Book
Share
Chapter

Chapter 15 Macroprudential Bank Solvency Stress Testing in FSAPs for Systemically Important Financial Systems

Author(s):
Li Lian Ong, and Andreas A. Jobst
Published Date:
September 2020
Share
  • ShareShare
Show Summary Details
Author(s)
Andreas A. Jobst, Li Lian Ong and Christian Schmieder 

This chapter is based on IMF Working Paper 13/68 (Jobst, Ong, and Schmieder 2013). The authors would like to thank Laura Kodres and country authorities for their helpful comments and IMF staff members who participate as stress testers in Financial Sector Assessment Programs (FSAPs) for their input.

The global financial crisis put bank stress tests squarely in the spotlight. Stress tests conducted in the lead-up to the crisis, including those by the IMF staff, were not always able to identify underlying risks and vulnerabilities and their magnitude. Since then, the IMF staff has developed more robust stress testing methods and adopted a more coherent and consistent approach. This chapter articulates the solvency stress testing framework that is commonly applied in the IMF’s surveillance of member countries’ banking sectors, and discusses examples of its implementation of Financial Sector Assessment Programs (FSAPs) in 18 countries with systemically important financial systems or that are in the Group of Twenty, in the latter part of the global financial crisis. In doing so, the chapter offers guidance for readers seeking to develop their own stress testing frameworks and country authorities preparing for FSAPs. A detailed stress test matrix comparing the stress test parameters applied in each of these major country FSAPs is provided, together with stress test output templates.

1. Introduction

The global financial crisis has placed the spotlight squarely on the stress testing of financial institutions, notably that of banks. On one hand, the crisis revealed the shortcomings of stress tests as a tool for detecting important vulnerabilities during the lead-up period, which forestalled possible mitigating actions being taken. On the other hand, the experience highlighted the usefulness of credible stress tests in restoring market confidence in the financial system— by shedding light on the potential magnitude of capital shortfall— as demonstrated by the successful Supervisory Capital Assessment Program exercise undertaken by the US authorities in 2009 (Bernanke 2010), which was subsequently transformed into the annual Comprehensive Capital Analysis and Review for large financial institutions. Ultimately, the crisis underscored that stress tests, irrespective of their level of sophistication or regularity of implementation, are not fail-safe, stand-alone diagnostic approaches, but need to be complemented with other tools, such as the Early Warning Exercise, which the IMF 2010d completes regularly together with the Financial Stability Board (FSB).1

Post mortems following the crisis show that the stress tests conducted by supervisory authorities, the IMF staf, and financial institutions themselves were not always able to identify imminent risks and exposures. As such, they frequently failed to provide sufficient early warning of potential vulnerabilities (Borio, Drehmann, and Tsatsaronis 2012). In some cases, the simulated shocks and resulting impacts were not sufficiently severe (often informed by historical stress events), reflecting the general reluctance to recognize the possible realization of extreme scenarios;2 in others, failure was attributable to the specification of the stress tests themselves, including inadequate models to capture complex f-nancial instruments, behavioral elements, or feedback effects. Elsewhere, inadequate data or weaknesses in scenario design, such as the exclusion or cursory treatment of certain types of risks and insufficient focus on spillover risks across different segments of the financial system— within a country as well as across borders— also contributed to the lack of robustness of the stress tests.

At the IMF, stress testing has become a central aspect of the staff’s macroprudential surveillance of financial systems. It is a key component of the Financial Sector Assessment Program (FSAP) and has evolved into an important part of the conjunctural analysis in the Global Financial Stability Report; it is also applied in annual Article IV consultation process and crisis program work. Also IMF member countries view stress testing as an essential aspect of supervision and financial stability analysis. In addition to microprudential (or supervisory) stress testing, some jurisdictions have established national macroprudential authorities, which engage in macroprudential stress testing. Countries are also increasingly requesting technical assistance on stress testing from the IMF as they seek to build or enhance their capacity in this area. These developments strengthen the case for a coherent and consistent approach to stress testing by the IMF staff in its engagement with the membership.

With more attention drawn to stress testing, exercises conducted by the IMF staff have come under considerable scrutiny. Consistency in the implementation of these stress tests is essential to enhanced disclosure and transparency, especially during volatile times. In this context, the completion of stress tests as part of the financial stability assessment in FSAPs has resulted in several important observations.

The variety of approaches, models, scenarios, and assumptions applied in the staff’s analyses has given rise to questions about the interpretation of the results and their consistent comparison across countries. The issue has been further complicated by the lack of generally accepted “best practice” principles (at least for some dimensions of stress tests) (see for example, Board of Governors of the Federal Reserve System, Federal Deposit Insurance Corporation, and Office of the Comptroller of the Currency 2012; BCBS 2018a) and evolving practices. The IMF staff has developed prescriptive guidelines in IMF-related stress testing exercises to ensure sufficient coverage and a modicum of uniformity, both within a financial system and, at the very least, across “peer” countries.

The communication of stress test results has also become an increasingly sensitive issue for the IMF’s membership. Both financial supervisors and financial institutions are struggling to balance the call for increased transparency with the need to avoid unduly alarming the markets and creating self-fulfilling prophesies, especially during periods of stress.

In the decade-and-a-half since stress testing premiered in the IMF’s surveillance toolkit, the scope of stress tests has expanded to include both banking and nonbank institutions, with a strong focus on the former. However, the number of FSAP stress tests of the insurance and pension funds sectors, while having increased, still trails that of the banking sector significantly. The IMF staff has developed more robust stress testing methods and models, especially since the crisis. Based on the IMF’s rich and diverse practical experience with stress tests through more than a decade-and-a-half of FSAPs, the staff has proposed a set of “best practice” principles for macro-financial stress testing (Chapter 2 and IMF 2012a). The principles cover areas such as the scope of stress tests, transmission channels, risk types, valuation methodologies, risk sensitivity, and communication strategies.

At the IMF, bank stress testing is most advanced, given the systemic importance of the banking sector in practically all member countries. The focus has been on solvency risk, and work to continually develop a comprehensive and robust framework is ongoing. Separately, the development of liquidity stress tests by the IMF staf, which is covered in Chapter 16, has also intensified in response to lessons learned from the crisis. This chapter complements Chapter 2 by providing an operational perspective of those “best practice” principles for bank solvency stress testing applied by the IMF staff. Specifically, the chapter:

  • Articulates the stress testing framework and demonstrates how best practice principles have been applied to key elements in the IMF’s surveillance of banking sectors in selected FSAPs. The sample group consists of 16 of the 25 countries with financial systems that had been identified at the time as being systemically important and were subject to mandatory assessments every five years (IMF 2010a, 2010b),3 plus two of the five Group of Twenty countries that are not among the systemic group (all hereafter “major countries”) (Table 15.1). These 18 jurisdictions participated in FSAPs during the latter part of the global financial crisis, between the 2010–13 IMF fiscal years, when the IMF staff was implementing significant changes to the stress testing framework. In focusing on these countries during this period, the comparisons capture how IMF stress tests evolved with the aim of improving their robustness, transparency, and consistency.
  • Presents the framework in a detailed cross-country stress testing matrix to compare the actual implementation across the major country FSAPs (Appendix 15.1). An abridged version of this stress test matrix for each country, which carries the technical reference “STeM,” is typically presented in the main FSAP report, the Financial System Stability Assessment (FSSA), to enhance the transparency of each exercise.
  • Aims to provide useful guidance for readers seeking to develop their own stress testing frameworks and for country authorities preparing for FSAPs. The chapter is illustrative in this regard in that it discusses the detailed setup of FSAP stress testing exercises.
Table 15.1S- 25 and Other G20 Countries: Status of FSAPs since Fiscal Year 2010
RankJurisdictionGroupingCompleted FSAPs since FY 2010RankJurisdictionGroupingCompleted FSAPs since FY 2010
1United KingdomS-25, G20, G7FY 201116Hong Kong SARS-25—**
2GermanyS-25, G20, G7FY 201117BrazilS-25, G20FY 2012
3United StatesS-25, G20, G7FY 201018Russian FederationS-25, G20FY 2011
4FranceS-25, G20, G7FY 201219KoreaS-25, G20—**
5JapanS-25, G20, G7FY 201220AustriaS-25FY 2013
6ItalyS-25, G20, G7FY 201321LuxembourgS-25FY 2011
7NetherlandsS-25FY 201122SwedenS-25FY 2011
8SpainS-25FY 201223SingaporeS-25—**
9CanadaS-25, G20, G7—**24TurkeyS-25, G20FY 2011
10SwitzerlandS-25—**25MexicoS-25, G20FY 2012
11ChinaS-25, G20FY 2010ArgentinaG20FY 2013
12BelgiumS-25FY 2013European UnionG20FY 2013*
13AustraliaS-25, G20FY 2013IndonesiaG20FY 2010
14IndiaS-25, G20FY 2012Saudi ArabiaG20FY 2011
15IrelandS-25South AfricaG20
Sources: IMF 2010a, 2010b, and 2013; and Monetary and Capital Markets Department, IMF.Note: Systemic- 25 (S-25) countries were ranked according to the size and interconnectedness of their financial systems. The IMF’s fiscal year (FY) runs from May 1 of the previous year to April 30 of the current year. G20 = Group of Twenty.*Stress tests were not conducted as part of the FSAP for the European Union.**FSAPs scheduled for completion in FY 2014.
Sources: IMF 2010a, 2010b, and 2013; and Monetary and Capital Markets Department, IMF.Note: Systemic- 25 (S-25) countries were ranked according to the size and interconnectedness of their financial systems. The IMF’s fiscal year (FY) runs from May 1 of the previous year to April 30 of the current year. G20 = Group of Twenty.*Stress tests were not conducted as part of the FSAP for the European Union.**FSAPs scheduled for completion in FY 2014.

Eight of the 18 countries in the sample published all the details of their respective FSAP stress tests. They comprise Australia, France, Germany, Japan, Spain, Sweden, the United Kingdom, and the United States. Of the remaining 10 countries, all but one consented to the inclusion of all information on their respective FSAP stress tests, some of which was not contained in previously published reports.

There has been some success in standardizing FSAP stress tests across countries, and improvements continue to be made in this area. However, in some instances, expert judgment might render “one size fts all” approaches less relevant. Moreover, it is important to recognize that surveillance stress tests are not fail-safe, stand-alone diagnostic tools, although the value of well-designed exercises should not be underestimated. Their usefulness critically depends on the availability and quality of data.

This chapter is organized as follows. Section 2 puts into context the nature of the stress testing work conducted by the IMF staff. It is followed in Section 3 by detailed coverage of the various components and elements of the stress testing framework and their application in FSAPs. Section 4 concludes.

2. IMF Stress Testing in Context

Stress testing is a forward-looking technique that attempts to measure the sensitivity of a portfolio, an institution, or even an entire financial system to events that have a very small probability of occurrence but a significant impact if they manifest. Scenario and/or sensitivity analysis are applied in “what if” exercises, which characterize what might happen if certain “ extreme-but- plausible” risks were to crystallize. In the decade and a half since the concept was first introduced, stress testing has been used by central banks, financial supervisors, and international organizations, such as the IMF, to identify vulnerabilities and incipient risks in the financial sector. Stress tests are commonly used for macroprudential, microprudential, and/or risk-management purposes (Figure 15.1).

Solvency Stress Testing Applications

Source: Authors.

Note: Top-down stress tests are completed based on either individual bank data (which are then aggregated) or aggregated portfolios; bottom-up stress tests are conducted by individual institutions using their own internal risk models and data. CCAR = Comprehensive Capital Analysis and Review; CEBS = Committee of European Banking Supervisors; EBA = European Banking Authority; FSAPs = Financial Sector Assessment Programs; SCAP = Supervisory Capital Assessment Program.

Stress testing conducted by the IMF staf, as part of the institution’s surveillance mandate, is completed mostly in FSAPs and for macroprudential purposes (IMF and World Bank 2003; Moretti, Stolz, and Swinburne 2008). It is aimed at assessing system-wide resilience to shocks over the medium term, uncovering vulnerabilities to any rapid deterioration in the macroeconomic environment, and, more generally, identifying potential threats to overall financial stability. In this context, stress tests for both solvency and liquidity risks tend to incorporate very severe scenarios to assess the capacity of the financial system to withstand tail risks. The findings of these exercises typically do not require management action by financial institutions; rather, they are used to inform policy discussions with country authorities about the frameworks in place to deal with systemic shocks.

The cooperation with country authorities has a crucial impact on the robustness and credibility of IMF stress tests (see Section 3). According to Article VIII of the Articles of Agreement of the IMF, member countries are under no obligation to disclose information about individuals or corporations. This means that the IMF cannot compel country authorities to provide the necessary confidential bank-by-bank data for the stress tests. In some cases, authorities have refused to share any prudential information, and the IMF staff has had to rely solely on publicly available data, which reduces the specificity of the results; in others, authorities have only consented to running the tests themselves, based on some agreed-upon parameters, and sharing the aggregated results. The recourse for the IMF staff is to ensure that the transparency of the process— or any limitations thereof— is clearly documented in the official documents.

The IMF’s objectives may be contrasted with the stress tests undertaken by supervisory authorities, usually for microprudential purposes (Fell 2006). Such exercises are normally embedded in the oversight process wherein supervisors would run stress tests involving individual institutions on a periodic basis to assess their financial soundness under adverse economic conditions, such as in the case of the annual US Comprehensive Capital Analysis and Review (Board of Governors of the Federal Reserve System 2012a, 2012b), which is completed together with a liquidity stress test (Comprehensive Liquidity Analysis and Review). Supervisory stress tests can be independent of an institution’s systemic relevance, where “failure” would typically require some form of management action, including recapitalization.

