Professional Documents
Culture Documents
Impact of Environmental
Regulations on Trade in the Main
EU Countries: Conflict or Synergy?
Roberta De Santis
Istat (Italian National Institute of Statistics) and University of Rome La Sapienza, Rome, Italy
1. INTRODUCTION
The views expressed in this paper are those of the author and do not necessarily represent the insti-
tutions with which the author is affiliated. Any error or mistake remains the author’s sole responsi-
bility. Thanks are due to Claudio Vicarelli for many helpful suggestions.
1
The so-called environmental Kuznets curve hypothesis (Copeland and Taylor 2003) states that as
far as an healthier natural environment is a normal good, demand for it tends to be higher in richer
countries that impose more stringent environmental regulations, as compared to poorer ones. For
this reason in the related literature often the difference in GDP per capita is used as a measure of
the regulatory gap between countries. See for example Grether and de Melo (2004).
2
In the existing literature as far as we know only Jug and Mirza (2005) focuses on EU countries
trade.
3
Launched by the European Commission in March 2010, the ‘Europe 2020’ Strategy is the over-
arching policy initiative which brings together all areas of EU competence and activity in order to
prepare the EU economy for the next decade.
Less widely recognized is the Porter hypothesis4 stating that stringent envi-
ronmental regulation does not necessarily deteriorate the industrial competitive-
ness of a country. Rather, stringent environmental policies – under the
condition that they are efficiently designed and employed – can further a
nation’s international competitiveness.
While there is a broad empirical literature on the impact of trade on environ-
ment, the empirical literature on the impact of environmental regulations on
trade flows is relatively scarce, is very heterogeneous and presents mixed
results. One of the main problems is that most studies are incomparable to
other ones with the consequence that results do not lead to a uniform conclu-
sion. Mainly, due to differences in model assumptions, methods employed and
data that are used as a comparison of results across studies are extremely
difficult.5
The differences in study outcomes are mainly related to three factors:6 (i)
different studies use different policy stringency indicators. These comprise
input- versus output-oriented indicators, costs versus physical measures, objec-
tive (observed) versus subjective (self-reported, expert judgemental) measures,
(ii) studies use different types of temporal data, (iii) different methods are
employed: simple statistical indicators or econometric studies; cross-section,
time series or panel data econometric studies; and studies at country, state, firm
or plant level.
Among the papers comparable, using a gravity setting with OECD data,7 the
most significant studies are those of Van Beers and van den Bergh (1997),
which test the impact of environmental stringency on bilateral exports. They
construct indicators of environmental stringency based mainly on energy
intensities and recycling rates and rank OECD countries according to
their stringency into a 0–1 index. Their main result confirms in a way the
4
Porter and van der Linde (1995). For a survey on empirical paper assessing Porter hypothesis see
Ambec and Lanoie (2007).
5
In Table A1 in the Appendix there is a comparison among the main empirical papers on this
issue.
6
The following categorization of empirical studies illustrates the wide diversity of approaches: (i)
trade-in-goods (Tobey, 1990; Van Beers and van den Bergh, 1997) versus factor content of trade
(Xu and Song, 2000),(ii) simple statistical indicators (Low and Yeats, 1992) versus (multivariate)
regression models (Tobey, 1990; Van Beers and van den Bergh, 1997) or applied equilibrium mod-
elling (Steininger, 1999), (iii) multilateral trade flows (Tobey, 1990) versus bilateral trade flows
(van Beers and van den Bergh), (iv) single country (Low and Yeats, 1992; studies in Fredriksson,
1999) versus multi-country or multiregion (Tobey 1990, van Beers and van den Bergh 1997, Xu
and Song, 2000), (v) static (Tobey, 1990; Van Beers and van den Bergh, 1997) versus dynamic
(Bjørn et al., (1997); Xu and Song, 2000), (vi) analysis at the individual firm level (Bjørn et al.,
1997) versus sector level (all of the other studies mentioned in this list).
