What makes a carbon leak?

23 December 2008

Leslie Nielson
Economics Section


What is 'carbon leakage'
  An Australian definition
  An academic definition
Why might carbon leakage happen?
How would you measure it?
Vulnerability to carbon leakage
  Assessing vulnerability
The relocation decision
Has it happened yet?
Might it happen in Australia?
Is it a bad thing?
Possible policy responses


In July 2008 the Minister for Climate Change and Water, Senator the Hon. Penny Wong, released the Carbon Pollution Reduction Scheme Green Paper.[1] The Green paper outlined the possible design of an Australian greenhouse gas (GHG) emissions trading scheme (ETS) and invited submissions from all interested parties by 10 September 2008. Numerous submissions were made.

Industry representatives expressed strong reservations about the impact of the proposed ETS and some suggested that if the impact is too severe, large scale resource and industrial projects will not go ahead. Further, the impact of the proposed ETS may see high GHG emission production relocate to a jurisdiction that is not subject to such a scheme or any other restrictions on GHG emissions.[2]  These concerns were repeated in response to the recent White Paper outlining the Australian Government’s final decisions on the design of the Australian ETS.[3]

The possible relocation of industry to another jurisdiction and the long-term delay to major resource projects in response to the imposition of an ETS is a manifestation of ‘carbon leakage’. The purpose of this background note is to define what exactly this term means, consider whether it has occurred anywhere else in the world, and outline the difficulties in actually measuring it.  The most important point for policy makers is to assess whether it might occur and some factors in this assessment are put forward. A list of considerations for assessing whether carbon leakage may occur, and possible policy responses, conclude this note.

What is ‘carbon leakage’

An Australian definition

A recent major Australian report on climate change and possible policy responses also defines the term:

carbon leakage—the shift of emissions-intensive industries from high-mitigation to low-mitigation countries[4]

‘carbon leakage’—a loss of competitiveness and relocation of trade-exposed, emissions-intensive industries as a result of carbon penalties applying in some countries but not others.[5]

Trade-exposed, emissions-intensive industries represent a special case. All other factors being equal, if such enterprises were subject to a higher emissions price in Australia than in competitor countries, there could be sufficient reason for relocation of emissions-intensive activity to other countries. The relocation may not reduce, and in the worst case may increase, global emissions. This is known as the problem of carbon leakage.[6]

As can be seen, the emphases in the above definitions are on the relocation of productivity capacity from Australia to other countries, due to the imposition of an ETS. This definition is repeated in recent government reports on the possible design of an Australian ETS.[7]

An academic definition

Carbon leakage can also be defined as the ratio of emissions increase from a specific sector in a second country (as a result of environmental policy affecting that sector in the first country) over the emission reductions in the sector in the first country.[8] As a formula, it looks like this:

Sector emissions increase in second country

-----------------------------------------------------   X 100

Sector emissions reduction in first country

The increase in GHG emissions in the second country is caused by increased emissions intensive mining activity or industrial production, where there is either a decline in the same activities - or GHG emissions are stable - in the first country. This increase can come about by industrial activities relocating to or expanding in the second country which has no ETS or other environmental policy constraints; or simply by the reduction or cessation of the same activities inside the first country with no production transfer to the second country. For this measure to be meaningful the overall result must be that global GHG emissions increase more than they otherwise would have.

To fall under the definition of carbon leakage these changes must come about through the application of environmental policy such as an ETS. Changes in the location of industrial production or minerals processing activities, due to the increasing economic attractiveness of a particular location (apart from the impact of environmental policy in the first country), and the resulting rise in that country’s emissions, is not carbon leakage. For example, it would be very difficult to say that increases in China’s industrial production have come about because of an ETS. Rather, the development of China’s internal market, the increasingly open international trading economy and the lower costs of production, has a far greater impact on the decision to locate industrial production in that country than the environmental regulations of any other country.

Why might carbon leakage happen?

There are several ways in which carbon leakage, due to the direct effects of an ETS on industrial facilities, may come about:

  • a short-term loss of competitiveness in one country due to the initial introduction of an ETS. This may lead to increases in production of an emissions intensive product in another location and an increase in the overall level of GHG emissions. For this to occur there must be unused capacity in the second country that can be brought into production at a cost of less than that of increasing production in the first country, or
  • an ETS may reduce the investment returns arising from certain industries in the first country compared to a second country that is not subject to emissions restrictions. This could cause investment capital to flow to that second country for the purposes of building or expanding an emissions intensive activity in that location.[9]

The first cause of carbon leakage may turn out to be temporary as production costs vary over time between locations. However, the second type of leakage causes a permanent increase in the carbon leakage rate.

