Non-linear relationship between tourism, economic growth, urbanization, and environmental degradation: evidence from smooth transition models

Abstract

The study aims to analyze two objectives: first is to explore the non-linear relationship between tourism development, economic growth, urbanization, and environmental degradation, and also to analyze the threshold level of the contribution of tourism development on environmental degradation in top tourist arrival destinations. We applied the newly proposed econometric method panel smooth transition regression (PSTR) framework with two regimes on yearly panel data from 1995 to 2017. Findings suggest that the relationship between tourism development and environmental degradation is non-linear and regime dependent. Furthermore, the findings indicated that the relationship above the threshold level is negative and significant, while below the threshold, tourism development is positive and significant effect on environmental degradation. Tourism development and environmental degradation also exhibit the inverted U-shape relationship meaning that at a particular point, increase in tourism development increases in environmental degradation but after a particular point, increase in tourism development decreases the environmental degradation. The economic growth and urbanization also portray a non-linear and regime-dependent relationship with environmental degradation. The study assists policies and empirical information.

Introduction

Since the past few years, some noticeable changes have taken place in the worldwide environment structure. As stated by the State of Climate states, the universal temperature of the surface has been 0.38 to 0.48 degree centigrade over 1981 until 2010 on average, which categorizes them as hottest years among which the second wave of hottest years starting from 2014. The statement highlights that the worldwide growth level of carbon emission is the leading cause of global warming that has increased four times since the 1960s. Noticeable changes, for example, drop in the level of snow and ice, increase in sea levels, and the Intergovernmental Panel observed the change in the span of seasons, deviation in the rainfall, and rainstorm system in humid areas was underlined in climate change statements on environmental changes (IPCC 2014, 2018). Bilgili et al. (2016) consider that these issues ultimately influence the economic system, government, standard of living, community, and geopolitical advancement. Also, Escobar et al. (2009) conclude that global warming and environmental deviations are the reason behind hunger, disease, flood, and water shortages from which billions of people are going through. Currently, there is consent among researchers that the leading cause for global warming and environmental degradation is the frequent rise in carbon usage that has increased from the last 50 years (Anderson et al. 2016).

Currently, global warming and environmental degradation are a few of the main problems observed by policymakers, activist groups, and researchers. To create awareness about overall environmental issues, the elements of carbon emission have become the most talked topic for the study. Findings of past studies indicate that carbon emission has a significant relationship with economic development, population, urbanization, trade, energy usage, foreign direct investment, and financial growth (Cetin et al. 2018; Raza et al. 2020b; Dong et al. 2018; Li and Lin 2015; Nasrollahi et al. 2018; Park et al. 2018; Ali et al. 2020). Solarin (2014) examines a significant relationship between tourism and environmental degradation, but the issue of a frequent rise in carbon emission has not been fully discussed in tourism studies. According to Scott et al. (2010), most of the literature highlights the negative impact of global warming and environmental degradation on the tourism and travel industry. Specifically, it is stressed that environmental degradation influences the tourism sector’s business, the selection of tourism destinations, and the contentment of traveling (Hoogendoorn and Fitchett 2018). Additionally, Amelung et al. (2007) state that adverse environmental situation works as an unfavorable element and slows the tourism sector growth; however, appropriate and favorable environmental circumstances are determined as an enhancing element. Also, Saint Akadiri et al. (2019a) explore that the effect of global warming and environmental degradation, particularly in coastal areas, is a threat to tourism business because of the rise in sea levels.

At the same time, tourism needs these environmental resources as key elements and strong background settings for the production and utilization of tourism contentment. Precisely, studies consider the usage of land and enhancement of infrastructure, natural resource mining, and waste production as the leading cause of tourism effects of environmental degradation (Mieczkowski 1995; Mathieson and Wall 1982; Williams 1994). In a few scenarios, the environmental effect was more severe due to the development taking place in sensitive ecological regions. Holden (2000) and Williams and Ponsford (2009) note that natural environments are the key factors that attract tourists that serve as an appealing feature. But these natural environments are, at the same time, misused and bargained to please the tourist’s exploitive demands. Certainly, the usage and production being part of tourism business occur at once and frequently in a specified geographical location; the environmental effects of tourist activities are more rapidly noticeable, for example, an imported produced product (Miller 2003). The adverse effect on the natural environment and ecological diversity because of urbanization and utilization of land, forming of accumulation of waste in the surrounding and water bodies, and the progressive seemingly outcomes of universal environment deviation remain to be considered main dangers to tourism spots. Though Shaw and Williams (2009) consider that in a few scenarios, tourism business plays a significant role in a large system of environmental impacts, ruined natural environments impact tourist contentment in unpleasant ways. Hence, the economic consequences for tourism spots with clearly drained or contaminated natural environments can be adverse. In some circumstances, tourism- induced impact on environmental resources may vanish the attraction of these spots. The long-term equilibrium between suitable tourism consumption and misuse remains an important issue challenging travel spots and their interested parties (McCool et al. 2001).

Similarly, economic growth is also considered to be the reason for the increase in carbon emission. According to Adams et al. (2016) and Adams and Klobodu (2017), world carbon emission has escalated steadily over the time from 171 million metric tons in 1964 to 446 million metric tons in 1979 and over 1100 million tons by 2010. It is declared that emerging countries usually undergo extensive environmental degradation due to lenient environmental laws (Temurshoev 2006). United Nations Development Programme (2014) stated that population growth and urbanization might be favorable in terms of education, health care, accommodation, and other vital requirements if well managed. Still, it can be a threat if not managed well. Another Foresight Africa Report (2016) reported urbanization as the key element that has impacted human societies, particularly from the last decade.

Additionally, the report states that urbanization and frequent demographic variations can be major reasons for the development. On the other hand, they can intensify challenges, such as forming slums, shortage of water, and sewage dumping, causing more pollution in the environment. There is no doubt that urbanization is an important factor for demographic inclination in Asia and Africa because it connects to the impact, specifically on the power and the environment (Ghosh and Kanjilal 2014). The recent observable trend of growth in urbanization suggests that future environmental quality is more likely to worsen if not organized and managed. Hence, there is a possibility of neglecting prospect sustainable growth in the area. A study that examines the effects of urbanization on the environment is needed based on the current circumstances.

