Environmental taxes: drivers behind the revenue collected

This paper investigates the determinants of environmental taxes revenue. While the effects of environmental taxations are well discussed across academic and political debate, less analysis have been specifically devoted to investigate the factors influencing the revenue variations. By using an Index Decomposition Techniques, the present paper separates the contribution of economic factors from the contribution provided by taxation rates and regulations. The data pertain to the period 2004-2016 and 25 European Member States are included in the analysis. Results show that stricter environmental taxation rates and regulations has been the main factor influencing the revenue increase just for 5 of the 25 countries considered in this paper. For the other Member States, economic growth and the role played in the European economic panorama have been the main drivers of variations. By providing a comparative analysis of the factors influencing the environmental taxes revenue of European areas, this paper contributes to identify how regulations and economic factors have influenced the sustainability paths of countries and can be used to support the design of policies across the EU. Structural change effect. It is specific country and the total gross domestic product the countries considered in the decomposition exercise G Economic activity growth effect. It reflects changes of the total economic activity and it is used to quantify the environmental taxes revenue generated by economic growth CH Cumulated change. It summarizes the aggregated contribution provided by every; decomposition factors


Introduction
Environmental taxation is playing an important role in the environmental policies of the EU.
Defined by a "tax base in physical unit (or proxy of it) of something that has a proven, specific negative impact on the environment" (UN et al., 2014), environmental taxes have been mainly introduced to internalise the negative effects of economic activities and to relieve some of the pressures on the environment (EEA, 2008). Since the 1990s, environmental taxation has also been used by Member States to shift the burden of taxation from growth-oriented factors, such as capital and labour, to welfare-reduction elements such as resources depletion and pollution (EEA, 2005;Ekins and Speck, 2009). Firstly prosed in the White Paper on Growth, Competitiveness and Employment (EC, 1993) the "green tax reform" is today an important element of the Europe 2020 strategy and environmental taxation is considered an enabling factor in achieving a smart, sustainable and inclusive economy (Eurostat, 2013).
Within the framework of EU regulations, Member States have generally freedom to design specific taxation systems. For this reason, a large panorama of environmental taxation rates and regulations presently exists among countries. For example, despite the application of the comprehensive energy taxation, established by the Directive 2003/96/EC, Member States are allowed to introduce tax reductions and exemptions. For the transport activities, the registration and the circulation taxes vary widely among countries. In addition, the pollution, waste and resource use taxes are differentiated, not just between Member States, but also among products and regions (EEA, 2016). If from one side, fiscal harmonisation can be useful to ensure good functioning of the internal market and to address the problems generated by cross-boundary pollution; On the other side, the freedom to design specific taxation systems is functional to help Member States in achieving domestic policy objectives. Within this context, the revenue collected, and the related differences existing over time and geographical areas, does not provide consistent information to evaluate the sustainability path of countries. That is because economic growth and more resource intensive activities, can produce a revenue increase similar to that generated by variations in the taxation rate and regulation.
The main research problem arising from the present situation is then related to the fact that the lack of information related to the drivers of revenue variations makes it difficult to identify if a revenue increase, taking place in a specific country during a specific period of time, is generated by an increasing level of pollution or by higher priority attributed to environmental protection. In fact, while the effects of environmental taxations and the possible impacts of allocation of the revenue collected are well discussed across academic and political debate, less studies have been specifically oriented to investigate how regulatory frameworks and economic variables are influencing the revenue collected (Ekins, 1999;Ekins et al., 2011;Carattini et al., 2018).
As highlighted by previous studies, the analysis of the drivers of revenue variations can provide important information to investigate the effect of taxation policies both on economic and environmental terms (Castiglione et al., 2018;Li and Masui, 2019). In addition, the comparative analysis of Member States, presently characterized by different economic and legislative structures, can support the design of consistent policies oriented to promote the sustainable path of countries. As previously highlighted by Castiglione et al. (2014) a better understanding of the main factors influencing environmental taxation can contribute to increase the effectiveness of this policy instrument.