The global financial crisis brought about a new concept of stress testing, that is, one with a crisis management objective, which the IMF staff refers to as “crisis stress testing” (see Chapter 14). Largely macroprudential, as the aim is to restore and sustain market confidence in the financial system, it can also be considered microprudential in that it examines the soundness of individual financial institutions, and “failure” would typically require recapitalization— or even resolution. Compared to the medium-term risk horizon of surveillance stress tests, crisis stress tests tend to have a near-term focus. In the United States, system-wide (solvency) stress testing of banks was used by the authorities in 2009 for crisis management purposes, through the Supervisory Capital Assessment Program exercise (Board of Governors of the Federal Reserve System 2009), the predecessor of the Comprehensive Capital Analysis and Review; the EU authorities also made a similar effort through the region-wide stress testing exercise conducted by the Committee of European Banking Supervisors in 2009 and 2010 and then by its successor, the European Banking Authority in 2011 (Committee of European Banking Supervisors 2010; EBA 2010, 2011a, 2011b), as did Ireland (Central Bank of Ireland 2011) and Spain (Banco de España 2012). IMF teams working on crisis countries may sometimes run stress tests to determine the condition of the banking sector as an input to the design of a potential program.

Separately, financial institutions regularly carry out stress tests for risk-management purposes. In these internal exercises, financial institutions develop and implement their own stress testing programs, which assess their ability to meet capital and liquidity requirements under stressed conditions. The IMF staff sometimes relies on banks’ stress testing infrastructure for the FSAP bottom-up stress tests (see Section 3). In some countries, supervisors issued guidance on stress testing to the financial institutions under their supervision before FSAPs took place (for example, the United Kingdom and the United States) (Board of Governors of the Federal Reserve System 2012c; Financial Services Authority 2009). However, this practice is not yet widely implemented, including in some of the world’s largest financial systems. The Basel Committee for Banking Supervision (BCBS) has also issued guidelines for stress testing by individual banks (BCBS 2009, 2017), followed up by a peer review of supervisory authorities’ implementation of those principles (BCBS 2012a).

3. A Framework for Bank Solvency Stress Testing

The objective of the bank solvency stress tests conducted by the IMF staff is to assess the soundness of banking sectors under adverse macroeconomic conditions. Tests are designed to determine banks’ resilience to the adverse impact of severe macro-financial stress over the short and medium terms (relative to a predefined baseline scenario). The aim is to identify the sector’s vulnerabilities and its capacity to absorb shocks.

Within the framework, the development of plausible and coherent tests requires a thorough understanding of the financial system in question and its institutions, which includes structural characteristics, such as differences in banks’ business models, their role in the domestic financial sector, and, increasingly, cross-border linkages. While the identification of transmission channels of stress impacting financial intermediation in smaller countries tends to be straightforward, more complex banks in larger economies and financial centers may create conceptual challenges for stress testing.

Up until the latter part of the global financial crisis, in 2013, the IMF had conducted assessments of about 140 advanced and developing countries. Of these, solvency stress tests had been conducted in practically all instances. Thus, the solvency stress tests in FSAPs must necessarily be adaptable to diverse financial systems, with consistent applications of assumptions and models but sufficient flexibility to accommodate vastly different circumstances (for example, normal or crisis times), systems (for example, sophisticated or basic), and regulatory regimes (for example, Basel I or Basel II/III) as well as be sensitive to when and how the outcomes are presented and communicated (Table 15.2). Further, the FSAP stress tests necessarily require trade-offs among the scope, scenario design, and methodologies applied in the context of staff and authorities’ resources and time constraints.

Table 15.2A Framework for Macroprudential Bank Solvency Stress Testing
Framework/ComponentsKey ElementsIllustrative Example
1.Scope
ApproachBottom-up (BU) Top-down (TD)By individual banks By authorities; by IMF
CoverageInstitutions

Market share
Number of banks

Percentage of banking sector assets
DataSource

Cutoff date
Banks’ own, supervisory, and public data

End of last fiscal year
Reporting basisUnconsolidated banking groups, domestic businesses only
2.Scenario design
Risk horizonMultiperiod

Instantaneous
1–5 years
ScenariosBaseline

Growth shocks
IMF World Economic Outlook projections

Double-dip recession and protracted slow growth
RisksKey risk(s)

Other risks covered in scenario analysis

Other tests/risks
Credit risk, market risk

Sovereign risk, funding risk, exchange rate risk

Sensitivity analysis of credit and market risks; network analysis of spillover risk
Factors that management controlsBalance sheet growth

Credit growth

Dividend payout rule

Other business strategy considerations
Consistent with nominal GDP growth

Based on satellite model

Historical payout ratio

No asset disposal allowed
Other assumptionsTaxesUniform (local corporate income) tax rate
3.Regulatory capital standards
Capital definitionDomestic InternationalLocal regulatory requirements Basel III transition
Capital adequacyMetrics

Hurdle rate(s)

Changes in risk-weighted assets
Amount of recapitalization required (in domestic currency); total capital, Tier 1 and core/common equity Tier 1

In line with Basel III transition schedule

Risk-weighted assets calculated using Basel II formula
4.Methodology
Stress test modelAccounting-based

Market price-based
Balance sheet approach (for example, Schmieder and others 2011)

Systemic Contingent Claims Analysis (Jobst and Gray 2013)
Modeling of macro-financial linkagesSatellite modelsEconometric models for credit losses, income, credit growth
5.Communication
Presentation of outputTemplate(s)Standardized output template for individual BU results provided to banks and authorities
PublicationMediumResults published in FSSA; Technical Note published
Source: Authors.Note: FSSA = Financial System Stability Assessment.
Source: Authors.Note: FSSA = Financial System Stability Assessment.

Scope

The scope of a stress testing exercise needs to be sufficiently comprehensive to capture the main characteristics of a particular financial system. Key considerations are: (1) the stress testing approach(es); (2) the coverage in terms of the institutions, their market shares, and the sources of their earnings and exposures; and (3) the granularity and timeliness of relevant data (and their reliability). In this regard, stress tests conducted by the IMF staff for financial surveillance purposes are typically undertaken in close collaboration with supervisory authorities. In many instances, the staff is given access to the necessary detailed supervisory data during FSAPs (on agreement of strict confidentiality); data quality is further enhanced when individual financial institutions participate in the exercise.

Approach

Surveillance stress testing of banks’ solvency risk in the context of FSAPs is usually based on a “ top-down” (TD) approach, which is sometimes combined with a “ bottom-up” (BU) exercise:

  • TD tests are carried out by the IMF staf, by the authorities, or by both, typically in close collaboration. In these exercises, tests are either conducted using the data of individual banks (which are then aggregated), or on an aggregated group of banks to analyze the impact of predefined, system-wide shocks. A common macro-financial environment is assumed, and a standardized set of behavioral assumptions is applied to all institutions. TD stress tests may be used as a stand-alone analysis or to complement the BU exercise, if one is conducted.
  • The BU approach is used by FSAP teams if authorities are supportive of having individual institutions conduct their own stress tests (and banks are prepared and have sufficient capacity to do so). Individual institutions use their own data and internal risk models. As with the TD approach, common macro-economic shocks and selected standardized assumptions are prescribed by the IMF staff to isolate the impact of shocks on banks’ financial soundness to identify specific vulnerabilities.

The IMF staff advocates conducting both BU and TD stress tests, as much as possible, to enrich the surveillance analysis in FSAPs. Each approach has its strengths and weaknesses and is considered complementary for cross-validation purposes, rather than being considered a substitute for the other. The process of reconciling the BU and TD results is usually an important learning process in itself, with any divergence in the results from the two approaches usually traced to differences in either the model design, the scope of the stress testing exercise (including the type of underlying data used), behavioral assumptions, and/or modeling of sensitivities. For instance, bank-specific assumptions and the application of internal models based on more granular data can lead to differences in the projection of profits and losses— and consequently the impact on the capital ratios— for individual banks under the various scenarios.

The decision as to whether BU stress tests are conducted to complement TD tests, or whether TD stress tests are performed by country authorities or by the IMF staff, or jointly, is mostly made on an ad hoc, country-by- country basis, depending on data and resource availability as well as the receptiveness and degree of involvement by authorities. Around half of the FSAPs for major countries between FY2010/13 comprised both BU and TD tests (for example, Australia, China, France, India, Indonesia, Japan, Mexico, Russia, Turkey, and the United Kingdom). TD tests were either conducted by the IMF team only (for example, Australia, India, Indonesia, Netherlands, Saudi Arabia, Sweden, and Turkey) or by the authorities only (for example, Japan, Luxembourg, and Russia), or in some cases, separately by both, using different methods (for example, China, France, Mexico, and the United Kingdom).

The solvency stress testing of the banking sector in the 2011 FSAP Update for the United Kingdom exemplifies the effective collaboration among country authorities, the IMF, and individual financial institutions in all aspects of the exercise (IMF 2011a), which involved both BU and TD solvency stress tests (together with TD liquidity risk stress tests). BU stress tests were run by the seven major UK banks, in close coordination with the FSAP team and the then Financial Services Authority (Figure 15.2). At the same time, TD tests were separately performed by the Bank of England using its Risk Assessment Model for Systemic Institutions and by the FSAP team using the Systemic Contingent Claims Analysis model, applying macroeconomic forecasts and projections from the IMF and FSAP, respectively, and satellite model outputs from the Bank of England.

Example of IMF Stress Testing Exercise: UK FSAP Update

Source: IMF 2011d.

Note: CCA = contingent claims analysis; FSA = Financial Services Authority; FSAP = Financial Sector Assessment Program; LCR = liquidity coverage ratio; NSFR = net stable funding ratio; RAMSI = Risk Assessment Model of Systemic Institutions.

Coverage

The coverage is crucial for the usefulness, and, thus, credibility of the stress test exercise. Ideally, surveillance stress testing for macroprudential purposes should include all institutions, if data availability and resources permit. Realistically, all systemically important institutions, as well as second-tier banks that are potentially systemic depending on circumstances, should be covered. Smaller banks that may be considered at risk could also be included.

FSAPs typically focus on stress testing the major commercial banks in their respective jurisdictions, with coverage usually determined in collaboration with the authorities. Where resource constraints dictate that only a small sample of banks can be considered, especially in the case of BU stress tests, the usual practice is to focus on the systemically important institutions. The market share of the banks included in the 18 major country stress testing exercises between FY2010/13 was at least 60 percent of the total assets of the sector; coverage was 100 percent in six of them (Brazil, India, Indonesia, Japan, Luxembourg, and Russia).

The identification of systemically important domestic banks is still not clear-cut. While some banks are of obvious systemic importance in their own respective countries and their selection for stress tests is indisputable, the difficulty has been in identifying those that are systemic at the margins, for example, some of the smaller institutions that may have the potential to become systemic depending on the environment at a particular point in time (IMF, Bank for International Settlements, and Financial Stability Board 2009). Thus, the definition of what constitutes a systemic bank remains somehow ad hoc in IMF-related stress testing exercises, and a more structured approach is desirable. The BCBS methodology for identifying global systemically important banks, which has been reviewed recently, has facilitated this process (BCBS 2011, 2018b; Financial Stability Board 2011). The guidelines on the implementation of supervisory measures for domestic systemically important banks and the policy recommendations by the Financial Stability Board (Financial Stability Board 2012) for their identification represent another positive step in this direction (BCBS 2012d), whereby many jurisdictions have identified these banks in their jurisdictions.4

Data

The credibility of stress test results depends on the availability and sufficiency of timely and reliable data. The quantity and quality of data not only determine the scope and risk coverage of the stress tests but also the type of models that can be applied (Howard 2008). As much as possible, FSAP stress tests utilize the latest audited and/or supervisory data alongside the latest macroeconomic projections, all of which determine the appropriate cutoff date. Supervisors typically make available to IMF teams the relevant data from prudential reporting and reviews, which are usually supplemented by publicly available information. Supervisory data were provided in almost all 18 major country FSAPs covered in this chapter; only in one instance was the staff wholly dependent on public information for the stress testing exercise. If there is no access to supervisory data, the use of publicly available data on individual banks may be less granular and timely.

There has been little standardization across FSAPs regarding the reporting level of bank data. While about half of the FSAPs in the sample used consolidated banking group data for the stress tests (for example, Australia, Brazil, China, France, Japan, Netherlands, Sweden, the United Kingdom, and the United States), the rest utilized unconsolidated, legal entity data (for example, Germany, India, Luxembourg, Mexico, Russia, Spain, and Turkey).

FSAPs typically focus on the domestic banking sector, which suggests that the data of banks’ local businesses should be utilized on a local-consolidated basis. Such data would avoid double counting local business operations. The use of consolidated level data would prevent the examination of ring-fencing of subsidiary profits, capital, and liquidity by host countries, which may be important for large international groups (Cerutti and Schmieder 2012). That said, the decision as to which type of data to use may sometimes be moot, as it could be constrained by the type of data that are collected for supervisory purposes.

There is an increasing use of forward-looking market data and other variables to complement accounting information, especially for data-rich advanced economies. Market data reflect important investor perceptions of how risks affect the actual valuation of reported book values. They can also be used as a benchmark for the calibration of banks’ own credit risk models under internal-ratings- based (IRB) approaches—and for cross-validating the quantification of other, difficult-to- model risks, namely, market and operational risks.

The use and interpretation of the data require caution. FSAPs do not conduct audits of banks’ accounts and cannot corroborate the quality of the reported data and valuations used in stress tests. Thus, quantitative approaches benefit greatly from discretionary expert judgment. In instances where the staff may be concerned about the effects of issues such as loan misclassifications and/or lender forbearance on the accuracy of the data, caveats are often explicitly noted (for example, in the Spain and the United Kingdom exercises).

Scenario Design

Risk Horizon

For surveillance purposes, the choice of a risk horizon should be consistent with the desired implications for the related policy discussions. Covering a longer time period offers several benefits, in particular: (1) major macro-financial shocks typically have a lasting impact over several years, especially in the case of credit risk; and (2) evolving regulatory reforms are likely to be protracted and take several years to implement (for example, the implementation of the post-crisis regulatory reforms in Basel III). A longer risk horizon entails greater uncertainty, but surveillance stress testing is not a forecasting exercise; rather, the exercise should adequately capture any medium-term impact. In contrast, sensitivity tests are usually applied to assess instantaneous shocks.

