In line with its mandate, the European Insurance and Occupational Pensions Authority (EIOPA) initiates and coordinates EU-wide stress tests (ST) to assess the resilience of Institutions for Occupational Retirement Provisions (IORPs or “pension funds”) and their impact on the financial stability and real economy.
A sample of IORPs is required to participate in this Stress Test exercise that started on 13 April 2022 and will run until 13 June 2022. The EIOPA is expected to publish the results by December 2022.
The purpose of the exercise is to assess the impact of three sources of risk: the aftermath of COVID-19, the increase in inflation rates and climate transition to a sustainable economy.
Insurers and other IORPs trying to assess the potential impact of climate risk on their balance sheet (for example as part of their OR(S)A) might find pertinent information in the material provided by the EIOPA. The new ORSA circular, published by National Bank of Belgium (NBB) on 23 March 2022, explicitly foresees climate risk as part of the own self-assessment to be performed by insurers and is applicable as of day one.
Undertakings with exposure to inflation risk (through-inflation sensitive exposures and operational expenses) and IORPs, concerned by the purchasing power of pension beneficiaries, might also be interested in the inflation component related to the assessment of an increase in inflation.
The objective of this paper is to provide the necessary insight to Insurers and IORPs, who are interested in climate and inflation risk approaches used by the EIOPA in the 2022 ST exercise.
Inflation risk – unprecedented high levels
The pandemic led to exceptional measures, such as the closure of parts of Member States’ key economic sectors and strong government intervention through financial support. This resulted in remarkably high fiscal deficits, as well as supply shortages. The situation led to escalating prices for commodities, and other items, such as petrol, energy, housing and food.
While there is still uncertainty at this stage whether this spike is temporary or long term, there are short-term concerns that strong GDP growth along with spending and consumption may drive inflation rates higher than those observed over the last decade. Without considering potential reactions from the Central Banks, inflation affects the real return on investments and may result in higher spreads of already highly indebted sovereigns and corporates.
Given that the stress test was conducted at the end 2021, the recent developments related to the Russian invasion of Ukraine, which has put additional pressure on energy and commodity prices, are not reflected in the scenario.
To measure inflation risk for IORPs, EIOPA has combined a qualitative and quantitative approach to assess mainly the impact on future retirement income and technical provisions.
The graphs below give an overview of the risk-free rates (RFR), inflation rates and real rates under the baseline and adverse scenario of the STs performed in 2017, 2019 and 2022:
We note the following:
- The baseline risk-free rate curve as of the 2022 ST is higher than in the 2021 insurance ST but still below the 2017-2019 levels.
- The 2022 ST adverse scenario is a significant up shock with an approximately 80 basis points (bps) parallel shift until the Last Liquid Point of 20 years before extrapolation. The upward shock goes in the same direction as the 2019 ST, but the shock magnitude is bigger (shock on the 10-year rate was +25 bps). These upward shocks contrast with 2017 ST when the shock direction fell (shock on the year rate was -51 bps). The upward shock also contrasts with the latest insurance stress tests, which consisted in down interest scenarios in the context of “a low for long” interest rate environment (not shown on the above graphs). We seem to be at a turning point at EOY21 with higher rates and concerns for potential stagflation.
- The inflation rates in the 2022 ST are calibrated based on inflation rates swaps as of EOY21 and indicate expected inflation of 3.5% in 2022, 2.6% in 2023 before approaching the 2% ECB target level sometime in 2032, while remaining slightly above this threshold over the whole liquid term structure. We further note the absence of an adverse scenario on inflation rates while observed inflation rates in the EU area over the first 3 months of the year 2002 are much higher than 3.5% with an average of approx. 6%, following the Ukraine invasion and energy supply shock.
- The real rate levels in 2022 ST are much lower than in 2019 and 2017 ST and remain negative, whereas the adverse real rate scenario in the current ST appears to be close to the 2017 adverse scenario.
Regarding the increase in nominal risk-free rate, we understand that the shocks aim at capturing globally the impact of climate risk (see section 3): the 80 bps up parallel shift on the risk-free rate is based on the modeled shock on government bonds (assumed proxy for the risk-free rate in the National Institute Global Econometric Model (NiGEM)). The inflation curve in the baseline is not subject to any further stress.
The combined effect of discounting and inflation will be entity specific depending on exposures to each risk and related duration. Compared to the 2019 IORP ST, the baseline scenario of the 2022 ST is characterized by lower interest rates and higher inflation rates and is already a more hostile environment than three years ago. The adverse scenario with higher discounting rates and similar inflation rates is likely to lead to different results and conclusions than the previous stress test. It is therefore essential to focus on the purchasing power of people and future retirees next to the financial position of financial institutions, as it can threaten the orderly functioning of an economy and its financial stability.
