statistic (MP_t) and three M-type optimum statistics:

M Z_{α} (λ^{0}) = (T^{- 1} {\tilde{y}}_{T}^{2} - s {(λ^{0})}^{2}) {(2 T^{- 2} \sum_{t = 1}^{T} {\tilde{y}}_{t - 1}^{2})}^{- 1}

(10)

MSB (λ^{0}) = {(s {(λ^{0})}^{- 2} T^{- 2} \sum_{t = 1}^{T} {\tilde{y}}_{t - 1}^{2})}^{\frac{1}{2}}

(11)

M Z_{t} (λ^{0}) = (T^{- 1} {\tilde{y}}_{T}^{2} - s {(λ^{0})}^{2}) {(4 s {(λ^{0})}^{2} T^{- 2} \sum_{t = 1}^{T} {\tilde{y}}_{t - 1}^{2})}^{\frac{1}{2}}

(12)where MZα, MSB and MZ_t are the optimum test statistics as they are computed using GLS-detrending methods.⁸³ The simulation runs of the test demonstrated that it is suitable to be applied in small sample cases. Therefore, we prefer to employ Carrion-i-Silvestre et al.⁸³ unit root test in the present study

Cointegration test

Conventional cointegration tests that identify the long-run relationship like Engle and Granger⁹⁷ and Johansen⁹⁸ are often criticized for not taking potential structural breaks into consideration. In response, Gregory and Hansen⁷⁹ introduced a new cointegration test which lets series to contain a single structural break. Hatemi-J⁹⁹ improved these tests by enabling two structural breaks to be included in each series. According to Maki⁸⁴ the main problem behind the mentioned tests is the necessity to include a specified number of breaks without having any priori information about the number of breaks. In other words, if a series is supposed to contain more than two breaks due to its volatile past values, then both Gregory and Hansen⁷⁹ and Hatemi-J⁹⁹ cointegration tests would perform poorly. Hence, Maki⁸⁴ proposes a residual based cointegration test which allows series to have structural breaks up to five. Monte Carlo simulations confirmed that when there are more than three structural breaks for a given cointegration relationship, the test of Maki⁸⁴ performs better than the others. The test can be conducted under four models which are written in a general regression equation as follows:

y_{t} = λ + \sum_{i = 1}^{b} λ_{i} P_{i, t} + α^{'} x_{t} + ε_{t}

(13)

y_{t} = λ + \sum_{i = 1}^{b} λ_{i} P_{i, t} + α^{'} x_{t} + \sum_{i = 1}^{b} α_{i}^{'} x_{t} P_{i, t} + ε_{t}

(14)

y_{t} = λ + \sum_{i = 1}^{b} λ_{i} P_{i, t} + γ^{t} + α^{'} x_{t} + \sum_{i = 1}^{b} α_{i}^{'} x_{t} P_{i, t} + ε_{t}

(15)

y_{t} = λ + \sum_{i = 1}^{b} λ_{i} P_{i, t} + γ^{t} + \sum_{i = 1}^{b} γ_{i} t P_{i, t} + α^{'} x_{t} + \sum_{i = 1}^{b} α_{i}^{'} x_{t} P_{i, t} + ε_{t}

(16)where y_t and x_t denote variables with their first differences, b gives the maximum number of breaks and P_i,t equals to 1 if t is larger than the time period of the breaks. Equation 13 presents level shifts (model 0), equation 14 presents regime shifts model (model 1), equation 15 presents regime shifts model with trend (model 2), and equation 16 presents structural breaks of levels, trends, and regressors (model 3). Given that Turkey had experienced frequent economic downturns and recoveries in its past, we expect that the series of the country contain a high number of structural breaks. Therefore, this study employs Maki⁸⁴ cointegration test under the null hypothesis of no cointegration through equations 4 to 8.

Estimations of the long-Run coefficients

After the long-run equilibrium relationship is identified, the long-run coefficient estimations are utilized by employing the Fully Modified Ordinary Least Squares (FMOLS) technique. The FMOLS method has initially proposed by Phillips and Hansen¹⁰⁰ to address the issues associated with the statistical inference in integrated processes. By modifying the Wald statistics through semiparametric corrections for serial correlation and endogeneity, the test statistics of the FMOLS corrects the cointegrated models by removing the nuisance parameter dependencies. This methodology also asymptotically removes the sample bias as long as the variables in a cointegration system are integrated at order one.¹⁰¹ The econometric equation model of FMOLS can be written as:

y_{t} = δ_{0} + δ_{1}^{'} X_{t} + ε_{t}, t = 1, 2, \dots ., n

(17)where y_t is a variable with an order of integration one and X_t is a (kx1) vector of non-cointegrating regressors.

During the long-run coefficient estimation of this study, the structural breaks found by the conducted cointegration test are inserted into the FMOLS as dummy variables.

Since the EKC framework hypothesizes an inverted U-shaped curve, the level of income per capita where the CO₂ emissions are the highest can be calculated.^18,102 This point is often referred as the “threshold” or “turning” point which can be derived as;

\partial = \exp [- λ_{1_{Y}} / (2 λ_{2_{Y^{2}}})]

(18)where ∂ stands for the level of income per capita,

λ_{1_{Y}}

and

λ_{2_{Y^{2}}}

shows the estimated long-run coefficients of GDP and GDP2.

Empirical results

We examined the integration properties of the variables by the Carrion-i-Silvestre et al.⁸³ unit root test Table 2 presents the outcome of the unit root test under five structural breaks. The upper panel of the table shows that the null hypothesis of non-stationarity is not rejected when the series are tested at their level forms. The series are tested one more time after their first differences are taken. In this case, it is revealed that the null hypothesis is rejected for all series as the test statistics of P_t, MP_t, MZ_α, MBT and MZ_t turn out to be statistically significant at five percent. Therefore, we determined that the variables under investigation are integrated order one, I(1).

