The EU's Proposed Carbon Border Adjustment Tax

Cost Challenges Ahead Signal Inflationary Pressures

The European Union (EU) may soon push forward with a carbon border adjustment mechanism (CBAM) if it can ensure compliance with the World Trade Organization (WTO).

The EU parliament backed the CBAM proposal in March, and the European Commission is set to make a formal legislative proposal for a CBAM shortly. I anticipate that Parliament and the Commission will begin negations in the second half of the year.

If implemented, a CBAM would levy tariffs on carbon-intensive imports such as aluminum, cement, and chemicals imported into the EU. The tariffs would be linked to the cost of emission allowances under the current EU carbon emission trading system (EU ETS), effectively imposing the EU ETS globally on select industries that export to the EU. In this way, the CBAM tackles two interrelated issues for EU member states: encouraging other nations to increase their carbon reduction and distributing the burden of carbon reduction in difficult to decarbonize sectors globally.

Thinking through the impact of a CBAM is tricky but essential. We do not yet know the scope or framework for implementation. That said, impressions to date are as follows:

1. It is not evident that a CBAM would be effective in reducing carbon leakage.
2. There is a real risk of continued trade flow deviations rather than decarbonization for the aluminum sector in the EU. Despite a growing global demand for aluminum, Europe has lost more than 30% of its primary production capacity since 2008 due to the indirect costs related to the ETS [1]. Furthermore, while aluminum imports may be more economically competitive than domestic production, the cost of a carbon tax is punitive even for low carbon producers.
3. CO² prices may prove to be inflationary and could undermine energy transition policies in severe scenarios. Furthermore, technologies that many expect to deflate in cost (solar PV, for example) may re-inflate.
4. An active CBAM may have unintended feedback loops should it inflate the cost of new energy technologies.

The aluminum industry is an excellent backdrop with which to evaluate the implications of a CBAM. EU aluminum imports make up 40% of total domestic consumption. 40% is considerably higher than cement (3% of consumption), steel (4% of consumption), and nitrogen (17% of consumption). [2]

The largest aluminum exporters to Europe are Russia, Mozambique, Canada, South Africa, and the UAE, accounting for 80% of the EU's total imports [3]. Of the EU imports in 2020, 1.2 Mt (45% of the total) depended on either captive, non-renewable sources of generation or non-renewable grid power. [4]

It is not evident that a CBAM would be effective in reducing carbon leakage.

Should a CBAM capture both Scope 1+2 emissions, it would be cost-prohibitive for a substantial portion of European imports (at least 45%), assuming costs are not passed down to the consumer. In theory, this should stabilize or even encourage an increase in domestic production. However, it is not likely that European producers re-capture the lost import tonnage. Roughly 80% of European smelters rely on grid power, which has fossil generation in some form, putting them at a disadvantage to many Canadian and Russian producers that utilize hydropower.

The current average global level of direct and non-direct energy emissions of aluminum production is 12.3 tones of CO² per tone of aluminum. RUSAL, a prominent Russian producer that attributes ~50% of gross revenue to European sales, had an emission intensity of 2.2 tones of CO² per tone of aluminum in 2019. [5] An 82% difference in carbon intensity relative to global averages.

One of the only producers in the EU with a comparable carbon footprint is Norsk Hydro, which utilizes extensive hydropower resources for electricity production in Norway.

It is telling that Norsk Hydro has come out publicly against a CBAM, citing the inability to address production capacity continuing to leave Europe. While I agree, the firm is also keenly aware that their profitability is in trouble should their free allowances be removed at the expense of a carbon price being placed on imports. Industries' position that a CBAM is not an adequate replacement of free allowances is understandable. Scope 1+2 emissions across global producers can vary dramatically, principally due to the electricity source powering the aluminum smelters. However, even for low carbon producers that can utilize hydroelectric power, removing free allowances and implementing carbon prices is costly. [6]

There is a real risk ahead of continued trade flow deviations rather than decarbonization for the aluminum sector in the EU. While aluminum imports may be more economically competitive than domestic production, the costs of a carbon tax are not trivial, even for low carbon production.

Both the US and China compete for aluminum from Canada and Russia, respectively. As a result, the EU ingot premium may need to substantially increase to divert flows away from the US and China to meet European aluminum demand.

A carbon tax on imports becomes a consumption tax on the consumer in a scenario in which carbon abatement is either not possible or excessively costly.

