## wp18197

## Source details

**Canonical URL:** [wp18197](https://www.imf.org/-/media/files/publications/wp/2018/wp18197.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2018/wp18197.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2018/wp18197.pdf.json)

---

### Preface — collaborative program, aims, and key messages
- Collaborative participants: IMF-MCM and the Systemic Risk Centre (SRC, LSE) with partner institutions including Bank of Canada (BoC), Bank of England (BoE), Bank of Japan (BoJ), Banco de México, European Central Bank (ECB), Hong Kong Monetary Authority (HKMA), Reserve Bank of India (RBI), and the U.S. Office of Financial Research (OFR). An Advisory Committee and additional contributors provided technical guidance; Felipe Nierhoff (IMF-MCM) provided research assistance.
- Aims of the exercise:
  - (i) present state-of-the-art approaches on macroprudential stress testing focusing on modeling and implementation challenges, including modeling of systemic risk amplification (SRA);
  - (ii) provide a roadmap for future research and practical implementations in stress testing; and
  - (iii) discuss potential uses of macroprudential stress tests to support macroprudential policy.
- Key messages on MaPSTs:
  - MaPSTs quantify losses from systemic risk amplification via macrofinancial feedback effects, contagion across financial entities, and market dynamics that can magnify moderate exogenous shocks into substantial negative outcomes.
  - To date, MaPSTs by national authorities have mostly been diagnostic and largely independent of tool calibration, though leading central banks in the United States and the United Kingdom have begun exploring links between stress-test results and calibration of macroprudential tools.
  - Properly designed MaPSTs can support surveillance of macrofinancial vulnerabilities, calibration and use of MaPP instruments, risk management in distress, and design of recovery and resolution frameworks.
- Modeling and implementation challenges:
  - Amplification mechanisms are diverse, complex, and time-varying.
  - Relevant data are often scarce; models constrained by available data are subject to model error.
  - A single integrated model is unlikely to capture the whole range of amplification effects.

### Encompassing Frameworks (EF) — approach and Box 2 (reduced-form EF)
- Purpose of EFs:
  - Integrate diverse data and modeling frameworks with different characteristics.
  - Maximize information content of heterogeneous data sources and minimize model error.
  - Capture SRA mechanisms at varying levels of comprehensiveness given country-specific data constraints.
- Box 2 — Reduced-form SRA-loss EF (overview and steps):
  - Supplements MiPST entity loss estimates with contagion losses estimated from market-based information via the “systemic risk amplification loss” (SRA-loss) module.
  - Key steps:
    1. Start from a macrofinancial scenario; estimate entity-level risk parameters (probabilities of default, loss given default, exposures at default).
    2. Use these parameters to estimate entity losses and profitability under MiPST modules.
    3. Use parameters to estimate the system’s multivariate density and distress dependence structure; compute conditional SRA losses via an asset pricing model and add to MiPST losses.
    4. Integrate nonbank financial intermediaries (banks, insurance companies, pension funds, investment funds, hedge funds) into the multivariate analysis.
- Modeling and inputs:
  - Entity risk parameters: probabilities of default, loss given default, exposures at default.
  - Multiple model choices allowed: Merton-type approaches, Value at Risk approaches, non-arbitrage models using CDS spreads, bond spreads, etc.
- Box 2 — Developed module (features, method, limitations):
  - Distress dependence estimation method (Segoviano 2006):
    - Uses market data on stock returns and PoDs, constructs a multivariate distribution consistent with Merton (1973) and observed PoDs, chooses the distribution closest to a prior calibrated to stock returns; updates dependence as PoDs evolve.
  - Operational advantages:
    - Incorporates market-updated views of direct and indirect contagion, including nonlinear increases in high volatility periods.
    - Quantifies SRA contagion losses without detailed supervisory granular data; timely updates consistent with market perceptions.
  - Limitations:
    - Reduced-form approach minimizes model error and simplifies implementation but provides limited insight into specific amplification mechanisms.
    - Complements structural MaPSTs and other contagion approaches for transmission-channel insights.