It is important to balance the consistency of the risk horizon across countries with the usefulness of the findings for individual country circumstances. As in other aspects of stress testing, some expert judgment is involved— major country FSAPs typically apply a five-year risk horizon, but exceptions may be made in cases where the staff is of the view that the application of a longer sample period may be unconstructive. As an example, the FSAP stress test for Spain during the global financial crisis applied a two-year risk horizon to accommodate the rapidly changing banking sector due to ongoing restructuring efforts (IMF 2012b). In most emerging market economies with less mature banking sectors (for example, China, Indonesia, Mexico, and Turkey), risk horizons of between one and three years were used.

Stress Scenarios

Stress tests are based on scenario shocks and/or sensitivity analysis. In scenario tests, a baseline scenario is first established; postshock assessments are made relative to the baseline scenario. In FSAPs, the IMF’s World Economic Outlook projections are typically used as the baseline for stress tests. Stress scenarios are then defined based on either (1) historical simulation (by defining the scope and severity of the scenario based on previous stress episodes and crisis periods), or (2) hypothetical scenarios (which have not yet happened but are particularly relevant given specific vulnerabilities in banks’ portfolios). In both cases, ad hoc expert judgment often influences the scenario design. The stress scenarios are then applied consistently across banks within the same sector.

The availability of data and the modeling capabilities govern how the appropriate stress scenarios are constructed for FSAP solvency tests. Scenarios either (1) reflect a hypothetical state of risk parameters under stress affecting solvency conditions (“direct approach”), which is often used in the case of ad hoc scenarios or historical simulation, or (2) are based on adverse macroeconomic scenarios, which need to be translated into financial stress parameters (“indirect approach”). The latter approach consists of the following elements:

  • Estimating relevant macroeconomic and financial variables conditional upon the chosen scenario. Common methods for predicting economic and financial variables conditional upon certain macroeconomic conditions include: (1) structural econometric models, (2) vector autoregressive methods, and (3) pure statistical approaches (Foglia 2009). As a general rule, these macro-financial linkages would need to be clearly documented and back-tested.
  • Converting these variables into financial risk parameters via various types of “satellite” (or auxiliary) models. This step links different macro-financial shocks, reflected in macroeconomic variables, to the main determinants of bank solvency, that is, preimpairment profit, impairments, and risk-weighted assets (RWAs), since macroeconomic models do not usually include financial balance sheet variables (and credit aggregates in particular). Common explanatory variables include:
    • Macroeconomic variables, such as economic growth, unemployment, short-and long-term interest rates, inflation, and exchange rates;
    • Sectoral (asset price) indicators, such as residential and commercial real estate prices, commodity prices, and equity market conditions (Figure 15.3); and
    • Microlevel data, such as bank-specific credit growth (for example, deleveraging under severe stress conditions), which could also be modeled as a macro-economic variable, operational/financial leverage, and funding gaps.

Example of Macroscenarios for Stress Testing: UK FSAP Update

Source: IMF 2011d.

Note: BoE fan charts are based on BoE, rather than WEO projections. BoE = Bank of England: CPI = consumer price index; DCLG = Department of Communities and Local Government; FSA = Financiall Services Authority; FTSE = Financial Times Stock Exchange; FSAP = Financial Sector Assessment Program; IPD = Investment Property Databank; WEO = World Economic Outlook.

FSAPs attempt to introduce consistently severe macro-economic shocks in the specification of scenarios in solvency stress tests. The aim is to facilitate the identification of other factors that drive differences across institutions and to facilitate comparisons between peer countries. Shocks to economic growth are defined in terms of historical volatility, usually one (mildly adverse) and/or two (severely adverse) standard deviations from the long-term average. Among the sample countries covered in this chapter, the four-standard-deviation shock imposed on the Australian banking sector was estimated over a 50-year period, whereas the two-standard-deviation shocks applied to several EU countries were calculated over 30-year periods. In about half the exercises, a prolonged slow growth scenario was also included as a separate stress (for example, Australia, Brazil, China, Germany, Japan, Sweden, Turkey, the United Kingdom, and the United States).

While the standardization of the scale of shocks has become a general rule of thumb for FSAPs (Hardy and Schmieder 2013), a certain flexibility in the scenario design remains key. The prevailing macroeconomic environment and market conditions as well as the main risks to financial stability determine the configuration of the most appropriate and credible tail-shock scenario(s). For example, the issue of overheating was a key risk for Turkey at the time of its 2011 FSAP and was therefore incorporated into the design of the stress scenario. For Spain, the one-standard- deviation shock applied during the global financial crisis included a revised baseline, which had already incorporated the rapidly deteriorating economic outlook and a fiscal adjustment (and, thus, implied a severity that was higher than that of a customary two-standard-deviation shock calibrated over a long-term average growth rate).

Nonetheless, spillover effects have remained largely un-addressed in the scenario design. For many years, the IMF’s stress tests did not consistently and comprehensively quantify the possible impact of scenarios on the macroeconomic conditions of other countries where stressed banks form part of the host banking sector. In such cases, the IMF staff often relied on the banks themselves to estimate the corresponding scenarios in relevant countries in BU exercises, potentially giving rise to inconsistent projections, and creating biased results of banks’ financial performance, possibly for the same countries. However, in the meantime, this aspect of IMF stress testing has been improved with the development of the Global Macro-Financial Model (Chapter 4).

FSAP stress scenarios emphasize the importance of tail risks. The tests are aimed at identifying the vulnerabilities of a country’s financial system and the ability of its supervisory and crisis-management frameworks to deal with the realization of extreme but plausible risks. In the 2011 UK FSAP, for instance, capital losses were estimated for a 0.1 percent probability event (IMF 2011d, 2011c)—the UK Financial Services Authority 2011 had ascribed a 2 percent probability to a two-standard- deviation shock to growth materializing, and the IMF’s model subsequently calculated capital losses at the 95th percentile of this scenario (that is, falling into the 5 percent tail of the loss distribution), which would have a probability of 0.05 × 0.02 = 0.001 percent. That said, it is sometimes difficult to convince national authorities of the importance of running extreme tail scenarios. A useful way forward may be to also run reverse stress tests, that is, stress tests that aim to determine scenarios that would cause a bank to become insolvent.

Separately, sensitivity tests provide useful information on the immediate impact of individual shocks. They are usually applied if there is little or no data (and/or data quality is in-sufficient), or to complement the scenario analyses conducted on more complex financial systems. Several risk factors could also be combined to determine the impact of concurrent multiple shocks to a system. Sensitivity analyses were conducted in most major country FSAP stress testing exercises on various market risk factors.

Risk Factors

The selection of main risk drivers, and the way they are integrated (or not), has significant bearing on the interpretation and potential implications of stress test results. The coverage of risk factors in FSAP stress tests has evolved and expanded over time, with significant enhancements in the technical analysis and specification of risks in the wake of the global financial crisis. FSAPs attempt to cover all relevant macro-financial risks affecting the performance and valuation of banks and the financial system at large. Prior to the global financial crisis, these tests focused largely on credit and market risks (for example, interest rates, exchange rates, and credit spreads as well as equity and commodity prices). While these risks remain the mainstay of FSAP solvency stress tests, additional facets of risk affecting different types of exposures have been included:

  • Exposures to sovereign and other previously low-default assets: Prior to the global financial crisis, exposures to sovereign debt did not figure prominently in stress tests, if at all. They were considered “risk free” and were typically assigned the lowest (often zero percent) risk weights for the calculation of regulatory capital requirements under the Basel framework. However, FSAPs have acknowledged rising sovereign risks by estimating the potential valuation losses of such exposures (and the commensurate impact on unexpected losses reflected in higher risk weights). The valuation loss can be derived from the negative impact of higher sovereign credit risk on the price of sovereign exposures, which would result in a corresponding valuation haircut (Chapter 10). The same approach can be applied to other previously low-default portfolios, such as holdings of bank debt (which are indirectly affected by higher sovereign risk). Shocks to sovereign exposures were incorporated into the FSAP stress tests for all larger EU Member States and Japan; the same treatment was also applied to portfolios of bank debt in cases where domestic banks were highly exposed to their local government.
  • Banking and trading books: For securities, stress tests had previously considered shocks to trading books only, largely because longer risk horizons were not covered. However, during the global financial crisis, many institutions moved their securities to their banking books, supported, in some cases, by regulatory forbearance. This change in the accounting treatment underscored the need for stress tests to cover also (largely unrealized) valuation losses of securities in the two components of the banking book: (1) the available-for- sale portfolio, where losses are absorbed by reserves as part of shareholders’ equity (unlike mark-to- market losses in the trading book, which are reflected in net income), and (2) the hold-to-maturity portfolio, where lower market values (and higher downgrade risk) require higher provisions (which weigh on net income). However, not all country authorities are receptive to a comprehensive application of shocks to banks’ all securities holdings. In the FSAP stress tests for France, Japan, the Netherlands, Sweden, and the United Kingdom during the global financial crisis, valuation haircuts were applied to both portfolios (excluding exposures to “AAA”-rated and/or own sovereigns in the hold-to-maturity portfolio), but only to the available-for-sale portfolio in the case of Russia and Spain.
  • Funding costs: The global financial crisis underscored the importance of assessing the impact of rising funding costs on bank solvency (as part of the simulation of income under stress more generally). Funding costs change disproportionately to changes in solvency conditions, rising sharply as a bank’s capital adequacy worsens (especially for banks with sizeable portions of wholesale funding, which is more risk sensitive than a large deposit base). Stress test calculations link net funding costs (simulating the impact on both assets and liabilities) to income, possibly linking both solvency and liquidity exercises. The explicit treatment of funding costs in FSAP solvency stress tests was incipient during the global financial crisis (for example, France, Germany, Sweden, and the United Kingdom) but has since become an important aspect of the exercise.
  • Off-balance-sheet items: The realization of contingent liabilities from explicit and implicit guarantees of investment vehicles (and the emergence of contingent assets related to related party lending during the global financial crisis) resulted in the sudden realization of large losses (and considerable cash outflows). Thus, incorporating off-balance-sheet positions that could give rise to such contingent liabilities (such as guarantees, commitments, and derivatives) is important to adequately capture the impact of extreme stress on all relevant exposures. However, such data are not as readily available, especially from public sources.
  • Cross-border exposures: Prior to the crisis, credit risk tests focused largely on banks’ exposures to domestic firms and households without consideration of foreign exposures (through branches and subsidiaries). Since then, FSAPs have incorporated spillover risks in the form of network (for example, Australia, France, Japan, and Spain) and ring-fencing (for example, Spain) analyses as separate modules in the TD approach. In some BU assessments, international banks have also considered shocks to the countries in which they are active (for example, the United Kingdom).

The estimation of impairment losses based on credit risk parameters— probability of default (PD) and loss given default (LGD)—took on significant importance during the crisis. Differences in banks’ respective business models and/ or specific risks tend to explain differences in the characteristics of nonaccrual events and impairment losses. The calibration approach of credit risk parameters— through-the-cycle or point-in-time PD (and LGD)—has a significant impact on the estimated net income of banks. Point-in- time risk parameters provide a more realistic assessment of expected credit risk, especially during stressed periods. Another key challenge in FSAPs is to ensure the availability of these parameters for all (or most) sample banks, and if necessary, develop proxy metrics, such as loan loss provisions (Schmieder, Puhr, and Hasan 2011).

Factors that Management Controls

Surveillance stress tests often include common assumptions about strategic decisions and behavioral adjustments of banks during times of stress. These assumptions (on factors that management controls) ensure that stress test findings can be analyzed in a consistent and comparable manner. In FSAPs, common assumptions are especially pertinent for BU stress tests that rely on banks’ internal models— at the expense of methodological flexibility and realism. Assumptions adopted in FSAPs are also typically (and appropriately) on the conservative side. The main behavioral variables include:

  • Balance sheet growth: This assumption determines the trend growth in core items on the assets and liabilities sides of banks’ balance sheets. FSAP stress tests typically assume that the balance sheet is either (1) constant (that is, growing with nominal GDP or some predefined rule); or (2) static (that is, not growing at all, possibly in combination with a constant credit portfolio). Indeed, the major country FSAPs in the sample were split almost evenly on the adoption of either assumption.
  • Credit growth: Assumptions about credit growth are usually based on either models (for example, Brazil, Spain, and Sweden) or descriptive empirical evidence (for example, Turkey), and, in many cases, also involve expert judgment. Banks under stress are likely to reduce lending in line with a slowdown or reversal in balance sheet growth, usually consistent with changes in nominal GDP.
  • Dividend payout: Dividends are generally assumed to be paid only by banks that satisfy all three measures of capital adequacy, as relevant (that is, total capital, Tier 1, and core/common equity Tier 1) after making adequate provisions for asset impairments and transfers of profits to statutory reserves, which banks must keep on hand to meet their obligations to depositors. In most of the major country FSAPs covered in this chapter, it was assumed that banks would suspend dividends under stress. For the others, assumptions included payouts based on Basel III capital conservation standards (for example, Sweden) or on historical ratios (for example, Brazil, France, and Japan).
  • Strategic changes and asset disposal: FSAP stress tests typically do not consider changes to business operations that require managerial involvement, such as plans to increase operational efficiencies. Moreover, nonrealized and/or strategic disposals (for example, loan books in runoff or sales of noncore businesses)5 and acquisitions are generally excluded. Firms are also expected to replace maturing exposures unless there is a sound basis for assuming that this will not happen (for example, deleveraging plans for banks in IMF program countries).

There is not necessarily a specific “best practice” associated with each assumption on the factors that bank management controls. However, general conservatism (often aligned with historical experience) should be an important consideration. FSAPs have sought to ensure some uniformity in their application where possible, and to match their specific relevance to the country in question. Detailed guidance on how these assumptions should be implemented is usually provided (Appendix 15.2).