Table 2.

Results of the quasi-GLS based unit root test.

Variable	Levels					Break Points
	P_T	MP_T	MZ_α	MSB	MZ_t
lnGDP	18.36 [8.55]	17.29 [8.55]	−22.06 [-44.03]	0.14 [0.10]	-3.30 [-4.69]	1977; 1982; 1990; 2000; 2007
lnGDP²	18.67 [8.55]	17.54 [8.55]	-21.76 [-44.03]	0.15 [0.10]	-3.28 [-4.69]	1977; 1982; 1990; 2000; 2007
lnEN	18.90 [8.58]	17.29 [8.58]	-23.07 [-45.65]	0.14 [0.10]	-3.39 [-4.77]	1965; 1978; 1993; 1998; 2007
lnFD	24.59 [9.01]	21.50 [9.01]	-19.34 [-45.81]	0.16 [0.10]	-3.11 [-4.77]	1973; 1980; 1985; 1997; 2002
lnINF	17.54 [8.84]	16.80 [8.84]	-24.86 [-46.10]	0.14 [0.10]	-3.51 [-4.80]	1968; 1978; 1993; 1998; 2007
lnCO₂	20.10 [8.45]	17.07 [8.45]	-22.14 [-43.79]	0.14 [0.10]	-3.30 [-4.66]	1977; 1987; 1994; 1999; 2004
lnGDP_INF	22.31 [9.02]	18.71 [9.02]	-22.33 [-45.59]	0.15 [0.10]	-3.32 [-4.76]	1974; 1979; 1991; 1996; 2005
lnEN_INF	22.61 [8.94]	19.86 [8.94]	-20.54 [-45.15]	0.16 [0.10]	-3.20 [-4.74]	1974; 1979; 1992; 2000; 2007
lnFD_INF	26.43 [9.13]	23.02 [9.13]	-18.28 [-45.65]	0.16 [0.10]	-3.02 [-4.76]	1974; 1980; 1985; 1997; 2004
	First Differences

ΔlnGDP	3.98* [5.54]	3.81* [5.54]	-25.69* [-17.32]	0.13* [0.16]	-3.53* [-2.89]	-
ΔlnGDP²	3.95* [5.54]	3.84* [5.54]	-25.77* [-17.32]	0.13* [0.16]	-3.53* [-2.89]	-
ΔlnEN	3.61* [5.54]	3.53* [5.54]	-25.85* [-17.32]	0.13* [0.16]	-3.59* [-2.89]	-
ΔlnFD	3.95* [5.54]	4.04* [5.54]	-23.00* [-17.32]	0.14* [0.16]	-3.37* [-2.89]	-
ΔlnINF	3.50* [5.54]	3.61* [5.54]	-25.62* [-17.32]	0.13* [0.16]	-3.57* [-2.89]	-
ΔlnCO₂	3.85* [5.54]	3.67* [5.54]	-25.72* [-17.32]	0.13* [0.16]	-3.56* [-2.89]	-
ΔlnGDP_INF	4.09* [5.54]	3.91* [5.54]	-25.47* [-17.32]	0.13* [0.16]	-3.51* [-2.89]	-
ΔlnEN_INF	3.73* [5.54]	3.61* [5.54]	-25.63* [-17.32]	0.13* [0.16]	-3.57* [-2.89]	-
ΔlnFD_INF	3.89* [5.54]	4.00* [5.54]	-23.43* [-17.32]	0.14* [0.16]	-3.40* [-2.89]	-

Note: Structural break points are acquired by applying the quasi GLS-based unit root tests by Carrion-i-Silvestre et al.⁸³ * symbolizes the rejection of the null hypothesis at 5% alpha level. Critical values are given in brackets.

After identifying the stationarity properties of the series, we employed Maki⁸⁴ cointegration test to detect possible equilibrium association amongst the variables. The test is conducted for the models that are specified through the equations 4 to 8. Table 3 presents the test statistics, critical values and structural break points for each model specification of Maki⁸⁴ The results of the cointegration test provide evidence for long-run equilibrium relationship. The structural break points acquired from the Maki⁸⁴ test can be used for the process of long-term coefficient estimations which is carried out by the FMOLS approach.

Table 3.

Results of the cointegration test.

	Test Statistics [Critical Values]	Structural Break Occurances
Model 1
Model 0	−6.26** [-5.76]	1971; 1978; 1990; 2000; 2007
Model 1	-6.88* [-6.53]	1971; 1980; 1988; 2000; 2006
Model 2	-8.70* [-7.84]	1966; 1972; 1980; 1990; 2000
Model 3	-12.54* [-8.71]	1970; 1977; 1984; 1992; 2006
Model 2
Model 0	-6.25** [-6.04]	1966; 1980; 1988; 1998; 2006
Model 1	-7.01* [-6.25]**	1970; 1980; 1988; 2000; 2006
Model 2	-8.34* [-8.11]	1969; 1978; 1988; 1998; 2007
Model 3	-11.57* [-8.80]	1970; 1978; 1984; 1992; 2007
Model 3
Model 0	-6.86* [-6.85]	1971; 1978; 1990; 2000; 2007
Model 1	-7.52* [-7.05]	1971; 1980; 1988; 1994; 2006
Model 2	-10.15* [-9.44]	1969; 1980; 1990; 2000; 2007
Model 3	-10.14* [-10.08]	1969; 1980; 1990; 2000; 2007
Model 4
Model 0	-6.80** [-6.31]	1971; 1978; 1988; 1995; 2001
Model 1	-7.97* [-7.05]	1971; 1978; 1988; 1999; 2005
Model 2	-12.09* [-9.44]	1966; 1969; 1980; 1992; 2002
Model 3	-12.57* [-10.08]	1971; 1978; 1984; 1993; 2005
Model 5
Model 0	-7.35* [-6.85]	1971; 1978; 1984; 1992; 2005
Model 1	-7.93* [-10.98]	1971; 1984; 1992; 2000; 2007
Model 2	-12.28* [-9.44]	1967; 1976; 1988; 1995; 2002
Model 3	-12.17* [-10.08]	1970; 1980; 1987; 1996; 2004

Note: ⁱCritical values at 5% provided by Maki⁸⁴ and given in the corner brockets. * denotes the rejection of the null hypothesis at 5%; while ** denotes the rejection of the null hypothesis at 10%.

