Counting the cost of climate transition risk — an Energy Finance approach
At CPI Energy Finance, scenario analysis is at the heart of our work on climate transition risk and as we expand this programme, communicating how we approach our work is becoming increasingly important.
Policy, technology and finance are all factors driving the global energy transition. In that transition, many companies, particularly producers of high carbon commodities, such as coal and oil, face the risk of decline in the value of their core businesses as demand for high carbon products drops. The change in the value of assets or companies in a low carbon world, relative to today’s business as usual path, is what we call climate transition risk.
As energy companies are some of the world’s largest corporations, a sharp decline in their fortunes could have serious knock-on consequences for global financial markets. In recent years, companies, investors and the central banks responsible for financial stability have started to pay closer attention to these risks. However, previous research by CPI EF and others has shown that more risk in commodity markets sits with the sovereign — ie, national, public balance sheets — than private investors.
CPI EF is working on a series of studies in collaboration with national governments and public institutions. In March this year, we launched the results of our first study, Understanding the impact of a low carbon transition on South Africa. In that report, we identified a serious risk to South Africa’s investment grade sovereign credit rating from climate transition risk of $124bn, mostly driven by $83.7bn in the export coal sector.
Our choice of model
For South Africa, we selected an approach to modelling international thermal coal and oil markets that would enable us to see how a global low carbon transition would affect supply and demand, and the prices that producers would receive. We then built South Africa-specific models to understand the impact of those price and volume changes on adjacent domestic industries, including coal mining and logistics (rail, road, ports), domestic coal users (mainly power generation and liquid fuel production) as well as the rest of the fuels value chain (including oil refining, transport and distribution, storage, wholesaling and retailing).
When combined, these models not only allowed us to allocate the explicit risk on an asset by asset basis but also how the risk is transferred between parties at an implicit level, ie, where the flow of risk may not be immediately obvious because of a contractual, legal or regulatory obligation. This flow of explicit and implicit risk is explained in more detail below.
Economists and policymakers around the world commonly use tools such as computable general equilibrium models to assess the long term impact of policies and shocks on the economy. These models balance supply and demand for goods, services and factors of production, such as labour and capital, across an economy as a whole, but they do not provide asset level granularity.
However, our models enable us to apply a much higher resolution of detail. For example, we were able to distinguish between the impacts on international fossil fuel majors and companies based mainly in South Africa. In some cases, these companies had a disproportionately large exposure to climate value at risk as compared with their international counterparts. Figure 1 shows the comparison in transition value at risk relative to market capitalisation between Exxaro, a coal miner with assets concentrated in South Africa and international miner Anglo American.
Scenario analysis
Once we had developed the range of models for South Africa, we then applied two main demand scenarios. First, the “business as usual” (BAU) scenario informed the “counterfactual”, ie what happens if there are no changes to current trajectories. This was largely derived from International Energy Agency (IEA) data. Second, our “well below 2°C scenario” (2DS) represented a more stringent low carbon transition — where demand for fossil fuels would decline globally because of carbon reduction targets. This was based on in-house research as well as IEA assumptions. For coal, actual demand data came from the IEA’s global forecast, which was adjusted or changed to reflect our assumptions in the case of the 2DS scenario as we explain later. We then forecast the price of seaborne thermal coal by balancing these two forecasts (BAU and 2DS) against the same supply curve, as obtained from research consultants WoodMackenzie.
With our oil models, we used IEA New Policies and Sustainable Development demand scenarios to forecast global oil prices in each case (BAU and 2DS). For South Africa, our assumption was that domestic policies continue on a BAU path and so we therefore allowed South African consumption of oil to adjust freely to the lower global price forecast in the 2DS.
When compared against each other, our two scenarios determined the difference between oil and gas prices and consumption in the BAU, versus a trajectory that reflected the most ambitious carbon reduction targets.
The seaborne coal trade optimisation model works by using linear programming, a mathematical technique used to find the best outcome of a set of variables with linear relationships. Supply curves are constructed for each region, with the range of coal available in the market ordered by the cost of exporting to each respective region. The linear programme then identifies the unique set of trade flows which minimise the level of cost in the system in each scenario and for each year. Import prices are then derived from the marginal cost, or cost of the most expensive coal, supplied to each region. The resulting revenues, from the trade flows and prices, and costs, from the underlying supply data, can be used to calculate value changes through time in each scenario.