### Stress testing evolution and objectives (MiPST vs MaPST)
- Evolution:
  - Originated in the 1990s for institution-level risk; supervisors adopted MiPSTs to prevent individual bank failure.
  - FSAP introduced stress testing for macrofinancial vulnerabilities in the late 1990s.
  - Post-2008 crisis, stress testing evolved toward macroprudential perspectives incorporating endogenous interactions and amplification.
- Objectives:
  - Microprudential: prevent failure of individual entities; examine balance sheets, capital, regulatory ratios; apply remedial measures where deficiencies found.
  - Macroprudential: assess impact of adverse scenarios on system-wide capital, profitability, and economic-supporting capacity; capture endogenous magnification via macrofinancial feedback and contagion across entities and markets.
- MaPSTs’ role in policy:
  - MaPSTs inform design and calibration of MaPP instruments, recovery and resolution frameworks, and systemic crisis management.
  - MaPSTs require coordination with microprudential supervision, monetary and fiscal policy due to complementarities and tensions among policies.

### Current and future stress testing developments and priorities
- Priorities for improvement:
  - Integration of liquidity and solvency stress tests to model feedbacks that provoke significant losses.
  - Development of dynamic balance sheet stress testing to capture behavioral responses (business strategy, portfolio composition) affecting solvency, liquidity, and profitability.
- Scope expansion:
  - Incorporate SRA mechanisms: macrofinancial feedback and contagion via direct and indirect interconnectedness.
  - Expand MaPSTs to nonbank financial entities (pension funds, insurance companies, asset managers, hedge funds, sovereign wealth funds) that can amplify or dampen systemic risk.
  - Proposed approaches include monitoring largest banks, asset classes, and counterparties (Duffie 2011) and SyRIN (Cortes and others, forthcoming).
- Data and model constraints necessitate research to improve quantification of SRA mechanisms to enhance MaPST policy usefulness.

### MaPP instruments (Box 1) — groups, calibration, and guided discretion
- Macroprudential instrument groups:
  - Broad-based buffers/capital tools: Dynamic provisioning requirements (DPRs); Countercyclical capital buffers (CCyBs); Countercyclical leverage ratio caps. Purpose: increase resilience to aggregate shocks and maintain credit supply; typically uniform.
  - Sectoral tools: sectoral capital requirements; limits on LTV, DSTI, LTI ratios; caps on foreign currency loans. Typical application: mortgages; purpose: address sector-specific lending-standard vulnerabilities.
  - Liquidity tools: differentiated reserve requirements; liquidity coverage ratio (LCR); net stable funding ratio; caps on loan-to-deposit ratio; liquidity charges on non-core funding. Purpose: contain liquidity risk build-up and prevent fire sales.
- Tools to contain contagion and resolvability:
  - Capital surcharges on systemically important institutions; systemic risk buffer under CRD IV; TLAC for resolvability.
  - Phasing example: starting to be phased in from 2016 to 2019 in equal steps of 25 percent.
- Calibration process:
  - Policymakers assess (i) economy-wide vulnerabilities from excessive leverage and credit/asset-price growth; (ii) sectoral vulnerabilities; (iii) maturity and FX mismatches; (iv) financial system structure and inter-linkages.
  - Multiple indicators and analyses used due to imperfection of any single indicator.
- Guided discretion:
  - Quantitative analysis increasingly used but judgment remains overriding because tools provide partial coverage and data limitations persist.
  - Policy actions’ influence on behavior and expectations is an area where quantitative approaches offer limited guidance.

### Calibrating the CCyB: BoE approach (Box 3)
- Role of stress tests:
  - BoE treats stress tests as one input among many for CCyB setting.
  - Annually subjects the seven largest U.K. banks (covering c.80 percent of lending to the real economy) to the annual cyclical scenario (ACS).
  - ACS severity varies with FPC’s view of system-wide risk.
- ACS-implied CCyB extraction:
  - BoE treats seven participants as a single bank; isolates U.K. cyclical impact at the system-wide CET1 ratio low point.
  - Numerator: total absolute change in CCyB-relevant capital resources under stress plus absolute change in capital requirements from start to low point.
  - Denominator: starting point CCyB-applicable RWAs base; U.K.-relevant RWAs allocation requires proxies/judgment.
  - Where stress impact exceeds the CCoB, residual informs ACS-implied U.K. CCyB; actual U.K. CCyB set in 0.25 percent increments.
- Data and allocation judgments:
  - Participants provide U.K.-element data (U.K. income/expenses, impairment charges, U.K. credit risk RWA).
  - Judgment used for partly idiosyncratic or cross-border items (dividends, available-for-sale assets, defined benefit pension impacts).
- Timeline:
  - ACS published end-Q1 each year.
  - Firms submit projections; BoE analysis over Q2 and Q3.
  - Results and decisions disclosed in Q4.
- Governance:
  - FPC uses ACS-implied CCyB as input; retains judgment and may adjust for changes in risk outlook between test setting and decisions.