7
Particularly interesting, in a more theoretical context, is the paper of Cole and Elliott (2003) This
paper examines the issue in the context of the Heckscher–Ohlin–Vanek (HOV) model of trade and
also in a ‘new’ trade model characterized by monopolistic competition and differentiated products,
improving the well-cited study by Tobey (1990) in many ways.
pollution-haven hypothesis, because they come to the finding that the OECD
countries’ exports are negatively and significantly affected by more stringent reg-
ulations. They also show that imports are negatively correlated with the importing
country’s stringency, which does not support the pollution-haven hypothesis.
Harris et al. (2002) slightly modify Van Beers and Van den Bergh’s tests by
adding up exporters and importers’ fixed effects as well as time effects to show
that the stringency variable does not confirm the first findings anymore. In a
gravity setting, Grether and de Melo (2004) represent stringency by a regula-
tory gap between countries, measured by difference in GDP per capita. How-
ever, when they control for different factors in their trade equation, they
conclude that the relationship between the regulatory gap and trade flows is not
robust.
The study by Jug and Mirza (2005) is the only other empirical study testing
the impact of environmental regulation on EU trade. It shows negative effects
of environmental stringency estimating a gravity equation. Their results show
that more stringent environmental regulations, when depicting a pure cost
effect, are reducing exports. The coefficient is even larger in the case where
exporting countries are Central and Eastern European countries, comparing to
the EU15. They also show that there is no significant difference in the impact
of regulations on trade in case of dirty and clean sectors.
3. EMPIRICAL STRATEGY
The aim of our empirical analysis is to estimate whether and how environ-
mental regulation exerts a negative impact on trade flows on EU-15 exports in
a gravity setting.
As one innovative feature of the paper is the introduction in the estimates of
MEAs as a proxy of environmental regulation, three major multilateral agree-
ments have been selected. In particular, the Montreal Protocol on Substances
that Deplete the Ozone Layer, the United Nations Framework Convention on
Climate Change and the Kyoto Protocol seem particularly suitable for our pur-
pose. They include, in fact, many trade-related measures8 covering a large
number of goods. Moreover, the aim of these MEAs is mainly related to emis-
sion reduction and could have had a strong impact on a broad range of produc-
tion techniques and industrial sectors’ competitive advantages also at an
aggregate level. The three agreements have been signed and implemented in
different years by the EU members and by many (but not all) OECD members.
In line with recent works, we augmented the gravity equation with a multi-
lateral trade resistance index. Starting from Anderson and Van Wincoop
8
For a detailed description see WTO, WT=CTE=W=160=Rev.1 14 June 2001.
where yW is the world income, country i’s world income share hi = yi=yW, and trade cost tij is a
function of border effect bij and distance dij. bij = 1 if there is no border barriers between country i
and j, otherwise equals one plus the tariff equivalent of the border barrier between the two coun-
tries. The model says that the trade between country i and j is determined by the share of the multi-
plier of both countries’ income to the world income, as well as trade cost adjusted for the price
indexes in both countries. The price index in country j is a function of the price indexes, income
shares, and the trade costs of all countries.
10
Ordinary least squares (OLS) suffer from heterogeneity bias in a gravity setting. The two most
widely used panel data models are the random effect model (REM) and fixed effect model (FEM):
both can control for heterogeneity. Their assumptions are different. REM models require that unob-
served bilateral effects are n.i.i. and orthogonal to the remaining part of the error term. regressors
have to be uncorrelated to individual effects and error term for all cross sections and time periods.
If the orthogonality conditions hold, the REM provides more efficient estimates than fixed effect
estimators. If explanatory variables are correlated with unobserved individual effects FEM is con-
sistent.
