Published as part of the ECB Economic Bulletin, Issue 2/2023.
In 2022 the ECB conducted a climate risk stress test of the Eurosystem balance sheet as part of its action plan to include climate change considerations in its monetary policy strategy.[1] The aims of this exercise were to (i) analyse the sensitivity of the Eurosystem’s financial risk profile to climate change; and (ii) enhance the Eurosystem’s climate risk assessment capabilities. The scope of the exercise covered a number of the Eurosystem’s monetary policy portfolios, namely its holdings of corporate bonds, covered bonds, asset-backed securities (ABSs), as well as its collateralised credit operations.
This climate risk stress test used scenarios developed by both the Network of Central Banks and Supervisors for Greening the Financial System (NGFS) and the ECB. It employed three NGFS Phase II long-term scenarios[2] that project macro-financial and climate variables over a 30-year horizon. The scenarios differ in terms of the extent to which climate policies are assumed to have been implemented (primarily in the form of a carbon tax) and the different types of climate risk that are expected to materialise as a result. The hot house world scenario entails severe physical risk but does not lead to transition risk, as it is based on the assumption that climate policies are not enforced. Under the disorderly transition scenario, the implementation of climate policies is delayed, leading to severe transition risk but only limited physical risk. The risks stemming from the disorderly transition and hot house world scenarios are analysed against those arising from the orderly transition scenario, which assumes that climate policies are implemented in a timely manner. In addition, the stress test exercise considered two further short-term scenarios designed by ECB staff: a flood risk scenario, which includes severe physical hazards materialising over a one-year horizon; and a short-term disorderly transition scenario, which frontloads sharp increases in carbon prices over a short-term (three-year) horizon. In view of the challenges associated with designing long-term climate scenarios, these two short-term scenarios provided useful additional input to the analysis, with the flood risk scenario setting out how a severe physical hazard could potentially materialise across the whole of Europe.
The methodology and scope of the exercise were aligned with the 2022 climate risk stress test[3] conducted by ECB Banking Supervision and the 2021 ECB economy-wide climate stress test[4]. Under all five scenarios, the exercise applied credit risk shocks using satellite models specific to each type of financial exposure. These shocks are based on the aforementioned 2022 climate risk stress test by ECB Banking Supervision as well as on NGFS data. In addition to credit shocks, the exercise used market shocks in the form of increases in risk-free interest rates and corporate bond spreads.
This climate risk stress test of the Eurosystem balance sheet used the Eurosystem’s financial risk assessment framework as the basis for its risk estimation, using the aforementioned shocks. This framework, which is also used for the Eurosystem’s regular financial risk assessment and reporting tasks, is based on a joint market and credit risk simulation model. The analysed results take the form of an expected shortfall[5] estimated at a 99% confidence level over a one-year horizon. Two different perspectives were considered: a standalone risk approach, which calculates the risk of each portfolio independently; and a risk contribution approach, which determines the contribution of each portfolio to the total risk for the Eurosystem. The cut-off date for the Eurosystem balance sheet and market data was 30 June 2022.
Table A
Overview of the scenarios and main results of the 2022 climate risk stress test of the Eurosystem balance sheet
The results of the exercise show that both types of climate risk – transition risk and physical risk – have a material impact on the risk profile of the Eurosystem balance sheet. The disorderly transition and hot house world long-term stress scenarios produce risk estimates that are between 20% and 30% higher than those under the orderly transition scenario. The hot house world scenario generates a higher risk impact, showing that physical risk has a greater impact on the Eurosystem balance sheet than transition risk. Integrating climate change risk into the Eurosystem’s regular risk assessment and provisioning frameworks should make it possible to modify risk control frameworks and build up financial buffers over time, thereby addressing such risks.
The aggregate result is driven mainly by outright holdings of corporate bonds, which under all scenarios make a larger contribution to the total risk increase than the other types of financial exposures included within the scope of this exercise. The impact of climate risk on corporate bonds is particularly concentrated in areas that are specific to each risk type. The impact of transition risk, for example, is primarily concentrated in a limited number of sectors that are particularly vulnerable to climate risk (and which have, on average, a high level of emissions as a percentage of revenue), whereas the impact of physical risk is concentrated in certain geographical areas.
The Eurosystem’s corporate bond holdings entail a similar degree of climate risk as the outstanding market volume of securities eligible for such purchases. This can be seen by performing the same stress test on a benchmark sample of securities that meet the Eurosystem’s eligibility criteria and are weighted by market capitalisation. Under the two adverse scenarios, the resulting risk increases do not significantly differ from the results obtained for the Eurosystem balance sheet. This outcome was expected owing to the fact that, at the cut-off date, the Eurosystem’s corporate bond purchases were determined by a market capitalisation benchmark, as climate change considerations were only incorporated into those types of purchases as of October 2022.
The relative risk increase for both covered bonds and ABSs is greater under the hot house world scenario than under the disorderly transition scenario. The relatively high sensitivity of these assets to physical risk is also reflected in the outcome of the flood risk scenario. Under this scenario, the increase in risk estimates for covered bonds and ABSs is much higher than that for corporate bonds, and it is also higher than under the long-term scenarios. As a result, the contribution of covered bonds to the total risk increase under this scenario is particularly significant. This is not the case for ABSs, however, as the portfolio is considerably smaller. Also, the result for the flood risk scenario highlights the importance of the house price channel in the transmission of climate risk, as covered bonds and ABSs secured by real estate are particularly exposed to fluctuations in housing market valuations.
Collateralised credit operations, meanwhile, make only a small contribution to the total risk increase despite the large size of the exposure. This exercise considered credit operations collateralised by corporate bonds, covered bonds, ABSs and uncovered bank bonds. The lower risk per unit of exposure of these lending operations can be linked to their double default nature: although climate risk stress is channelled through both the counterparty and the collateral, the risk only materialises under scenarios whereby the counterparty defaults and the value of the collateral falls below the level of protection offered by applicable valuation haircuts. This typically occurs in instances when the collateral issuer also defaults. Climate risk is therefore concentrated in exposures to specific counterparties, especially under the hot house world scenario, in which certain institutions and the collateral they have posted are both located in regions that are severely affected.
Climate risk stress tests of the Eurosystem balance sheet are expected to be carried out on a regular basis in future. These future exercises should provide an opportunity to further enhance the methodology and expand the scope of the financial exposures covered. Looking ahead, climate risk considerations should also become an integral part of the existing risk management framework, which involves an analysis of the total financial risk for the Eurosystem against the existing financial buffers.
For further details, see the press release “ECB presents action plan to include climate change considerations in its monetary policy strategy”, ECB, 8 July 2021. The climate risk stress test was conducted by the ECB’s Directorate Risk Management in cooperation with the Eurosystem’s Risk Management Committee.
For further details, see “NGFS Climate Scenarios for central banks and supervisors”, Network for Greening the Financial System, June 2021.
For further details, see “2022 climate risk stress test”, ECB Banking Supervision, July 2022.
The approach is described in Alogoskoufis, S. et al., “ECB economy-wide climate stress test”, Occasional Paper Series, No 281, ECB, Frankfurt am Main, September 2021.
This expected shortfall is a tail measure of the distribution of the losses on the Eurosystem balance sheet, which are computed based on relative price differences between the snapshot date and one year later: the shortfall is computed as the average of the worst 1% of losses in the distribution.