The Global Crisis in 2008 demonstrated that existent financial regulations were inadequate. The authorities missed the excessive amounts of risk-taking and the increasing fragility of banks, which in turn created the conditions that allowed the crisis to happen. Since then, a number of measures have been put in place under the rubric of macroprudential policy.
In the words of Claudio Borio here on Vox in 2009, “we are all macroprudentialists now”.
The aim of macroprudential policy is to contain systemic risk (Danielsson and Zigrand 2015) while ensuring the efficient functioning of the financial system. It sits aside microprudential policy, focused on the conduct of individual financial institutions.
An important tool in the policymakers' toolkit is stress testing of banks, which has emerged in the past ten years as the main way we assess the strength and weaknesses of the banking sector. Initially, the stress tests were microprudential in nature, focused on individual banks’ resilience to exogenous shocks. For many banks, passing their annual stress tests has been the binding regulatory constraint that has driven capital policy and has helped to shape major supervisory decisions. This experience is reviewed in a Vox eBook (Anderson 2016).
However, the Global Crisis in 2008 showed us that relatively small initial losses in the financial system can be magnified to systemic dimensions and can threaten to bring financial system to a halt; thus, a microprudential perspective that focuses on losses at the individual institution level were proven not enough.
In response, the financial stability authorities are increasingly turning to ‘system-wide stress testing’ as a tool to be better prepared for handling systemic risks. Such stress tests need to capture systemic risk, and hence be focused on the stability of the entire financial system (e.g. Constâncio 2017).
What is required from true macroprudential stress tests (MaPSTs)? They should recognise that systemic risk is driven by the interaction of the variety of financial institutions that make up the system and amplification factors to which such institutions are exposed. Thus, interactions and the severity amplification factors define the way shocks are amplified or dampened. This means that systemic risk is endogenous, as noted by Danielsson et al. (2009); hence, stress tests that are supposed to capture systemic risk need to take into account its endogenous nature.
Designing such MaPSTs, however, is not straightforward.
- First, they need to consider the interaction of all the various entities that make up the financial system, from regulated banks to other types of financial institutions, including in the domestic and global financial system.
- Second, they need to account for amplification factors to which institutions might be exposed to. These usually manifest as financial imbalances in the form of excessive leverage, liquidity and maturity mismatches and exposures to overvalued assets.
To identify the current state of development in MaPSTs, the Monetary and Capital Markets (MCM) Department at the IMF joined forces with the Systemic Risk Centre (SRC) at the London School of Economics to produce a definitive report on MaPSTs (IMF-LSE 2018).
Purpose of MaPSTs
A properly designed MaPST has the potential to be of considerable benefit to both policymakers and practitioners. It provides a quantitative, forward-looking assessment of the resilience of individual banks as well as the financial system as a whole to adverse shocks.
For policymakers, MaPSTs are useful for macroprudential policy, including the determination of bank capital buffers and possibly the calibration of other instruments, including loan-to-value and debt-to-income ratios. However, MaPSTs could also provide useful information for certain aspects of microprudential policymaking, the design of recovery and resolution frameworks, and systemic crisis management.
Similarly, private sector financial institutions can use MaPSTs for risk management, asset allocation, and decision making in times of systemic distress. They also benefit financial institutions – such as banks, pension funds, hedge funds, sovereign wealth funds and insurance companies – concerned about systemic risk.
Evolution of stress tests
Stress testing emerged in the 1990s as a tool to assess financial institutions’ exposure to tail risks. After the 2007 crisis, the financial stability authorities have embraced stress tests as means for ascertaining the strength of individual institutions to exogenous shocks (Dent et al. 2016).
However, the soundness of individual banks is not necessarily critical to overall financial stability, as systemic risk arises from the endogenous interaction of market participants, implying that a stress test designed to capture systemic risk needs to be based on the dynamic interaction of the entire financial system (Alla et al. 2018).
Systemic risk amplification
Systemic risk arises from the interaction of all the market participants that make up the financial system. While banks are the ultimate manifestation of a systemic crisis, they interact with other entities, including pension funds, insurance companies, asset managers, hedge funds, and sovereign wealth funds. These react to stress in different ways, acting as amplifiers and dampeners, depending on the state of their cycles and exposures to diverse amplification factors. Thus, under certain conditions, a relatively small initial shock can be magnified to systemic dimensions due to systemic risk amplification (SRA) mechanisms (e.g. Adrian and Liang 2016).
