Environmental Taxation and Import Demand for Environmental Goods: Theory and Evidence from the European Union

Abstract

In this paper, we study the impact of environmental taxation on trade in environmental goods (EGs). Using a trade model in which demand for and supply of EGs are endogenous, we show that the relationship between environmental taxation and demand for EGs follows a bell-shaped curve. Above a cutoff tax rate, a higher tax rate can reduce bilateral trade in EGs because there are too many low-productivity EG suppliers. Based on trade data from 1995 to 2012 across the EU-27 countries, our empirical results are in accordance with the predictions of our model when we use the Asia-Pacific Economic Cooperation (APEC) list of EGs. We find that environmental taxation (measured as the ratio of environmental tax revenoe to GDP) has a monotonically positive impact on the number of trading partners. Furthermore, we show that if countries were to apply an environmental tax rate equal to $3.96\%$ (e.g., the tax rate maximizing international trade in EGs), then trade in EGs across the EU-27 members would experience an increase of 25.33 percentage points. The results are mixed when we analyse the EGs on the OECD list. While the results for the the number of trading partners are confirmed when we use this list, there is no effect of environmental taxation on import demand.

Appendices

Appendix A. The Structural Gravity Equation

We need to specify the production technology used by firms of the polluting industry and market structure to obtain the structural trade equation. The profit of a polluting firm located in country j producing variety v is given by

$$\begin{aligned} \pi _{j} (v) =\sum _{k}p_{j k} (v) q_{j k} (v) -c_{j} (v) -g_{j} (v) \end{aligned}$$

(32)

where $p_{j k}$ the output price prevailing in country k, $q_{j k}$ the output quantity consumed in country k with $q_{j} =\sum _{k}\tau _{j k} q_{j k}$ and with $\tau _{j k}$ being the iceberg bilateral trade cost and $c_{j} (v)$ the production cost. Each firm produces its variety under monopolistic competition.

Consumers have identical Cobb-Douglas preferences over differentiated products (supplied by the polluting industry) and a (non-tradable) homogeneous good (provided by a non-polluting industry). The homogeneous good is produced with a unit requirement in labour so that its price is equal to one. We posit a CES sub-utility function for the differentiated products. Hence, the utility function is given by

$$\begin{aligned} U_{k} =h_{k}^{1 -\mu } \left[ \int _{\Omega _{k}}q_{j k} (v)^{1 -\varepsilon }\text {d}v\right] ^{\frac{\mu }{1 -\varepsilon }} \end{aligned}$$

(33)

where $\varepsilon$ is the constant elasticity of substitution and $1>\mu >0$. The Cobb-Douglas upper tier of utility implies that consumers spend $h_{j} =(1 -\mu _{j}) R_{j}$ on homogeneous goods, where $R_{j}$ is the total income in country j. Demand for a variety v can be expressed as $q_{j k} (v) =p_{j k} (v)^{ -\varepsilon } P_{k}^{\varepsilon -1} E_{k}$, where $P_{k}$ is the price index, given by

$$\begin{aligned} P_{k} =\left[ \int _{\Omega _{k}}p_{j k} (v)^{1 -\varepsilon }\text {d} v\right] ^{\frac{1}{1 -\varepsilon }} \end{aligned}$$

(34)

where $\Omega _{k}$ is the set of varieties available in country k and $E_{k}$ is the expenditure level for the final good produced in country k (with $E_{k} =\mu _{j} R_{j}$). Hence, the sales of a firm producing in country j are given by

$$\begin{aligned} \sum _{k}p_{j k} (v) q_{j k} (v) =\sum _{k}p_{j k} (v)^{1 -\varepsilon } P_{k}^{\varepsilon -1} E_{k} \end{aligned}$$

(35)

In each country, we assume that the production technology requires a single input, labour. Labour demand $\ell _{j}$ is given by $\ell _{j} =q_{j}/\kappa _{j} +f_{j}$, where the parameter $\kappa _{j}$ represents the technological parameter and $f_{j}$ is the fixed requirement in labour. The cost associated with production is given by $c_{j} (v) =\sum _{k}(\tau _{j k} q_{j k}/\kappa _{j}) +f_{j}$. We assume that $\tau _{j j} =1 <\tau _{j k}$ . Serving the domestic market implies lower trade costs.