Subsequently, the current statistics published by the World Travel and Tourism Council (WTTC) reports that the travel and tourism (T&T) sector contribution to gross domestic product is $8272.3 billion (2017), which is likely to increase from 10.4 to 11.7% by 2028. Thus, it will elevate the travel and tourism sector’s economic share by $12,450.1 billion after an era. The direct share of this industry in the gross domestic product is $2570.1 billion (2017) and possible to be $3890.0 billion by 2028. The travel and tourism sector provided 330 million jobs globally and produced 10.3% global gross domestic product in 2019. The rapid growth in the tourism sector influences the level of export and investment globally. Likewise, in various countries, international tourism can balance extreme trade imports to balance payment via providing jobs or improvement and others (Lanza et al. 2003; Balsalobre-Lorente et al. 2020). The study of Lee and Chang (2008) supports the previous study point of view that tourism not only promotes the development of its industry but also encourages the economic development of the entire economy. Many policymakers and governments have carried out movements to support tourism to enhance economic growth (Ahmad et al. 2019).

However, the study results also indicate that an increase of 1% in international tourist arrivals leads to 0.129% increase in carbon emission per capita metric tons. The reason is in top travel destinations; the tourism industry contributes approximately 4.1% to gross domestic product for economic development, eventually escalating carbon emission, which emerges from the usage of energy, for example, coal production and fossil fuel usage produced by the automobiles, electricity production, and industrial activities that increase environmental degradation (Saint Akadiri et al. 2019b). Furthermore, Dogan and Aslan (2017) state another perspective that the consumption of fossil fuels and automobiles in countries pollutes the air, which leads to greenhouse gas generation. Additionally, the accumulation of tourist arrivals decreases the rate of carbon emission. Thus, the findings are supported by Katircioglu et al. (2014) and Zhang and Gao (2016). The reason might be the usage of renewable energy and environmentally friendly products in their activities. To enhance the environmental quality or decrease environmental degradation, a bicycle intended tourism and tourism transportation by environmentally friendly automobiles can be promoted and opted by replacing diesel-powered and environmentally unfavorable mode of transport. Secondly, the recommendations at times are certainly not executed; therefore, it is required that these topics need to be investigated time by time as environmental indicators and the requirement of analysis; there is obvious confidence that research can be of significant importance in supporting policymakers and superiors.

In the framework of environmental degradation and tourism connection, it is vital to consider the following queries. What is the effect of the tourism sector on carbon emission, although global warming and environmental changes have harmful effects on the tourism industry? Second, does the growth of the tourism industry influence worldwide environmental degradation? Bach and Gößling (1996) discussed the role of tourism in worldwide greenhouse gases from a theoretical perspective, and the focus was on the aviation industry’s contribution to greenhouse gas emissions. The versatile nature of the topic needs to be discussed more often with different viewpoints (Scott et al. 2010).

Moreover, it is considered that the current study does not provide ample evidence on the effects of the tourism industry on environmental degradation (Nepal et al. 2019; Shakouri et al. 2017). Though, the study of Lenzen et al. (2018) provides a perspective on the level of carbon emission due to the tourism industry, which is a valuable contribution to evaluate the tourism-related carbon flow. Based on the carbon footprint evaluation and estimation worldwide, tourism sector elevates greenhouse gas emission up to 8% worldwide. Secondly, Paramati et al. (2017) examine that few studies have highlighted the environmental impact of tourism from a qualitative perspective. Hence, this study examines the effect of the tourism sector on environmental degradation for the top 20 tourism destinations for the period of 1995–2017 utilizing the annual data collected from the World Bank and Our World in Data. This study aims to focus and contribute to the literature in the following ways:

Firstly, the tourism industry has been performing tremendously since the last 4 years worldwide as this industry plays an important role in economic growth for developed and developing countries (Gössling et al. 2015; Meo et al. 2018; Cannonier et al. 2019). The tourism industry provides employment opportunities, provides access to foreign exchange and outward investment, enhances infrastructure, and contributes to the growth of manufacturing, agriculture, and service industries (Zaman et al. 2017; Zuo and Huang 2018). The present indicators, international tourist arrivals augmented by 6% to 1.4 billion in 2018. Similarly, international tourism receipts accumulated by 1.479 billion with a 3% change in real terms. Similarly, international tourism export provides $1733 billion from international tourism receipts and passenger transport. Additionally, UNWTO (2018) reported that the tourism sector is placed on the third level for worldwide export returns following chemicals, fuel, and automotive product in 2017. The significance of the tourism industry within the economy, along with the contribution of carbon emission leading to environmental degradation, is vital to plan suitable yet appropriate tourism policy, determining that tourism is affected adversely by global warming.

Secondly, examining the association between the tourism sector and environmental degradation for most visited destinations (France, Spain, USA, China, Italy, Turkey, Mexico, Germany, Thailand, UK, Japan, Austria, Greece, Hong Kong, Malaysia, Russia, Portugal, Canada, Poland, Netherland) is the second contribution of this study as the top tourist destinations perfectly display the position of the world tourism industry. The share of these top 20 destinations for international tourism arrival and international tourism receipts was 190.6% and 196.5% in 2018. Effective tourism policies for these destinations will assist with universal tourism policies. Based on the stated background, this study focuses on exhibiting significant results with empirical analysis.

Thirdly, this study uses the panel smooth transition model to further information on the changing aspects of variables that exhibit their values during transition duration. It governs the non-linearities and shifting of the regime. The technique of PSTR is used to attain robust results. The other traditional panel data estimators including ordinary least square (OLS), fixed effects (FE), random effects (RE), generalized method of moments (GMM), dynamic least squares (DOLS), and full modified least squares (FM-OLS) depend on the hypothesis which is used in the various study (Bai et al. 2009; Bilgili et al. 2017a; Liddle 2014; Ulucak and Bilgili 2018), whereas the current technique PSTR is an advanced panel technique which lets the coefficients of the exogenous variables differ between the variables and above time; dependent on the alterations in the threshold variable, it would help to closely determine the relationship between the variables (Mensi et al. 2019).

