Climate change will unevenly impact the European financial system
By Paul Hiebert
Climate change will impact those parts of the financial system most exposed to its disruptive effects. This column analyses a new financial stability risk mapping for the EU financial system, linking financial exposures of thousands of banks, insurance companies, and investment funds to millions of firms subject to climate risk. It highlights a high level of risk concentration, both in European regions subject to climate hazards as well as economic sectors with diverse carbon emission intensities. Long-term scenario analyses suggest that the risks will be best addressed through proactive policies that directly contain global temperature rises.
Accurate measurement of the prospective impacts of climate change on financial stability rests on a detailed mapping of its disruptive potential, both over space and time. This requires first a careful linking of climate risk drivers, and their granular region-, sector- and firm-specific impacts on the economy, to financial exposures (Cruz and Rossi-Hansberg 2021 underscore the unequal impacts of climate change across regions). It also requires modelling long-dated risk of a possibly insidious nature, assessing the prospective benefits of foregone financial sector losses from natural hazards accompanying the pace of global temperature rises, against costs in the form of measures taken to mitigate them (Bolton et al. 2021 argue for a combination of public interventions and private sector mitigation strategies to reduce the long-term implications of climate-related events).
The ECB and European Systemic Risk Board (ESRB) have joined forces to measure financial stability risks from climate change for the EU. Initial work focused on how existing data and models of central banks and supervisors could be deployed for climate risk analysis (ESRB–ECB 2020). A new report focuses on deepening granular measurement of climate risk drivers, while extending the horizon of models for long-term scenario analysis (ESRB–ECB 2021a). In particular, a granular topology of current financial risk exposures stemming from both physical and transition risk aspects of climate change has been constructed for millions of global firms, as well as thousands of financial intermediaries in the EU. Long-dated scenario analysis has leveraged this risk topology to model future credit and market risk losses, to gain insights into path dependence of climate hazards associated with an ongoing rise in global temperatures on the stringency of action to mitigate carbon equivalent emissions.
Results from this work suggest that, while the aggregate financial system seems able to weather first order impacts of climate change, its resilience will both vary spatially and evolve temporally.
Financial vulnerability comes from compound physical risk events, notably river flooding risks combined with potential for pronounced heat and water stress in some regions
A concentration of vulnerabilities across EU regions related to physical climate risk implies stranding risks in case of a coalescing of hazards. Looking over the next 20 years, a combination of riverine floods with wildfires, heat stress, and water stress could generate strong localised impacts, and leave a collective 30% of euro area bank exposures to firms at risk. Rising sea levels could significantly add to credit exposures later this century. Losses from climate-related hazards might be amplified by protection gaps – relating either to compromised physical collateral (backing the majority of secured exposures), or insurability (noting a starting point of only 35% of economically relevant climate losses estimated to be currently insured in the EU).
Figure 1 Physical climate risks
A) Maximum firm exposure to physical hazards
B) Share of euro area banks’ credit exposures to firms by corporate physical risk level (percentages of total bank exposures to firms)
Sources: Four Twenty Seven, an affiliate of Moody’s, AnaCredit and ECB calculations.
Notes: The location of firms’ headquarters and that of their largest subsidiaries are used as proxies for firm location. Data coverage varies by country, selected firms may not be representative of all firms within the country. Top panel: Based on 1.5 million firms in Europe. Each dot stands for one firm, its colour refers to the maximum exposure level across six hazards, including hurricanes, sea level rise, floods, water stress, heat stress and wildfires. Bottom panel: Bank loan exposure is taken from AnaCredit and matched with Four Twenty Seven data at corporate level. Credit exposures to NFCs above €25,000 are considered; total exposures amount to €4.2 trillion. 31% of exposures can be matched directly, 58% are matched using postcode-level aggregates of the Four Twenty Seven corporate level indicators and 11% cannot be matched this way due to missing geo-locational information in AnaCredit (“no information” in bottom panel).
Vulnerability to financial market repricing not only across sectors, but also within sectors
A holistic view of carbon emission intensity of firms (including also downstream emissions) suggests a concentration of vulnerabilities both across and within sectors. Exposures to highly emitting firms occupy 14% of collective euro area banking sector balance sheets, and are concentrated in the manufacturing, electricity, transportation, and construction sectors. Perhaps more importantly, 10% of bank balance sheets exposed to firms whose emissions efficiency varies enormously could be vulnerable to credit ratings downgrades, should carbon prices rapidly adapt to Paris-aligned levels (Klusak et al. (2021) warn of climate-driven ratings downgrades in the next decade).
Climate risk exposures are also concentrated within certain financial institutions
There is also concentration of vulnerabilities at the level of financial institutions. For transition risks of climate change, nowhere does the vulnerability seem higher than for investment funds, where over 55% of investments are tilted toward high-emitting firms, and an estimated alignment with the EU Taxonomy at only 1% of assets. While direct holdings of institutional investors such as insurers appear to be manageable given well diversified portfolios, indirect losses could result from investment fund cross-holdings estimated at around 30%. As for physical risks of climate change, exposures to hazards appear to be concentrated in the hands of only two dozen banks, holding up to 70% of exposures to areas of high or increasing risk over the next two decades. Perhaps more worryingly, such exposures may aggravate pre-existing vulnerabilities, with physical risk exposures skewed toward weakly capitalised and/or less profitable banks in the euro area.