Below is a chart that maps the carbon intensity per tone of aluminum, by producer, from 2015 - 2020. The dotted lines indicate the planned carbon reduction roadmap for producers that have made their targets public. The shaded blue area represents the 2015 Paris pledges, testing the alignment of company emission targets against the UN Paris Agreement goals. [7]

Figure 1: Carbon Intensity of Major Aluminum Producers

Should a Scope 1+2 import tax be implemented, RUSAL is in a comparatively strong position. EBIT margins for RUSAL over the last three years are ~6%. However, should the carbon price on imports be roughly equivalent to the EU ETS carbon price today, average EBIT margins would drop to -8.3%. At a $100 per tone CO² price and zero-carbon abatement, margins drop to -22% unless costs are passed through to the consumer. This is an unsustainable cost profile for a major producer.

I should note that if EU ETS allowances are scaled back faster under an active CBAM, then the auction clearing price for carbon is likely only to increase, not decrease. A $100 per tone price on carbon is very much in reach.

For producers that do not have access to hydropower, the prognosis under an active CBAM is grim. For example, Hillside Aluminum in South Africa (100% owned by South32) has roughly 6x the carbon intensity of RUSAL. Using the same scenarios as RUSAL, margins are unsustainable for prices less than $50 per tone.

Figure 2: Margin Impacts on Aluminum Producers

Interestingly, Hillside Aluminum has a slightly superior 3-year average EBIT margin compared to RUSAL when a carbon price is not applied. Over a more extended period, one could say that the two businesses have similar earnings margins. Hillside's carbon intensity, though, is 6.3x that of RUSAL. As a result, once a carbon price is applied, the similarities in the two businesses ' earnings power evaporates and the availability of free cash flow to invest back in the business changes dramatically.

If the less carbon-intensive capacity is not profitable and most domestic production is even less profitable, I find it unlikely that the EU can wind down domestic allowances. Should the EU choose to accelerate the removal of domestic allowances with a CBAM in place, EU aluminum prices will be forced to appreciate substantially. The price of aluminum may increase by ~40% under a $100 per tone carbon price scenario. [8] If they appreciate to an equilibrium that compensates the average EU producer, it will likely attract export production back into the member states at the expense of the domestic output. In short, there is a significant quantity of non-EU aluminum production that is less carbon-intensive than domestic production. Any attempt to price carbon makes the EU less competitive than global producers all else equal.

CO² prices may prove to be inflationary and could undermine energy transition policies in severe scenarios. Furthermore, technologies that many expect to deflate in cost (solar PV, for example) may re-inflate.

CO² prices are a form of subsidy. The IEA's recent roadmap for net-zero calls for a global CO² price of $130 per tone by 2030 and $250 per tone by 2050. Any technology that can decrease CO² is effectively subsidized. Companies with fewer emissions are also rewarded relative to peers. But CO² prices may not always drive decarbonization.

One of the challenges of CO² prices is that if a $100 per tone CO² price is applied across the entire economy, and CO² emissions only fall by 20%, then the cost of CO² abatement is $500 per tone, not $100 per tone. [9] The abatement cost is simply the cost per unit of decarbonization technology minus the cost per unit of the incumbent technology, divided by the CO² reduction achieved from switching from the incumbent to the decarbonization technology. It is a helpful tool to compare different technologies in different industries.

On some level, this is an intuitive but essential observation: if the goal of CO² prices is to incentivize carbon reduction but fails to do so, then the C02 price is simply a consumption tax. Should CO² prices continue to increase with no effect on emissions, then the cost of removing (abating) emissions is very high. The CO² price only equals the cost of abatement if 100% of emissions are abated.

An active CBAM may have unintended feedback loops should it inflate the cost of new energy technologies.

Close to 90% of solar components heading into the EU are manufactured in China, which has a grid intensity close to 0.7 kg /kWh. It takes an average of 5MWh of energy to manufacture 1 kW of solar panels. This embodied carbon suggests that a $100 price would add ~$0.35/W to construction costs against an average baseline of $1.3/W, a ~30% increase.

CAPEX is the most significant driver of the levelized cost of energy for solar (measuring the total cost of ownership), at ~ 57%. This means that the cost of ownership increases by 17% should CAPEX increase by 30%.

This could be a problem for asset developers and the proliferation of renewable energy deployment. Encavis (ECV), the largest publicly traded European renewable asset developer, reported an operating return on assets of 7.2% for their solar projects in 2018. [10] Given these returns, a 17% increase in the total cost of ownership is problematic for developers.

The cost of solar has deflated a remarkable 90% over the last decade to ~4-7 cents per kWh. Technology improvements will undoubtedly continue to place downward pressure on that price, but future cost deflation may not be as extreme. [11] All else equal, an active CBAM would likely have inflationary consequences to solar panel imports coming into Europe.