To this purpose, the objective of this paper is to use a decomposition analysis to identify the main drivers of variations in the environmental taxes revenue collected by 25 European countries. By considering the contribution of different explanator factors, decomposition approaches have been previously used to analyse variations in the revenue collected (Aparicio et al., 2013;Baležentis and Kriščiukaitienė, 2015). However, as far as we know, no previous studies specifically focused on the revenue from environmental taxation.
By using the Index Decomposition Technique (IDA) proposed by Sun (1998), this paper analyses the main drivers that have influenced the variations in the environmental taxes revenue for the time period 2004-2016. The decomposition approach adopted in this paper have been previously applied to the energy and environmental field and have proven to be a suitable methodology for cross-country comparisons and time-series analysis (Ang et al., 2015;Andreoni and Galmarini, 2017).
The present paper contributes to the existing literature in different ways. Firstly, it uses, for the first time, an index decomposition technique to investigate the contribution that tax intensity, the economic structure and growth have played in the variations of the environmental taxes revenue. Secondly, it considers a case study of 25 European countries for the time period 2004-2016, that is particularly interesting as: i) it starts in 2004 when most of the eastern European countries joined the EU. Within this context, the results of the paper can contribute to the existing debate around the impact that European integration is playing in the reduction of the environmental pressures (EC, 2012); ii) it covers a time period including years both before and after the 2008 financial crisis. Most of the decomposition exercises performed so far, have been mainly focused on times of GDP growth and limited analysis have included the period that followed the global financial crash (Roinioti and Koroneos, 2017;. From a theoretical perspective, the 2008 crisis represents an opportunity to investigate the role that economic crisis has provided to the variations of the taxation structures and to the variations of the revenue collected; iii) it includes the most recent available information, up to 2016. Finally, from a theoretical perspective, this paper provides a reverse standpoint of analysis. Up to now, most of the studies have investigated how environmental taxes can be used to drive economic and environmental changes. On the contrary, in this paper, the economic, the structural and the regulatory elements are used to explain the variations in the revenue collected. By shifting the focus of the analysis from "what can be done by collecting the revenue" to "what has been done to generate the revenue", the present paper provides a contribution to the design policies oriented to avoid, rather than to reduce or to repair, the impacts on environment.
The paper is structured as follow: Section 2 summarises the data. Section 3 provides a literature review of the main decomposition techniques and presents the decomposition approach used in this study. The results of the analysis are discussed in Section 4 while the conclusions are presented in Section 5.

Data
As reported above, the objective of this paper is to investigate the main factors influencing the variations in the environmental taxes revenue (ETR). To this purpose, a decomposition analysis is performed for 25 European countries. Since some data for specific variables or years are not available for Malta, Luxembourg and Croatia, these countries have not been included in the study. The period considered is between 2004-2016 that, as reported above, includes the most recent available information. The data have been taken from Eurostat that provides consistent information on environmental taxes and Gross Domestic Product (GDP). The harmonise index consumer price has also been used to harmonise the environmental taxes revenue and the GDP reported by Eurostat at current prices. According to Eurostat, the environmental taxes considered in this paper includes the energy taxes, the transport taxes and the pollution and the resource taxes (Eurostat, 2013).
As reported in Figure 1, the environmental taxes collected in the 25 European countries largely increased between 2004 and 2016. In the two years that followed the global financial crisis the revenue collected decreased by around 14,5 million Euro. However, in 2010, the environmental taxes revenue was already back to the level before crisis. From there, a sustained revenue increase has taken place for all the years up to 2016. When considering the growth rate of environmental taxes revenue and GDP, similar trends can be identified between 2004 and 2016.
In particular, as reported in Figure 2, before 2008, the income generation grew quicker than the ETR. After that, the ETR had a growth rate slightly above GDP. As reported above, revenue variations and GDP are strictly connected.
