Carbon Leakage and Competitiveness of Cement and Steel Industries Under the EU ETS: Much Ado About Nothing.

• Author: Branger, Frederic

1. INTRODUCTION

   With international climate negotiations at a standstill, a world of fragmented regional climate policies is emerging (Rayner, 2010) and the perspective of a worldwide carbon price has been postponed. The main regional climate pricing experiment so far is the European Union Emissions Trading System (EU ETS), which is presented as the EU flagship climate policy. The EU ETS entered its third phase in 2013, after a first test phase (2005-2007) and a second phase corresponding to the period of compliance for the Kyoto protocol (2008-2012). Different carbon markets have been implemented since then (Hood, 2010; World Bank & Ecofys, 2013) but they remain modest in their ambition. In the two main emitting countries (China and the US), only a minority of states and cities are covered by an ETS. Even in the EU, the current economic downturn has created a massive surplus of allowances, approximating 2 billion tonnes, i.e. one year of emissions (European Commission, 2012), and reduced the C[O.sub.2] allowance price to 5 euros in the beginning of phase III vs. around 25 until mid-2008. The European institutions have not yet agreed to reduce the surplus of allowances.

   Although there are different reasons for this worldwide lack of ambitious climate policies, among the main ones is the possible threat to the competitiveness of heavy industries and the resulting carbon leakage. The argument goes as follows: carbon-constrained industries may face additional costs vis-a-vis their foreign competitors. This comparative disadvantage may induce immediate losses of market share to the benefit of foreign competitors (operational leakage) or at a longer term location of energy-intensive industries in regions with more favorable climate policies (investment leakage) (Reinaud, 2008b). As a result, emissions would rise in non-constrained countries ("carbon leakage" (1)), weakening or nullifying climate policy efficiency. Moreover, the additional cost generated by the climate policy may reduce the domestic industry's market share, destroy jobs and reduce profits. Such adverse effects are grouped together under the heading of a loss in "competitiveness," a term the popularity of which is inversely proportional to its clarity.

   Many policies have been proposed to deal with leakage, mostly at the local level, but also at the global level (Bosetti and Frankel, 2011). In practice, not only has the threat of carbon leakage reduced the environmental ambition of climate policies (increasing the global cap in the EU ETS), it has pushed public authorities to distribute a large part of allowances for free, limiting the use of allowance revenue to reduce preexisting taxes or to produce public goods.

   Unfortunately, the existing evidence about the amount of carbon leakage and losses in competitiveness that can be expected from a given climate policy is not conclusive (cf. section 2 below). Among ex ante studies, general equilibrium models point to a positive but limited leakage at the aggregate level (typically from 5% to 25%) while for some carbon-intensive sectors like steel or cement, a higher leakage rate is sometimes forecast (Oikonomou et al., 2006; Demailly and Quirion, 2006). Moreover, the few existing ex post studies do not afford consistent conclusions.

   The present paper aims at filling this gap by econometrically assessing the operational leakage (2) over the first two phases of the EU ETS, in the two most emitting manufacturing industry sectors: cement and steel. The methodology is to econometrically estimate a relationship, obtained via an analytic model, between net imports (imports minus exports) and the carbon price, controlling for other factors that may influence net imports such as economic activity in and outside Europe. Using two different econometric techniques that provide consistent results, we conclude that net imports of cement and steel have been driven by domestic and foreign demand but not by the C[O.sub.2] allowance price, falsifying the claim that the ETS has generated leakage, at least in the short run.

   The remainder of the article is as follows. Section 2 provides a review of the literature on empirical studies focusing on environmental regulations and trade. Section 3 gives an overview of the industry contexts of the different studied sectors. Section 4 explains the methodology (model and data). Section 5 details the results discussed in section 6.