Table 4 demonstrates the long-term coefficient estimation results of the FMOLS test To observe the impact of each independent variable on the conventional EKC framework, five separate models have been produced through the testing process. Model 1 shows the results of the conventional EKC hypothesis in which GDP and GDP squared are regressed with the CO₂ emissions. In this basic model, it is revealed that GDP conveys a positive while GDP squared has a negative effect on CO₂ emissions. In other words, the inverted U-shaped EKC holds when it is tested through the conventional variables. Energy consumption variable is included in model 2 which reveals that CO₂ emissions are anticipated to increase as the level of energy consumption rises. Model 3 demonstrates that the coefficient of the financial development turns out to be negative and statistically significant at 5% level. As the informal economic activities are taken into consideration in model 4, it is shown that the financial development coefficient becomes positive. The effect of the informal economy on global emissions is also found to be positive. Finally, in model 5, the moderating role of the informal economic activities through GDP, energy consumption and financial development on the global emissions are estimated. As reported, lnGDP_INF has a negative coefficient and indicating that the official economy is decreased by the informal one which in turn leads CO₂ emissions to be reduced. The moderating role of the informal economy through financial development has also found to be statistically significant, and it produced a similar result with the case of GDP. The negative coefficient of the lnFD_INF implies that financial development is decreased by the informal economic activities and as a result, CO₂ emissions are expected to be decreased. For the lnENG_INF variable, the informal economy has a moderating role through the level of energy consumption and the effect is positive. In this case, the informal economic activities promote greater levels of energy use and consequently CO₂ emissions also increase.

Table 4.

Results of the FMOLS estimations.

Dependent Variable: LnCO₂	Model 1			Model 2			Model 3			Model 4			Model 5
Independent Variables	Coefficient	Std. Error	P-value	Coefficient	Std. Error	P-value	Coefficient	Std. Error	P-value	Coefficient	Std. Error	P-value	Coefficient	Std. Error	P-value
Intercept	-11.989**	5.151	0.024	-10.400**	4.379	0.022	-24.399*	4.005	0.000	-19.615*	3.715	0.000	-31.274*	3.070	0.000
Trend	0.030*	0.005	0.000	0.034*	0.004	0.000	0.049*	0.003	0.000	0.072*	0.005	0.000	0.078*	0.003	0.000
lnGDP	5.573*	1.559	0.000	4.275*	1.359	0.003	7.725*	0.870	0.000	5.935*	0.987	0.000	12.127*	1.091	0.000
lnGDP²	-0.321*	0.117	0.008	-0.250**	0.101	0.017	-0.455*	0.077	0.000	-0.382*	0.072	0.000	-0.987*	0.096	0.000
lnEN				0.452*	0.143	0.002	0.596**	0.288	0.045	0.569*	0.096	0.000	0.676*	0.143	0.000
lnFD							-0.158**	0.062	0.015	0.105*	0.023	0.000	0.045	0.038	0.239
lnINF										4.592*	0.798	0.000	14.535*	1.786	0.000
lnGDP_INF													-2.233*	0.265	0.000
lnENG_INF													0.282***	0.145	0.059
lnFD_INF													-0.087***	0.044	0.055
D1	0.055***	0.033	0.098	0.031	0.028	0.273	0.052	0.059	0.386	0.036**	0.017	0.038	0.045*	0.007	0.000
D2	0.111*	0.035	0.003	0.028	0.028	0.319	-0.141**	0.061	0.025	-0.050*	0.017	0.006	-0.009	0.007	0.186
D3	-0.045	0.033	0.174	-0.026	0.027	0.343	0.017	0.057	0.758	-0.017	0.016	0.304	-0.002	0.006	0.775
D4	-0.003	0.032	0.919	-0.005	0.027	0.850	0.026	0.058	0.656	0.007	0.016	0.651	-0.011***	0.006	0.092
D5	-0.017	0.032	0.606	-0.015	0.029	0.602	-0.091	0.058	0.127	-0.035**	0.017	0.051	-0.049*	0.007	0.000
R²	0.998			0.998			0.996			0.999			0.999
Adj. R²	0.997			0.997			0.995			0.999			0.999
S.E. of Reg.	0.037			0.030			0.055			0.020			0.022
Long run variance	0.0010			0.0007			0.0031			0.0002			0.0000

Note: *** and ** and * indicate rejection of null hypothesis at 10% and 5% and 1% significance level respectively. The letter “D” denotes for dummy variables that are acquired from the Maki⁸⁴ cointegration test's structural break points. In all six estimation, Hannan-Quinn Information Criteria is used.