The analysis of data, however, does not provide specific information about the sustainability path of countries. In particular, when the growth rate of environmental taxation lies above the growth rate of GDP, contrasting factors could have taken place. For example, stricter environmental regulations could have generated a growth variation similar to that generated by shift of production toward more resource intensive activities. In addition, the different regulation and economic structures of MS largely influenced the aggregated data presented below. Within this context, the decomposition of the main drivers of variations, is then an important step to identify how policies and characteristics of countries can influence the sustainability path of MS. To this purpose, a decomposition exercise is performed in the following sections. The Shift Share and the Growth Accounting techniques have been largely applied in economics and have been used to decompose variations on employment, inequalities and growth (Chu and Cozzi, 2016;Chen and Zhang, 2018). On the contrary, the Index and the Structural Decomposition techniques have been mainly related to the energy and the environmental fields.
The Structural Decomposition Analysis (SDA), firstly developed by Rose and Miernyk (1989), distinguishes the major sources of change by splitting an identity into its components. Based on the use of input-output tables, the SDA is suitable to offer a wide sectoral perspective and has been largely used to identify variations across sectors of activities (Su and Ang, 2012;Hawkins et al., 2015). Cansino et al. (2016), Liu and Liang (2017)  The Index Decomposition Analysis (IDA) has been developed from the index theory of economics. It mainly uses aggregated sector information and has been widely applied to perform cross-country comparisons of the drivers of changes between time 0 and time T. Over the last few decades, a large set of IDA extensions have been proposed and "perfect" decomposition techniques have been specifically designed to address the problem of the residuals rising in the decomposition results (Ang and Choi, 1997;Sun, 1998;Ang and Zhang, 2000;Chung and Rhee, 2001). The IDA and the related extensions have been largely used to investigate the driving factors of different environmental elements, such as emissions (Cansino et al., 2015;Ang and Su, 2016;Shahiduzzaman and Layton, 2017;Wang and Zhou, 2018) and energy and material use (Mulder and De Groot, 2012;Ang et al., 2015;Wang and Li, 2016;Wang et al., 2018a). The IDA has also been used to investigate decoupling trends Leal et al., 2019) and to analyse emissions and energy use of economic activities, such as transports (Andreoni and Galmarini, 2012a;Jennings et al., 2013;Lyu et al., 2016) and industrial sectors (Timma and Blumberga, 2014;Wang et al., 2018b).
Despite the methodological differences existing between SDA and IDA (Hoekstra and van den Bergh, 2003; Su and Ang, 2012) some attempts have been oriented to integrate the two decomposition approaches (Choi and Ang, 2012;Nie and Kemp, 2013;Choi and Oh, 2014) and to include the production-theoretical decomposition techniques (PDA) to analyse the role of production technology, such as technical efficiency and technological changes (Chen and Duan, 2016;Liu et al., 2018;Chen et al., 2019).
The decomposition techniques reported above, have been widely applied both as a diagnostic and simulation tool and the related discussions have been used to investigate the possible coexistence of environmental protection and growth. In addition, decomposition approaches have also been applied to support the forecasting analysis oriented to investigate the patterns of energy and natural resource uses (Meng et al., 2015;Meng et al., 2018;Li and Qin, 2019).
Within this context, the decomposition of the environmental taxes revenue performed in this paper, and the analysis of the contribution played by tax intensity, economic structure and growth, can be used to investigate if European countries have moved toward increasing level of pollution or toward an higher environmental protection. In addition, the cross-country comparison can be used to identify successful examples of taxation structure and to support the design of sustainability policies both within and across European countries.