11
The test statistic of 128.87 is greater than the chi-squared critical value with 11 degrees of
freedom therefore the null hypothesis that the REM is consistent is rejected.
the time-varying variables uncorrelated with lij, X2 are the time-varying vari-
ables correlated with lij, Z1 are time-invariant variables uncorrelated with lij
and Z2 are time-invariant variables correlated with lij.
lij is the part of eijt, including all the bilateral characteristics not specifically
modelled in X1, X2, Z1 and Z2. It also includes the unobserved trade resistance
variables, both bilateral- and country-specific:
lij ¼ gij þ ki þ kj þ xt ; ð3Þ
where gij are the bilateral-specific effects, ki and kj are importer and exporter
country characteristics, x are optional time effects.
The presence of X2 and Z2 causes correlation with unobserved individual
effect. HT model uses variables already included in the model to instrument X2
and Z2.12
In the empirical literature, there are different selection procedures to select the
variables correlated with lij. It is possible to select instruments on the basis of econ-
omic intuition (Hausman and Taylor (1981) or following different procedures.13
The dependent variables in the estimates are the EU-15 bilateral export
flows. The equation is estimated for the EU-15 countries as exporting countries
and 24 countries (15 EU members14 + 10 OECD) as trading partners; the time
span is 1988–2008.15
We introduced three sets of variables into the gravity equation: (i) standard
gravity variables; (ii) variables proxing multilateral trade resistance index; and
(iii) dummy variables for trade and environmental agreements.
1. Standard gravity variables. Bilateral distance as a proxy of transport costs
and the product of the importer’s and exporter’s GDP as proxies of the
‘mass’.16
12
In detail, X2 can be instrumented by deviation from the group means of X2; Z2 can be instrumen-
ted by deviation from the group means of X2. The model is identified as long as the number of
variables in X1 is greater than the number of variables in Z2.
13
See for instance Walsh (2006).
14
Since data for Belgium and Luxembourg have been supplied together since 1999, they will be
included in the estimates together.
15
The dataset is taken by OECD (STAN DTB) for bilateral exports in value terms and environ-
mental stringency indicators, World Bank WDI for GDP in US$ and population, WTO and MEAs
membership are taken by WTO and OECD websites, distance is taken from http://www.cepii.fr/
anglaisgraph/bdd/distances.htm. Summary statistics of the data are provided in Table A2 in the
Appendix.
16
The mass, as conventionally calculated in the gravity equation, is the product of the gross
domestic product of the exporting and importing countries.
where
1. Ln is the natural logarithm, i is the exporting country, j is the importing
country and t is the year;
2. EXPijt is exports in value from country i to country j;
3. MASSijt is the product of the gross domestic product of the exporting and
importing countries, a proxy of the ‘mass’, that is the size of the coun-
tries involved in bilateral trade;
4. Distij is the great circle distance between i and j;19 this formula approxi-
mates the shape of the earth as a sphere and calculates the minimum
distance along the surface;
5. Similijt is the similarity index of the two trading partners’ GDP as a
measure of relative country size; it is built as:
17
Rose and van Wincoop (2001).
18
Dummies for common language, shared borders, currency, islands, countries, See Marques and
Spies (2006), Melitz (2005), Subramanian and Wei (2003), Rose (2002).
19
To calculate great circle distances you need the longitude and latitude of the capital or ‘econ-
omic centre’ of each economy in the study. The apply the following formula to obtain the distance
measure in miles: Dij = 3962.6 arccos([sin(Yi) Æ sin(Yj)] (6) + [cos(Yi) Æ cos(Yj) Æ cos(Xi Xj)]),
where X is longitude in degrees multiplied by 57.3 to convert it to radians and Y is latitude multi-
plied by 57.3 (assuming it is measured in degrees West).
" 2 2 #
GDPit GDPjt
Ln 1 ; ð5Þ
GDPit þ GDPjt GDPit þ GDPjt
20
FEM results were taken as benchmarks: the within estimator is a consistent estimator of parame-
ters, controlling for any source of correlation between regressors and unobserved individual effects.