Discussion of the results

Impact of energy consumption, financial development and informal economy on CO₂ emissions can be analyzed by the estimated coefficients under model 4. Consistent with the previous models, GDP and GDP squared variables identified as positive and negative, respectively. This implies that as the Turkish economy grows, CO₂ emissions will increase until a certain threshold level is reached. This threshold level is calculated to be $2364.51 (constant 2010$) per capita. After this level, further economic growth is expected to make a negative impact on CO₂ emissions and reduce the environmental pollution. In line with the theoretical structure of the EKC hypothesis, empirical evidence found in this study can be inferred that economic policies in Turkey tend to ignore the environmental hazards of fast economic growth during earlier industrial stage. As the economic growth continues, pollutant-induced industrial stage is expected to make a transition toward technologically improved and more efficient production, which in turn reduces the global pollution. This phenomenon has also been verified by number of previous studies for the case of Turkey (e.g..^78,85,87,90 The results produced by model 4 also show that the coefficient of energy consumption is positive which indicates an increase in energy consumption level leads to an increase in CO₂ emissions. A feasible explanation behind this particular result is fossil fuel resources play a dominant role in Turkey's energy production. Halicioglu¹⁰³ Yavuz⁸⁰ and Isiksal, Samour and Resatoglu¹⁰⁴ also observed the same impact of energy consumption on CO₂ emissions and argued that Turkey should diversify its energy resources by investing more on renewable energy alternatives. Model 4 further shows that both financial development and informal economy variables contribute the CO₂ emissions positively. It can be inferred that the financial sector improvement in Turkey is not environmentally efficient and the accessible funds through the financial market are invested in carbon intensive projects. This outcome is in line with the results of Pata⁹⁰ but it is contradictory to the findings of Dogan and Seker¹⁰⁵ since a negative coefficient estimated between financial development and CO₂ for Turkey. One of the possible reasons behind these contradictive results is the difference in data ranges used for the analyses. While the current study employs a time span from 1960 to 2013, Dogan and Seker¹⁰⁵ used a time span from 1985 to 2011. Another plausible reason is the proxy preferences for the financial development variable. As it was mentioned in section 2.2, the current study uses a compound financial development index which aims to reflect the development of the financial sector completely. However, the study of Dogan and Seker¹⁰⁵ measures the financial development by “domestic credits to private sector” only. It is also found that the informal economy has a deteriorating effect on the environment. The operations of underground companies are not subject to any environmental rules and regulations of the governments. Production of pollution intensive goods and services such as metals, chemicals, and illegal urban transportation might be among the reasons behind the estimated coefficient.

Model 5 is the extension of model 4 which includes the moderating effect variables; lnGDP_INF, lnENG_INF and lnFD_INF. Coefficients obtained from common variables in both models show consistency as they retain their signs and significances. Model 5 estimates that the coefficient signs of lnGDP_INF and lnFD_INF are negative. To put it another way, CO₂ emissions are reduced because of the informal economy's interaction with the formal economy and financial sector development. Considering the income growth and financial development increase the level of global pollution, it can be assumed that these variables are negatively affected by the activities of the informal economy. Hence, as the informal economy's exposure on GDP and financial sector surges, the level of CO₂ emissions decreases in the country. This interaction relationship is in parallel with the arguments of Loayza⁶¹ Schneider and Enste⁵⁹ According to their view, the informal economy makes public services less available for everyone in an economy and it adversely affects economic growth.

Conclusion

Based on the empirical ambiguity in the literature of environmental economics, this study examines the long-run equilibrium between economic growth, energy use, development of the financial sector and informal economy. Although the linkage between income growth and environmental degradation has investigated by a vast number of papers, the role of the informal economy has not captured in most of them. Current study doesn't only include this factor into account, but it also examines the moderating role of the informal economy on global emission level for the case of Turkey. Constructed empirical models are based on the EKC framework which has generated essential conclusions on the dynamic relationships between income growth, energy, informal economy, financial development, and global pollution in the case of Turkey. The estimations acquired from the Maki⁸⁴ cointegration test have identified the long-term equilibrium among the variables. After the equilibrium linkage is determined, the long-term coefficients have generated by using the FMOLS methodology. Starting from the conventional EKC model, five different models have constructed to analyze the impact of the variables in the EKC setting. The long-term coefficient results have supported the EKC hypothesis by providing evidence on the increasing level of CO₂ emissions at the initial stages of income growth which is then followed by reduced emissions at the later phases of income growth. The test results have also shown that financial development and the informal economy in Turkey produce higher CO₂ emissions. Moreover, the moderating role of the informal economy through financial development and income growth has verified to be negative on the CO₂ emissions which highlight the deteriorating influence of the informal economy on the official economy and financial sector. Finally, the moderating effect of the informal economy on energy consumption has demonstrated to be positive and leading to higher CO₂ emissions.

Policy recommendations

In the existence of the given empirical evidence, Turkish authorities might consider employing several policies to achieve a sustainable economy and contain environmental degradation. Turkey is in a very rich geographical location in terms of producing solar, wind, geothermal and hydroelectric energy. Hence, hazardous influences of carbon-based energy can be minimized by encouraging investments in renewable alternatives. In addition to environmental benefits, renewable energy production might aid the economy by achieving energy diversification. The study of Dogan, Taspinar and Gokmenoglu¹⁰⁶ estimated that renewable energy consumption decreases environmental pollution for the case of Turkey. Policymakers may stimulate the deployment of alternative energy production facilities by granting incentives for such investments. For example, tax credits and green loans may help the investors in establishing new energy systems.

Decreasing the total amount of fossil fuel in the energy market enables higher resistance against price shocks. Financial sector might play a crucial role in governments’ attempts to reduce global pollution. In this regard, sustainable and environmentally friendly projects should be implemented without financial barriers preventing feasible projects to emerge. Entrepreneurs of such projects can be aided by government guarantees which allocate risk in privately financed infrastructure ventures. Moreover, financial market instruments can be used as a tool for mitigation of risk and reducing the cost of green investments. For example, `climate bonds` can be prominently promoted by the Turkish regulators to achieve an effective flow of funds toward sustainable projects.