Decomposition methodology
The Kaya identity and a Perfect Decomposition Technique proposed by Sun (1998) are used in this paper to investigate the main factors influencing the environmental taxes revenue variations. The Kaya identity, generally used to express a variable as the products of different explanatory factors, has been largely used both for diagnostic and simulation purpose, and has been previously applied to a wide range of economic and environmental relationships (Kaya, 1990;Wu et al., 2016). The Perfect Decomposition Technique, proposed by Sun (1998) as a revised version of the Index Decomposition Analysis, has been specifically developed to eliminate the role of the residual in the decomposition results and has been largely applied for cross-country comparisons and time series analysis (Ang et al., 2003;Shyamal and Bhattacharya, 2004). Based on the principle of "jointly created and equally distributed" (Sun, 1996), Sun (1998) proposed to split the residual terms, among the variables that created them (Hoekstraa, and van der Bergh, 2003;Zhang, 2009). By suppressing the role of the residual, this decomposition technique is able to perform a more reliable and accurate analysis than the conventional Layspeyres and Divisia techniques (Ang and Liu, 2001;Shyamal and Bhattacharya 2004;. By adjusting the Kaya identity and the Index Decomposition Technique, the variations in the environmental taxes revenue are expressed in this paper as the product of the three explanatory factors reported in Table 1. Following Eurostat (2013) and OECD (2015), that identify taxation rate, economic structure and growth as some of the main elements influencing the revenue collected, the drivers of variations are expressed in this paper as the tax intensity effect (TI), the economic structure effect (SC) and the economic activity growth effect (G).

Table 1. Explanatory factors Factor Abbreviation Description ETR t i / GDP t i TI
Environmental taxes intensity effect. It is defined by the ratio of environmental taxes revenue and GDP production. It reflects changes in the tax rate or in the introduction/abolition of new taxes.
GDP t i /GDP t SC Structural change effect. It is calculated as the ratio between the gross domestic product of a specific country (GDPi) and the total GDP generated by the countries considered in the decomposition exercise. The SC effect has been previously used by Shyamal and Bhattacharya (2004) and Andreoni and Galmarini (2012b) to quantify the change in relative shares of different economic sectors. Since in this paper we are considering the economic activity of every country as a single aggregate, the approach previously adopted in Andreoni and Galmarini (2017) is used to account for the role that a specific country plays in the overall economic panorama. The aim is to identify the contribution provided by every country to the total GDP generation. Based on this equation, the SC provides useful information to identify the countries responsible for the main variations in the underlying structure of the European economic system. GDP t G Economic activity growth effect. It reflects changes of the total economic activity and it is used to quantify the environmental taxes revenue generated by economic growth.
In particular, equation (1) expresses the environmental taxes revenue of the ith country at time t (ETR t i) as the product of the intensity, the structural and the economic activity growth effects.
where for a specific year (t), GDPi t refers to the value added of the ith country and GDP t refers to the aggregated value added of all the countries considered in this paper. Since expression (1) runs from a base year 0 to a target year t, we can calculate the variation in the environmental taxes revenue over the period of time as (2): The calculation of each component reported in equation (2) are expressed by the following equations where the sums are intended over the individual country values and where the fractional multipliers are used according to Shyamal and Bhattacharya (2004) to equally distribute the residual among the decomposition factors.
Equation (3) calculates the environmental taxes intensity effect: Equation (4) calculates the structural change effect: Equation (5) calculates the economic activity growth effect:

Results and discussion
The decomposition technique presented above has been used in this paper to identify the main Following the approach adopted by previous studies (Castiglione et al., 2014;Mouthino et al., 2015;Vehmas et al., 2018), that aggregates the European countries based on homogenous factors, the present paper analyses the similarities existing between Member States by considering two main groups of countries. Based on the data reported below, similarities can be identified among the groups of countries considered in this paper. The main objective is to identify the characteristics of the environmental taxes revenue variations by proving a comparative analysis of Member States.
In Tables 1a and 1b,

Eastern European Countries
In the group of countries classified as north and central, the economic activity growth effect (G) has been the main driver contributing to the ETR increases. For these countries, the increasing production and the related enlargement of the taxation base have been the main elements influencing the positive variations in the revenue collected.  Makinen (2019), the financial stability measures, adopted after the banking crisis of the 1990s, enable the Swedish economy to recover quickly than the other European countries.
