Transition to alternative energy wrecks GDP in oil and gas producing economies

The distribution of ownership of transition risk associated with stranded fossil-fuel assets remains poorly understood. We calculate that global stranded assets as present value of future lost profits in the upstream oil and gas sector exceed US$1 trillion under plausible changes in expectations about the effects of climate policy. We trace the equity risk ownership from 43,439 oil and gas production assets through a global equity network of 1.8 million companies to their ultimate owners. Most of the market risk falls on private investors, overwhelmingly in OECD countries, including substantial exposure through pension funds and financial markets. The ownership distribution reveals an international net transfer of more than 15% of global stranded asset risk to OECD-based investors. Rich country stakeholders therefore have a major stake in how the transition in oil and gas production is managed, as ongoing supporters of the fossil-fuel economy and potentially exposed owners of stranded assets.

The transition to a global low-carbon economy entails deep and fast structural change that poses challenges for economic adjustment everywhere1,2. One key challenge both for the real economy and financial markets is the fast phase-out of fossil-fuel production, which will necessitate the write-down of major, functioning capital assets and reserves reflected as assets on fossil energy companies’ balance sheets. But while over 100 studies have analysed scenario-contingent early retirement of fossil-energy supply facilities3, this retirement has not been linked to financial ownership. As a result, academic and regulator studies undertaking stress tests of the financial system start from synthetic shocks to financial assets, rather than the underlying real assets4,5,6. The distribution of financial ownership and exposure to loss risk remains insufficiently understood.

Asset stranding is the process of collapsing expectations of future profits from invested capital (the asset) as a result of disruptive policy and/or technological change7,8. This loss of value in fossil-fuel assets is reflected in investor expectations of enterprise value and therefore market prices, including—where listed—stock market indices. Such price corrections lead to a wealth loss for the ultimate owners of these assets; additionally, further losses can propagate to other entities indirectly through highly connected financial networks.

Asset stranding becomes a social concern where these effects destabilize financial markets with negative repercussions in the real economy such as on pensions and government finances9,10. The (premature) obsolescence of capital stock is a recurring feature of dynamic, capitalist economies, as new products and industries replace old ‘sunset’ ones, and is not typically associated with systemic financial risks because the financial sector is buoyed by the new ‘sunrise’ sectors2. Yet, in the case of the low-carbon transition, the rate of industrial change required for achieving a 2 °C, let alone 1.5 °C, goal is so large11 that the rapid collapse of fossil-fuel ‘sunset’ industries presents major transition risks6,12.

Here we map comprehensively the current global financial geography of stranded oil and gas asset risk for equity ownership. We trace potential losses from extraction sites through corporate headquarters and their immediate shareholders (including banks and fund managers) all the way to the ultimate owners (government and individual shareholders) for oil and gas extraction companies worldwide. We comprehensively link fossil-fuel stranded assets and transition risk studies at the asset level for the transmission channel of equity mispricing. We distinguish both geographic and functional characteristics of the organizations along the equity ownership path. We find that exposure to wealth losses is more evenly shared geographically than the distribution of oil and gas production assets may suggest. Therefore, private investors in rich countries have both a larger stake in continued fossil-fuel production and greater exposure to stranded assets than the literature has so far suggested.

Estimating stranded assets and wealth losses

We operationalize asset stranding as the effect of a change in expectations on the present value of discounted future profit streams. We calculate profits given expectations per asset. Energy is supplied from 43,439 oil and gas production assets based on Rystad’s Ucube dataset. Whether an asset is expected to supply demand depends on its present-day production cost and reserve profile in relation to the expected market-clearing oil price. If investor expectations for total demand for oil and gas fall, some assets must become unprofitable relative to initial expectations; that is, the oil or gas price falls below the break-even price for those assets.

Once the stranded assets are determined, we establish a four-stage description of who bears the loss. At stage 1, asset stranding is attributed to the country where sites are located. Stage 2 aggregates the ownership of stranded assets by fossil-fuel company. Each asset is owned by one or more oil companies (we count 69,990 ownership links). The loss is allocated to the country where the parent company has its headquarters. Out of the 3,113 active oil and gas parent companies reported in the Rystad database, our analysis identifies 1,759 as owning 93.4% of all losses. The 1,772,899 company nodes in the global equity ownership network are curated from Bureau van Dijk’s ORBIS database. At stage 3, this allows us to further trace the financial losses through the directed graph of ownership using a network model. Losses pass through 33,836 separate corporate ownership and fund management nodes, including most of the world’s large financial companies, to 16,171 ultimate corporate owners. At stage 4, we track all losses to their ultimate owners, governments and individuals, as shareholders or outright owners of companies or investors in funds, including pension funds. To account for company-level losses, we subtract losses from shareholder equity on the balance sheet reported in ORBIS in the most recent year (typically 2019). We detail our stranding and loss propagation models in Methods.

To quantify profit losses from changing expectations, we use an initially expected (baseline) scenario of global oil and gas demand and prices, upon which prior financial value has been estimated, and a revised scenario representing updated expectations resulting from climate policies (policy scenario). We call the expectations shift a realignment.

Our focus is on the medium realignment, in which the baseline scenario follows IEA’s WEO 2019 current policies scenario, consistent with 3.5 °C median warming in the 21st century. We refer to this baseline as investor expectations, InvE. The policy scenario, termed EUEA, incorporates the stated policies of the European Union (EU) and East Asia (EA) to reach net-zero greenhouse gas/CO2 emissions by 2050/2060, respectively, noting that non-CO2 emissions are exogenous and follow RCP2.6 (Methods). The EUEA scenario has a median warming of 2 °C. In line with the IEA’s projections1, the policy scenario features sell-off (SO) behaviour, whereby companies operating at low-cost fields in the Middle East supply a larger and increasing share of the market as the global oil and gas demand peaks and declines and low-cost producers scramble to capture the declining market.

We assume expectations to realign in 2022. Because the expectations underlying current asset prices vary, evolve continuously and are extremely difficult to quantify, we consider three other possible realignments, each yielding a magnitude and distribution of risk ownership.

Oil and gas demand and prices in baseline and policy scenarios are produced by the E3ME-FTT-GENIE integrated assessment modelling framework, which like CGE models provides sufficient sectoral disaggregation13,14, while remaining theoretically consistent with asset stranding. It couples a macroeconometric model of the economy that distinguishes 43 sectors and 61 regions and their trade (E3ME), an evolutionary energy technology model distinguishing 88 supply and demand-side technologies (FTT)15,16,17 and a carbon cycle and climate system model of intermediate complexity (GENIE)18. The embedded energy market model determines which assets supply demand (Methods). The overall model and alternative scenarios are presented in Supplementary Notes1 and 2.

We illustrate our calculation of total stranded assets in the medium realignment (Fig. 1a). Annual revenue in the InvE baseline grows, while in the EUEA policy scenario it reaches an early peak and falls steadily due to both lower quantities and prices. Upstream oil and gas lost profits, a subset of lost revenue after subtracting labour and intermediate input cost, is represented by the light green wedge. We discount differences in expected profits by 6% y−1 (Fig. 1b) to calculate the present value of stranded assets, which sums to US$1.4 trillion (see Supplementary Note 3 for a sensitivity analysis about the choice of discount rate). That is, investors realign their expectations of the ability of assets to generate profits from the baseline to the policy scenario in 2022 over a 15 yr horizon of profits, and present-value accounting translates deflated profit expectations into lower asset value.