How would you measure it?

The above academic definition of carbon leakage, while generally accepted amongst commentators, is not that helpful in measuring whether it is happening. How would governments or markets know accurately if emissions were increasing in another jurisdiction that may have a less than adequate GHG emissions control regime, given that such jurisdictions may not adequately measure such emissions?

A conceptually better way of indicating whether carbon leakage occurs may be to measure the international trade flows of industries affected by an ETS.[10] A drop in exports of an emissions intensive product from the country with an ETS may signal that carbon leakage is happening, all other things being equal. That is, all other demand and supply factors remain the same.

In practice this latter condition may prove impossible to determine because factors driving international trade (currency factors, demand from internal verses external markets, transport costs e.t.c) frequently change. Disentangling all these latter influences from the effects of an ETS in a particular country may be very difficult indeed. A further difficulty is that this measure can tell only what has taken place. This is not all that much help to policy makers seeking to assess whether carbon leakage will take place under a particular ETS.

Another way to determine whether carbon leakage may happen is to rely on statements from the affected industries themselves. But, in a time where emissions trading regimes are being developed, such statements may be hard to distinguish from the manoeuvring of individual firms to gain the most favourable set of conditions possible for their operations.

Economic modelling may help governments to assess whether a particular set of climate policies may result in carbon leakage. The trouble is that some economic modelling studies have overestimated the extent to which carbon leakage would have happened with policies currently in place in Europe and in compliance with the Kyoto Protocol.[11] This does not inspire confidence in the use of economic modelling. That said, more realistic models at the sector level may be a helpful guide for policy makers when assessing where carbon leakage may occur.

Vulnerability to carbon leakage

One of the key determinants of whether a sector, or a particular firm, is open to carbon leakage is whether it has the capacity to pass any cost increases through to its customers. If a firm or sector can pass the cost increases on to its customers then their vulnerability to carbon leakage is lower. The reverse is also true; if a firm or sector cannot pass on their costs then their vulnerability to carbon leakage is higher. Many factors bear on this decision and studies for each particular firm would have to be carried out. Some of these factors are outlined below:

  • supplier structure: is the market competitive? One or two suppliers may well be able to recover price rises due to the imposition of an ETS; where many suppliers exist this may not be possible as competition in the affected market will be stronger
  • proportion of overall commodity/product traded: the higher the proportion of the commodity traded, the more open a particular sector may be to carbon leakage, due to the increased trading of the product in question between markets. For example, only about six per cent of all cement is internationally traded, whereas a much higher percentage of aluminium is traded internationally. A more traded commodity may be supplied to the end customer from many alternative sources and the resultant market would be more competitive. Consequently there may be a reduced capacity to pass on cost increases to end customers
  • trade barriers: markets that are open to imports are more competitive and suppliers are less able to recover cost increases from domestic and external customers
  • long-term contractual agreements: a particular facility may be locked into supplying its customers at fixed prices that may not be able to accommodate cost increases arising from an ETS. Of course, this may be a temporary influence as most long-term contracts eventually end and are renegotiated
  • commodity price cycle:  if prices still afford profitable production in a jurisdiction subject to an ETS, then the competitive impact of alternative supplies of the same commodity may be further delayed. The higher the price the less vulnerable a producer may be to carbon leakage
  • energy costs as a proportion of total production costs: generally, the higher the proportion of energy costs are of the total costs of production, the higher the impact of an ETS on that activity. In other words, there would be a relative price effect. This will increase the amount of cost that has to be passed on. If the price cannot be passed on then the producing firm or sector’s vulnerability to carbon leakage increases, and
  • substitution effect: there may be a point at which the higher relative costs of a particular item trigger a decision by consumers to substitute a less expensive alternative. For example, the use of steel in place of aluminium in packaging.[12] The greater the substitution effect the greater the vulnerability to carbon leakage, as firms seek to lower their costs to regain competitiveness against the substitute.