For all of other north and central European countries, the structural change effect (SC) also contributed to increase the environmental taxes revenue collected between 2004 and 2016. The only exception has been United Kingdom that had a GDP growth rate generally lower than the European average. According to studies published by Griffith and Miller (2015) and by ONS (2018), the productivity reduction, that largely characterised the financial and the service sectors, has been the main factor contributing to reduce the role of UK in the European economic panorama.
When considering the Mediterranean countries, the following trends can be identified: the structural change effect (SC), and the related reduction of the contribution to the European economy, have generated a negative impact on the overall environmental taxes revenue increases. This result is consistent with the fact that Greece, Italy, Portugal and Spain have been among the Member States most affected by the global financial crash, with a recovery that has been slower than the other European countries (Quiggin, 2012). The reduction of international competitiveness, together with the low level of productivity and the extensive debt, largely contributed to reduce the European market share of Greece, Italy, Portugal and Spain (Lin et al., 2013). As pointed out by Vegetti and Adascalieti (2017) Table 1 of the appendix, the largest structural change variations have taken place in the years that followed the annexation to EU, and that preceded the global financial crash (2004)(2005)(2006)(2007)(2008). According to Campos et al. (2018) the increased level of trade and financial integration, together with common market and regulations contributed to increase productivity and growth of the countries that joined the EU.
In addition, the reduction of the technology gap and the inflows of direct investments from the central European institutions improved the quality of human capital with positive effects on exports and growth (Tang, 2016). For all the eastern European countries, the economic growth activity effect (G) also contributed to increase the revenue collected.
When analysing the taxes intensity effect (TI) different trends can be identified. In particular, in Slovenia, Estonia and Latvia the TI effect has largely contributed to increase the overall environmental taxes revenue. According to Eurostat (2018) Slovenia, Estonia and Latvia, the tax intensity effect (TI) has been the most important driver. These five countries have moved toward a more sustainable system. In particular, they have reduced the relative contribution that economic factors have played in the generation of the environmental taxes revenue, and have increased the role of taxation rate and regulation. However, it is important to highlight that some of the countries that moved toward a more sustainable path, such as Slovenia, Estonia and Latvia had an environmental regulation structure generally lower than the European average. On the contrary, the countries that mostly reduced the tax intensity effect, such as Germany, Denmark and Sweden have some of the highest environmental standard in EU.

Conclusions
The environmental taxes revenue collected by countries, and the related differences over time and geographical areas, does not provide exhaustive information around the sustainability path of countries. That is because economic growth and more resource intensive activities, can produce a revenue increase similar to that generated by changes in the taxation rate and For this reason, further analysis would be needed to investigate if economic activities have moved toward more sustainable practices or to increased environmental pressure.
Decomposition analysis, investigating the environmental taxes revenue collected from different economic sectors will be functional to that. In addition, the analysis of the revenue collected from households can be used to investigate the impacts of income, sustainable consumption and efficiency gains, to the country-scale level of energy and natural resource demand. The decomposition of the revenue across different categories of environmental taxes (such as the energy, the transport and the pollution and the resources use) can also be used to forecast the impact of policies both at European and Member States level.
The main implication for theory and practice are related to the fact that the decomposition exercise performed in this paper provides a reverse standpoint of analysis. Up to know, most of the studies have focused on the economic and environmental impacts of revenue collection.
On the contrary, the present paper investigates how economic and taxation elements have influenced the revenue collected. By shifting the focus of analysis from the impacts, to the drivers of revenue, this paper provides a better understanding of environmental taxation and can be used to increase the effectiveness of this policy instruments. In addition, the comparative analysis of European Member States contributes to identify how different economic and regulatory framework can influence the sustainability path of countries.
By providing a disaggregated analysis of the drivers influencing the environmental taxes revenue, this paper contributes to the existing debate around environmental protection and economic growth and highlights how the environmental taxes revenue collected by Member States does not provide exhaustive information around the sustainability paths of countries.
Within this context, the main recommendation relates to the fact that the use of disaggregated analysis, oriented to decompose the role of economic and regulatory factors, would need to be adopted in the design policies oriented to reconcile growth and environmental protection.