The paper is structured as follows: the “Literature review” section discusses the findings and results of past literature. The “Methodology” section consists of discussion on the technique applied and data utilized for estimations. The “Data analysis” section displays and explains the estimations and the “Conclusion” section summarizes the findings.

Literature review

The theoretical background of the study

Tourism is an important growth element due to the advanced economic system that emphasizes sustainable development (Ahmad et al. 2019). Tourism provides various advantages to the local economy by creating jobs, enhancing the standard of living, and supporting native culture. Xie and Zheng (2001) consider that these advantages are availed by compromising the quality of the environment due to the expected growth, particularly leading to increase usage of energy and an effective environment, which raises carbon emission. The relationship between tourism and environmental degradation could be significantly negative, slightly negative, and positive (Pigram 1980). The study of Bi and Zeng (2019) supports the claim that tourism increases carbon emission and environmental degradation that the growth of the tourism sector refers to the development of infrastructure which affects environment as it would eventually enhance the infrastructure of transportation, information, and accommodation resulting in greater production of carbon emission. According to the Intergovernmental Panel on Climate Change (2001), tourism infrastructure was classified to be the main reason of land alteration procedure. Changes in the land are considered to be the leading cause of the production of greenhouse fumes. Subsequently, the pressure on the area leads to produce soil erosion, higher pollution, and discharge into the sea, natural habitat loss, and greater pressure on rare species and intensified exposure to forest fires. It rapidly stresses on resources of water, and it can cause the local population to race for the consumption of scare resources. Tourism growth stresses the natural resources when it rises usage in areas where resources are limited, such as the shortage of water supply, land deprivation, and the reduction of other resources.

The extensive usage of energy within the tourism sector comprises of three elements named as transportation, accommodation, and other businesses. Also, the power usage required for tourism activities typically relies on fossil fuels. As stated above, fossil fuels are the leading cause of carbon emission. So, for environmental degradation, the theoretical explanation is that tourism businesses depend on the greater level of energy usage and carbon emission. Some of the empirical findings support the theoretical explanation (Scott et al. 2010). Tourism has both biophysical and socio-cultural environment effects. For example, tourism leads to polluting the environment because of smoke production, sulfur dioxide, nitrogen oxides, and other hazardous gas discharge. According to Jiang (1996), tourist businesses might harm the natural environment and its attraction. An increase in wastage can form attractive areas into junk points. Also, Zhong et al. (2011) analyze that tourism considerably increases noise pollution, which involves environmental noise and transportation usage. Additionally, airplanes, hotel accommodation, and mechanical water activities are the reason for the increase in carbon emission (Kuo and Chen 2009; Wu and Shi 2011; Liu et al. 2011).

Indeed, environmental degradation is determined to be the leading factor behind change in atmosphere and global warming; numerous international organizations and establishments have been creating the awareness about global warming because of the increase in fossil fuel usage and rise in carbon emission rate throughout the world (Hoogendoorn and Fitchett 2018). However, the study of Otgaar (2012) claims that the growth of the tourism sector can decrease carbon by promoting technological advancement, such as power plants. It is vital to consider that the financing in the tourism industry simultaneously attains two purposes that are enhancing the infrastructure associated with tourism and improving the quality of the environment by financing renewable energy ventures. However, the study of Lu et al. (2019) examine that huge investment in the tourism business can assist in constructing hotels, eateries, and other infrastructure named as energy efficiency technologies, solar energy, and others which can enhance the value to mend quality of the environment, and all these elements positively influence the development of tourism sector. Then, the greater renewable energy leads to a reduced rate of carbon emission. However, the rise in tourism investments helps the tourism organization to construct new infrastructure services, which all are of significant importance to support the growth of the tourism sector. These results suggest that the various destinations’ policymakers should advance more and reinforce the strategies that benefit these destinations to decrease the utilization of fossil fuel and appeal to greater tourism investments. The need for energy for tourism business emerges from clean energy plants such as renewable energy usage, which is essential to decrease the hazardous effect of the tourism sector on the quality of the environment and reduce environmental degradation through minimizing the use of fossil fuel and decreasing carbon emission.

Tourism and environmental degradation

The rise in the tourism industry’s significance requires frequently adjusting according to the customers’ demand and catering to the issues related to change in the environment as each scenario requires new solutions such as encouraging different ways to reduce carbon emission and enhance economic development. Gössling (2002) and Gössling et al. (2002) state that tourism activities depend on developed infrastructure, which affects the environment. Besides, tourism is connected with the need for the energy required for transportation, accommodation, and tourist attractions, which escalates environmental degradation (Becken et al. 2003; Gössling 2002). Past studies have examined the connection between economic growth and the environment, but there is still a lack of literature that examines the connection between economic development and environmental degradation. However, various theoretical perspectives indicate that the rise in the demand for energy consumption is due to the growth in the tourism industry, which is hazardous for environmental quality. Thus, some of the studies investigated the dominance of tourism industry on environmental degradation (Katircioglu 2014; Katircioglu et al. 2014; Lee and Brahmasrene 2013; Solarin 2014; Dogan and Aslan 2017; Paramati et al. 2017; Gao and Zhang 2019; Mishra et al. 2019; Koçak et al. 2020). In terms of empirical analysis, Liu et al. (2011) study analyze that tourism is the key element through which economic condition can be enhanced in terms of income generation. Still, it also escalates the consumption of energy.

Lee and Brahmasrene (2013) consider that tourism and carbon emission increases economic development; tourism has a negative effect on the environment in European countries, while it is also analyzed that tourists from the region with great energy and change in climate responsiveness prefer to shift towards more efficient energy infrastructures and renewable energy resources (Tsagarakis et al. 2011). Past studies based on the linear relationship of international tourism and environment indicate conflicting empirical findings (Jatuporn et al. 2011; Rasekhi and Mohammadi 2015; Sherafatian-Jahromi et al. 2017). The second perspective determines that tourism is a source of environmental degradation, but Brida et al. (2016) support the claim of previous studies by concluding that tourism and the travel sector usually increase economic development. Amzath and Zhao (2014) state that an increase in economic development due to the tourism industry increases energy capacity and usage, resulting in higher carbon emissions.