Because firms produce under monopolistic competition, each producer sets its price and its demand for the EG, treating the price index $P_{k}$ as given. The first-order conditions, given by d$\pi _{j}/$d$p_{j k} =0$ and d$\pi _{j}/$d$a_{j} =0$, lead to

$$\begin{aligned} p_{j k} (v) =\frac{\varepsilon }{\varepsilon -1} \left( \kappa _{j}^{ -1} +t_{j} \xi _{j}\right) \tau _{j k} \end{aligned}$$

(36)

The price is given by a constant markup $\varepsilon /(\varepsilon -1)$ over the marginal cost of producing $1/\kappa _{j} +t_{j} \xi _{v}$ times the marginal cost of exporting $\tau _{j k}$. As expected, a higher tax rate raises the marginal cost and, in turn, the prices set by firms. Note that that the price of the final product ($p_{j k}$) does not vary among polluting firms located in the same country, even if their levels of emissions differ. Indeed, we assume that the marginal impact of production on emissions ($\xi _{j}$) does not vary among firms and that they have an identical technological parameter ($\kappa _{j}$).

We assume that the mass of labour units in each country is given by $L_{j}$ and that $1 -\mu _{j}$ is large enough that all countries produce this good in the open economy equilibrium. Hence, the mass of labour allocated to the production of the non-polluting good is $(1 -\mu _{j}) L_{j}$. In addition, we consider that labour is mobile across industries and is inelastically supplied. These assumptions imply a unit wage.

The free-entry condition in the downstream industry implies that $\pi _{j} (v) =0$. Firms adopting an abatement technology have higher profits than do other firms, and firms enter the market as long as their profits without an abatement technology reach zero (we allow the two types of firms to coexist in equilibrium). Hence, $\pi _{j} (v) =0$ implies $\sum _{k}\left[ p_{j k} q_{j k} -(1/\kappa _{j} +t_{j} \xi _{j}) \tau _{j k} q_{j k}\right] =f_{j}$ . Using (36) yields

$$\begin{aligned} q_{j} =\frac{(\varepsilon -1) f_{j}}{1/\kappa _{j} +t_{j} \xi } \end{aligned}$$

(37)

It is worth stressing that in equilibrium, the pollution intensity of a firm with an abatement activity is given by

$$\begin{aligned} \frac{e_{j} (v)}{q_{j} (v)} =\xi _{j} -\frac{\theta _{j} (v)}{q_{j} (v)} =\xi _{j} \left[ 1 -\frac{\varphi }{1 -\alpha } \left( {\begin{array}{c}t_{j}\\ z_{j}\end{array}}\right) ^{\frac{ 1 -\alpha }{\alpha }} \frac{1/\kappa _{j} +t_{j}}{(\varepsilon -1) f_{j}} \right] \text {,} \end{aligned}$$

(38)

so that its pollution intensity decreases with productivity.

We now determine the mass of polluting firms in each economy. We assume that there is no eco-industry in country j. Therefore, part of the total labour force in country j allocated to the polluting industry is $\mu L_{j}$. The labour market clearing condition in country j implies that $M_{j} (q_{j}/\kappa _{j} +f_{j}) +M_{j}^{e} \phi _{j} +M_{j} f_{e} =\mu L_{j}$ with $M_{j}^{e} =\overline{\varphi }_{j}^{ -\gamma } M_{j}$. Using (37) and the labour market clearing condition in country j implies that the mass of firms is given by

$$\begin{aligned} M_{j} =\frac{\mu L_{j}}{\frac{\varepsilon +\kappa _{j} t_{j} \xi _{j}}{1 +\kappa _{j} t_{j} \xi _{j}} f_{j} +\overline{\varphi }_{j}^{ -\gamma } \phi _{j} +f_{e}} =\frac{\mu L_{j}}{\frac{\varepsilon +\kappa _{j} t_{j} \xi _{j} }{1 +\kappa _{j} t_{j} \xi _{j}} f_{j} +\alpha \gamma f_{e}} \end{aligned}$$