Assessing vulnerability

In practice both the European Union and the Australian government have outlined rules for assessing whether a particular firm or industry is potentially vulnerable to carbon leakage. In Europe, these rules are:

  • a sector is deemed to be exposed to a significant risk of carbon leakage if the sum of both direct and indirect additional costs arising from environmental controls leads to an increase in production costs exceeding 5 per cent of its gross value added and, if the total value of its exports and imports divided by the total value of its turnover and imports exceeds 10 per cent, or
  • if the sum of the direct and indirect additional costs arising from environmental controls leads to production costs exceeding 30 per cent of its gross value added or if the total value of its exports and imports divided by the total value of its turnover and imports exceeds 30 per cent.[13]

The proposed Australian ETS will provide assistance to emission intensive trade exposed firms or sectors. Trade exposure will be defined by having a trade share (defined as the ratio of the value of imports and exports to the value of domestic production) of greater than 10 per cent in any year between 2004-05 and 2007-08, or a demonstrated lack of capacity to pass through costs due to the potential for international competition.[14]

As can be seen, in practice the degree of trade exposure is the key part in assessing whether a sector or entity is likely to be vulnerable to carbon leakage. The European approach is more sophisticated that the Australian one. Only time will show which approach yields the best results.

The relocation decision

Should a company or sector find itself vulnerable to carbon leakage, and the associated falling profits, then it may decide to relocate its activities to a jurisdiction that is not subject to an ETS. In no particular order some of the considerations when relocating may be:

whether a facility is the highest cost facility in a globally integrated industry. For example, the global aluminium industry is dominated by a few major multinational companies with aluminium smelters in many locations. Such companies would relocate production from the highest cost smelters first

  • the availability of an alternative location with the necessary physical attributes, particularly for achieving the necessary scale of production if establishing a new facility
  • whether the alternative location can physically expand production if the activity in question is already present? This may require the establishment of additional power stations e.t.c
  • the availability of less expensive energy in the alternative location
  • the availability of a trained workforce in the alternative location
  • if there are strong nearby markets with low import barriers: one of the reasons why emission intensive activities are located where they are close to their markets. This is particularly important for products where transport costs to markets are a major cost. Transport costs are not so important for low volume high value commodities such as aluminium
  • whether there is a stable government willing to host the activity in question without extracting a disproportionate ‘rent’ in the form of fees and charges
  • whether there are less stringent tax and environmental regimes in the alternative location, both now and over the economic life of the new investment
    • as the income of developing countries increases, they may well demand enhanced environmental controls. Further, participation in any new international agreement to limit GHG emissions may radically affect the decision to relocate
  • whether it is less expensive to upgrade an existing facility compared to the investment required to relocate,  and
  • whether by relocating, a company risks a major negative impact on its reputation.[15]

The answers to these questions will be different for each particular firm and activity. But the important point is that a decision to relocate an activity due to the impact of emissions trading is not a simple decision. Some firms may choose to absorb the additional costs and continue operating within a region subject to an ETS.

There appears to be no substitute for assessment on a case-by-case basis before deciding that a firm or sector may be vulnerable to carbon leakage. However, governments are seldom able to respond on a case by case basis when implementing such a broad economic policy such as the introduction of an ETS.

Has it happened yet?

If carbon leakage was going to occur it would have happened in Europe, where environmental taxes have been applied for some years and where a multi country ETS is in operation. To date there has been little, if any, observed carbon leakage from areas subject to emissions trading schemes or environmental taxes.[16]

The reasons for this lack of ‘leakage’ in Europe to date may be due to the particular design features of the European Union Emissions Trading Scheme (EU ETS) during its first trading period (2005–2007). The very high proportion of free emission permits issued, the scheme’s low coverage (only 40 per cent of the emissions of one GHG, carbon dioxide), and the over-allocation of permits, blunted its economic impact. Further, many environmental taxes in Europe were not overly severe.[17] Another reason may be that during the recent buoyant economic circumstances, energy intensive industries’ returns were sufficient to postpone a decision on whether or not to relocate.

Thus, in Europe there has not yet been a strong economic incentive for carbon leakage to occur. This may change as the EU ETS evolves and other countries, including Australia, implement their own GHG emissions trading schemes.

Might it happen in Australia?

Some doubts have been raised whether Australian emissions intensive industries are really vulnerable to carbon leakage under the proposed Carbon Pollution Reduction Scheme (CPRS).  Recent Treasury modelling suggests that there may be a minor amount of carbon leakage at most expected prices of a tonne of carbon dioxide or equivalent. Where carbon prices are double the highest expected price range per tonne of carbon dioxide or equivalent, significant leakage may occur.[18] Treasury also concluded that recent concerns raised about carbon leakage, based on private economic modelling, may be exaggerated.[19]

However, these conclusions reflect Treasury’s modelling assumptions. One important assumption is that other countries also adopt GHG emissions control measures within a few years of the start of the Australian ETS in 2010. If other countries also adopt significant GHG emission control measures the scope for carbon leakage is lower. Treasury did not model the outcome in respect of carbon leakage should this not occur.