21
Several HT specifications were implemented in order to select the appropriate instrumental vari-
ables we selected HT1 with the following: instruments (Simil, lndist, lnmassa, border).
22
This hypothesis was resumed by Helpman and Krugman (1986). They asserted – using a model
derived from a standard monopolistic competition framework- that the theory behind comparative
advantages (i.e. the Heckscher–Ohlin model) does not predict the relationships in the gravity model.
Deardoff (1998) suggested that the basic gravity model can be derived from H-O, and so too can
the Linder, Helpman–Krugman hypothesis. Reconciliation is provided by Evenett and Keller
(2002) who find that ‘factor endowments and increasing returns explain different components of the
international variation of production patterns and trade volumes’.
among the comparable studies, with those of Grether and de Melo (2004) and
Jug and Mirza (2005) and seems to support a pollution-haven view.23
Interestingly, the dummies of three MEAs have positive and significant
coefficients.24 This evidence seems to be in contrast with the previous result.
However, the average positive variations of exports (of EU-15 towards 24
OECD countries) induced by signing UNFCCC, Kyoto and Montreal agree-
ments could be explained by a possible trade creation effect among members
of the environmental agreements (owing for example to homogeneous environ-
mental standards) and a consequent trade diversion effect with respect to coun-
tries that did not sign the agreements (and whose products have different
environmental standard).
TABLE 1
The Impact of Trade Agreements and MEAs on EU-15 Export Flows
Zij
Common language 0.4 2.6
Common currency 0.2*** 0.2*** 0.2***
Border 1.0*** 8.1***
Island 0.1 1.1*
Constant 10.4*** 19.13*** 8.6***
Hausman test v2 (11) 1128.87***
F test F(11, 5339) = 551.02***
Over-identification test: v2 (3) 2.94**
Notes:
(i) ***Significant at 1%; **Significant at 5%; *Significant at 10%.
23
The result is robust to the inclusion of exporters and importers’ fixed effects as well as time
effects. For a selection of robustness checks see Table A3 in the Appendix.
24
As the coefficient of the dummy UNFCCC is 0.20, the variation of exports induced by signing
this agreement (UNFCCC = 1) with respect to the case of not signing (UNFCCC = 0) is given,
other things being equal, by [(exp0.20 1=exp0.20 0) 1] 100 = 22 per cent.
TABLE 2
Interaction Effects
Coefficient
Notes:
(i) ***Test: v2 (3): 497.27.
6. CONCLUSIONS
TABLE A1
Selection of Previous Studies Investigating the Impact of Environmental Regulations on Trade
Van Beers and van Gravity 21 OECD Total Qualitative, STR (exporter)
den Bergh Cross- countries Dirty output-oriented negative,
(1997) section 1975; 1992 Footloose 7 broad measures, significant
OLS 2 narrow measures STR (importer)
(energy intensity) negative, even
more significant
Harris et al. Gravity 24 OECD Total Similar to Van Beers STR signs not
(2002) countries Dirty and Van den Bergh; always what
1990–96 Footloose own indices constructed expected – When
(based on energy, accounting for FE,
consumption and STR not significant
energy supply)
Grether and Gravity 52 countries Dirty: pulp and paper, Regulatory gap between Dirty: STR not
de Melo Panel (North, industrial chemicals, countries (difference significant
(2004) revealed South) 1981–98 Nonmetallic in GDPpc) Footloose: STR
comparative minerals, Iron & Steel, significant
advantage Nonferrous metals sectoral analysis: not
(RCA) - Clean all significant
IMPACT OF ENVIRONMENTAL REGULATIONS ON TRADE
811