Sherafatian-Jahromi et al. (2017) and Zaman et al. (2016) examined the non-linear connection between economic development and carbon emission and an inverted U-shape relationship. The study examined the association between tourism and carbon emission for Southeast Asian countries for 1979–2010; the results indicated an inverted U-shape relationship that verifies an environmental Kuznets curve (EKC) for the Southeast Asian tourism sector. It suggests that tourism can also reduce the level of carbon emission once environmental laws have been announced in all the related industries. When the tourism industry achieves a certain level of growth, such as an increase in international tourism expenditures, it does not always tend to raise carbon emission, which affects environmental quality.

Urbanization and environmental degradation

The various theoretical point of views are utilized to illustrate the urbanization and environment relationship, but the most common ones are the urban transition, ecological modernization, and the compact theories (Kasarda and Crenshaw 1991; McGranahan et al. 2001; Poumanyvong and Kaneko 2010a). Urban transition theory proposed by McGranahan et al. (2001) is based on research state that there is a connection between urban environmental challenges and the rise in affluence. In the course of prosperity escalation, environmental issues become more distributed, hindered, and additionally change in kind. According to Marcotullio and Lee (2003), low-income countries’ environmental issues are linked with urbanization that are restricted, instant, and health concerning; thus, for prosperous ones, environmental issues are universal, cross-generational, and environment alarming. The researchers, although, noticed that these trends are disposition slightly than determined results. Sadorsky (2014) suggests that the net impact of urbanization cannot be certain a priori. Generally, the significance of the urban environmental transition theory includes three parts. At first, it explains the association between development (prosperity) and the urban environment. It then argues about the regions that have suffered a chain of environmental issues (McGranahan et al. 1996). Finally, the theory concludes at the scale of environmental effect at the focal point of the policy engagement.

The ecological modernization theory claims that at a low level of growth, societies prefer economic development over environmental sustainability. As societies progress towards prosperity, they tend to be more anxious about environmental degradation and try to discover the solution to lower it. The outcome, changes within an economy and society, takes place through, such as foreign direct investment, technological advancement, urbanization, and transfer from secondary sector to tertiary sector (Crenshaw and Jenkins 1996; Gouldson and Murphy 1997; Mol and Spaargaren 2000; Ahmed 2014). Bornschier and Chase-Dunn (1985) and Ajayi (2006) consider that dependency theorists disapprove of the modernization theory because an economy reliance on external agents does not approve for organic development.

The central point of the compact city theory is high-density growth narrowly to or included in the city core with a combination of housing, workplaces, and shops. Elliott et al. (2015) determine that the focus of production and usage in comparatively small geographical regions should offer chances for economies of scale, which can enhance the entire usage of energy. Moreover, high-density growth promotes several other features that are advantageous for sustainable energy consumption such as low power consumption for housing, efficient advance transportation system, effective remote heating systems, and considerably less carbon emission (Holden and Norland 2005; Jacobs 1961; Newman and Kenworthy 1989; Enwicht 1992; McLaren 1992). Thus, the study of Naess (2001), Frey (1999), and Holden and Norland (2005) considers that the followers of the dispersed region dispute over the compact city due to its damaging impact on the quality of the environment, negligence of rural, suburban societies, and insufficient open space.

Barnes et al. (2005) and Poumanyvong and Kaneko (2010) explain urbanization as an economic advancement; it is a demographic sign that escalates the urban population and changes human behavior, whereas affecting household power consumption structure. As countries advance, infrastructure, transportation, and individual resource utilization in urban regions have to meet the requirements and needs. Hence, issues such as environmental degradation and an increase in carbon emission noticeably rise. Zhang et al. (2017) determine that change in the demand and responsiveness of private households is because of urbanization. People in urban regions mostly depend on profitable products and services. According to Clancy et al. (2008), an increase in heating installation and electric appliances produces higher greenhouse gases. Also, urbanization increases the demand for traveling between the cities and within the city transportation (Jones 2004). Greater private and public transport usage harms the quality of the environment.

Moreover, it triggers the further requirement for infrastructure in the course of urbanization and development in countries. Madlener and Sunak (2011) analyze that building public infrastructures such as highways, bridges, and sanitation systems requires a greater amount of power consumption, leading to an increase in carbon emission. Additionally, throughout the procedure of urbanization, the existing space becomes highly limited. Construction of many level buildings is usually costly and requires more power consumption, and the borders of the regions need to be extended. Generally, the land and area save for agricultural purposes may be used. These systems collectively lead to a rise in carbon emission ratio.

Similarly, theoretical statements and the empirical findings do not provide a reliable outcome for the impact of urbanization. For instance, in a study of emerging countries for 1967–1985, it was analyzed that urbanization is more considerably associated with carbon dioxide than the gross domestic product per capita (Parikh and Shukla 1995). Zhang and Lin (2012) state that urbanization has a positive impact on energy and carbon emission usage through regional-level panel estimation for 1995–2010. Moreover, it was analyzed that urbanization and energy intensity connection can be influenced by economic structure for provinces of China for 1997–2010 (Elliott et al. 2015). Dogan and Turkekul (2016) explore the connection between urbanization and carbon emission in the USA for 1960–2010 through the ARDL technique. The results suggested that the variables are cointegrated. Energy and urbanization usage raises environmental degradation in the long run, whereas trade enhances the environment’s quality. Additionally, the past study does not back up the rationality of the EKC hypothesis for the USA. Another study for 80 countries was conducted for 1983–2005 indicating that normally, the flexibility of urbanization and carbon emission connection is 0.95%, which denotes that an increase by 1% in urbanization can cause to 0.95% increase in carbon emission (Ponce de Leon Barido and Marshall 2014).

The other perspective of urbanization argues that with high urban strength, the influence of urban groups tends to reduce carbon emission. Due to technological advancement, green technological inventions and advancement are constantly supported throughout the urbanization process. Poumanyvong and Kaneko (2010b) claim that production has become more eco-friendly because of the enhancement of labor productivity. Also, the economies of significant scale contributions are noticeable. Based on the compact city theory, densely urban populated regions with diversified land utilization can resolve urban slump and environmental issues. For instance, the utilization of basic infrastructure is effective and long-lasting and tends to decrease the usage of power. Also, the focus on environmental quality is improved extensively. The increase in living standard measures and spirit civilization, residents of urban regions, tend to be more responsive towards the environment, particularly towards the signs of global warming. It supports and creates awareness about adopting environmentally friendly ways and then decreasing the rate of carbon emission.