Inflation risk is complex with different drivers and spillover effects, impacting insurers, pension funds and Member States in different ways.
While there is always some basis risk of inflation, it would be good, from a risk management perspective, to assess the inflation exposures and underlying drivers on the balance sheet, whether the retained best estimate assumption differs significantly from the one used by the EIOPA with some possible adjustments following the recent observations, and whether the stress test considered is sufficiently severe.
Climate risk – sudden and disorderly transition scenario
The EIOPA climate scenario analysis consists of four key steps:
- Scenario narrative: define the context of the study in terms of global warming limitation ambitions, climate policy design and technological progress
- Integrated Assessment Model (IAM): model environmental and energy variables accordingly (i.e. energy demand, energy capacity, energy prices, emissions trajectories, etc.)
- Macroeconomic model: further specify broad economic outputs such as GDP, employment, inflation and interest rates pathways
- Financial model: assess financial impact across economic sectors and countries based on sensitivity to climate-related risks or even more granularly at the individual firm level
The following diagram gives a high-level view of how the stress test shocks for the 2022 ST IORP have been derived:
2.1 Scenario narrative – NGFS
The climate scenario is based on the June 2021 framework published by the Network for Greening the Financial System (NGFS). Consisting of Central Banks and Supervisors, their aim is to establish a reference framework to analyze climate risk.
Once the limit of temperature increase has been defined, the related reduction in greenhouse gas emissions and the carbon price trajectory, needed to achieve these, are determined.
Typically, NGFS considers four different situations as illustrated below:
The situation looked at by the EIOPA is a disorderly transition, where new climate policies are only introduced by 2030 and availability of carbon removal technologies is assumed to be low. To make up for inaction, stronger policies are needed to limit global warming below 2°C in line with the Paris Agreement. This leads to a higher transition risk than under the orderly transition where the abrupt implementation of policies affects the energy sector, as well as several areas of the real economy. Given the short time horizon, the following elements are disregarded: physical risks, monetary or fiscal policy change and benefits from the green transition.
2.2 Integrated Assessment Model – REMIND-MAgPIE
The scenario narrative dynamic is modeled using the REMIND-MAgPIE framework. This integrated assessment model simulates, in a forward-looking fashion, the dynamic between the macro-economic system and the energy and climate systems.
2.3 Econometric model - NIGEM
Once the environmental and energy variables have been defined by NGFS, it is necessary to assess the financial impacts. The NiGEM works this step out by providing consistent macroeconomic projections on interest rates, house price variables and equity prices. The NiGEM methodological approach is a reference that has also been used in the ACPR and DNB stress tests.
2.4 Specificity to this 2022 IORP ST – frontloading macro-economic impacts
As this is the first climate ST for IORPs, a pragmatic approach has been retained where projected variables over the years have been translated in a shock at time to:
This Stress Test assumes an increase in the carbon price to occur immediately in 2022 (e.g. frontloading in 2022 the 2031-2033 NiGEM calibration), along with an instantaneous shock on financial investments.
This makes any comparison of the resulting shock with other STs performed by local supervisors spurious.
2.5 Resulting instantaneous shocks
The modeled variables and (some of) the resulting shocks to the different scenario variables – as provided by EIOPA - are presented in table below:
As expected with the disorderly transition narrative, interest rate tends to increase (+0.8% on 10-year risk-free rate), reflecting the inflationary pressure created by increasing energy and carbon prices, as well as the increased investment demand required by the transition needs.
Shocks provided by the EIOPA indicate government bond yields would increase for most of the EU Area countries (10-year yield of Belgium +122 bps, Germany +120 bps and Italy +95 bps). Conversely, the yield for Polish government bonds would decrease. One possible explanation could be that the shocks provided by EIOPA would be calculated based on the difference between the future inferred spot rates starting from end-of year 2020 interest rates, as per NGFS scenario projection (see page 7 of ESRB document – based on NiGEM), and the current spot rates, as observed end-of year 2021 (i.e. starting point values).
Using the example of Poland, the 10Y rate was 3.7% on 31 December 2021, while the corresponding rate in the disorderly transition scenario is projected to be 2.7% (ESRB document). The resulting differences of -1% can be reconciled with the shock provided by the EIOPA shown in the table above (-100 bps). Following the same logic, one can reconcile, on one hand, the shock on 10Y Belgium bonds as provided by EIOPA shown in the table above (+122bp) and, on the other hand, the difference between the 10Y rate of Belgium bonds of 0.2% as of 31 December2021 and the corresponding yield in the disorderly transition scenario of 1.4% (ESRB document).
In the absence of NGFS scenarios being calibrated based on EOY21 market conditions, an alternative approach could have been to take the difference in yields between 2031 and 2033 in the disorderly scenario, “frontload” the shock and add the difference to the starting point yield EOY21.