14679701, 2012, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1467-9701.2012.01450.x by CAPES, Wiley Online Library on [05/12/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
812
TABLE A1 Continued
Ederington et al. Panel IMP and EXP Total US: ratio of PACE to Total trade – STR not
(2003) Industry and from the Dirty total costs of materials significant
time FE United States Footloose other: indices US -developing: STR
1978–92 (World Bank) significant,
Footloose: STR
significant
Large costs: STR not
significant
Ederington and Panel US industries US 4-digit SIC-level PACE (US); ratio of STR significant
Minier Industry and 1978–92 manufacturing industries PACE to total costs
(2003) time FE of materials
Levinson and Panel US net imports from US 133 4-digit PACE (US) as a STR significant
Taylor Theoretical Mexico, Canada; SIC-level fraction of (0.05–0.27)
R. DE SANTIS
(2003) model other OECD countries; manufacturing value added STR significant
non-OECD industries; (1.1–4.4)
countries 1974–86 differentiated
by most =
least polluting
Jug and Panel Importers: 12 9 sectors: by ISIC; Eurostat – current STR significant
Mirza Structural EU countries, differentiated environmental (0.3); larger
(2005) gravity exporters: 19 into dirty and clean; expenditures by for Candidate
EU+CC 1996–99 homogeneous industry (total countries
and differentiated manufacturing) STR even
higher (3)
TABLE A2
Summary Statistics of the Data
TABLE A3
Robustness Checks. The Impact of Trade Agreements and MEAs on EU-15 Export Flows
HT1
Note:
(i) ***Significant at 1%, *Significant at 10%.
REFERENCES
Ambec, S. and P. Lanoie (2007), ‘When and Why Does It Pay to be Green?’, Working Papers
200704 (Grenoble: Grenoble Applied Economics Laboratory).
Anderson, J. E. and E. Van Wincoop (2003), ‘Gravity with Gravitas: A Solution to the Border
Puzzle’, American Economic Review, 93, 1, 170–92.
Antweiler, W., B. R. Copeland and M. Scott Taylor (2001), ‘Is Free Trade Good for the
Environment?’, American Economic Review, 91, 877–908.
Baltagi, B. H., P. Egger and M. Pfaffermayr (2003), ‘A Generalised Design for Trade Flows
Models’, Economic Letters, 80, 391–7.
Bjørn, E., R. Golombek and A. Raknerud (1997), ‘Environmental Regulation and Plant Exit:
A Logit Analysis Based on Establishment Panel Data’, Environmental and Resource
Economics, 11, 3, 35–59.
Cole, M. A. and R. J. R. Elliott (2003), Do Environmental Regulations Influence Trade Patterns?
Testing Old and New Trade Theories, The World Economy, 26, 8, 1163–86.
Copeland, B. R. and M. Scott Taylor (2004), ‘Trade, Growth, and the Environment’, Journal of
Economic Literature, 42, 1, 7–71.
De Santis, R., S. De Nardis and C. Vicarelli (2008a), ‘‘The Single Currency’s Effects on
Eurozone Sectoral trade: Winners and Losers?’’, Economics Journal Kiel, 2, 17, 1–34.
De Santis, R., S. De Nardis and C. Vicarelli (2008b), ‘‘The Euro’s Effects on Trade in a
Dynamic Setting’’, European Journal of Comparative Economics, 5, 1, 73–85.
De Santis, R. and C. Vicarelli (2007) ‘The ‘‘Deeper’’ and the ‘‘Wider’’ EU Strategies of Trade
Integration. An Empirical Evaluation of Common Commercial Policy Effects’, Global
Economy Journal, 7, 4, art.4.
Deardoff, A. V. (1998), ‘‘Determinants of Bilateral Trade: Does Gravity Work in a Neoclassical
World?’’ in J. A. Frankel (ed) The Regionalisation of the World Economy (Chicago, IL:
University of Chicago Press for NBER), 7–22.
Ederington, J. and J. Minier (2003), ‘Is Environmental Policy a Secondary Trade Barrier? An
Empirical Analysis’, The Canadian Journal of Economics, 36, 1, 137–54.
Ederington, J., A. Levinson and J. Minier (2005), ‘Footloose and Pollution-Free’, The Review of
Economics and Statistics, 87, 1, 92–9.