Further, recent equity analysis reports on the impact of the proposed ETS on Australia’s largest companies (including Woodside) suggest that its impact should not have a significant effect on their financial positions.[20] This conclusion was repeated in the wake of the release of the Australian government’s recent white paper outlining the final design of the proposed Australian ETS.[21]

The above analysis does not appear to apply to the aluminium smelting industry. The Garnaut Climate Change Review noted that Australia’s aluminium smelting industry may well eventually decline due to the introduction of an ETS in Australia. The review suggested that this industry would relocate to take advantages of cheaper energy in Africa, Asia and the island of New Guinea.[22]

Is it a bad thing?

Whether carbon leakage is a good or bad thing will depend on what criteria is used in this assessment.

Where carbon leakage occurs through a stagnation of production in Australia, or its relocation overseas, it may cause an undesirable loss of employment and a reduction in economic output. During a period of subdued economic activity this would not be a desirable outcome.

In Australia, the aluminium industry is one of the most vulnerable to carbon leakage. While some of the existing smelters are relative new, others are of an older design. It may be appropriate that some older Australian facilities should close, and that amount of production be relocated, so that newer, less emission intensive facilities are constructed.[23] If new facility was coupled with a less emissions intensive power source (say hydro electricity) from an environmental point of view it may be a very good thing indeed. If that production is relocated to a developing nation, then a useful transfer of skills may also occur.

The validity of the above argument depends on the alternative location having a low emissions power source. It also depends on the alternative country have a sufficiently robust emissions control regime to ensure that the lowest possible emitting technology is deployed in the alternative location. Neither of these two conditions can automatically be assumed to exist.

Possible policy responses

Before governments take specific action to deal with carbon leakage, care should be taken to clarify that the activity in question is either ceasing production or relocating solely because of the impact of emissions trading. A comprehensive study of this issue in 2004 noted that industrialised countries have been losing production of emissions intensive production for most of the last few decades.[24] Factors, other than the possible impact of an emissions trading regime, have been driving these decisions. These factors are not going to disappear simply because an ETS commences.

Should governments wish to restrict carbon leakage, the available policy options may include the following:

  • implementing as wide a global agreement on emissions trading as possible
    • the greater the number of countries where a global agreement applies, the less the possible alternative locations for carbon leakage. The wider the coverage in terms of GHG covered and industries included, the lower the overall cost of compliance, and
  • encourage technical advances in low emission production so that affected industries don’t have to move
    • part of this approach may be to maintain as high an emissions permit price as required to foster technical innovation.

Should the adoption of a global emissions trading scheme be uneven in terms of the number of countries participating, sector specific measures may be required, such as the allocation of free emissions permits to those sectors most at risk of carbon leakage, direct financial assistance and policies to advance technical solutions for further emissions reduction.

Finally, should all else fail, the imposition of discriminatory tariffs on environmental grounds on the imports of products from jurisdictions with less stringent environmental controls (these are known as ‘border measures’ in the jargon surrounding this topic) is a possible policy response.

The best option is the adoption of a robust global agreement. But this may not be practicable for some time yet. The last option is the least desirable option as it may well lead to the imposition of counter tariffs between countries and an increase in industrial protectionism - both practices in which nobody ultimately wins.


[1].    Senator the Hon. Penny Wong, Minister for Climate Change and Water, Carbon Pollution Reduction Scheme Green Paper, Canberra, July 2008.

[2].    Woodside Energy Limited, Submission on the proposed carbon pollution reduction scheme, No. 485, Part A. See also press articles, for example ‘Woodside chief threatens to take gas plant to East Timor’, Sydney Morning Herald, 17 November 2008, p. 6, and Alcoa Australia, ‘Alcoa’s position on climate change’, accessed 20 November 2008.

[3].    Australian Government, Carbon Pollution Reduction Scheme – Australia’s Low Pollution Future, Canberra, December 2008. See Lenore Taylor, ‘Industry revolt on green plan’, Australian, 17 December 2008, p. 1. This latter article cites comments by representatives of the Australian cement industry to locate new production in either Indonesia or Thailand.

[4].    Garnaut Climate Change Review, Final Report, Canberra, October 2008, p. 7.

[5].    ibid, p. 230.

[6].    ibid, p. 316.