Conflicting to the past ones which indicate a positive effect of urbanization on the environment, the study of Sharma et al. (2011) exhibits that urbanization, trade openness, and energy consumption per capita have a negative effect on carbon emission and environment for 69-panel counties for 1985–2005. Another study supports Sharma et al. (2011) findings that there is a negative effect of urbanization on developing countries’ carbon emissions for 1990–2013 (Saidi and Mbarek, 2016). Hence, those studies that neither supports the positive nor negative impact of urbanization, Hossain and Miyata (2012), examine that an increase in energy usage causes environmental degradation. But economic development, trade openness, and urbanization do not affect the quality of the environment in the long term for Japan by analyzing the data from 1960 to 2009 through the Johansen cointegration test. Likewise, the relationship between urbanization and carbon emission in Nigeria was examined for 1971–2011. The findings suggest that there is no relationship between the variables. Thus, the researchers determine that policymakers should not consider urbanization while planning and formulating policies to conquer environmental challenges. The rise in the urban population is not connected to rising carbon emissions (Ali et al. 2016).

Based on the contradiction in the findings of numerous studies on urbanization, carbon emission, and environmental degradation, we aim to contribute to the literature by analyzing urbanization effect on environmental degradation where urbanization acts as a control variable.

Economic development and environmental degradation

More than two eras, the connection between economic growth and environmental degradation has been examined empirically, and it is considered that the relationship between environmental degradation and economic growth exists. Few studies claim that an inverted U-shaped relationship between carbon emission and economic growth is present, termed the EKC hypothesis. For instance, past studies argue that economic development damages the quality of the environment at an early stage, but after a certain level of economic development, environmental improvement occurs according to the EKC hypothesis (Grossman and Krueger 1991; Selden and Song 1994). However, the findings of Shafik (1994) present a conflicting point of view that carbon emission increases in a similar direction with economic development. Another study states that carbon emission begins to reduce when the economy reaches a particular income level (Stern et al. 1996). According to, income and environmental degradation are negatively and positively correlated in low and high-income countries. In the case of Sub-Saharan Africa, the change in climate and sustainable growth displays a positive relationship between change in environment and growth. The results were analyzed through the cointegration method (Joseph 2010). Furthermore, the study of Usenobomg and Chukwu (2011) presents findings that contradict previous results that there exists an N-shaped connection between economic growth and environmental degradation in Nigeria. The study concludes that policymakers should take measures to protect the environment and enforce strict environmental laws.

In terms of the positive relationship between economic development and environmental degradation, it is examined that the production of electricity produces less pollution compared with other energy sources (Al Khathlan and Javid 2013). Ozturk and Acaravci (2013) analyze the validity of the EKC hypothesis for the Turkish economy. Likewise, the study of Hamdi and Sbia (2014) also found the validity of the EKC curve hypothesis for Gulf Cooperation Council regions in the long term although, examining the causality between carbon emission, energy consumption, and real output in these regions. Muftau et al. (2014) studied the relationship between environmental degradation and economic growth for West-African counties by analyzing the cointegration method. The results suggested a long-term equilibrium relationship between environmental quality and growth as an N-shaped connection exists between income and environmental degradation. At the same time, the EKC hypothesis does not support for West Africa. In Bangladesh, a short and long-term connection between industrial development and environmental degradation exists (Rahman and Kashem 2017). According to Rahman (2017), unidirectional causality is present for populated Asian countries between economic development and carbon emission. Whereas Mbarek et al. (2018) support Rahman’s (2017) claim, short and long-term effects of economic growth on environmental degradation can occur in the case of Tunisia. Another study concludes that carbon emission negatively impacts economic growth in 58 countries (Saidi and Hammami 2016).

Population and environmental degradation

When population size increases, it triggers the growth of human production and usage; population size is determined to be closely connected with energy consumption and environmental quality (Birdsall 1992; Kaya 1989). Cui et al. (2019) consider that effect of population size and variation on provincial-level carbon emission is of significant importance for researchers. Various studies conclude that there was a short-run unidirectional causality (Knapp and Mookerjee 1996), long-run bi-directional causality (Asumadu-Sarkodie and Owusu 2016), or scaling proportion (Fragkias et al. 2013) between population growth and carbon emissions, or that population size had a significantly positive impact on regional carbon emissions (Anser 2019; Gonzalez et al. 2014; Jorgenson and Clark 2010). Some studies consider that population size is not an important factor in the escalation of carbon emission (Tian et al. 2015; Zhu and Peng 2012). Li and Lin (2015) state that the findings’ contradiction might be due to the different techniques, dataset, and time. The provincial heterogeneity displays the fact that population size has a different effect on environmental degradation over the countries with varied socio-economic growth levels (Fang et al. 2015), different income groups, or different mechanical growth rate of population (Shi 2003). Cui et al. (2017) analyze the effect of the population on carbon emission along with other variable gross domestic product and urbanization levels; hence, it was proved that these variables increase carbon emission, which affects the quality of the environment. The study in China’s context supports the finding of Cui et al. (2017) that economic development, energy production, and consumption, along with population growth, affect environmental degradation (Cui et al. 2018).

Methodology

The method used for the data estimation is the panel smooth transition regression (PSTR) method proposed by González et al. (2005). This method supports dual application which can be used for non-linear homogenous panel or a linear heterogeneous panel dataset. Therefore, the method is a simplification of the panel threshold regression (PTR) technique proposed by Hansen (2000). It is a technique for any change they cause to individual or variable which remains the same with exogenous regressors and resolves the issues related to heterogeneity particularly in a non-linear model (Raza et al. 2020a, b). Although, it is panel framework in which the coefficients vary with the time and period for the countries and provides assistance for heterogeneity in the regression coefficients. Consequently, suppose the certainty that coefficients are constant of an observable variable through a bounded function denoted to the transition function of such variable and modification between maximum conditions. The general PSTR function with two patterns is as follows:

$$ {y}_{i,t}={u}_i+{\beta}_0{x}_{i,t}+\beta {\prime}_1{x}_{i,t}g\left({q}_{i,t},\upgamma, \mathrm{c}\right)+{\varepsilon}_{i,t} $$

(1)

The i = 1, …, N, T = 1, …, T; N represents the numbers of cross-sections and T refers to the time dimensions; y_{i, t} demonstrates the dependent variable; u_i represents the fixed individual effect; x_{i, t} denotes the vector of explanatory and control variables; g(q_{i, t}, γ, c) refers to the transition function and is determined by q_{i, t} which is the threshold variable; C is the threshold parameter; γ is the parameter which depends on the slope of the transition function; ε_{i, t} represents the error term.