Evenett, S. J. and W. Keller (2002), ‘‘On Theories Explaining the Success of Gravity Models’’,
Journal of Political Economy, 110, 2, 281–316.
Feenstra, R. (2004), Advanced International Trade: Theory and Evidence (Oxford: Princeton
University Press).
Fredriksson, P. G. (ed.), (1999), Trade, Global Policy, and the Environment. World Bank
Discussion Paper, No. 402 (Washington DC.: The World Bank).
Grether, J.-M. and J. de Melo (2004), ‘‘Globalisation and Dirty Industries: Do Pollution Havens
Matter?’’ in R. E. Baldwin and A. Winters (eds.), Challenges to Globalisation: Analyzing the
Economics (Chicago, IL: University of Chicago Press), 167–208.
Harris, M. N., L. Kónya and L. Mátyás (2002), ‘Modelling the Impact of Environmental
Regulations on Bilateral Trade Flows: OECD, 1990–1996’, The World Economy, 25, 3, 387–
405.
Hausman, J. A. and W. E. Taylor (1981), ‘Panel Data and Unobservable Individual Effects’,
Econometrica, 49, 6, 1377–98.
Helpman, E. and P. Krugman (1986), Market Structure and Foreign Trade (Cambridge, MA:
MIT Press).
Jug, J. and D. Mirza (2005), Environmental Regulations in Gravity Equations: Evidence from
Europe, The World Economy, 28, 11, 1591–615.
Levinson, A. and M. Scott Taylor (2004), ‘Unmasking the Pollution Haven Effect’, NBER
Working Papers 10629 (Washington, DC: National Bureau of Economic Research).
Low, P. and A. Yeats (1992). Do ‘‘Dirty’’ Industries Migrate?, in P. Low (ed.), International
Trade and the Environment (Washington D.C.: World Bank Discussion Papers) 159, 89–103.
Marques, H. and J. Spies (2006), ‘‘Trade Effects of the Europe Agreements,’’ Discussion Paper
274=2006 (Hohenheim: Department of Economics, University of Hohenheim).
Melitz, J. (2005), ‘North, South and Distance in the Gravity Model’, Discussion Paper 5136
(London: CEPR).
Porter, M. C. (1991), ‘‘America’s Green Strategy’’, Scientific American, 264, 4, 168.
Porter, M. E. and C. van der Linde (1995), Green and Competitive: Ending the Stalemate,
Harvard Business Review, 9, 4, 120–34.
Rose, A. K. (2002), Do We Really Know that the WTO Increases Trade?, Working Paper 9273
(Washington, DC: NBER).
Rose, A. K. and E. van Wincoop (2001), ‘National Money as a Barrier to International Trade:
The Real Case for Currency Union’, American Economic Review, 91, 2, 386–90.
Steininger, K. W. (1999), General Models of Environmental Policy and Foreign Trade, in J. C. J.
M. van den Bergh (ed.), Handbook of Environmental and Resource Economics (Cheltenham:
Edward Elgar), 120–34.
Subramanian, A. and S. Wei (2003), The WTO Promotes Trade but Unevenly, NBER Working
Paper 10024 (Washington, DC: NBER).
Tobey, J. A. (1990), The Effects of Domestic Environmental Policies on Patterns of World
Trade: An Empirical Test, Kyklos, 43, 2, 191–209.
Van Beers, C. and J. C. J. M. van den Bergh (1997), An Empirical Multi-Country Analysis of
the Impact of Environmental Regulations on Foreign Trade Flows, Kyklos, 50, 1, 29–46.
Walsh, K. (2006), ‘Trade in Services: Does Gravity Hold? A Gravity Model Approach to
Estimating Barriers to Services Trade’, Mimeo.
Xu, X. and L. Song (2000), Regional Cooperation and the Environment: Do ‘‘Dirty’’ Industries
Migrate?, Weltwirtschaftliches Archiv, 136, 1: 137–57.