Corporate exposures (bond and equity)
Disorderly transition is also expected to affect corporate exposures differently, depending on their economic activities. Companies that decarbonize less efficiently will most likely be affected to a greater degree than those that decarbonize more efficiently. These aspects are reflected in the stress test, where, for example, equity and corporate shocks are the highest on companies whose principle economic activity is a) mining and quarrying and b) manufacturing of coke and refined petroleum products.
2.6 Limitations of the climate exercise
We note the following limitations of this first climate stress test:
- Scenario calibration date differs with ST starting point: climate scenario is considered from EOY20 (based on NGFS June 2021) versus EOY21 calibration dates characterized by another inflation and interest rate environment. This can potentially lead to undesirable effects.
- Frontloading climate shocks make any comparison with other reference stress tests spurious (e.g. ACPR, BCE, BoE).
- Same credit spread shock and equity shock are applied to counterparties belonging to the same sector, but possibly with a different credit rating and transition risk profile. Therefore, the shock does not consider differences in the financial situation/creditworthiness of counterparties nor company specific climate strategy and actions. This is an important limitation of this stress test, in particular for insurers/IORPs, who consider using shocks in the context of their Own Risk (and Solvency) Assessment.
- Assess the impact of alternative plausible scenarios (see for example the other NGFS reference scenarios), which might complete the picture of the company’s financial resilience to different plausible views of the future (as identified by the experts who have contributed to the elaboration of those reference scenarios).
- Better capture the chronology of events and assess their impact on various projection horizons (possibly contemplating change, over time, in business strategy) – as opposed to an instantaneous shock applied to the company’s static balance sheet as of 31 December 2021.
2.7 Possible challenges
In the table below, we highlight some of the possible challenges for participants to meet the Stress Test requirements:
This 2022 IORP ST gives interesting insights on inflation and climate risk even with the perceived limitations with regard to inflation uncertainty and its underlying drivers, as well as necessary proxies on climate shocks for a first exercise.
The high volatility and risk interconnectedness in our current environment further supports the need for strong risk management that combines quantitative and governance skills with sufficient agility at both the entity and societal levels, in a collaborative way.
Financial institutions have a key role to play in a sustainable society. They must preserve value for investors and future retirees, fund climate transition and offer protection from physical risk. As those topics are continually evolving, it is important to combine and leverage the reference studies, while identifying the areas for improvement.
Authors: Erik Van Camp, Partner Risk & Regulatory; Daphné de Leval, Director Risk & Regulatory; Samuel Silber, Senior Manager Risk & Regulatory.
- “Consistently with previous IORP stress tests, only EEA countries with material IORP sectors, exceeding EUR 500m in assets at year-end 2020, are required to participate. Therefore, Belgium is in the scope of this exercise and FSMA will choose the sample of IORPs participating in this exercise.” – see EIOPA Technical Specifications – 2022 IORP Stress Test (4 April 2022).
- See EIOPA issues Opinion on the supervision of the use of climate change risk scenarios in ORSA | Eiopa (europa.eu) for Insurers, EIOPA issues opinions on governance and risk management of pension funds | Eiopa (europa.eu) for pension funds.
- See 20220323_nbb_2022_09.pdf
- l.e. calibration dates EOY16, EOY17, EOY21.
- See ba153bc6-c1aa-f6e5-785b-21c83f5319e5 (europa.eu): 5.1% in January, 5.9% in February, 7.5% in March.
- E.g. Insurers are exposed to expenses (incl. salary) inflation, medical expenses indexation, claims indexation, reinsurance recoverables following applicable indexation clause but also some management fees depending on the contracts with the asset managers. IORPs are also exposed to expenses, management fees, salary increases, benefits indexation. The resulting purchasing power of active citizens and future retirees will further impact demand and GDP.
- E.g. by comparing with the SCR expense under Solvency 2 (10% immediate shock on the expense basis and 1% parallel shift on the trend) or by applying a 70% factor to the SCR interest shock percentage similarly to the QIS for IORPs where nominal rate both inflation and real interest rates are assumed to contribute to the same global volatility level of the nominal rate and are independent.
- See a description here: NGFS (2021): NGFS Climate Scenarios for central banks and supervisors, https://www.ngfs.net/sites/default/files/media/2021/08/27/ngfs_climate_scenarios_phase2_june2021.pdf.
- Integrated assessment models (IAMs) refer to models which integrated knowledge from two or more domains (ex: climate sciences and economy) into a single framework.
- See eiopa-22-315_2022_iorp_st_inputs_helper_tool.xlsx available on EIOPA website: Climate stress test for the occupational pensions sector 2022 | Eiopa (europa.eu)