Hence, in this study, the objective is to examine the relationship between tourism development and environmental degradation by analyzing the annual panel data of the top 20 tourist arrival destinations from 1995 until 2017. The list of countries is reported in Table 1. It is supposed that the nexus among the two is non-linear, so the non-linear method is exercised to confirm to validate the non-linearity of the connection between tourism development and environmental degradation. Furthermore, the study analyzes economic growth, population, and urbanization as the control variables that affect the environmental degradation (Koçak et al. 2020). Thus, the general PSTR function is as follows:

$$ {\mathrm{END}}_{i,t}={u}_i+\mathrm{GDP}{x}_{i,t}+T{R}_1^{\prime }{x}_{i,t}g\left({q}_{i,t},\upgamma, \mathrm{c}\right)+\upalpha {\mathrm{POP}}_{i,t}+\zeta {\mathrm{UR}}_{i,t}+{\varepsilon}_{i,t} $$

(2)

where i denotes the number of cross-sections (in this study, it is top 20 tourist arrival destinations), t is the time frame (1995–2017). END is environmental degradation which is measured by per capita CO₂ emission, GDP is economic growth which is measured per capita gross domestic product, POP is population count in millions, and UR is urban population as percentage of total population. g(q_{i, t}, γ, c) refers to the transition function and is determined by q_{i, t} which is the threshold variable; C is the threshold parameter; γ is the parameter which depends on the slope of the transition function; ε_{i, t} represents the error term.

Similarly, Fouquau et al. (2008) and González et al. (2005) stated the following function which is:

$$ g\left({q}_{i,t},\upgamma, \mathrm{c}\right)=\frac{1}{1+\exp \left[-\upgamma \left({q}_{i,t}-c\right)\right]} $$

(3)

The threshold parameter is estimated with C and the slope of the transition function is estimated by γ > 0, while the transition function is determined to change into an indicator function when γ → ∞. Further, the (q_{i, t}, γ, c) = 1, if q_it ≥ c and g(q_{i, t}, γ, c) = 0 if q_it < c. The framework of PSTR considers a panel model with fixed effects when γ → 0. Due to increase in the threshold variable (tourism receipt and tourism arrival), the coefficients vary efficiently and steadily from initial condition (β₀) similarly to small levels of tourism receipt and arrivals to second condition (β₀ + β₁) same as to greater levels of tourism receipt and arrival. Though, the PSTR framework of the parameter relies on the threshold variable and changes according to the destinations and duration. However, for the mentioned level of q (tourism receipt and arrivals), the understanding of tourism sector to environmental degradation for mentioned tourism destinations (i) and time (t) is illustrated:

$$ {\varepsilon}_{it}={\beta}_0+{\beta}_1\ xg\left({q}_{i,t},\upgamma, \mathrm{c}\right) $$

(4)

The next step process is implemented to examine the parameter of the PSTR framework. At the initial step, the linearity of the framework is determined. The test illustrates that whether the connection between the tourism receipt and arrival with environmental degradation is appropriately described by the standard linear models (simple model) or by non-linear model (PSTR method). The H₀ is the linear model which is adequate; thus, the H₁ is the PSTR with two conditions or one transition is suitable. For the null hypothesis (H₀ : γ = 0) is examined besides the alternative hypothesis. As the existence of unidentified nuisance in the parameter of the null hypothesis, the correlated test is non-standardized (González et al. 2005). In order to resolve this problem, a regression function is proposed in which the transition function (q_{i, t}, γ, c) in Eq. (1) is modified through the initial order Taylor expansion around γ = 0 and the tailored regression is as follows:

$$ {y}_{i,t}={u}_i+{\beta}_0^{\ast }{Z}_{it}+{\beta}_1^{\ast }{Z}_{it}{q}_{it}+{\beta}_2^{\ast }{Z}_{it}{q}_{it}^2+\dots +{\beta}_m^{\ast }{Z}_{it}{q}_{it}^m+{\varepsilon}_{\mathrm{it}}^{\ast } $$

(5)

The mentioned equation parameters ${\beta}_0^{\ast}\dots {\beta}_m^{\ast }$ are accumulate of γ, and ${u}_{it}^{\ast }={u}_{it}+{R}_m{\beta}_1{Z}_{it}$ in which R_m exhibits the remainder of the Taylor function. Within the presented situation, the testing of H₀ : γ = 0 in Eq. (1) is same as to test the H₀ in Eq. (5) ${H}_0^{\ast }={\beta}_1^{\ast }=\dots ={\beta}_m^{\ast }$. The Fischer LM test, Wald test, and likelihood test are applied to estimate the null hypothesis of the linearity. The estimations are illustrated as follows:

$$ \mathrm{Fischer}\ \mathrm{LM}\ \mathrm{test}={\mathrm{LM}}_{\mathrm{f}}=\frac{\frac{{\mathrm{SSR}}_0-{\mathrm{SSR}}_1}{K}}{\frac{{\mathrm{SSR}}_0}{\mathrm{NT}-N-K}} $$

(6)

$$ \mathrm{Wald}\ \mathrm{LM}\ \mathrm{test}={\mathrm{LM}}_{\mathrm{W}}=\frac{\mathrm{NT}\left({\mathrm{SSR}}_0-{\mathrm{SSR}}_1\right)}{{\mathrm{SSR}}_0} $$

(7)

$$ \mathrm{Likelihood}\ \mathrm{ratio}\ \mathrm{test}=-2\ \left[\log \left({\mathrm{SSR}}_1\right)-\log \left({\mathrm{SSR}}_0\right)\right] $$

(8)

Though, in null hypothesis (H₀), the sum of squared residuals is described by SSR₀; in alternative hypothesis (H₁), the sum of squared residuals is denoted by SSR₁. F(K, NT − N − K) distribution is applied in the Fischer LM test, within which the number of explanatory variables is referred by K, the number of destinations is denoted by N, and time is denoted by T. x²(K) distribution is followed in Wald and likelihood test.