Capital Standards

The assessment of bank solvency in stress tests requires a consistent definition and appropriate measurement of capital standards, which comprise two elements: (1) the definition of capital, and (2) the calculation of capital adequacy (based on the choice of capital metric(s), hurdle rate(s), assumptions on RWAs, and the nature of data consolidation). Any capital shortfall under stress indicates the amount of potential recapitalization. Since the definition of capital varies across jurisdictions but directly impacts the scale of capital shortfall, there needs to be full disclosure of the composition of capital (Appendix 15.3), along with sufficient information about the implications of the planned adoption of regulatory changes— for example, phasing out of some types of eligible capital (BCBS 2010c, 2010a)—over the stress test risk horizon. With the finalization of the second stage of Basel III in late 2017 (BCBS 2017a), the transitional arrangements continue through 2027, and will have to be considered for the forthcoming stress tests (BCBS 2017b).

In FSAPs, the definition of capital is usually aligned with the one applied by country authorities. Over recent years, the capital definitions used in FSAP stress tests were guided by the jurisdictions’ implementation status of the Basel framework. All major countries have adopted Basel II (BCBS 2012b, 2012c) and, more recently, also Basel III capital standards (FSB 2017).6 They either:

  • Followed Basel II requirements (for example, Australia, India, Indonesia, Luxembourg, Netherlands, Turkey, Russia, and the United States);
  • Changed in line with the Basel III transition schedule for those that were moving to the new regime (for example, Brazil, France, Japan, Spain, and Sweden); in a couple of cases, their own national transitional schedules were applied (for example, Brazil and Japan);
  • Used benchmark parameters from the BCBS’s Sixth Quantitative Impact Study (BCBS 2010c)—a comprehensive study to evaluate the impact of the initial Basel III package agreed to in 2010—to simulate the likely impact of regulatory reforms on bank solvency (for example, Germany and the United Kingdom, where a separate and additional transitioning arrangement was also included for the BU exercise in the form of the interim capital regime); or
  • Applied a separate local regulatory capital definition (for example, Mexico).

Also capital metrics (and appropriate hurdle rates) tend to vary across countries. For countries with solvency regimes based on the Basel II capital definition, FSAP stress tests apply total regulatory capital to determine the hurdle rate. In cases when the Basel III (or a national modified version) applies, the metrics usually comprised the fully phased-in total capital, Tier 1 capital, and common Tier 1 capital (as of 2019 and 2022, respectively), along with the associated Basel III hurdle rates (Table 15.3) or (higher) national requirements, respectively. In some instance, the hurdles rates have included the capital conservation buffer (for example, France and Japan), while the loss absorbency requirement for global systemically important banks was applied rarely (for example, France). In a few cases, hurdle rates were set in line with existing regulatory standards (for example, Australia, the Netherlands, and the United Kingdom as an additional benchmark). In one instance, the 2019 Basel III target for common Tier 1 was applied as a supplementary benchmark for crisis credibility purposes (that is, Spain).

Table 15.3Original Basel III Transition Schedule
Source: Basel Committee for Banking Supervision (BCBS).Note: See BCBS 2010a, 2010b, 2017a, and Appendix 15.3 for capital definitions. Revisions to the liquidity risk framework under Basel III (BCBS 2013) resulted in the graduated introduction of the Liquidity Coverage Ratio (LCR). Specifically, the LCR was introduced as planned on January 1, 2015, but the minimum requirement began at 60 percent, rising in equal annual increments of 10 percentage points to reach 100 percent on January 1, 2019. CET1 = Common Equity Tier 1; DTAs = deferred tax assets; MSRs = mortgage servicing rights. For the most recent transitional arrangements see BCBS 2017b.
Source: Basel Committee for Banking Supervision (BCBS).Note: See BCBS 2010a, 2010b, 2017a, and Appendix 15.3 for capital definitions. Revisions to the liquidity risk framework under Basel III (BCBS 2013) resulted in the graduated introduction of the Liquidity Coverage Ratio (LCR). Specifically, the LCR was introduced as planned on January 1, 2015, but the minimum requirement began at 60 percent, rising in equal annual increments of 10 percentage points to reach 100 percent on January 1, 2019. CET1 = Common Equity Tier 1; DTAs = deferred tax assets; MSRs = mortgage servicing rights. For the most recent transitional arrangements see BCBS 2017b.

The capital assessment is also influenced by the change in RWAs over the risk horizon. Economic capital models and the credit risk assumptions underpinning the credit risk treatment in the Basel II/III framework imply a positive relationship between unexpected losses implied by RWAs (that is, potential worst-case losses) and default risk (and the resulting recovery rates). However, actual practice in stress tests for major country FSAPs has been varied:

  • RWAs were kept constant (for example, China, Japan, and Mexico).
  • RWA weights were kept constant, but the total RWA amounts were adjusted for credit growth and/or credit losses. This method corresponds approximately to the evolution of RWAs for banks using the standardized approach Basel II (for example, Russia, Saudi Arabia, and Spain).
  • RWAs changed under stress due to changes in the risk profile, in addition to the effects from asset growth (for example, France, Germany, Japan, Luxembourg, the Netherlands, and the United Kingdom). These changes are consistent with the rules for risk weights according to the Basel framework, which are either implicitly captured (for example, based on information from the Basel Committee for Banking Supervision’s Sixth Quantitative Impact Study, such as for Germany and the United Kingdom), or are treated more explicitly (for example, France). In other words, the evolution of RWAs is determined by changes in the estimated PDs and LGDs of a firm and/or portfolio level for IRB banks, subject to the evolution of total credit exposure under stress. For some countries, implicit IRB risk weights were simulated, to reflect the economic risk profile of banks that are still under the standardized approach (for example, Brazil).
  • RWAs for operational and market risks are often assumed to either (1) remain unchanged, or (2) change proportionally to the changes in RWAs for credit risk (mainly for market risk). FSAP stress tests are typically based on the assumption that the asset structure of banks remains the same during the stress test horizon, that is, there are portfolio rebalancing and/or substitution effects, such as replacing maturing loans with securities that attract different (usually lower) risk weights. Going forward, projecting RWAs during times of stress would need to also reflect revised specifications in the finalized Basel III framework as of late 2017 (BCBS 2017b).

Method

Once the key elements of the stress testing framework have been determined, various quantitative methods can be applied to estimate capital adequacy under projected financial stress. However, the stress testing literature provides little guidance on the selection and application of appropriate models in different circumstances. This issue has given rise to questions about the consistency and comparability of FSAP stress test results across countries and their implications for the associated stability analysis.

A comprehensive FSAP solvency stress testing exercise would preferably consist of three components: a balance sheet module, a market-based portfolio model, and spillover analysis. Balance-sheet- based methods cover a wide range of risks and exposures, often based on granular, reliable data, which tend to be most accessible, and, thus, represent the core of solvency stress tests. Market-based portfolio models are better able to reflect changes in actual risks (as reflected in investor perceptions) and the dependencies across multiple risks (possibly with greater flexibility to quantify point estimates at higher levels of statistical confidence) (see “Stress Test Models” section). Spillover analysis, which captures contagion risk and feedback effects, has become an important element of solvency stress tests in increasingly interconnected financial systems; however, the development of robust models in this area remains nascent (for example, Espinosa-Vega and Solé 2011), in large part due to data limitations.

Macroeconomic and Satellite Modeling

System-wide stress tests that are informed by adverse macro-economic conditions affecting the profitability and solvency of banks require satellite models that help project the impact of key sources of risk. Under each scenario, these models determine how changes in macroeconomic and financial sector variables impact impairments, various income components, such as net interest income (including funding costs), noninterest income, and trading income as well as credit growth as input to the solvency stress tests (Figure 15.4). Satellite models can be run at the economy level, sectoral level, and also at the level of individual banks (or of one of their specific portfolios).

General Representation of Satellite Modeling in Bank Solvency Stress Testing

Source: Authors.

Note: P&L = profit and loss statement; RWAs = risk-weighted assets.

The construction of satellite models typically comprises three key steps: (1) the choice of the estimation method, (2) the selection of the dependent variable and a set of potential explanatory variables that form the initial model specification, and (3) the iterative process of fitting the model (and completing robustness checks).

Various types of modeling may be used. These include time series analysis, regression models (for example, ordinary least squares regression, logistic regression, and panel data analysis), and structural models (Foglia 2009; Drehmann 2009). Most major country FSAP stress tests have typically relied on the country authorities’ satellite models, on the basis that these models would have undergone repeated calibrations and robustness checks over time. The FSAP team sometimes cross-validates with the IMF staff’s own satellite models in parallel TD tests (see Figures 15.5 and 15.6 for application in the FSAP Update for the United Kingdom [IMF 2011d]).

Example of Satellite Model Estimations for Bank Solvency Stress Testing: 2011 UK FSAP Update

Source: IMF 2011d.

Note: BU = bottom up; BoE = Bank of England; CPI = consumer price index; FSA = Financial Supervisory Authority; FSAP = Financial Sector Assessment Program; RAMSI = Risk Assessment Model of Systemic Institutions; TD = top down; WEO = World Economic Outlook.

Example of Application of Satellite Model Outputs to Top-Down Bank Solvency Stress Test Models: UK FSAP Update

Source: IMF 2011d.

Note: CCA = contingent claims analysis; CPI = consumer price index; CRE = commercial real estate; FSAP = Financial Sector Assessment Program; RAMSI = Risk Assessment Model of Systemic Institutions.

Stress Test Models

Given the rapidly evolving characteristics of financial systems in most countries, there is no stress testing model that is perfectly suited for a particular financial system. However, the chosen modeling approach should adequately capture the complexity, uniqueness, and idiosyncrasies of that system, subject to data availability. In FSAPs, the choice of the appropriate stress test model(s) can vary significantly depending on the characteristics of the respective financial system. For simple financial systems (with predominantly local banks), stress tests are normally less resource-intensive and can be completed with simpler models. In contrast, more complex systems require correspondingly more advanced stress testing methods to capture a wide range of material risks.

The stress testing methods are based on either a deterministic or stochastic framework. Deterministic approaches are predicated on prudential information in balance-sheet-based stress test specifications, while stochastic frameworks incorporate uncertainty around these accounting identities using historical volatility and/or market information, usually in the context of portfolio-based models. Both approaches can accommodate scenario and sensitivity analyses.

It is important to be aware of the differences between different stress testing methods and their implications for the interpretation of the results. As a general rule, the more sophisticated the model, the higher the estimation uncertainty. At the same time, simpler methods might be inadequate for highly interconnected and complex banking sectors with large credit and market risk exposures. When different approaches (TD, BU) and models are used in FSAPs, the results are cross-validated and the differences reconciled, which includes a discussion of the assumptions and caveats associated with the different models.

Most stress testing approaches do not fully reflect the impact of macro-financial dynamics during times of stress. For instance, existing FSAP stress tests do not adequately capture feedback effects beyond the initial impact of macroeconomic shocks on the banking sector, notwithstanding some work in this area (for example, Vitek and Bayoumi 2011; Catalán and others 2017; Krznar and Matheson 2017). The literature and the actual use of stress test models that include macro-financial feedback effects remain very limited (Alfaro and Drehmann 2009). An important reason is that the interaction between adverse macroeconomic scenarios, such as changes in credit aggregates, and firm-level financial soundness complicates the specification of feedback effects (BCBS 2009).

The IMF staff deploys a suite of stress testing models for surveillance stress testing. They can be categorized into two broad strands, supplemented by a third, and are discussed elsewhere in this section. These approaches are not mutually exclusive in that there are overlaps in the types of data utilized (Table 15.4 and Figure 15.7). The IMF staff has catalogued models developed within the institution to improve transparency in the models used in FSAPs and other areas of IMF work (Ong 2014).

Table 15.4Scorecard on Data and IMF Stress Test Models
Data Quality
Basic DataDetailed Data
Data TypeBalance sheet
Macroeconomic
Supervisory
Market
Interbank
Source: Ong 2014.Note: For descriptions of models, see: Espinosa-Vega and Solé 2011 for the network approach; Chan-Lau, Mitra, and Ong 2012 for the extreme value theory approach; Jobst and Gray 2013 and Gray and Jobst 2011 for Systemic Contingent Claims Analysis; Chan-Lau 2010 for CoRisk; and Segoviano and Padilla 2006 for the distress dependence framework.
Stress Test ElementsMethodologyAccounting basedAccounting based incorporating macro-financial modelsAccounting basedMarket-price basedMarket-price basedAccounting-and market-price based incorporating macro-financial modelsAccounting based incorporating macro-financial modelsMarket-price based incorporating macro-financial models
ModelBalance sheet approachBalance sheet approachNetwork approachExtreme-value theory approachCoRiskSystemic Contingent Claims AnalysisSatellite modelsDistress dependence framework
ShockSensitivity analysisMacroscenariosSensitivity analysisSensitivity analysisSensitivity analysisMacroscenariosMacroscenariosMacroscenarios
Source: Ong 2014.Note: For descriptions of models, see: Espinosa-Vega and Solé 2011 for the network approach; Chan-Lau, Mitra, and Ong 2012 for the extreme value theory approach; Jobst and Gray 2013 and Gray and Jobst 2011 for Systemic Contingent Claims Analysis; Chan-Lau 2010 for CoRisk; and Segoviano and Padilla 2006 for the distress dependence framework.
Source: Ong 2014.Note: For descriptions of models, see: Espinosa-Vega and Solé 2011 for the network approach; Chan-Lau, Mitra, and Ong 2012 for the extreme value theory approach; Jobst and Gray 2013 and Gray and Jobst 2011 for Systemic Contingent Claims Analysis; Chan-Lau 2010 for CoRisk; and Segoviano and Padilla 2006 for the distress dependence framework.

