The first application of stress tests can be traced back to the early 1990s, where individual banks were using them for their internal risk management processes.  Over the past decades, stress tests have become an inseparable instrument of risk management: supervisory authorities use them to measure a bank’s exposure to various types of risks, ranging from “fat-finger” errors, to rogue traders to cybercrimes. Today, stress testing frameworks are becoming instrumental for institutions worldwide to assess the risks stemming from climate change. In fact, the World Economic Forum 2020 Global Risk report identified a list of climate related risks that are more likely to happen than the risk of a cyber-attack. These climate related risks are, for example, extreme weather events, failure of climate change mitigation and adaptation and natural disasters.
Adaptation of the stress testing frameworks
Within the boundaries of their mandates, central banks and supervisors have been focusing on the development of appropriate tools to address climate related risks. Including the readaptation of traditional stress testing frameworks to capture the unique features of these new risks. The development of the stress test scenarios is usually broken down into two major categories: 1) scenarios to measure physical risks; and 2) scenarios to measure transition risks.
The Network for Greening the Financial System (NGFS) published its Guide to Climate Scenarios Analysis for Central Banks and Supervisors covering a set of reference scenarios following three alternative assumptions around the transition to the low-carbon economy:
- Orderly: Early, ambitious action to a net zero CO2 emission economy;
- Disorderly: Action that is late, disruptive, sudden and/or anticipated;
- Hot house world: Limited actions lead to a hot house world with significant global warming and, as a result, strongly increased exposure to physical risk.
Various national authorities have developed and published their guidelines or expectations on climate related to stress testing. In 2019, PRA issued a paper covering the Climate Biennial Exploratory Scenarios (CBES), later followed by various updates to the initial approach. The main goal was to measure the resilience of major banks and insurers in the UK to climate-related risks. The approach considers a scenario horizon of 30 years and is built around three climate scenarios (comparable to the NGFS). The results of this exercise are expected to be released in 2021.
Likewise, the Autorité de Contrôle Prudentiel et de Résolution (ACPR), together with the Banque de France, have been the first national supervisors to engage in a climate stress test exercise on a large sample of institutions using a bottom-up approach, i.e., using assessments made individually by each entity based on common macroeconomic assumptions defined by the supervisor. The pilot exercise conducted from June 2020 to April 2021 illustrated what is the exposure of French banks and insurers towards the climate-related risks, following the NGFS framework and taking into account the macroeconomic projections of the Eurosystem.
The ECB is conducting its economy-wide exercise, the results of which will be released in June 2021. The exercise included around four million companies worldwide, on the basis of climate risk data as well as financial information. It is based on the aggregate trajectories for transition and physical risk embedded into the NGFS scenarios. The study covers almost all monetary financial institutions in the Euro area.
Challenges of climate stress tests
The construction of the climate risk stress tests imposes various methodological challenges in the industry as it deviates from the current stress testing processes. The current regulatory stress testing framework aims to measure the impact on the regulatory capital from a set of macroeconomic scenarios over a pre-defined time period. The current stress testing frameworks are based on the short-term projections (3-5 years) and are built using sufficient historical data, allowing the backtesting of the results.
Climate stress tests usually cover a period of 30 to 50 years. The measurement of any outcomes for such a long period of time is very unstable as it will include a lot of uncertainties in loss projections. In addition to this, there is a lack of historical climate-related data. This means that banks will need either to look for various external data or/and to make a lot of assumptions about the impact of climate change on the institutions’ losses for each sector. As a result, the outputs of such climate stress tests will be very volatile and will be difficult to compare. In fact, a study conducted by BCBS (2020) on climate-related risks states that the most common challenge associated with climate change is the availability of data followed by methodological constraints.
Other challenges associated with climate stress tests are the difficulty in identifying the number and granularity of the scenarios, dynamic balance sheet assumptions for bank projections, the difficulty in translating the climate risks into financial impacts, and so on.
Driven by supervisory expectations and industry expected losses resulting from climate change, the development of a robust climate stress testing framework will be one of the key subjects in the upcoming years.
More information about Avantage Reply Luxembourg here