The null hypothesis of linear relationship is rejected which explains that the connection between the variables is non-linear and can be apprehended by the PSTR with minimum of two conditions. The second step involves the null hypothesis of no remaining non-linearity is estimated. In this estimation, the relation of variables is examined in terms of non-linearity which is apprehended by PSTR with two conditions or no conditions. H₀ is PSTR comprising of two higher conditions which is suitable while the H₁ is PSTR with minimum of three conditions which is suitable. The frame is stated below:

$$ {y}_{i,t}={u}_i+{\beta}_0{Z}_{it}+{\beta}_1{Z}_{it}{g}_1\left({q_{it}}_{,}^{.}{\gamma}_1,{c}_1\right)+{\beta}_2{Z}_{it}{g}_2\left({q_{it}}_{,}^{.}{\gamma}_2,{c}_2\right)+{\varepsilon}_{it} $$

(9)

The following equation includes the null hypothesis which is estimated as H₀ : γ₂ = 0. Similarly, there is an issue of identification same as before which is resolved by applying the Taylor expansion of ${g}_2\left({q_{it}}_{,}^{.}{\gamma}_2,{c}_2\right)$ around γ₂ = 0. This provides the following equation:

$$ {y}_{i,t}={u}_i+{\beta}_0^{\ast }{Z}_{it}+{\beta}_1^{\ast }{Z}_{it}{g}_1\left({q_{it}}_{,}^{.}{\gamma}_1,{c}_1\right)+{\beta}_{21}^{\ast }{Z}_{it}{q}_{it}+\dots +{\beta}_{2m}^{\ast }{Z}_{it}{q}_{it}^m+{\varepsilon}_{\mathrm{it}}^{\ast } $$

(10)

The equation mentioned above, the null hypothesis H₀ : γ₂ = 0 of the PSTR framework with one transition or two conditions, is mentioned again as ${H}_0^{\ast }:{\beta}_{21}^{\ast }=\dots ={\beta}_{2m}^{\ast }=0$. The tests conducted are applied with Wald, Fischer, and likelihood estimations. In the scenario of acceptance of the null hypothesis, the process is finished and summarizes that the PSTR with one transition and two conditions is suitable to study the relationship between the variables. In case of rejection of null hypothesis, the process is conducted again until the null hypothesis of no remaining non-linearity is accepted. After the selection of the conditions in the final step, the non-linear least square technique is applied to estimate the parameter of the framework.

Data

To study the relationship between the tourism development, economic growth, urbanization, and environmental degradation, the yearly annual data is obtained from 1995 to 2017 for top 20 tourist arrival destinations. The reason for choosing these years is dictated by data availability and the information related to the countries is mentioned in Table 1. CO₂ emission per capita, economic growth which is measured per capita gross domestic product, population count in millions, and urban population as percentage of total population. Tourism development was measured through two different proxies which are tourism receipts that is measured in US dollar and number of tourist arrival. The data of all variables is acquired through the official websites of World Bank and Our World in Data.

Table 1 List of countries (alphabetical order)

Data analysis

Descriptive statistics

The first estimation which is exhibited by applying the technique is the descriptive statistics. The results describe the general properties of the dataset. The estimations of descriptive statistics are displayed in Table 2. The average value of environmental degradation is 8.414 with a maximum value of 21.280 and a minimum value of 2.555. The average tourist receipt is 29,628.930 million US dollars, with a maximum value of 251,544 million US dollars and a minimum value of 3237 million US dollars. The average tourism arrivals are 27.025 million US dollars, with a maximum value of 86.758 million US dollars and a minimum value of 3.345 million US dollars. The GDP displays the average value of 266,988.800 per capita with a maximum value of 4,297,534 per capita and a minimum value of 9449.803 per capita. The population displayed an average value of 128.799, with a maximum value of 1386.395 and a minimum value of 6.156. Urban population displayed an average value of 73.327%, with a maximum value of 100% and a minimum value of 30.961%.

Table 2 Descriptive statistics (before taking logarithm)

Cross-sectional dependence

After descriptive statistics, the cross-sectional dependence test (CD) given by Pesaran (2004) is applied. The CD is employed to examine whether the data has cross-sectional independence or not because as stated by Dogan and Seker (2016) one should check the presence of cross-sectional dependence when a panel study is carried out. Table 3 depicts the result of the CD test. The outcome shows the associated p values for all the variables are less than 0.1 meaning that the alternative hypothesis of cross-sectional independence is accepted. This implies that the variables under investigation have cross-sectional dependence.

Table 3 Results of Pesaran (2004) cross-sectional dependence test

Unit root test

The unit root test is used to analyze the stationary features of the variables. Table 4 exhibits the estimations of the unit root test. According to the results, variables are non-stationary at level whereas a change to stationary at first difference.

Table 4 Results of stationary analyses

Panel smooth transition regression (PSTR technique)

The first step of the PSTR technique is to apply a linear test. Linear test determines the relationship between the variables which is apprehended by the linear model hence the standard panel model with fixed effect or by the non-linear model which is the PSTR framework. The estimations of Table 5 indicate that the null hypothesis is rejected and the alternative hypothesis is accepted. This proposes that tourism development exerts a non-linear effect on environmental degradation which can be captured by the PSTR framework.

Table 5 Linearity test

The next step of the data analysis is to confirm the actual number of thresholds by applying the test of no remaining non-linearity. The null hypothesis H₀ of the test is that a PSTR framework with one threshold or two regimes is suitable to apprehend the connection of variables which are tourism development and environmental degradation; hence, the alternative hypothesis H₁ is that a PSTR framework with at least two thresholds or three regimes is favorable. The findings exhibited in Table 6 suggest that the null hypothesis cannot be rejected while the PSTR model with one threshold or two regimes is suitable to apprehend appropriately the non-linear relationship between tourism development and environmental degradation.