Stress Test Models Developed by the IMF Staff

Sources: Ong 2014; and authors.

Note: CCA = contingent claims analysis; EVT = extreme value theory.

Accounting-Based (balance sheet) Models This approach is most pervasive and has the longest history of use given its relative simplicity (for example, simulations could be done in a spreadsheet). It has the added attraction of directly producing results in terms of regulatory variables (for example, capital adequacy ratios). A variety of stress testing tools for banking sector analysis at various levels of development have been deployed for FSAPs and other surveillance work and technical assistance (Čihák 2007; Ong, Maino, and Duma 2010; Schmieder, Puhr, and Hasan 2011). The balance-sheet- based approach remains the cornerstone of FSAP stress testing and continues to be applied even in the largest, most systemic f-nancial systems, including in all major country FSAPs to date. The network model used for spillover analysis in FSAPs (Espinosa-Vega and Solé 2011) can also be considered an accounting-based approach.

Market-Price-Based Models The market-price-based models are often built on portfolio risk-management techniques and typically derive concise “systemic risk measures” from estimated dependencies among different risk factors. These risks (for example, sovereign, credit, and market) are typically excluded when modeling the default risk of each institution in isolation (Segoviano and Padilla 2006; Gray, Jobst, and Malone 2010; Chan-Lau 2010; Gray and Jobst 2011; Jobst and Gray 2013). Unlike accounting values, risk-based measures of solvency include additional considerations to inform the assessment of capital adequacy during times of financial stress (Figure 15.8):

  • The possibility that institutions may fail simultaneously (joint default risk): Most conventional stress tests are agnostic to default dependencies across institutions, that is, when one risk factor increases the likelihood of realization of other risk factors (with common shocks affecting multiple firms at the same time), especially under stressful conditions. The joint default risk of banks within a system depends on the individual bank’s propensity to cause and/or propagate shocks due to adverse changes in one or more risk factors (a distribution-based approach). Given that large shocks are transmitted across entities differently from small shocks, measuring nonlinear dependence in stress testing can provide important insights into the joint tail risks that arise in extreme loss scenarios. This would also include measuring the differential effects of the joint realization of multiple risk factors, which affects system-wide capital adequacy.
  • The sensitivity of stress test results to the historical volatility of risk factors (risk-based capital adequacy): Prudential information based purely on accounting identities observed at a certain point in time reflects the outcome of a stochastic process rather than a discrete value. Given that individual (and joint) default risk varies over time (and depends on joint effects across banks as discussed above), distribution-based approaches to capital assessments based on the historical volatility of risk factors represent a clear conceptual departure from conventional balance-sheet-based stress testing techniques. Unlike RWAs, risk-based measures of solvency (such as market-implied expected losses and the corresponding capital shortfalls) consider the actual historical dynamics of default risk, such as value at risk or expected shortfall (that is, the average density of extreme losses beyond value at risk at a selected percentile level [or conditional tail expectation]). Hence, in this distribution-based approach, the capital adequacy assessment reflects the variability of both assets and liabilities at different levels of statistical confidence.

Key Conceptual Differences in Loss Measurements between the Accounting-Based and Market-Price- Based Approaches

Source: Jobst and Gray 2013; and authors.

The market-based stress tests are likely to involve varying approaches of linking valuation methods to prudential standards, which could make them less tractable across countries due to flexible modeling. They usually do not show direct links to key regulatory ratios, which need to be derived in separate, additional steps. Prices of certain market instruments (for example, equity prices and credit-default- swap spreads) as essential inputs to market-based stress tests are not always readily available. Thus, data limitations have rendered these approaches supplementary to accounting-based approaches in FSAPs. For instance, distress dependence models were used in only a handful of major country FSAPs where the necessary data were available for robust estimation and credible implementation (for example, Germany, Mexico, Spain, Sweden, the United Kingdom, and the United States), and provided interesting insights into the interlink-ages of the default risk of sample banks.

Macro-Financial Models Macro-financial models represent the third strand of stress testing approaches. They are focused on the system-wide impact of different transmission channels between macroeconomic and financial conditions. By specifying certain macroeconomic situations, stress testers would apply consistent combinations of multiple shocks (for example, GDP, employment, inflation, exchange rate, interest rates, and asset prices), which could simultaneously affect various segments of banks’ businesses and exposures, and hence potentially result in overall capital losses. Macro-financial stress testing can be implemented by means of both accounting-and market-price-based models, by estimating additional macro-financial linkages affecting risk parameters used in the simulation exercises. The market-based models that fall into this category include the Systemic Contingent Claims Analysis and distress dependence (due to some structural elements in the specification of changes in joint default risk), while satellite models may also be classified as macrofinancial in nature.

Communication

Presentation of Outputs

Stress tests are aimed at drawing attention to the significance of key vulnerabilities during times of adverse macroeconomic and/or market conditions and deliver findings that may trigger closer supervisory review of certain financial activities, if necessary. Thus, results need to be presented in an accessible manner to appropriately convey the findings, which would allow country authorities to draw policy recommendations that inform appropriate prudential actions. In FSAPs, stress test results, especially those generated via the BU approach, are often aggregated for confidentiality reasons, which means that the design of a meaningful presentation format for analysis by the FSAP team is essential (Figures 15.915.11). Specifically, the presentation of the aggregated results by the authorities should be consistent with local regulatory requirements and, where relevant, any transition to a new regulatory regime (for example, Basel III). It should also be sufficiently granular to cover the presentation of the following information:

  • Name of each institution or peer group (if constrained by confidentiality) included in the stress testing exercise;
  • Dispersion of capital adequacy levels (before and after the application of different stress scenarios), such as the minimum, the maximum, and the interquartile range (for example, the 25th, 50th, and 75th percentiles) if they are not presented by institution;
  • Outcome for each year of the risk horizon;
  • Capital shortfall if one or more institutions fail to meet the predefined hurdle rate of capital adequacy (in absolute terms, as a percentage of GDP, and as a percentage of total sector assets within the scope of the exercise);
  • Details on the contributions of different drivers (for example, profitability, credit/trading losses, RWA) of the results; and
  • Assumptions and limitations of the design and implementation of the stress test(s).

Example of Bottom-Up Bank Solvency Stress Test Output Template Provided to Banks: UK FSAP Update1

Source: Authors.

Note: Results should be reported for hurdle rate assumptions without capital buffers (lines 10–12). Alternative stress test results should be based on hurdle rates that either ignore the capital phase-in/phase-out provisions of Basel III (lines 14–22) or include a capital buffer (lines 26–34). These results, however, have no impact on other sections of the main stress test and serve merely as a basis for sensitivity analysis. All other results reported in the spreadsheet (from line 39 onward) are based on the main results obtained from a stress testing set-up consistent with the Basel III treatment of capital but without capital buffers (lines 1–9). AfS = available for sale; FSAP = Financial Sector Assessment Program; FX = foreign exchange; GBP = UK Pound Sterling; LGD = loss given default; PD/NPL = probability of default/nonperforming loan; RWAs = risk-weighted assets.

1 The actual template is available on the IMF eLibrary at https://www.elibrary.imf.org/page/stress-test2-toolkit.

Example of Bottom-Up Bank Solvency Stress Test Output Template Provided to Authorities: UK FSAP Update1

Source: Authors.

Note: Results should be reported for hurdle rate assumptions without capital buffers (lines 10–12). Alternative stress test results are based on either hurdle rates that ignore the capital phase-in/phase-out provisions of Basel III (lines 14–22) or include a capital buffer (lines 26–34). These results, however, have no impact on other sections of the main stress test and serve merely as a basis for sensitivity analysis. All other results reported in the spreadsheet (from line 39 onward) are based on the main results obtained from a stress testing set-up consistent with the Basel III treatment of capital but without capital buffers (lines 1–9). AfS = available for sale; FSAP = Financial Sector Assessment Program; FX = foreign exchange; GBP = UK Pound Sterling; LGD = loss given default; PD/NPL = probability of default/nonperforming loans; RWAs = risk-weighted assets.

1 The actual template is available on the IMF eLibrary at https://www.elibrary.imf.org/page/stress-test2-toolkit.

Example of Bottom-Up Bank Solvency Stress Test Summary Template Provided to Authorities: UK FSAP Update1

Source: Authors.

Note: FSAP = Financial Sector Assessment Program; ppts = percentage points.

1 The actual template is available on the IMF eLibrary at https://www.elibrary.imf.org/page/stress-test2-toolkit.

The findings of the stress tests are then used to (1) provide quantitative support for the FSAP’s stability risk assessment, and (2) facilitate policy discussions with the authorities on risk-mitigation strategies and crisis preparedness.

Publication

As a final step in the FSAP stress testing process, the main findings are published. The disclosure of the stress test results (in addition to the standards assessment) forms a substantial part of public accountability but is often a very sensitive issue (especially during times of macro-financial challenges) and requires supervisory discretion (if the effective implementation of remedial actions might be compromised). In addition to providing a meaningful judgment on the outcome of the test (for instance, the fact that no bank fails a test does not mean that vulnerabilities do not exist), findings should be appropriately nuanced to ensure that the information does not promote a false sense of security or cause undue alarm. In FSAPs, this objective is commonly achieved through:

  • Clear documentation of definitions, assumptions, models, and limitations of stress tests in Technical Notes and/or as supplementary information in the FSSA report;7
  • Mandatory summaries of the stress testing exercises in a standardized format (that is, the STeM) as part of the FSSA to improve transparency and facilitate cross-country comparisons (Table 15.5); and
  • Disclosure of the aggregated results of stress tests after conclusion of the FSAP, including a minimum amount of information such as the relevant post-stress ratio(s) and the respective amount(s) of capital shortfall.8
Table 15.5Example of Stress Test Matrix (STeM) for Bank Solvency Risk: Spain FSAP Update
DomainAssumptions
Top Down by AuthoritiesTop Down by FSAP Team
Institutions included
  • Commercial banks and intervened savings banks
  • All publicly listed banks with sufcient pricing history
Market share
  • Over 96 percent of the banking sector, excluding foreign branches
  • About 45 percent of the banking sector, excluding foreign branches
Data and baseline date
  • Supervisory data as at end-2011
  • Scope of consolidation: legal entity as at end-2011
  • Risk horizon of two years, under crisis conditions
  • Publicly available market and statutory data. Scope of consolidation: legal entity as at end-2011
  • Risk horizon of two years, under crisis conditions
Methodology (for example, included in scenario analysis linking solvency and liquidity, separate test using ad hoc model/balance sheet)
  • BdE macro-financial panel regression model (estimates capital shortfall) without behavioral adjustments
  • IMF balance sheet approach (estimates capital shortfall)
  • Systemic CCA model (estimates expected losses, capital shortfall, and contingent liabilities)
Risks (for example, funding liquidity shock, market liquidity shock, both)
  • “Double-dip” recession (severe and short-term) scenario of one standard deviation from the IMF-projected baseline GDP growth trend over a two-year risk horizon—without positive adjustment dynamics toward the end of the (short) risk horizon
  • The second, more adverse scenario further escalates the macroeconomic shock by increasing the shock to two-year real GDP growth by another 2.5 percentage points
  • Sovereign risk refected in valuation haircut to AfS and trading book debt holdings
  • Extra provisioning and capital add-on due to regulatory changes
Regulatory standards
  • Basel II transitioning to Basel III and Basel III capital requirements slightly exceeded (4 percent Core Tier 1 hurdle rate for both years)
  • Basel III capital definition
  • RWAs remain constant but are subject to changes due to deleveraging by banks in both 2012 and 2013
Results
  • Postshock, more than a third of all banks in the system would not be able to comply with Basel III hurdle requirements until end-2013 irrespective of the choice of top-down model
  • The BdE model reveals projected impairment losses of around € 73 billion under the IMF adverse scenario, which generates capital shortfall of about € 18 billion compared with a Core Tier 1 capital hurdle rate of 4 percent
  • Based on the Systemic CCA results, challenges exist from the realization of low probability tail risk of multiple firms experiencing a dramatic escalation of losses. In the IMF adverse scenario, the largest (and publicly listed) banks would experience a market-implied capital shortfall of more than € 14 billion on average (with a peak in excess of € 21 billion at end-2012) at a statistical probability of 5 percent or less (expressed as “tail risk”)
Source: IMF 2012b.Note: AfS = available for sale; BdE = Banco de España; Systemic CCA = Systemic Contingent Claims Analysis; FSAP = Financial Sector Assessment Program; RWAs = risk-weighted assets.
Source: IMF 2012b.Note: AfS = available for sale; BdE = Banco de España; Systemic CCA = Systemic Contingent Claims Analysis; FSAP = Financial Sector Assessment Program; RWAs = risk-weighted assets.

In the case of the FSAPs covered in this chapter, all sample countries have published their FSSAs. In almost all cases, Technical Notes on the respective stress testing exercises were completed (with the exception of Australia and Spain, where the details are described in appendices to the respective FSSAs), but only a few countries (that is, Germany, Sweden, the United Kingdom, and the United States) consented to their publication (see IMF 2010c, 2011b, and 2011a, respectively).

Conclusion

The lead-up to the global financial crisis illustrated that surveillance stress tests are not fail-safe, stand-alone diagnostic tools. Conceptually, the implementation of stress tests is very challenging due to: (1) diverse business models and activities of sample banks; (2) varying degrees of estimation uncertainty associated with models, based on assumptions that may not be sufficiently robust to capture all the relevant risks; (3) binding constraints to data availability and quality; and (4) prudential concerns and/or political sensitivities affecting the formulation of credible stress scenarios. The complexity of running stress tests is magnified during crises when rapidly changing financial conditions and heightened market expectations require a carefully planned communication strategy.

At the IMF, significant efforts continue to be made to address the identified shortcomings. Some of the steps taken include: (1) standardizing the shock scenarios across countries, where possible, and making nascent attempts to quantify the likelihood of the realization of specific scenarios; (2) applying more encompassing stress tests (that is, complementary accounting-and market-price-based models) with a wider coverage of risks; and (3) ensuring a more consistent and cohesive presentation of assumptions and results to support an effective comparability of implementation and findings across different countries.