The modelling of losses from SRA is challenging. Amplification mechanisms are diverse and complex, and can vary in structure and magnitude at different points in time. The relevant data are usually scarce, and models constrained by data availability are often subject to model error.
Furthermore, the financial grid structure matters. Cross (direct) holdings and indirect inter-linkages can work as transmitters across the system, and various categories of entities will naturally amplify or absorb risk. This dynamic implies that vulnerabilities embedded in the macroeconomy and financial system, including leverage, lack of liquidity, and information asymmetries can interact to magnify and accelerate amplification.
The successful development of macroprudential stress testing requires a conceptual taxonomy of mechanisms including macrofinancial feedback effects as well as contagion stemming from direct and indirect interconnectedness across financial entities and markets.
These can change across time and at different points of the financial and economic cycles.
The empirical modelling of MaPSTs is likewise difficult. Hyman Minsky, famously said that “stability is destabilising”, arguing that economic agents react to a low-risk environment by taking on more risk, which eventually culminates in a crisis. A considerable time lag between observed risk and eventual adverse outcomes needs to be incorporated in any modelling.
A major hindrance for stress testing is the availability of the appropriate data. Although large amounts of publicly available and confidential supervisory-level data can be used for stress testing, the most promising sources of data are typically inconsistent, fragmented, geographically restricted, costly, and therefore hard to assimilate. Determining the relative importance of each data set is, consequently, a difficult conceptual problem.
Governance and communication
A key element of the effectiveness of MaPSTs depends on the governance framework and not just models and the mechanics of applying specific tests. The wider macrofinancial environment matters, both in the design and the application of stress tests as well as the wider decision-making process.
Many assumptions are made entailing much uncertainty, and judgements are formed on possible behavioural reactions, systemic interactions, and feedback effects. Based on these, a choice is made of the type of prudential instruments to be deployed, when, on whom, and how.
The authorities need to have policies in place for communication of MaPST outcomes, finding the appropriate balance between early warning and credibility against the risk of panics.
Consequently, a strong governance framework for MaPSTs should be fully integrated into the MaPP institutional framework.
The joint IMF-LSE report surveys the current state of research and policy considerations in macroprudential stress tests, identifying some trends in theoretical and empirical modelling.
MaPSTs are just beginning to be implemented and much work is left to be done. While the fundamental theoretic underpinnings of SRAs are by now well understood, the empirical foundations are still evolving. Data remain a challenge.
MaPSTs have the potential to be of considerable benefit to both policymakers and practitioners. Financial policy, including the determination of bank capital and the structure of the financial system will be informed by MaPSTs. Similarly, private sector financial institutions will get better risk management and assessment of tail risk.
Adrian T and N Liang (2016), “Monetary policy, financial conditions, and financial stability”, VoxEU.org, 14 August.
Alla, Z, R Espinoza, Q H Li and M Segoviano (2018), "Macroprudential Stress Tests: A Reduced-Form Approach to Quantifying Systemic Risk Losses", IMF Working Paper, forthcoming.
Anderson, R W (ed.) (2016), Stress Testing and Macroprudential Regulation: A Transatlantic Assessment, A VoxEU.org eBook.
Anderson, R W, C Baba, J Danielsson, H Kang, U S Das and M Segoviano (2018), "Macroprudential Stress Tests and Policies: Searching for Robust and Implementable Frameworks", Joint IMF and LSE report.
Borio, C (2009), "The macroprudential approach to regulation and supervision", VoxEU.org, 14 April.
Constâncio, V (2017), "Macroprudential stress tests: A new analytical tool", VoxEU.org, 22 February.
Dent, K, B Westwood and M Segoviano (2016), “Stress testing of banks: an introduction”, Bank of England Quarterly Bulletin, Q3.
Danielsson, J, H S Shin and J-P Zigrand (2009), “Modelling financial turmoil through endogenous risk”, VoxEU.org, 11 March.
Danielsson, J, and J-P Zigrand (2015), "A proposed research and policy agenda for systemic risk", VoxEU.org, 7 August.