2. LITERATURE REVIEW

   Whereas carbon pricing is relatively new, environmental regulations on local pollutants have a much longer history. For example the Clean Air Act was implemented in the US during the 1970s, well before climate change was on the agenda. Therefore the first studies empirically assessing the impacts of environmental regulations on trade dealt with local pollution issues and tested the pollution haven hypothesis/effect (Kalt, 1988; Tobey, 1990; Grossman and Krueger, 1993; Jaffe et al., 1995). The migration of dirty industries to countries with lower environmental standards (pollution havens) depends both on the environmental regulatory gap and on trade tariffs. In the pollution haven hypothesis (respectively effect), the first (respectively the second) factor is held constant. (3) The pollution haven hypothesis was a major concern during the negotiations of the North American Free Trade Agreements in the 1990s (Jaffe et al., 1995), but as the decrease in trade tariffs has seemed to slow down, the pollution haven effect has become a more relevant concern (and carbon leakage due the EU ETS would be a "carbon haven effect" (Branger and Quirion, 2014a)).

   The prevailing conclusion of the pollution haven literature is that environmental regulations have a small to negligible impact on relocations (Oikonomou et al., 2006). After a first wave of inconclusive works (Eskeland and Harrison, 2003), a second generation of studies have statistically demonstrated significant but small pollution haven effects using panels of data and industry or country fixed effects (Levinson and Taylor, 2008). Many reasons have been invoked to explain why the widely believed fear of environmental relocations was not observed. Some have pointed out that environmental regulations are not a main driver of relocations contrary to economic growth in emerging countries (Smarzynska, 2002), or that pollution abatement represents a small fraction of costs compared to other costs or barriers that still favor production in industrialized countries (Oikonomou et al., 2006) such as tariffs, transport costs, labor productivity, volatility in exchange rates and political risk. Others highlight that heavy industries are very capital-intensive and tend to be located in capital-abundant countries, or that their capital intensity makes them less prone to relocate than "footloose" industries (Ederington et al., 2005). Finally, the Porter hypothesis (Porter and Van der Linde, 1995), implying that regulations bring cost-reducing innovations, has also been cited.

   The pollution haven literature is mostly related to command-and-control regulations for local pollutants, whereas the EU ETS is a cap-and-trade system for carbon emissions. Some studies evaluate policies which are closer to the EU ETS such as environmental taxation in some European countries. Miltner and Salmons (2009) studied the impact of environmental tax reforms (ETR) on competitiveness indicators for seven European countries and eight sectors and found that, out of 56 cases, the impact of ETR on competitiveness was insignificant in 80% of cases, positive in 4% and negative for only 16% of the cases (Miltner and Salmons 2009). However, energy-intensive sectors benefited from exemptions and lower rates of taxation. Costantini and Mazzanti (2012) used a gravity model to analyse the impact on trade flows of environmental and innovation policies in Europe and revealed a Porter-like mechanism: when the regulatory framework is followed by private innovation, environmental policies seem to foster rather than undermine export dynamics.

   The question of carbon leakage was also a relevant issue for the Kyoto protocol. Aichele and Felbermayr (2012) assessed the impact of the Kyoto protocol on C[O.sub.2] emissions, C[O.sub.2] footprint and C[O.sub.2] net imports, using a differences-in-differences approach with the International Criminal Court participation as an instrumental variable for Kyoto ratification. They concluded that the Kyoto protocol has reduced domestic emissions by about 7%, but has not changed the carbon footprint (C[O.sub.2] net imports increased by about 14%). Though they do not explicitly formulate it, their results lead to a carbon leakage estimation of about 100%, contrasting with the other empirical studies. However, two caveats are in order. First, China became a member of the World Trade Organization (WTO) in 2002, just when most developed countries were ratifying the protocol. Since most C[O.sub.2] net imports are due to trade with China, the rise in net imports may well be due to China WTO membership rather than to the Kyoto protocol. Second, apart from those covered by the EU ETS, countries with a Kyoto target have not adopted significant policies to reduce emissions in the manufacturing industry. Hence, if Kyoto had caused leakage (through the competitiveness channel), it should show up on the C[O.sub.2] net imports of countries covered by the EU ETS rather than on C[O.sub.2] net imports of countries covered by a Kyoto target. However, when the authors include both EU membership and the existence of a Kyoto target in the regression, they report that EU membership does not increase C[O.sub.2] imports.

   Some papers use econometric models to empirically investigate the impact of climate policies on heavy industries ex ante, using energy prices as a proxy. Gerlagh and Mathys (2011) studied the links between energy abundance and trade in 14 countries in Europe, Asia and America. They found that energy is a...