Building on the progress so far, IMF stress tests will require continuous innovation and adjustment to adequately capture relevant risks in an evolving and an exceedingly complex international financial system. Important areas for improvement that will ensure that FSAP stress tests are ft for their purpose include: (1) the integration between solvency and liquidity risks; (2) spillover analysis, both within a financial system and across borders; and (3) the incorporation of feedback loops between the real economy and financial sector. The IMF staff has published a set of “best practice” principles on macro-financial stress testing, drawing on the accumulated experience of more than a decade of FSAPs (Chapter 2). This chapter, in turn, illustrates the application of these principles by reviewing key elements of IMF stress tests—specifically, in the major country FSAPs—during the global financial crisis and their actual implementation.

While greater harmonization of methods and approaches have helped enhance the consistency and comparability across countries, complete standardization of FSAP stress tests remains elusive. Qualitative analysis and expert judgment are, and will continue to be, indispensable for what amounts to an art form rather than an exact science. Given the many “moving parts” of stress tests, sufficient flexibility not only ensures that their design and implementation remain relevant amid a constantly evolving spectrum of risks but also helps absorb innovative methodologies to adequately capture them— subject to considerable variations in local regulatory requirements and the political sensitivities affecting the communication of stress test results. That said, these challenges should not undermine the value of well-designed stress tests.

Appendix 15.1. FSAP Solvency Stress Tests FY2010–FY13
Appendix Table 15.1.1Financial Sector Assessment Program Solvency Stress Tests FY2010–FY13: Stress Test Matrix for S- 25 and Other G20 Countries
Timing of FSAP1 United States FY20102 Indonesia FY20103 China FY20114 Luxembourg FY20115 Netherlands FY20116 Germany FY20117 United Kingdom FY20118 Turkey FY20119 Russia FY2011
Stress Testing Framework
1. Scope
Approach
Bottom-up
  • No.
  • Bottom-up (BU) by banks, in collaboration with authorities and IMF.
  • BU by banks in collaboration with authorities and IMF.
  • No.
  • No.
  • No.
  • BU by banks in collaboration with authorities and IMF.
  • BU by banks in collaboration with authorities and IMF.
  • BU by banks, in collaboration with.
Top-down
  • 3 top-down (TD) tests by IMF.
  • TD by IMF in collaboration with authorities.
  • TD by authorities.
  • TD by IMF.
  • TD by authorities.
  • TD by IMF in collaboration with authorities.
  • TD by IMF in collaboration with authorities.
  • TD by IMF.
  • TD by authorities.
  • TD by IMF.
  • TD by IMF in collaboration with authorities.
  • TD by authorities.
Coverage
Institutions
  • 54 bank holding companies (BHCs) using balance sheet (B/S) approach.
  • 36 BHCs using the Consistent Information Multivariate Density (CIMDO) methodology.
  • 14 Systemically Important Financial Institutions using Systemic Contingent Claims Analysis (SCCA)
  • TD: All 121 commercial banks, excl. rural banks (115 for scenario analysis; all for sensitivity analysis).
  • BU: 12 largest banks (8 for scenario analysis, all for sensitivity analysis).
  • 17 banks (5 large commercial, 12 joint-stock commercial)
  • 108 subsidiaries and branches.
  • 7 banks.
  • 3 banking groups, 16 largest German banks (14 SIFIs plus two Landesbanken), the savings banks (Sparkassen), and the other cooperative banks; very small banks were excluded from the sample.
  • 14 SIFIs (SCCA)
  • 6 largest banks + largest building society.
  • 9 largest banks.
  • BU: 15 largest banks.
  • TD: All commercial banks (1,012)
Market share
  • B/S: 85 percent.
  • CIMDO: 59 percent.
  • SCCA: 70 percent.
  • TD: 100 percent.
  • BU: 60 percent.
  • 66 percent of total banking sector assets (86 percent of commercial banking sector assets).
  • 100 percent.
  • 85 percent.
  • B/S: 85 percent.
  • SCCA: 60 percent.
  • 88 percent.
  • Over 80 percent.
  • BU: 56 percent.
  • TD: 100 percent.
Reporting basis
  • Consolidated banking groups.
  • Unconsolidated banking groups.
  • Consolidated banking groups.
  • Unconsolidated local entities.
  • Consolidated banking groups.
  • Unconsolidated domestic businesses.
  • Consolidated banking groups.
  • Unconsolidated domestic.
  • Unconsolidated local entities.
Data
Source
  • Publicly available data.
  • BU: Banks’ own data.
  • TD: Supervisory and publicly available data.
  • BU: Banks’ own data.
  • TD by authorities: Supervisory and publicly available data.
  • TD by IMF: Publicly available data.
  • Supervisory data.
  • Supervisory data.
  • Supervisory and publicly available data.
  • BU: Banks’ own data.
  • TD: Supervisory and publicly available data.
  • BU: Banks’ own data.
  • TD: Supervisory data.
  • BU: Banks’ own data.
  • TD: Supervisory data.
Cut-offs date
  • End-2009
  • As at Sep 2009
  • End-2009
  • As at Jun 2010
  • As at Jun 2010
  • Hybrid.,:
    • — End-2009 for B/S positions.
    • — Sep 2010 for P&L.
  • End-2010
  • End-2010
  • End-2010
2. Scenario Design
Risk horizon
  • 2010–14 (5 years)
  • 2009Q4–2012Q4 (3 years)
  • Scenario: 2010 (1 year)
  • Sensitivity: 1Q, 1 year or 2 years
  • 2011–12 (2 years)
  • 2011-15 (5 years)
  • 2011–15 (5 years)
  • 2011–15 (5 years)
  • BU by banks: 2011 (1 year)
  • TD: 2011–13 (3 years): Sudden stop.
  • TD: 2011–14 (4 years): Boom and bust.
  • Instantaneous
  • 2011 (1 year)
Scenarios
Baseline
  • WEO Apr 2010.
  • WEO Apr 2009.
  • N/A.
  • WEO Oct 2010.
  • WEO Oct 2010.
  • WEO Oct 2010.
  • WEO Oct 2010.
  • WEO Feb 2011.
  • Slightly below WEO Jan 2011.
Growth shocks (calculated per Committee of European Banking Supervisors for Scenario Designs unless indicated otherwise)
  • Combined impact of four adverse shocks.,:
    • (i) sizeable and persistent shock to growth rate of potential output;
    • (ii) an additional short run demand shock, reflecting high unemployment, weak credit, and continued fall in housing prices;
    • (iii) further near-term fiscal stimulus to support near-term growth; and (iv) rising inflation expectations.
    • Output gap falls by 23 percentage points relative to baseline in adverse scenario.
    • Output gap falls by 33 percentage points relative to baseline in alternative adverse scenario.
  • ≈ 1/3 output loss experienced during Asian crisis.
  • GDP growth down from 12 percent to:
    • 7 percent (mild)
    • —5 percent (medium)
    • —4 percent (severe)
  • 1 standard deviation (SD)
  • 1 SD.,
  • 2 SD.
  • 1.5 SD (1 SD and 2 SD run by FSAP team for internal comparisons)
  • 1 SD.
  • 2 SD.
  • Sharp contraction over four quarters followed by a sluggish recovery over the next 12 quarters.
  • A two-year boom in growth and credit followed by a sharp contraction over four quarters and then a sluggish recovery.
  • 1 SD.
  • 1.7 SD.
Slow growth scenario.,
  • Yes
  • No
  • Yes
  • No
  • No
  • Yes
  • Yes
  • Yes
  • No
Risks
Key risk(s)
  • Credit risk.
  • Credit risk.
  • Credit risk associated with rapid loan growth.
  • Credit risk.
  • Credit risk.
  • Credit risk.
  • Credit risk.
  • Credit risk.,.
  • Credit risk.
  • Adjustments for regulatory forbearance.
Other risks covered in scenario analysis
  • N/A.
  • N/A.
  • N/A.
  • Sovereign risk, in both trading and banking books (CEBS model)
  • Sovereign risk, in both trading and banking books (CEBS model)
  • Off-balance sheet exposures.
  • Sovereign risk in trading book only (IMF models); application of sovereign haircuts on banking book in sensitivity analysis completed separately by IMF staff.,.
  • Sovereign and banking risks in both trading and banking books (IMF models)
  • Funding risk.
  • N/A.
  • Sovereign and ings, in trading book and AfS in banking book.
  • Propagation channel through network effects.
  • Liquidity stress measured by its solvency impact (losses from fire sales of liquid assets)
Other tests/risks
  • Sensitivity tests: Credit and m arket risks.
  • Sensitivity tests: Credit and market risks.
  • Sensitivity tests: Credit and market risks, including.,:(i) largest individual exposures;:(ii) real estate sector exposures;:(iii) exposures to local government financing platforms (LGFPs);:(iv) exposures to overcapacity industries; and:(v) exposures to export sectors.
  • Contagion risk.
  • Reverse stress test.
  • Sensitivity tests: Credit and market risks.
  • Sensitivity tests: Credit and market risks.
  • N/A.
  • N/A.
  • Sensitivity tests: Credit and m arket risks.
  • Spillover risk through domestic network effects (included in macro scenarios)
Source: Compiled by authors with contributions from respective FSAP stress testers.Note: The IMF fiscal year runs from May 1 to April 30. The table presented here is a representation only; a full-sized version is available as an MS Excel® file on the IMF eLibrary at www.elibrary.imf.org/page/stress-test2-toolkit. AfS = available for sale; CEBS = Committee of European Banking Supervisors; SIFI = systemically important financial institution.
Source: Compiled by authors with contributions from respective FSAP stress testers.Note: The IMF fiscal year runs from May 1 to April 30. The table presented here is a representation only; a full-sized version is available as an MS Excel® file on the IMF eLibrary at www.elibrary.imf.org/page/stress-test2-toolkit. AfS = available for sale; CEBS = Committee of European Banking Supervisors; SIFI = systemically important financial institution.
Appendix 15.2. Example of Summary of Key Assumptions Applied in Solvency Stress Testing
Appendix Table 15.2.1Example of Summary of Key Assumptions Applied in Solvency Stress Testing Exercise: UK FSAP Update
DomainElementSpecific Rules/Assumptions
(Risk) factors assessedLoss rates Profitability Fixed income holdings FX shock Taxes
  • Credit losses based on satellite models developed by firms depending on scenario.
  • Profit (interest income, interest expenses, net fee, and commission income, and operating expenses) should be based on firm’s satellite models (or expert judgment). For end-2010, net profit before tax should be adjusted for extraordinary income/losses to avoid misleading results.
  • Trading income based on satellite model or statistical matching of both trading income and GDP growth using a parametric ft of their historical distribution (for example, a decline in GDP growth is assumed to reduce trading income).
  • Funding costs based on satellite model for interest expenses, including a nonlinear effect. Changes in funding costs due to different solvency conditions cannot be smaller than the one generated by either some general funding cost sensitivity or results from suggested CCA-based approach (Appendix 15.3, Option 2). These changes are unaffected by possible balance sheet deleveraging.
  • Mark-to-market impact on fixed-income holdings: Focuses on the projection of haircuts for holdings of both sovereign and bank debt based on IMF approach. These haircuts will be applied to both trading and banking book.
  • Sovereign and financial sector debt holdings: Haircut on holdings in the banking and trading books based on market expectations over five years after controlling for changes of market valuation during 2010 as developed by the IMF staf. Cash at central banks, repos, and asset swaps where there is no economic interest in the security (for instance, instruments held against assets pledged to the Bank of England) are excluded. Moreover, haircuts are applied only to issuers that are non-”AAA” rated.
  • FX shock: Firms are asked to report separately the marginal impact of the following FX shock of the following currencies on net open positions: US dollar, euro, and Japanese yen. The shock for each currency should be twice the standard deviation of the respective FX volatility during 2010 and impact the trading book in 2011 (100 percent) and 2012 (50 percent) only.
  • Tax assumption: 25 percent in case of positive profits, zero otherwise.
Behavioral adjustment of banksDividend payout rules (similar to Basel III minima)