[7].    Senator the Hon. Penny Wong, op. cit., p. 27 and 293; Australian
Government, Treasury, Australia’s Low Pollution Future – The economics of climate change mitigation, Canberra, October 2008, p. 3 and 169.

[8].    Julia Reinaud, International Energy Agency, ‘Issues behind competitiveness and carbon leakage – Focus on heavy industry’, IEA Information Paper, October 2008, p. 3. See also B Metz, O R Davidson, P R Bosch, R Dave, L A Meyer (Eds), IPCC Climate Change 2007: Mitigation. Contribution of Working Group III to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge Univ. Press, Cambridge and New York, 2007.

[9].    A third way is that a fall in fossil fuel energy demand (not prices) in the country subject to the ETS may cause an increased demand in a country not subject to an ETS. All other things being equal, the unconstrained country would use additional amounts of fossil fuels. See Julia Reinaud, ‘Issues behind competitiveness and carbon leakage’, ibid, pp. 3 and 4.

[10]. Julia Reinaud, ‘Issues behind competitiveness and carbon leakage’, ibid, p. 6.

[11]. See discussion in Reinaud, op. cit., p. 38 and following. This point was recently made in the report of the outcome of recent Treasury modelling. See Australian Government, Treasury, ibid, p. 170.

[12]. Julia Reinaud, ‘Issues behind competitiveness and carbon leakage’, ibid, p. 43 and following.

[13]. Council of the European Union, ‘Note from the General Secretariat of the Council to Delegations on energy and climate change – Elements of the final compromise’, 17215/08, Brussels, 12 December 2008, pp. 3-4.

[14]. Australian Government, Carbon Pollution Reduction Scheme – Australia’s Low Pollution Future, Executive Summary, Canberra, December 2008, p. xxxiv.

[15]. See discussion in J P M Sijm, ibid, p. 160 and Reinaud ibid.

[16]. Reinaud, ibid., p. 6 and T Barker, S Junanker, H Pollitt and P Summerton, ‘Carbon leakage from unilateral environmental tax reforms in Europe 1995–2000’, Energy Policy, No. 35, 2007, p. 6291, and J P M Sijm, O J Kuik, M Patel, V Oikonomou, E Worrell, P Lako, E Annevelink, G J Nabuurs and H W Elbersen, ‘Spillovers of Climate Policy – An assessment of the incidence of carbon leakage and induced technological change due to CO2 abatement measures’, Netherlands Research Program on Climate Change, Report 500036 002, December 2004, Appendix C, p. 153, and Julia Reinaud, International Energy Agency, ‘Climate Policy and Carbon Leakage – Impact of the European Emissions Trading Scheme on Aluminium’, IEA Information Paper, October 2008, p. 5. This last study also noted that it was too early to say with confidence that the EU ETS would not have a long term effect on the location of aluminium production.

[17]. Barker et al, ibid.

[18]. Australian Government, Treasury, ibid, p. 169. The price of carbon or its equivalent amount in other greenhouse gases depends on the ETS implemented and the rate at which GHG in the atmosphere are stabilised. The greater the rate of GHG emissions reduction and the faster the time table for stabilization as a set rate (measured in GHG parts per million in the atmosphere), the higher the carbon cost. Treasury has modelled the costs of stabilizing GHG concentrations at 450 parts per million using what is known as the Garnaut – 25 scenario. The expected cost of carbon permits in these circumstances is $US47 in 2013, rising to $US158 in 2050 in then year terms (Australian Government, Treasury op. cit., p. 93). Double this figure is $US94 in 2013 and $US316 in 2013 and 2050 respectively. These costs are for a tonne of carbon dioxide or its equivalent in other GHGs. The equivalent of a tonne of carbon dioxide is the amount of one or more of the five other GHG mentioned in the Kyoto Protocol to the United Nations Framework Convention on Climate Change in terms of their capacity to contribute to global warming.

[19]. Australian government, Treasury, ibid, p. 170.

[20]. The Climate Institute, Clearing the Air – Clean energy investments to power a low carbon future – and the myths polluters sue to stall progress, Sydney, December 2008, p. 12.

[21]. Paul Gravey, ‘Minimum impact on the worst polluters’, Australian Financial Review, 17 December 2008, p. 5.

[22]. Garnaut Climate Change Review, Final Report, Canberra, October 2008, p. 497.

[23]. ibid.

[24]. J P M Sijm, ibid, p. 147.





For copyright reasons some linked items are only available to members of Parliament.

Facebook LinkedIn Twitter Add | Email Print