Presented empirical outcome indicated that economic growth is another major concern for global pollution. Relevant measures should be enforced by the government to achieve sustainable economic growth. In the short term, industrial production can be carried out safer by regulating the level of maximum emissions that factories generate. A set of technical standards should be established for each energy-intense industry. Specifying the type of filters that must be installed by factories and achieving efficient industrial waste management might improve the global environmental performance of the country without jeopardizing its economic growth. In the long term, technological innovation of energy production must be targeted by the policymakers. Consistent innovation policies may play an important role in accomplishing successful transition from traditional fossil-based production to technologically advanced environment friendly production. Supporting research and development programs, establishing institutions for energy technologies, funding clean energy production projects and transfer of knowledge can be used as strong policy tools to promote sustainable economic growth.

During the design process of the environmental policies, the impact of the informal economy shouldn't be underestimated. As the test results of the current study have highlighted, unrecorded economic activities lead to higher CO₂ emissions and it also adversely affects the economic performance of the country. There are several policy options to mitigate environmental adversities of the informal economic activities. First, is employing a green subsidy policy which incentivizes firms to use clean energy inputs. Granting such a subsidy is expected to decrease production costs which in turn may encourage previously unregistered firms to shift into official economy.¹⁰⁷ Second, is to pursue investment policies towards technologies in industrial production. It is argued that technological innovations promote environmental sustainability as the production process minimizes pollutant wastes. Advancements in production should also improve cost efficiency. Therefore, firms including the ones operating under informal economy, will be willing to abandon obsolete production methods and adapt cleaner and cheaper alternatives as those alternatives are available.¹⁰⁸ Finally, effective enforcement of labor regulations can be vital in preventing both employers and employees to operate informally. Considering the likelihood that informal firms are ignoring the environmental regulations, political authorities must incentivize working under registered firms. In this respect, governmental actions such as improving the employment protection laws can make workers to feel more secure and motivate them to work under registered firms.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

ORCID iD

Baris Memduh Eren

Salih Katircioglu

Notes

Baris Memduh Eren is an Assistant Professor of Finance in the Department of Finance and Banking at University of Kyrenia, North Cyprus. He received his PhD in Finance from Eastern Mediterranean University, North Cyprus in 2019. His teaching and research interests lie on Energy and Environmental Economics, Energy Finance, Applied Financial Statistics and Applied Econometrics. He has published several research papers in international peer-reviewed journals. Amongst these include Science of the Total Environment, Environmental Science and Pollution Research, The Journal of International Trade & Economic Development.

Salih KATIRCIOGLU is Professor of Economics at Eastern Mediterranean University, Famagusta, Northern Cyprus, Via Mersin 10, Turkey. He graduated from the same institution and earned his Ph.D. from Uludag University, Turkey. His research interests include applied time series econometrics, international trade, international finance, tourism economics, and economic growth issues. He extensively publishes articles in the journals such as Tourism Management, The World Economy, Applied Economics, Renewable & Sustainable Energy Reviews, Economic Modeling, Energy, and Energy Policy among others.

Korhan K Gokmenoglu graduated from Middle East Technical University, Brandeis University, Essec University and worked as visiting scholar at University of California San Diego. Currently, he is an Associate Professor at the Department of Banking and Finance in the Eastern Mediterranean University. His research field focuses on applied econometrics, energy economics, environmental degradation, and macroeconomics. His research studies have been published in many international peer-reviewed journals, including Computation and Economic Cybernetics Studies and Research, Croatian Operational Research Review, Economic Systems, Energy, Energy & Environment, Environmental Science and Pollution Research, Science of the Total Environment, Singapore Economic Review, The Journal of International Trade & Economic Development, and International Trade & Economic Development. Also, he has served as a referee for many prestigious academic journals.

Appendix

Glossary of Terms

References

WMO, (2018). Greenhouse gas bulletin: November 2018. Retrieved from. https://library.wmo.int/index.php?lvl=notice_display&id=3030#.XLcHCIkzbcd (accessed: 20 October 2019).

GISTEMP Team, 2018: GISS surface temperature analysis (GISTEMP). NASA Goddard Institute for Space Studies. Retrieved from. https://data.giss.nasa.gov/gistemp.

IPCC, (2018). Summary for policymakers: October 2018. Retrieved from. https://www.ipcc.ch/site/assets/uploads/sites/2/2018/07/SR15_SPM_version_stand_alone_LR.pdf.

UNFCCC, (2019). Understanding the UN climate change regime: October 2018. Retrieved from. https://unfccc.int/resource/bigpicture/#content-the-paris-agreemen.

Tiris

Atagunduz

Dincer

. Energy, economy and environment modelling: applications for Turkey. Energy 1994; 19: 1005–1009.

Hawdon

Pearson

. Input-output simulations of energy, environment, economy interactions in the UK. Energy economics 1995; 17: 73–86.

ZhiDong

. An econometric study on China's Economy, energy and environment to the year 2030. Energy Policy 2003; 31: 1137–1150.

Oliveira

Antunes

. A multiple objective model to deal with economy - energy - environment interactions. Eur J Oper Res 2004; 153: 370–385.

Kapusuzoglu

. Causality relationships between carbon dioxide emissions and economic growth: results from a multi-country study. International Journal of Economic Perspectives 2014; 8: 5–15.

10.

Anatasia

. The causal relationship between GDP, exports, energy consumption, and CO₂ in Thailand and Malaysia. International Journal of Economic Perspectives 2015; 9: 37–48.

11.