Credit growth Asset disposal Capital raising
  • Balance sheets are assumed to be constant (that is, they grow in line with nominal GDP).
  • Dividend payout depends on capitalization under stress: dividend payout only if firm reports profits over the past year; if total capital ratio is above 8 percent (after the envisaged dividend payout and, at the same time, exhibits sufficient Tier 1 and common equity Tier 1 capitalization) but below the 10.5 percent threshold (which reflects the magnitude of the proposed “capital conservation buffer” under Basel III), the firm is considered capital-constrained and needs to follow a defined payout schedule.
  • Credit growth in line with nominal GDP for banks with a Tier 1 capital buffer of 2.5 percentage points above the regulatory minimum (that is, hurdle rate); credit growth decreases by 2 percentage points for each decrease in Tier 1 capital by 1 percentage point once the capital buffer is less than 2.5 percentage points above the Tier 1 capital hurdle rate. Hence, growth becomes negative when capitalization is at the minimum capital ratio unless nominal GDP grows by more than 5 percent.
  • Other business strategy considerations: Asset disposals or acquisitions over time should not be considered, except where legally binding commitments under EU State aid rules exist. Maturing exposures are assumed to be replaced. Any interim capital raising until end- 2010 can be considered in calculations.
Source: IMF 2011d.Note: CCA = contingent claims analysis; FX = foreign exchange rate; FSAP = Financial Sector Assessment Program.
Source: IMF 2011d.Note: CCA = contingent claims analysis; FX = foreign exchange rate; FSAP = Financial Sector Assessment Program.
Appendix 15.3. Example of Comparison Table on Relevant Core Tier 1 Capital Definitions
Appendix Table 15.3.1Example of Comparison Table on Relevant Core Tier 1 Capital Definitions: UK FSAP Update
Capital ComponentBasel IIBasel IIIEBAFSA General Prudential SourcebookFSA Interim Capital Regime/FSAP Bottom-Up Stress Testing
Core Tier 1 (CT1)
  • Ordinary shares.
  • Retained earnings and reserves.
  • Share p remium account.
  • Minority interests.
  • Externally verified interim net profits.
  • Ordinary shares.
  • Retained earnings and reserves.
  • Share premium account relating to CT1 instruments.
  • Minority interests (subject to limits).
  • Interim net profits.
  • Ordinary shares.
  • Retained earnings and reserves.
  • Share premium account relating to CT1 instruments.
  • Minority interests.
  • Externally verified interim net profits.
  • Existing government support measures counted as CT1.
  • Ordinary shares.
  • Retained earnings and reserves.
  • Share premium account relating to CT1 instruments.
  • Minority interests.
  • Externally verified interim net profits.
  • Ordinary shares.
  • Retained earnings and reserves.
  • Share premium account relating to CT1 instruments.
  • Minority interests.
  • Externally verified interim net profits.
Core Tier 1 Filters
  • Cash-flow hedge reserve not fair valued on balance sheet.
  • Gain on sale related to securitization transactions.
  • Cumulative gains and losses due to changes in own credit risk on fair valued financial liabilities.
  • Existing national flters -see FSA GENPRU column for UK flters.
  • Pension deficit net of DRA (if approach chosen).
  • Unrealized gains on AfS equities.
  • Unrealized gains on Investment property.
  • Unrealized gains on land and buildings.
  • Unrealized losses (gains) on AfS debt.
  • Cash-flow hedge reserve not fair valued on balance sheet.
  • Gain on sale related to securitization transactions.
  • Cumulative gains and losses due to changes in own credit risk on fair valued financial liabilities.
  • Pension deficit net of DRA (if approach chosen).
  • Unrealized gains on AfS equities.
  • Unrealized gains on investment property.
  • Unrealized gains on land and buildings.
  • Unrealized losses (gains) on AfS debt.
  • Cash-flow hedge reserve not fair valued on balance sheet.
  • Gain on sale related to securitization transactions.
  • Cumulative gains and losses due to changes in own credit risk on fair valued financial liabilities.
Deductions from Core Tier 1
  • Interim net losses.
  • Interim net losses.
  • Intangibles including goodwill (limited recognition of mortgage servicing rights).
  • Investments in own shares.
  • Shortfall of the stock of provisions to expected losses.
  • Defined benefit pension fund assets and liabilities (include liabilities in full, deduct assets).
  • Deferred tax assets (limited recognition allowed).
  • Reciprocal cross holdings in the common stock of banking, financial, and insurance entities.
  • Investments in the common stock of banking, financial, and insurance entities that are outside the scope of regulatory consolidation and where the bank does not own more than 10 percent of the issued common share capital.
  • Significant investments in the common stock of banking, financial, and insurance entities that are outside the scope of regulatory consolidation (limited recognition).
  • Interim net losses.
  • Intangibles including goodwill.
  • Investments in own shares.
  • 50 percent shortfall in stock of provisions to expected losses.
  • 50 percent of certain securitization exposures.
  • Certain qualifying holdings.
  • 50 percent material holdings in financial institutions (excluding material insurance holdings).
  • 50 percent free deliveries.
  • Interim net losses.
  • Interim net losses.
  • Intangibles including goodwill.
  • Investments in own shares.
  • 50 percent shortfall in stock of provisions to expected losses.
  • 50 percent of certain securitization exposures.
Source: IMF 2011d.Note: AfS = available for sale; DRA = deficit reduction amount; EBA = European Banking Authority; FSA = Financial Supervisory Authority; FSAP = Financial Sector Assessment Program; GENPRU = general prudential sourcebook for banks, building societies, insurers, and investment firms.
Source: IMF 2011d.Note: AfS = available for sale; DRA = deficit reduction amount; EBA = European Banking Authority; FSA = Financial Supervisory Authority; FSAP = Financial Sector Assessment Program; GENPRU = general prudential sourcebook for banks, building societies, insurers, and investment firms.
References

    AlfaroRodrigo andMathiasDrehmann. 2009. “Macro Stress Tests and Crises: What Can We Learn.” BIS Quarterly Review (December): 2941. https://www.bis.org/publ/qtrpdf/r_qt0912e.htm.

    • Search Google Scholar
    • Export Citation

    Banco de España. 2012. “Bank Recapitalisation and Restructuring Process: Results of the Independent Evaluation of the Spanish Sector: Presentation by the Deputy Governor.” http://www.bde.es/bde/en/secciones/prensa/infointeres/reestructuracion/valoracionesind/.

    • Search Google Scholar
    • Export Citation

    Basel Committee on Banking Supervision (BCBS). 2009. “Principles for Sound Stress Testing Practices and Supervision.” BCBS Publication 155 Bank for International Settlements Basel Switzerland May. https://www.bis.org/publ/bcbs155.htm.

    • Search Google Scholar
    • Export Citation

    Basel Committee on Banking Supervision (BCBS). 2010a. “Basel III: A Global Regulatory Framework for More Resilient Banks and Banking Systems.” BCBS Publication 189 Bank for International Settlements Basel Switzerland December. http://www.bis.org/publ/bcbs189.htm.

    • Search Google Scholar
    • Export Citation

    Basel Committee on Banking Supervision (BCBS). 2010b. “Basel III: International Framework for Liquidity Risk Measurement, Standards and Monitoring.” BCBS Publication 188 Bank for International Settlements Basel Switzerland December. https://www.bis.org/publ/bcbs188.htm.

    • Search Google Scholar
    • Export Citation

    Basel Committee on Banking Supervision (BCBS). 2010c. “Results of the Comprehensive Quantitative Impact Study.” BCBS Publication 186 Bank for International Settlements Basel Switzerland December. http://www.bis.org/publ/bcbs186.htm.

    • Search Google Scholar
    • Export Citation

    Basel Committee on Banking Supervision (BCBS). 2011. “Global Systemically Important Banks: Assessment Methodology and the Additional Loss Absorbency Requirement.” BCBS Publication 207 Bank for International Settlements Basel Switzerland November. http://www.bis.org/publ/bcbs207.htm.

    • Search Google Scholar
    • Export Citation

    Basel Committee on Banking Supervision (BCBS). 2012a. “Peer Review of Supervisory Authorities’ Implementation of Stress Testing Principles.” BCBS Publication 218 Bank for International Settlements Basel Switzerland April. http://www.bis.org/publ/bcbs218.htm.

    • Search Google Scholar
    • Export Citation

    Basel Committee on Banking Supervision (BCBS). 2012b. “Report to G20 Leaders on Basel III Implementation.” BCBS Publication 220 Bank for International Settlements Basel Switzerland June. http://www.bis.org/publ/bcbs220.htm.

    • Search Google Scholar
    • Export Citation

    Basel Committee on Banking Supervision (BCBS). 2012c. “Progress Report on Basel III Implementation.” BCBS Publication 232 Bank for International Settlements Basel Switzerland October. https://www.bis.org/publ/bcbs232.htm.

    • Search Google Scholar
    • Export Citation

    Basel Committee on Banking Supervision (BCBS). 2012d. “A Framework for Dealing with Domestic Systemically Important Banks.” BCBS Publication 233 Bank for International Settlements Basel Switzerland October. https://www.bis.org/publ/bcbs233.htm.

    • Search Google Scholar
    • Export Citation

    Basel Committee on Banking Supervision (BCBS). 2013. “Basel III: The Liquidity Coverage Ratio and Liquidity Risk Monitoring Tools.” BCBS Publication 238 Bank for International Settlements Basel Switzerland January. http://www.bis.org/publ/bcbs238.htm.

    • Search Google Scholar
    • Export Citation

    Basel Committee on Banking Supervision (BCBS). 2017a. “Basel III: Finalising Post-crisis Reforms.” BCBS Publication 424 Bank for International Settlements Basel Switzerland January. https://www.bis.org/bcbs/publ/d424.htm.

    • Search Google Scholar
    • Export Citation

    Basel Committee on Banking Supervision (BCBS). 2017b. “Basel III Transitional Arrangements, 2017–2027.” Bank for International Settlements Basel Switzerland December.

    • Search Google Scholar
    • Export Citation

    Basel Committee on Banking Supervision (BCBS). 2018a. “Stress Testing Principles.” BCBS Publication 450 Bank for International Settlements Basel Switzerland October. https://www.bis.org/bcbs/publ/d428.htm.

    • Search Google Scholar
    • Export Citation

    Basel Committee on Banking Supervision (BCBS). 2018b. “Global Systemically Important Banks: Revised Assessment Methodology and the Higher Loss Absorbency Requirement.” BCBS Publication 445 Bank for International Settlements Basel Switzerland December. https://www.bis.org/bcbs/publ/d445.htm.

    • Search Google Scholar
    • Export Citation

    BernankeBenS.2010. “The Supervisory Capital Assessment Program— One Year Later.” Speech delivered at the 46th Annual Conference on Bank Structure and Competition Federal Reserve Bank of Chicago Chicago May. http://www.federalreserve.gov/newsevents/speech/bernanke20100506a.htm.

    • Search Google Scholar
    • Export Citation

    Board of Governors of the Federal Reserve System. 2009. “The Supervisory Capital Assessment Program: Overview of Results.” Board of Governors of the Federal Reserve System Washington DC May.

    • Search Google Scholar
    • Export Citation

    Board of Governors of the Federal Reserve System. 2012a. “Comprehensive Capital Analysis and Review.” Board of Governors of the Federal Reserve System Washington DC March. http://www.federalreserve.gov/bankinforeg/ccar.htm.

    • Search Google Scholar
    • Export Citation

    Board of Governors of the Federal Reserve System. 2012b. “Comprehensive Capital Analysis and Review 2012: Methodology for Stress Scenario Projection.” Board of Governors of the Federal Reserve System Washington DC March. http://www.federalreserve.gov/newsevents/press/bcreg/20120312a.htm.

    • Search Google Scholar
    • Export Citation

    Board of Governors of the Federal Reserve System. 2012c. “Supervisory Guidance on Stress Testing for Banking Organizations with More Tan $10 Billion in Total Consolidated Assets.” Board of Governors of the Federal Reserve System SR Letter 12–7 Washington DC May.

    • Search Google Scholar
    • Export Citation

    Board of Governors of the Federal Reserve System Federal Deposit Insurance Corporation and Office of the Comptroller of the Currency. 2012. “Guidance on Stress Testing for Banking Organizations with Total Consolidated Assets of More Tan $10 Billion.” SR Letter 12–7 Board of Governors of the Federal Reserve System Washington DC May. https://www.federalreserve.gov/supervisionreg/srletters/sr1207.htm.

    • Search Google Scholar
    • Export Citation

    BorioClaudioMathiasDrehmann andKostasTsatsaronis. 2012 Stress-testing Macro Stress Testing: Does It Live Up to Expectations?BIS Working Paper 369 Bank for International Settlements Basel Switzerland January. http://www.bis.org/publ/work369.htm.

    • Search Google Scholar
    • Export Citation

    CatalánMarioDaleGrayLauraValderramaandTengTeng Xu. 2017. “ Macro-Financial Feedbacks in Stress Testing.” Presentation at Joint IMF-EBA Colloquium— New Frontiers on Stress Testing London March 1.

    • Search Google Scholar
    • Export Citation

    Central Bank of Ireland. 2011. “The Financial Measures Programme Report.” Central Bank of Ireland Dublin Ireland March. https://www.centralbank.ie/publication/financial-measures-programme.

    • Search Google Scholar
    • Export Citation

    CeruttiEugenio andChristianSchmieder. 2012. “The Need for “ Un-consolidating Consolidated Banks’ Stress Tests.” IMF Working Paper 12/288 International Monetary Fund Washington DC. https://www.imf.org/en/Publications/WP/Issues/2016/12/31/The-Need-for-Un-consolidating-Consolidated-Banks-Stress-Tests-40151.

    • Search Google Scholar
    • Export Citation

    Chan-LauJorge. 2010. “Regulatory Capital Charges for Too-Connected-To-Fail Institutions: A Practical Proposal.” IMF Working Paper 10/98 International Monetary Fund Washington DC. https://www.imf.org/en/Publications/WP/Issues/2016/12/31/Regulatory-Capital-Charges-for-Too-Connected-to-Fail-Institutions-A-Practical-Proposal-23753.

    • Search Google Scholar
    • Export Citation

    Chan-LauJorgeSrobonaMitra andLiLian Ong. 2012. “Identifying Contagion Risk in the International Banking System: An Extreme Value Theory Approach.” International Journal of Finance and Economics17 (4): 390406.

    • Search Google Scholar
    • Export Citation

    ČihákMartin. 2007. “Introduction to Applied Stress Testing.” IMF Working Paper 07/59 International Monetary Fund Washington DC. https://www.imf.org/en/Publications/WP/Issues/2016/12/31/Introduction-to-Applied-Stress-Testing-20222.

    • Search Google Scholar
    • Export Citation

    Committee of European Banking Supervisors (CEBS). 2010. “2010 EU Wide Stress Testing.” July. London: Committee of European Banking Supervisors.

    • Search Google Scholar
    • Export Citation

    DrehmannMathias. 2009. “Macroeconomic Stress-testing Banks: A Survey of Methodologies.” In Stress Testing the Banking System: Methodologies and Applications edited by MarioQuagliariello. New York: Cambridge University Press.

    • Search Google Scholar
    • Export Citation

    Espinosa-VegaMarcoA. andJuanSolé. 2011. “Cross-border Financial Surveillance: A Network Perspective.” Journal of Financial Economic Policy3 (33): 182205.