A peek under the hood
The 2DS scenario drives this change in value. It is therefore important to understand that there are actually a multitude of different pathways that could all lead to limiting global warming to well below 2 degrees above pre-industrial levels. Their implications on demand forecasts for specific commodities can therefore be materially different — despite allowing for a similar level of cumulative GHG emissions between them.
An example of this would be the IEA’s Sustainable Development Scenario which assumes a higher amount of carbon capture and storage (CCS) uptake, especially among thermal power plants. The IEA therefore forecasts higher “2 degree” global coal consumption in their World Economic Outlook report than we do given our more bearish CCS outlook because its use in the power sector would be more limited than is generally assumed.
Differences such as these can have a large impact on studies like our South Africa transition risk work, where we dig down to see the impacts of the energy transition on specific industries and market players.
Explicit and implicit risk allocation
What sets apart our transition risk work in South Africa from other studies is how we looked at the impact of transition on the sovereign by assessing the direct risk exposure that would be apparent from ownership and policy, as well as the flows of risk that may not be immediately apparent. We therefore separated the climate transition risk of the South African government into explicit risks and implicit risk transfers and contingent liabilities. Explicit risks are those relating to contractual, legal or regulatory obligations, current policy, taxation and royalties, among other factors. Implicit risks, however, are distributed from businesses and other participants who cannot bear the risk and transfer those risks as liabilities to the government.
We found that while 16% of explicit risk flowed to the government, once implicit risks had been factored in that level of exposure rose to more than half, ie by $20bn to $66.8bn.
Behind the Sankey flowchart above, however, lies multiple steps in the modelling. Excel-based sectoral risk allocation models first assess the explicit allocation of this quantified risk between the government, investors, consumers and workers. Equity ownership of assets helps allocate profits in the industry between investors (including private and public investors). Fiscal regimes determine the exposure to local and national governments through taxes, royalties and levies as differences in government revenue between BAU and energy transition scenarios would present a value at risk to governments. Contractual arrangements within value chains would also need to be accounted for.
For instance, you would expect a loss in value (compared to BAU) for coal infrastructure providers in an exporting country in a scenario with less international coal demand. However, as we found in the South African example, contractual arrangements between mining companies and the main state-owned rail company protect the infrastructure provider from a fall in volume and shift additional explicit risk onto the mining companies.
A crucial part of the next exercise is calculating the amount of risk which could be reallocated in the economy in the form of implicit risk transfers, as companies and their investors seek to transfer as much risk as possible to other parties. For instance, falling asset values and profits could lead to companies facing bankruptcy and defaulting on loans, this could lead to a transfer of risk from the private sector to the sovereign in the form of contingent liabilities from the sovereign guarantees of loans, support for unemployed workers or the need for the government to take on decommissioning and abandonment liabilities of energy assets which can no longer be covered by financially insolvent owners.
The last step was to assess how these explicit and implicit flows of risk affect the sovereign’s overall risk profile. This stage allows us to step back and assess the aggregate impacts and risk exposure on the sovereign government and how these impacts could alter the sovereign’s financial standing and credit rating. We can then evaluate different policy responses by their ability to de-risk and mitigate the exposure of the sovereign, which can then help guide our overall policy recommendations.
Next steps
We are now striving for greater resolution and accuracy through the development of in-house probabilistic scenarios. These will take a more all-encompassing view of the range of possibilities for technology development and structural demand in markets that could influence different paths being taken in reaching the same 2D scenario.
This will help us develop corporate assessments for investment and management strategies and also those for the finance sector at the portfolio level. As we grow our climate transition risk workstream, we will over a greater range of countries, including emerging market commodity importers, developed nations and oil and gas exporters.
Assessments of transition risk are not the whole solution to tackling climate change, they are a vital step on the way and a crucial path towards incentivising the most transformative economic agents, such as financial institutions and governments, to begin leading the global decarbonisation effort.
Muhammed Anwar is a consultant at Climate Policy Initiative Energy Finance and led the modelling for the transition risk work in South Africa.