Additionally, future deflation will be offset by rising curtailment levels as grids become more saturated with renewables. This, too, will start re-inflating the levelized cost of solar. When wind and solar supply ~40% of the generation to the grid, ~12% of new plant generation will fail to dispatch. The intermittency characteristics of renewables effectively overload the grid at certain times of the day if new capacity is added. This means that for each kW of capacity added, a smaller percentage of the potential kWh (energy) can be monetized. Less dispatch means less market revenue, driving up the levelized cost of energy under a steady cost profile. At a 40% penetration level, curtailment of renewables will raise levelized cost by about one (1) cent per kWh, or 25%. If renewables approach 60% of the grid's capacity, about 66% of renewable generation will fail to dispatch, raising levelized costs by ~ten (10) cents per kWh, a 250% increase. [12]

The recent IEA roadmap to net-zero has wind and solar providing 68% of power generation by 2050. To re-iterate, a significant portion of renewable power will need to be curtailed at that penetration level due to intermittency, which raises the required power price by ~200+% for developers looking to hold rates of return constant. [13]

Each $100 per tone of CO² abatement cost inflates the CAPEX of solar by 30%, wind by 6%, and hydrogen by 5-10% [14]. If the cost of new energies inflates, then the costs of CO² abatement grows. It is not clear what the breaking mechanism is on this feedback loop. The energy transition makes it tough to raise interest rates as renewables are hit 4x harder than fossil fuel generation. Wind and solar today are generally financed with a meager cost of capital (~5%), so small increases in rates disproportionately impact renewable financing terms. Additionally, renewables are more capital intensive than other asset classes [15]. Depreciating upfront CAPEX absorbs 33% of renewables revenue, versus ~15% for average project/plant in different industries. [16]

None of the above is meant to detract from the importance of achieving carbon reduction in the global economy. I do, however, believe that costs matter, and many of the deflationary trends and forecasts for new energy technologies that are often required for accelerated adoption, may need revaluation. The implementation of a CBAM in the EU could have a meaningful price impact. Moving forward, investors should be mindful of the following considerations:

• In a scenario of lower imports, do European producers prioritize price or market share?
• Does a CBAM come with lower limits on free allocations or accelerate the removal of free allowances altogether?
• Where could imports fall, and what would be the cost position and utilization rates of domestic players in that local region?
• Is it possible to scale recycling infrastructure in the timeline commensurate with a rollout of carbon prices? Today, secondary supply accounts for ~25% of aluminum. What happens if that can be raised to 50%?
• Finally, EUA prices. The price of carbon in the EU ETS market, if used as a proxy for a CBAM price, will have a significant impact on eventual outcomes.

[1] Morgan Stanley, “EU Carbon Border Adjustment: A “Notional ETS”?”, May 2021

[2] Ibid.

[3] Eurostat.

[4] Captive, non-renewable sources of generation could include thermal coal while non-renewable grid power could include natural gas, by way of example.

[5] RUSAL Annual Report, 2020.

[6] To-date, the EU has attempted to prevent leakage by allocating free allowances to sectors at risk. While the EU views a CBAM as an opportunity to reform allowance allocation (a potential acceleration of the drawdown of free allowances), industry almost uniformly views the CBAM as a necessary addition to free allocation, not a replacement.

[7] The Transition Pathway Initiative uses the Sectoral Decarbonization Approach (SDA) which was created by CDP, WWF & WRI in 2015 and is also used by the Science Based Targets Initiative. For more information on SDA, please refer to “A method for setting corporate emission reduction targets in line with climate science”, May 2015.

[8] A base case price of $2,400 per ton, with ~ 9.7 kg of CO²per kg of Aluminum, raises the price to $3,370: a 40% increase.

[9] If we have 100 tons of CO²emissions in the economy, 20 tones have been abated, and emitters of the remaining 80 tones are each paying $100 per tone of additional costs this would yield: 80 tones * $100 / 20 tones abated =$500 per tone.

[10] This is the last year ECV choose to publicly disclose operating returns for this segment. A 7.2% return for 2018 is roughly their eight-year average which saw a peak of 9.8% and a trough of 5.9%.

[11] Some commentators call for 2 cents per kWh by 2050.

[12] Thunder Said Energy, June 2021

[13] Grid scale storage is most certainly an option to avoid curtailment. However, curtailment is often the least expensive option. Grid scale storage today with daily charging/discharging requires a 22 cents per kWh spread. Green hydrogen (used to back up intermittent renewables) is even more costly at 64 cents per kWh. It is not trivial to get around the second law of thermodynamics which states that any energy transferring process will come at an efficiency loss (10-20% for batteries, 50-60% for hydrogen.

[14] Thunder Said Energy, June 2021.

[15] As a proportion of total costs, capital costs make up a significant fraction as the marginal costs of production is often close to zero.

[16] Ibid.