    • Search Google Scholar
    • Export Citation

    European Banking Authority (EBA). 2010. “Aggregate Outcome of the 2010 EU-wide Stress Test Exercise Coordinated by CEBS in Cooperation with the ECB.” European Banking Authority London July. https://www.eba.europa.eu/documents/10180/15938/Summaryreport.pdf/95030af2-7b52-4530-afe1-f067a895d163.

    • Search Google Scholar
    • Export Citation

    European Banking Authority (EBA). 2011a. “2011 EU-Wide Stress Testing Exercise.” European Banking AuthorityLondonMarch.

    European Banking Authority (EBA). 2011b. “Overview of the 2011 EU-Wide Stress Test: Methodological Note.” European Banking Authority London October. https://eba.europa.eu/documents/10180/26923/Sovereign-capital-shortfall_Methodology-FINAL.pdf/acac6c68–398e-4aa2-b8a1-c3dd7aa720d4.

    • Search Google Scholar
    • Export Citation

    FellJohn. 2006. “Overview of Stress Testing Methodologies: From Micro to Macro.” Presentation to the Korea Financial Supervisory Commission/Financial Supervisory Service-International Monetary Fund Seminar on Macroprudential Supervision Conference on Challenges for Financial Supervisors Seoul November.

    • Search Google Scholar
    • Export Citation

    Financial Services Authority (FSA). 2009. “Stress and Scenario Testing” Policy Statement PS09/20 London December.

    Financial Services Authority (FSA). 2011. “Prudential Risk Outlook.” Financial Services Authority London March. http://www.fsa.gov.uk/pages/library/corporate/pro/index.shtml.

    • Search Google Scholar
    • Export Citation

    Financial Stability Board (FSB). 2011. “Policy Measures to Address Systemically Important Financial Institutions.” Financial Stability Board Basel Switzerland November. http://www.fsb.org/2011/11/r_111104bb/.

    • Search Google Scholar
    • Export Citation

    Financial Stability Board (FSB). 2012. “Extending the G-SIFI Framework to Domestic Systemically Important Banks.” Progress Report to the G- 20 Finance Ministers and Central Bank Governors Financial Stability Board Basel Switzerland April. http://www.fsb.org/2012/04/r_120420b/.

    • Search Google Scholar
    • Export Citation

    Financial Stability Board (FSB). 2017. “Implementation and Effects of the G20 Financial Regulatory Reforms: Third Annual Report.” Progress Report to G20 Leaders Financial Stability Board Basel Switzerland July.

    • Search Google Scholar
    • Export Citation

    FogliaAntonella. 2009. “Stress Testing Credit Risk: A Survey of Authorities’ Approaches.” International Journal of Central Banking5 (3): 945. http://www.ijcb.org/journal/ijcb09q3a1.htm.

    • Search Google Scholar
    • Export Citation

    GrayDaleF. andAndreas A.Jobst. 2011. “Modelling Systemic Financial Sector and Sovereign Risk.” Sveriges Riksbank Economic Review (2): 68106.

    • Search Google Scholar
    • Export Citation

    GrayDaleF.Andreas A.Jobst andSamuelMalone. 2010. “Quantifying Systemic Risk and Reconceptualizing the Role of Finance for Economic Growth.” Journal of Investment Management8 (2): 90110.

    • Search Google Scholar
    • Export Citation

    HardyDanielC. andChristianSchmieder2013. “Rules of Tumb for Bank Solvency Stress Testing.” IMF Working Paper 13/232 International Monetary Fund Washington DC. https://www.imf.org/en/Publications/WP/Issues/2016/12/31/Rules-of-Thumb-for-Bank-Solvency-Stress-Testing-41047.

    • Search Google Scholar
    • Export Citation

    HowardStacia. 2008. “Stress Testing with Incomplete Data: A Practical Guide.” In Proceedings of the IFC Conference on Measuring Financial Innovation and Its Impact. Basel: Bank for International Settlements. www.bis.org/ifc/publ/ifcb31.htm.

    • Search Google Scholar
    • Export Citation

    International Monetary Fund (IMF). 2010a. “Integrating Stability Assessments under the Financial Sector Assessment Program into Article IV Surveillance.” IMF Policy Paper Washington DC August. https://www.imf.org/en/Publications/Policy-Papers/Issues/2016/12/31/Integrating-Stability-Assessments-Under-the-Financial-Sector-Assessment-Program-into-PP4477.

    • Search Google Scholar
    • Export Citation

    International Monetary Fund (IMF). 2010b. “Integrating Stability Assessments under the Financial Sector Assessment Program into Article IV Surveillance— Background Material.” IMF Policy Paper Washington DC August. https://www.imf.org/en/Publications/Policy-Papers/Issues/2016/12/31/Integrating-Stability-Assessments-Under-the-Financial-Sector-Assessment-Program-into-PP4478.

    • Search Google Scholar
    • Export Citation

    International Monetary Fund (IMF). 2010c. “United States: Technical Note on Stress Testing.” IMF Country Report 10/244 Washington DC July. https://www.imf.org/en/Publications/CR/Issues/2016/12/31/United-States-Publication-of-Financial-Sector-Assessment-Program-Documentation-Technical-24101.

    • Search Google Scholar
    • Export Citation

    International Monetary Fund (IMF). 2010d. “The IMF-FSB Early Warning Exercise: Design and Methodological Toolkit.” Washington DC Augusthttps://www.imf.org/en/Publications/Policy-Papers/Issues/2016/12/31/The-IMF-FSB-Early-Warning-Exercise-Design-and-Methodological-Toolkit-PP4479.

    • Search Google Scholar
    • Export Citation

    International Monetary Fund (IMF). 2011a. “Germany: Technical Note on Stress Testing.” IMF Country Report 11/371 Washington DC December. https://www.imf.org/en/Publications/CR/Issues/2016/12/31/Germany-Technical-Note-on-Stress-Testing-25461.

    • Search Google Scholar
    • Export Citation

    International Monetary Fund (IMF). 2011b. “Sweden: Financial Sector Assessment Program Update— Technical Note on Contingent Claims Analysis Approach to Measure Risk and Stress Test the Swedish Banking Sector.” IMF Country Report 11/286 Washington DC September. https://www.imf.org/en/Publications/CR/Issues/2016/12/31/Sweden-Financial-Sector-Assessment-Program-Update-Technical-Note-on-Contingent-Claims-25241.

    • Search Google Scholar
    • Export Citation

    International Monetary Fund (IMF). 2011c. “United Kingdom: Financial System Stability Assessment.” IMF Country Report 11/222 Washington DC July. https://www.imf.org/en/Publications/CR/Issues/2016/12/31/United-Kingdom-Financial-System-Stability-Assessment-25111.

    • Search Google Scholar
    • Export Citation

    International Monetary Fund (IMF). 2011d. “United Kingdom FSAP Update: Stress Testing the Banking Sector Technical Note.” IMF Country Report 11/227 Washington DC July. https://www.imf.org/en/Publications/CR/Issues/2016/12/31/United-Kingdom-Stress-Testing-the-Banking-Sector-Technical-Note-25119.

    • Search Google Scholar
    • Export Citation

    International Monetary Fund (IMF). 2012a. “ Macro-financial Stress Testing: Principles and Practices.” IMF Policy Paper Washington DC August.

    • Search Google Scholar
    • Export Citation

    International Monetary Fund (IMF). 2012b. “Spain: Financial System Stability Assessment.” IMF Country Report. 12/137 Washington DC June. https://www.imf.org/en/Publications/CR/Issues/2016/12/31/Spain-Financial-System-Stability-Assessment-25977.

    • Search Google Scholar
    • Export Citation

    International Monetary Fund (IMF). 2013. “Mandatory Financial Stability Assessments under the Financial Sector Assessment Program: Update.” IMF Policy Paper Washington DC November. http://www.imf.org/en/publications/policy-papers/issues/2016/12/31/mandatory-financial-stability-assessments-under-the-financial-sector-assessment-program-pp4838.

    • Search Google Scholar
    • Export Citation

    International Monetary Fund (IMF). 2018. “ IMF-FSB Early Warning Exercise.” IMF Fact Sheet Washington DC. March 8. https://www.imf.org/en/About/Factsheets/Sheets/2016/08/01/16/29/IMF-FSB-Early-Warning-Exercise.

    • Search Google Scholar
    • Export Citation

    International Monetary Fund Financial Stability Board and Bank for International Settlements. 2009. “Guidance to Assess the Systemic Importance of Financial Institutions, Markets and Instruments: Initial Considerations.” Report to the G- 20 Finance Ministers and Central Bank Governors October.

    • Search Google Scholar
    • Export Citation

    International Monetary Fund (IMF) and World Bank. 2003. “Analytical Tools of the FSAP.” International Monetary Fund Washington DC February. http://www.imf.org/external/np/fsap/2003/022403a.htm.

    • Search Google Scholar
    • Export Citation

    JobstAndreasA. andDale F.Gray. 2013. “Systemic Contingent Claims Analysis— Estimating Market-Implied Systemic Solvency Risk.” IMF Working Paper13/54 International Monetary Fund Washington DC. https://www.imf.org/en/Publications/WP/Issues/2016/12/31/Systemic-Contingent-Claims-Analysis-Estimating-Market-Implied-Systemic-Risk-40356.

    • Search Google Scholar
    • Export Citation

    JobstAndreasA. andHirokoOura. Forthcoming. “Sovereign Risk— Haircut Estimation and Bank Solvency Stress Testing.” IMF Working PaperInternational Monetary FundWashington, DC.

    • Search Google Scholar
    • Export Citation

    JobstAndreasA.LiLian Ong andChristianSchmieder. 2013. “A Framework for Macroprudential Bank Solvency Stress Testing: Application to S- 25 and Other G20 Country FSAPs.” IMF Working Paper 13/68 International Monetary Fund Washington DC. https://www.imf.org/en/Publications/WP/Issues/2016/12/31/A-Framework-for-Macroprudential-Bank-Solvency-Stress-Testing-Application-to-S-25-and-Other-G-40390.

    • Search Google Scholar
    • Export Citation

    KrznarIvo andTroyMatheson. 2017. “Towards Macroprudential Stress Testing: Incorporating Macro Feedback Effects.” IMF Working Paper 17/149 International Monetary Fund Washington DC. https://www.imf.org/en/Publications/WP/Issues/2017/06/30/Towards-Macroprudential-Stress-Testing-Incorporating-Macro-Feedback-Effects-44955.

    • Search Google Scholar
    • Export Citation

    MorettiMarinaStéphanieStolz andMarkSwinburne. 2008. “Stress Testing at the IMF.” IMF Working Paper 08/206 International Monetary Fund Washington DC. https://www.imf.org/en/Publications/WP/Issues/2016/12/31/Stress-Testing-at-the-IMF-22275.

    • Search Google Scholar
    • Export Citation

    OngLi Lian editor. 2014. A Guide to IMF Stress Testing: Methods and Models. Washington, DC: International Monetary Fund.

    OngLi LianRodolfoMaino andNombuleloDuma. 2010. “Into the Great Unknown: Stress Testing with Weak Data.” IMF Working Paper 10/282 International Monetary Fund Washington DC. https://www.imf.org/en/Publications/WP/Issues/2016/12/31/Into-the-Great-Unknown-Stress-Testing-with-Weak-Data-24488.

    • Search Google Scholar
    • Export Citation

    SchmiederChristianClausPuhr andMaherHasan. 2011. “Next Generation Balance Sheet Stress Testing.” IMF Working Paper 11/83 International Monetary Fund Washington DC. https://www.imf.org/en/Publications/WP/Issues/2016/12/31/Next-Generation-Balance-Sheet-Stress-Testing-24798.

    • Search Google Scholar
    • Export Citation

    SegovianoMiguel andPabloPadilla. 2006. “Portfolio Credit Risk and Macroeconomic Shocks: Applications to Stress Testing under Data-Restricted Environments.” IMF Working Paper 06/283 International Monetary Fund Washington DC. https://www.imf.org/en/Publications/WP/Issues/2016/12/31/Portfolio-Credit-Risk-and-Macroeconomic-Shocks-Applications-to-Stress-Testing-Under-Data-20064.

    • Search Google Scholar
    • Export Citation

    VitekFrancis andTamimBayoumi. 2011. “Spillovers from the Euro Area Sovereign Debt Crisis: A Macroeconometric Model Based Analysis.” CEPR Discussion Paper 8497 Centre for Economic Policy Research London. https://ideas.repec.org/p/cpr/ceprdp/8497.html.

    • Search Google Scholar
    • Export Citation
1

In November 2008, the G20 asked the IMF and the FSB to collaborate on a regular Early Warning Exercise (EWE). The EWE examines unlikely but plausible, high-impact risks that would necessitate policy recommendations that could differ from those related to baseline projections presented in the World Economic Outlook, Global Financial Stability Report, and the Fiscal Monitor. The EWE does not attempt to predict crises. Rather, it seeks to identify (1) potential vulnerabilities that could precipitate systemic crises; and (2) suitable risk-mitigating policies, including those that would require international cooperation. It integrates macroeconomic and financial perspectives on systemic risks, drawing on a wide range of quantitative tools and broad-based consultations. (See also IMF 2018.)

2

Ultimately, authorities will have to define a specific level of risk tolerance for their financial system (Hardy and Schmieder 2013).

3

In late 2013, the IMF increased the number of systemically important financial systems, from 25 in 2010 to 29 (IMF 2013).

4

In 2012, the Basel Committee developed a set of principles on the assessment methodology and the higher loss absorbency requirement for domestic systemically important banks. The framework takes a complementary perspective to the global systemically important bank framework by focusing on the impact that the distress or failure of banks will have on the domestic economy.

5

Except when there are legally binding commitments under competition rules, for example, as agreed with the European Commission in the case of EU Member States.

6

The Basel I definition is rarely used among the major countries (except for banks in Indonesia and specific groups of banks in the United States).

7

However, the publication of these documents is voluntary for country authorities.

8

Country authorities rarely agree to the publication of the results of individual banks.

    Other Resources Citing This Publication