Price

Keppo

. Modelling to generate alternatives: a technique to explore uncertainty in energy-environment-economy models. Appl Energy 2017; 195: 356–369.

12.

Demir

Cergibozan

Gök

. Income inequality and CO₂ emissions: empirical evidence from Turkey. Energy & Environment 2019; 30: 444–461.

13.

Kuznets

. Economic growth and income inequality. The American Economic Review 1955; 45: 1–28.

14.

Agras

Chapman

. A dynamic approach to the environmental kuznets curve hypothesis. Ecol Econ 1999; 28: 267–277.

15.

Katircioglu

. Investigating the role of oil prices in the conventional EKC model: evidence from Turkey. Asian Economic and Financial Review 2017; 7: 498–508.

16.

Dinda

. Environmental kuznets curve hypothesis: a survey. Ecol Econ 2004; 49: 431–455.

17.

Tao

Zheng

Lianjun

. An empirical test of the environmental kuznets curve in China: a panel cointegration approach. China Economic Review 2008; 19: 381–392.

18.

Stern

. The rise and fall of the environmental kuznets curve. World Dev 2004; 32: 1419–1439.

19.

Grossman

G. M.

Krueger

A. B.

(1991). Environmental impacts of a north American free trade agreement (No. w3914). Cambridge: National Bureau of Economic Research.

20.

Panayotou

(1993). Empirical tests and policy analysis of environmental degradation at different stages of economic development (No. 992927783402676). Geneva 22, Switzerland: International Labour Organization.

21.

Selden

Song

. Environmental quality and development: is there a kuznets curve for air pollution emissions? J Environ Econ Manage 1994; 27: 147–162.

22.

Moomaw

Unruh

. Are environmental kuznets curves misleading us? The case of CO 2 emissions. Environ Dev Econ 1997; 2: 451–463.

23.

Lindmark

. An EKC-pattern in historical perspective: carbon dioxide emissions, technology, fuel prices and growth in Sweden 1870–1997. Ecol Econ 2002; 42: 333–347.

24.

Managi

Jena

. Environmental productivity and kuznets curve in India. Ecol Econ 2008; 65: 432–440.

25.

Ozturk

Acaravci

. CO₂ Emissions, energy consumption and economic growth in Turkey. Renewable Sustainable Energy Rev 2010; 14: 3220–3225.

26.

Pao

Tsai

. Modeling and forecasting the CO₂ emissions, energy consumption, and economic growth in Brazil. Energy 2011; 36: 2450–2458.

27.

Aslan

Destek

Okumus

. Bootstrap rolling window estimation approach to analysis of the environment kuznets curve hypothesis: evidence from the USA. Environmental Science and Pollution Research 2018; 25: 2402–2408.

28.

Gokmenoglu

Taspinar

. Testing the agriculture-induced EKC hypothesis: the case of Pakistan. Environmental Science and Pollution Research 2018; 25: 22829–22841.

29.

Katircioglu

Gokmenoglu

Eren

. Testing the role of tourism development in ecological footprint quality: evidence from top 10 tourist destinations. Environmental Science and Pollution Research 2018; 25: 33611–33619.

30.

Zilio

Recalde

. GDP And environment pressure: the role of energy in latin america and the Caribbean. Energy Policy 2011; 39: 7941–7949.

31.

Arouri

MEH

Youssef

M'henni

, et al. Energy consumption, economic growth and CO₂ emissions in Middle East and north african countries. Energy Policy 2012; 45: 342–349.

32.

Chandran

VGR

Tang

. The impacts of transport energy consumption, foreign direct investment and income on CO₂ emissions in ASEAN-5 economies. Renewable Sustainable Energy Rev 2013; 24: 445–453.

33.

Ozcan

. The nexus between carbon emissions, energy consumption and economic growth in Middle East countries: a panel data analysis. Energy Policy 2013; 62: 1138–1147.

34.

Al-Mulali

Weng-Wai

Sheau-Ting

, et al. Investigating the environmental kuznets curve (EKC) hypothesis by utilizing the ecological footprint as an indicator of environmental degradation. Ecol Indic 2015; 48: 315–323.

35.

Baek

. Environmental kuznets curve for CO₂ emissions: the case of Arctic countries. Energy Economics 2015; 50: 13–17.

36.

Ganda

. Carbon emissions, diverse energy usage and economic growth in South Africa: investigating existence of the environmental kuznets curve (EKC). Environ Prog Sustain Energy 2019; 38: 30–46.

37.

Perman

Stern

. Evidence from panel unit root and cointegration tests that the environmental kuznets curve does not exist Australian Journal of Agricultural and Resource Economics 2003; 47: 325–347.

38.

Frankel

Romer

. Does trade cause growth? American Economic Review 1999; 89: 379–399.

39.

Lundgren

. A real options approach to abatement investments and green goodwill. Environmental and Resource Economics 2003; 25: 17–31.

40.

Stern

. Biomass and China's Carbon emissions: a missing piece of carbon decomposition. Energy Policy 2008; 36: 2517–2526.

41.

Sadorsky

. Financial development and energy consumption in central and eastern european frontier economies. Energy Policy 2011; 39: 999–1006.

42.

Katircioğlu

Taşpinar

. Testing the moderating role of financial development in an environmental kuznets curve: empirical evidence from Turkey. Renewable Sustainable Energy Rev 2017; 68: 572–586.

43.

Tamazian

Rao

. Do economic, financial and institutional developments matter for environmental degradation? Evidence from transitional economies. Energy Economics 2010; 32: 137–145.

44.

Jalil

Feridun

. The impact of growth, energy and financial development on the environment in China: a cointegration analysis. Energy Economics 2011; 33: 284–291.

45.