Figure 2 Transition climate risks
A) Firm-level emission intensities within and across climate policy relevant sectors in the euro area (x-axis: scope 1,2, and 3 emissions in tonnes of CO2 equivalents per million US dollar revenue; y-axis: NACE 1 sectors)
Note: Only firms directly reporting emissions are considered (approximately 3,000 European firms).
B) Share of EU fund portfolios by ‘green’ firms compared with that of ‘high-emitting’ firms
Sources: Morningstar, Refinitiv, and ESMA.
Notes: Percentage share of each individual fund’s equity and corporate bond portfolio (vertical axis) that is allocated to firms classified according to their portfolio emissions: firms with emissions that are below the 33rd percentile for the data sample (‘Green firms’); firms with emissions greater than or equal to the 67th percentile (‘High-emitting firms’); firms with emissions that fall between these two groups (‘Typical firms’); and also firms for which no emissions information is available. The horizontal axis denotes individual funds, sorted according to the percentage share of exposures to green firms in the portfolio (from lowest to highest share).
Long-horizon scenario analysis can shed light on financial system losses resulting from climate change
This granular vulnerability mapping can be used to gain insights on the prospective evolution of losses in the financial system as global warming gains further momentum. While a rise in temperatures accompanying carbon emissions appears inexorable (IPCC 2018), its pace and scale will be determined by the timeliness, stringency, and effectiveness of remedial action (NGFS 2020). This strong path dependence is borne out by climate scenario analysis, suggesting financial stability costs accrue over time from insufficiently orderly policy and effective technologies to limit global temperature rises.
- A mapping of Network for Greening the Financial System (NGFS) scenarios to 55 economic sectors and numerous regions suggests that credit and market risk could cumulate from a failure to effectively counteract global warming. Notwithstanding uncertainties around methodologies analysing such long dated horizons, scenarios indicate that physical risk losses – particularly for high-emitting firms – would become dominant in around 15 years in the event of an insufficiently orderly climate transition, with falls of up to 20% in global GDP by the end of the century should mitigation prove to be insufficient or ineffective.
- First-order direct losses appear manageable for European financial system, but are concentrated, and possibly at risk of amplifying features.
- EU banking sector credit risk losses under adverse climate scenarios could amount to 1.60-1.75% of risk-weighted assets over a 30-year timeframe. Such a magnitude, around half that of adverse scenarios in conventional macroeconomic stress test exercises (albeit with a far shorter horizon), would be concentrated the electricity and real estate sectors.
- EU insurance sector market risk revaluation losses taking into account production plans of firms over the next 15 years are only 5% on average – but could be material in key climate-sensitive sectors such as oil, gas, and vehicles.
- Uneven EU investment fund exposures could also lead to large losses from direct exposures of up to 14% in the next 15 years, despite limited direct aggregate asset write down risk of only 1.2% in holdings of around €4.8 trillion in equity and corporate bond exposures.
Figure 3 Probability of firm defaults under the ECB’s top-down stress test (percentages)
A) Across sectors
B) Across time
Notes: Differences in firms’ default probabilities under the two adverse scenarios with respect to the orderly transition scenario, by sector and group of firms (mean firms and firms mostly exposed to physical risk). Top panel: The bars represent the median changes in default probabilities over the next 30 years; the dots report the changes in default probabilities when considering the firms that are most exposed to physical risk (95th percentile based on firms’ physical risk score). Bottom panel: solid line is median across all firms in the sample, dashed line is the average of most exposed/vulnerable firms in the sample.
Further measurement and modelling refinements would enhance foundations of evidence-based policy
Notwithstanding continued progress in measuring and modelling climate-related risk, much remains to be done. The ECB and ESRB will further invest in broadening and firming up analysis in the coming year. On the side of data, the heterogeneity of climate-related disclosures among firms and financial institutions implies that the granular and country-level results will be subject to refinements as progress is made in addressing data gaps and obtaining more complete climate-relevant reporting (see ESRB–ECB 2021b for a detailed exposition on data advances, and remaining challenges). On the side of modelling, the incorporation of higher-order amplifying aspects, including assumptions on dynamic balance sheet adjustments as well as sharper aspects related to so-called dual materiality impacts of bank lending impacts on climate outcomes, will help to obtain a clearer financial stability view beyond direct impacts (see Altunbaş et al. 2021 for estimates of bank credit to polluting firms, as well as De Haas and Popov 2018 on the role of finance in reducing pollution).
While still subject to many uncertainties, advances already in empirical understanding of risks provides valuable evidence, laying the groundwork to support nascent macroprudential policy considerations, in an increasingly heated policy debate.