(39)

It follows that $M_{j}$ rises with $t_{j}$ as $\varepsilon -1 >0.$

Total income in each country $R_{j}$ is given by $L_{j} +\Psi _{j}^{e}$, where $\Psi _{j}^{e}$ is the total net gain associated with the use of EGs, given by $\Psi _{j}^{e} =M_{j}^{e} \psi _{j}^{e} -M_{j} f_{e}$. Because $\overline{\varphi }_{j}^{ -\gamma } \psi _{j}^{e} =f_{e}$ and $M_{j}^{e} = \overline{\varphi }_{j}^{ -\gamma } M_{j}$ in equilibrium, we have $R_{j} =L_{j}\text {. }$

By inserting (9), (25), and (11) into (15 ), we obtain the export sales of EGs

$$\begin{aligned} z_{j} a_{i j} =\frac{Y_{i}}{\Pi _{i}} \frac{M_{j} (t_{j}) t_{j}^{1/\alpha }}{ [z_{j}^{ *}(t_{j})]^{1/\alpha }} \frac{\alpha \gamma \overline{\varphi } _{j}^{1/\alpha }}{\phi _{j}} f_{e} m_{i j} \end{aligned}$$

(40)

Appendix B. Data Description

This study covers the period 1995-2012. The data cover bilateral trade flows of the EU-27 members and were collected at the HS6-digit level. Trade data on EGs are obtained from the UN Comtrade database referring to the EGs lists proposed by APEC and the OECD.²⁷ EGs trade is defined at the six-digit level using the harmonized system (HS6). As we exclude services, our sample includes 112 goods for the OECD list, 54 for the APEC list and 138 for the composite list (see Table 8).

Table 8 Number of environmental goods identified in the APEC and OECD lists

Previous studies have found that trade elasticities with respect to transport costs and other transaction cost variables are sensitive to the method used to proxy transport costs (Head and Mayer 2002). We use the indicator suggested by Head and Mayer (2002) to proxy transport costs

$$\begin{aligned} d_{ij}=\sum _{g\in i}\left( \sum _{h\in j}\omega _{h}d_{gh}\right) \omega _{g} \end{aligned}$$

where $d_{gh}$ is the distance between the two subregions $g\in i$ and $h\in j$, while $\omega _{g}$ and $\omega _{h}$ represent the economic activity share of the corresponding subregion. The Centre d’Études Prospectives et d’Informations Internationales (CEPII) uses the above formula to create a dataset. Data on language, legal system and sharing a common border also come from the CEPII database. Total consumption of EGs is calculated using the formula

$$\begin{aligned} y_{j}=\text {Production}_{j}-\text {Export}_{j}+\text {Import}_{j} \end{aligned}$$

where $\hbox {Production}_{{j}}$ is industrial production in the EGs industry located in country j, $\hbox {Export}_{{j}}$ is total exports of EGs and $\hbox {Import}_{{j}}$ is total imports of EGs. Data on production come from the United Nations Industrial Development Organization (UNIDO) Statistical Databases.²⁸ Our dataset for environmental treaties is constructed using the Environmental Treaties and Resource Indicators (ENTRI) dataset produced by Columbia University.²⁹ The GDP, population, land area, and trade openness index variables are collected from the World Development Indicators Database of the World Bank.³⁰ Table 9 reports some descriptive statistics.

Table 9 Summary statistics

Appendix C. Alternative Measures of Extensive Margin of Trade

Table 10 Number of (HS6 digit) products imported

Table 11 Number of partner country-product pairs (number of shipments)

Appendix D

Table 12 Intra-EU trade in non-EGs (N-EGs) and total bilateral trade (All goods)

Table 13 Imports of non-EGs from all trade partners (Imports from 155 countries)

Table 14 Intra-EU trade in EGs with treaty instrumented