Shahbaz

Hye

QMA

Tiwari

, et al. Economic growth, energy consumption, financial development, international trade and CO₂ emissions in Indonesia. Renewable Sustainable Energy Rev 2013; 25: 109–121.

46.

Shahbaz

Tiwari

Nasir

. The effects of financial development, economic growth, coal consumption and trade openness on CO₂ emissions in South Africa. Energy Policy 2013; 61: 1452–1459.

47.

Bekhet

Matar

Yasmin

. CO₂ Emissions, energy consumption, economic growth, and financial development in GCC countries: dynamic simultaneous equation models. Renewable Sustainable Energy Rev 2017; 70: 117–132.

48.

Saidi

Mbarek

. The impact of income, trade, urbanization, and financial development on CO₂ emissions in 19 emerging economies. Environmental Science and Pollution Research 2017; 24: 12748–12757.

49.

Abbasi

Riaz

. CO₂ Emissions and financial development in an emerging economy: an augmented VAR approach. Energy Policy 2016; 90: 102–114.

50.

Shahzad

SJH

Kumar

Zakaria

, et al. Carbon emission, energy consumption, trade openness and financial development in Pakistan: a revisit. Renewable Sustainable Energy Rev 2017; 70: 185–192.

51.

Sheng

Vochozka

. Pollution caused by finance and the relative policy analysis in China. Energy & Environment 2017; 28: 808–823.

52.

Cetin

Bakirtas

. The long-run environmental impacts of economic growth, financial development, and energy consumption: evidence from emerging markets. Energy & Environment 2020; 31: 634–655.

53.

Ozturk

Acaravci

. The long-run and causal analysis of energy, growth, openness and financial development on carbon emissions in Turkey. Energy Economics 2013; 36: 262–267.

54.

Imamoglu

. (2016), Re-estimation of the size of underground economy in european countries: mIMIC approach, International Journal of Economic Perspectives, Volume 10, Issue , pp. 171–193.

55.

Imamoglu

Katircioglu

Payaslioglu

C.E

. (2018), Financial services spillover effects on informal economic activity. The Services Industries Journal, Volume 38, Issue , pp. 669–687.

56.

Williams

C. C

. (2008), Cross-National variations in undeclared work: results From a survey of 27 european countries, International Journal of Economic Perspectives, Volume 2, Issue , pp. 46–63.

57.

Williams

C. C

. (2011), Blurring the formal/informal economy divide: beyond a dual economies approach, International Journal of Economic Perspectives, Volume 5, Issue , pp. 347–362.

58.

Smith

P. M

. The underground economy: Global evidence of its size and impact. Canada: The Fraser Institute, 1997.

59.

Schneider

Enste

. Shadow economies: size, causes, and consequences. J Econ Lit 2000; 38: 77–114.

60.

Baksi

Bose

. Environmental regulation in the presence of an informal sector. University of Winnipeg Department of Economics Working Paper 2010; 3: 1–28.

61.

Loayza

. The economics of the informal sector: a simple model and some empirical evidence from latin america. North-Holland: The World Bank, 1999.

62.

Biswas

Farzanegan

Thum

. Pollution, shadow economy and corruption: theory and evidence. Ecol Econ 2012; 75: 114–125.

63.

Abid

. The close relationship between informal economic growth and carbon emissions in Tunisia since 1980: the (ir)relevance of structural breaks. Sustainable Cities and Society 2015; 15: 11–21.

64.

Berdiev

Saunoris

. Financial development and the shadow economy: a panel VAR analysis. Econ Model 2016; 57: 197–207.

65.

Capasso

Jappelli

. Financial development and the underground economy. J Dev Econ 2013; 101: 167–178.

66.

Elgin

Uras

. Is informality a barrier to financial development? SERIEs: Journal of the Spanish Economic Association 2013; 4: 309–331.

67.

Blackburn

Bose

Capasso

. Tax evasion, the underground economy and financial development. Journal of Economic Behavior and Organization 2012; 83: 243–253.

68.

Bose

Capasso

Wurm

. The impact of banking development on the size of shadow economies. Journal of Economic Studies 2012; 39: 620–638.

69.

CCPI, (2019). Country scorecard statistics. Retrieved from. https://www.climate-change-performance-index.org/sites/default/files/documents/turkey_scorecard_ccpi_2019.pdf.

70.

World Bank Development Database, (2019). World development indicators. Retrieved from. http://databank.worldbank.org

71.

Gokmenoglu

Kirikkaleli

Eren

. Time and frequency domain causality testing: the causal linkage between FDI and economic risk for the case of Turkey. The Journal of International Trade & Economic Development 2019; 28(6): 649–667.

72.

OECD, (2018). OECD economic surveys: Turkey. July 2018. Retrieved from. http://www.oecd.org/eco/surveys/Turkey-2018-OECD-economic-survey-overview.pdf.

73.

Savasan

. Modeling the underground economy in Turkey: randomized response and MIMIC models. The Journal of Economics 2003; 29: 49–76.

74.

Schneider

Savasan

. Dymimic estimates of the size of shadow economies of Turkey and of her neighbouring countries. International Research Journal of Finance and Economics 2007; 9: 126–143.

75.

Elgin

Öztunalı

. Environmental kuznets curve for the informal sector of Turkey (1950–2009). Panoeconomicus 2014; 61: 471–485.

76.

Imamoglu

. Is the informal economic activity a determinant of environmental quality? Environmental Science and Pollution Research 2018; 25: 29078–29088.

77.

Lise

. Decomposition of CO₂ emissions over 1980–2003 in Turkey. Energy Policy 2006; 34: 1841–1852.

78.

Shahbaz

Ozturk

Afza

, et al. Revisiting the environmental kuznets curve in a global economy. Renewable Sustainable Energy Rev 2013; 25: 494–502.