Table 6 Test of no remaining non-linearity

The estimations of the PSTR method are exhibited in Table 7. In the PSTR model, the estimated sign is valued more than the estimated values as they are not directly explainable (Fouquau et al. 2008). The findings propose that tourism receipts have an insignificant and positive effect on the environmental degradation regime. As the top 20 destinations move from a low environmental degradation regime to a high environmental degradation regime, the connection between the tourist receipt and environmental degradation becomes negative and significant. The minimum threshold value of the environmental degradation above which tourism increases the environmental degradation is 10.301. In other words, the average environmental degradation for exhibiting the effects of tourism should be higher than 10.3% of environmental degradation. The slope parameter (C) indicates the value of 77.802 which determines that the transition from a weak regime to a strong regime is sharp. The fundamental concept from the estimation is that there is a threshold point of environmental degradation above which tourism raises gross domestic product.

Table 7 PSTR model estimations with tourism receipts

In low tourism regime the relationship between the two variable is that when tourism development is low environmental degradation becomes high. Koçak et al. (2020) determine that the tourism sector enhances the quality of the environment when it generates less carbon emission comparative with other industries. Additionally, tourism receipts have no direct impact on environmental degradation due to carbon emission. When the economy generates a high amount of revenue from the tourism, the government encourages spending more on the safety and quality of the environment. When tourism percentage declines or revenue reduces, the government tends to focus less on environmental degradation which eventually makes the environmental condition and quality neglected.

To conclude, in a high tourism regime, the connection between tourism and environmental degradation is negative which proposes that the tourism development plays an important role in reducing the effects of carbon emission (Lee and Brahmasrene 2013; Naradda Gamage et al. 2017; Paramati et al. 2017). The decrease in carbon dioxide emission ultimately affects environmental degradation as it enhances the economic development which aggregates the production of carbon dioxide; hence, tourism development has a significant effect on environmental degradation by decreasing carbon dioxide emission. According to Zhang and Gao (2016), the explanation for the findings suggests that the tourism sector is an integral component of the tertiary sector, which requires less power and is more hygienic than other sectors such as farming. Various studies state that other sectors such as agriculture and industrial involve activities due to which increase carbon emission comparative to tourism sector (Begum et al. 2014; Bilgili et al. 2017b; Shahbaz et al. 2016). The increase in tourism development reduces environmental degradation as this sector brings financial stability and allows the government to earn a hefty amount of revenue. The earning as then utilized for the benefit of the environment such as by introducing environmentally friendly products and investing more in the research and development sector to reduce carbon emission.

The threshold variable of tourism development also exhibits the similar findings. The estimations indicate that the relationship between tourism development and environmental degradation is inverse in the first regime but becomes positive in the second regime. The framework work is in accordance with the number of studies. The threshold value of tourism development is 10% that reveals the transition from the regime to another.

The control variable population displays a constant estimation in both regimes. The population has a significant and positive effect on environmental degradation. The results suggest that when the population will increase, it will increase the environmental degradation. According to Dogan and Aslan (2017), populated areas have more usage of energy-consuming products, such as refrigerator and dryer, and public transportation as there is a significant relationship between carbon emission and energy consumption (Koçak et al. 2020). Similarly, the increase in energy consumption increases the carbon emission which damages the environment increasing environmental degradation.

Furthermore, the urbanization effect on environmental degradation is positive and significant in the low regime and it applies a negative effect in the high regime. The negative effect of urbanization may be explained by the lack of appropriate finance and of appropriate policies of the urban regions to effectively incorporate the urban activities in the environment and enhance the awareness of the recent immigrants about the significance of the environment as a vital element for the population (Tiba 2019).

To apprehend the findings for estimation, tourism uses another element which is tourism arrivals. The finding is illustrated in Table 8 and is likewise to the earlier results. The effect of tourism arrival on environmental degradation is positive and insignificant while the threshold value is 7.310%. Thus, the effects of tourism on environmental degradation become positive and significant when the threshold value increases from 7.310%. The finding is Koçak et al. (2020), as it is stated that arrivals do not need to lead to an increase in usage of fossil fuel resources such as for transportation which increases the carbon emission. The findings suggest that the relationship tourism development and environmental degradation is positive in the first regime but becomes negative and significant in the second regime. The population and urbanization both have a positive and significant effect on environmental degradation. The objective of the study remains the same that there is a threshold point of tourism development beyond which tourism decreases environmental degradation.

Table 8 PSTR model estimations with tourist arrivals

Conclusion

The connection between tourism development and environmental degradation has been of immense importance in the current scenario. In this framework, this study analyzes tourism development and environmental degradation relationship for the top 20 tourism arrival destinations worldwide with yearly data from 1995 to 2017 determined on the PSTR model. The data of the variables involved in this study was acquired from two sources which were World Bank and Our World in Data. The PSTR model supports both methods that are non-linear homogenous panel and a linear heterogeneous panel dataset. Findings of the study are concluded as first when tourism development decreases, the environmental degradation increases. Secondly, when tourism development increases, environmental degradation decreases. Last, the population increases environmental degradation while urbanization has a positive and significant impact on environmental degradation but later, the relationship changes to negative. The reason behind the findings indicates that when tourism sector development takes place, it increases the revenue for the government. The rise in earnings makes the financial sector powerful, and the government spends more on promoting environmentally friendly goods and promotes consideration of environmental quality as a result of which environmental degradation is declining. Whereas, when tourism development decreases, the environmental degradation increases as the government earns less and allocates a small amount of budget for environment safety and security. Moreover when population and urbanization increase, it ultimately increases transportation, the need for accommodation, and numerous infrastructure facilities that contribute to an increase in carbon emission.

It is the responsibility of the government and tourism sector to introduce and formulate such policies that can reduce the environmental degradation. Although, the tourism sector involves activities that generate carbon dioxide and usage of energy consumption increases which contributes towards environmental degradation. Therefore, the tourism sector is also liable to promote and introduce less fuel consumption transportation and products which can enhance the quality of the environment.