79.

Gregory

Hansen

. Residual-based tests for cointegration in models with regime shifts. J Econom 1996; 70: 99–126.

80.

Yavuz

. CO₂ Emission, energy consumption, and economic growth for Turkey: evidence from a cointegration test with a structural break. Energy Sources, Part B: Economics, Planning, and Policy 2014; 9: 229–235.

81.

Bölük

Mert

. The renewable energy, growth and environmental kuznets curve in Turkey: an ARDL approach. Renewable Sustainable Energy Rev 2015; 52: 587–595.

82.

de Vita

Katircioglu

Altinay

, et al. Revisiting the environmental kuznets curve hypothesis in a tourism development context. Environmental Science and Pollution Research 2015; 22: 16652–16663.

83.

Carrion-i-Silvestre

Kim

Perron

. GLS-based unit root tests with multiple structural breaks under both the null and the alternative hypotheses. Econ Theory 2009; 25: 1754–1792.

84.

Maki

. Tests for cointegration allowing for an unknown number of breaks. Econ Model 2012; 29: 2011–2015.

85.

Dogan

. Agriculture and environmental kuznets curves in the case of Turkey: evidence from the ARDL and bounds test Agricultural Economics 2016; 62: 566–574.

86.

Gökmenoğlu

Taspinar

. The relationship between CO₂ emissions, energy consumption, economic growth and FDI: the case of Turkey. J Int Trade Econ Dev 2016; 25: 706–723.

87.

Gozgor

Can

. Export product diversification and the environmental kuznets curve: evidence from Turkey. Environmental Science and Pollution Research 2016; 23: 21594–21603.

88.

Katircioglu

. Testing the role of urban development in the conventional environmental kuznets curve: evidence from Turkey. Appl Econ Lett 2018; 25: 741–746.

89.

Ozcan

Apergis

Shahbaz

. A revisit of the environmental kuznets curve hypothesis for Turkey: new evidence from bootstrap rolling window causality. Environmental Science and Pollution Research 2018; 25: 32381–32394.

90.

Pata

. Renewable energy consumption, urbanization, financial development, income and CO₂ emissions in Turkey: testing EKC hypothesis with structural breaks. J Cleaner Prod 2018; 187: 770–779.

91.

Baron

Kenny

. The moderator–mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. J Pers Soc Psychol 1986; 51: 1173–1182.

92.

Elgin

Oztunali

. Shadow economies around the world: model based estimates. Bogazici University Department of Economics Working Papers 2012; 5: 1–48.

93.

Ang

. CO₂ Emissions, research and technology transfer in China. Ecol Econ 2009; 68: 2658–2665.

94.

Chen

. The economy, tourism growth and corporate performance in the Taiwanese hotel industry. Tourism Management 2010; 31: 665–675.

95.

Chen

Myagmarsuren

. Exploring the moderating effects of value offerings between market orientation and performance in tourism industry. International journal of tourism research 2013; 15: 595–610.

96.

Elliot

Rothenberg

Stock

. Efficient tests of the unit root hypothesis. Econometrica 1996; 64: 13–36.

97.

Engle

R. F.

Granger

C. W

. (1987). Co-integration and error correction: representation, estimation, and testing. Econometrica: Journal of the Econometric Society 55(2): 251–276.

98.

Johansen

. Estimation and hypothesis testing of cointegration vectors in Gaussian vector autoregressive models. Econometrica: Journal of the Econometric Society 1991; 59(6): 1551–1580.

99.

Hatemi-j

. Tests for cointegration with two unknown regime shifts with an application to financial market integration. Empir Econ 2008; 35: 497–505.

100.

Phillips

Hansen

. Statistical inference in instrumental variables regression with I(1) processes. Rev Econ Stud 1990; 57: 99–125.

101.

Narayan

. The saving and investment nexus for China: evidence from cointegration tests. Appl Econ 2005; 37: 1979–1990.

102.

Steger

T. M.

Egli

(2007). A dynamic model of the environmental kuznets curve: turning point and public policy. In Sustainable resource Use and economic dynamics (pp. 17–34). Springer, Dordrecht.

103.

Halicioglu

. An econometric study of CO₂ emissions, energy consumption, income and foreign trade in Turkey. Energy Policy 2009; 37: 1156–1164.

104.

Isiksal

Samour

Resatoglu

. Testing the impact of real interest rate, income, and energy consumption on Turkey's CO 2 emissions. Environmental Science and Pollution Research 2019; 26: 20219–20231.

105.

Dogan

Seker

. The influence of real output, renewable and non-renewable energy, trade and financial development on carbon emissions in the top renewable energy countries. Renewable Sustainable Energy Rev 2016; 60: 1074–1085.

106.

Dogan

Taspinar

Gokmenoglu

. Determinants of ecological footprint in MINT countries. Energy & Environment 2019; 30: 1065–1086.

107.

Chen

Hao

, et al. The impact of environmental regulation, shadow economy, and corruption on environmental quality: theory and empirical evidence from China. J Cleaner Prod 2018; 195: 200–214.

108.

Blackman

. Informal sector pollution control: what policy options do we have? World Dev 2000; 28: 2067–2082.

109.

Akbostancı

Türüt-Aşık

Tunç

Gİ

. The relationship between income and environment in Turkey: is there an environmental kuznets curve? Energy Policy 2009; 37: 861–867.

110.

Beck

Demirgüç-Kunt

Levine

. A new database on the structure and development of the financial sector. The World Bank Economic Review 2000; 14(3): 597–605.

111.

Levine

Loayza

Beck

. Financial intermediation and growth: causality and causes. Central Banking. Journal of monetary Economics 2000; 46(1): 031–77.