### Institutional frameworks for MaPP (Box 4)
- Institutional models:
  - Model 1: Main macroprudential mandate assigned to the central bank (Board or Governor).
  - Model 2: Main mandate assigned to a dedicated committee within the central bank.
  - Model 3: Main mandate assigned to an interagency committee outside the central bank with central bank participation; can include MoF.
- Country examples by model (selected):
  - Model 1 (Central Bank): Argentina; Belgium; Brazil*; Hong Kong (SAR); Norway; Singapore; Switzerland; etc.
  - Model 2 (Internal Committee): Algeria; Malaysia*; South Africa; United Kingdom; etc.
  - Model 3 (Committee outside central bank): Canada; U.S.; France; Germany; India; Mexico; Turkey; etc.
  - Notes: “*” indicates additional coordinating councils; “(C)” or “(M)” indicates whether council is chaired by central bank or government minister.
- Accountability and communication:
  - Clear mandate plus accountability mechanisms ensure resource allocation for stress testing and public explanation of judgments.
  - Transparency tools: testimony to parliament, financial stability reports, policy statements, meeting records; some legally required.
- Benefits and trade-offs of publishing stress-test results:
  - Benefits: enables external judgment of resilience; boosts market discipline; can shore up market confidence and preemptively mobilize private capital; enhances accountability and credibility.
  - Costs: may undermine risk sharing or liquidity provision, disincentivize private information acquisition, encourage gaming; poor contingency plans can undermine confidence; tests that are insufficiently severe can create false confidence.
- Best-practice transparency measures:
  - Publish policy strategy, periodic risk assessments, ex-post assessments, meeting records and voting where appropriate, and clear narratives linking stressed losses to plausible mechanisms.

### SRA modeling, contagion, feedbacks, liquidity, and network approaches (Section IV & implementation)
- Macrofinancial feedback / second-round effects — model types and examples:
  - DSGE with financially-constrained households and oligopolistic banking; data needs include stress test template data, PD, LGD, and historical capital ratios (Darracq-Paries and others 2010).
  - MCS-GVAR for EU28 to assess bank capital shock propagation; requires macrofinancial time series across EU28, US, consolidated banking data (Gross and others 2016).
  - IDHBS household balance sheet model using HFCS survey data for borrower-based instrument sensitivity; data gaps include employment duration parameters (Gross and others 2016).
  - Financial Macroeconometric Model (FMM) — BoJ: medium-sized macro model with financial sector panel estimation (Kitamura and others 2014).
  - HKMA frameworks: empirical default-rate models with Monte Carlo simulation and VAR-mapping to banking model; empirical finding: excluding macrofinancial feedback underestimates risks (HKMA 2016; Virolainen 2004).
- Liquidity and leverage interactions — examples:
  - BoC modules: leverage module (banks deleverage by selling securities), liquidity risk module (funding withdrawal via global game).
  - U.S. Federal Reserve: CCAR/DFAST emphasis on capital adequacy; research planned to integrate funding shocks into capital stress tests.
  - BoJ: analysis of fire-sales of cross-border loans and nonlinear disposal discounts when many institutions sell simultaneously (BoJ 2016 FSR).
- Contagion and network models:
  - Interbank network models using bilateral exposures and clearing algorithms (Eisenberg and Noe 2001); data-intensive and calibration-challenging (BoC, ECB, Banco de México, RBI, HKMA).
  - Banco de México constructs multi-layer network models (derivatives; loans/deposits; FX; securities); found loans and deposits accounted for roughly 10% of expected systemic risk in one application.
  - Indirect contagion via fire sales and price-mediated mechanisms: price-impact calibration requires granular asset-holding data and market liquidity/time-scale assumptions; agent-based and network-based fire-sale models explored by ECB, academic authors, and central banks.
- Data needs and recurrent modeling challenges:
  - Granular bank balance-sheet data (assets, liabilities, maturities, encumbrance).
  - Bilateral interbank exposure data.
  - Historic market data on asset prices and volumes.
  - Household survey data for micro-macro models.
  - Macrofinancial time series for GVAR models.
  - Recurrent challenges: steady-state calibration, market liquidity and price-impact calibration, missing employment-duration parameters, reconstructing bilateral interconnections when data are missing, modeling behavioral responses and endogenous network formation, integrating capital and liquidity stress testing, and endogenizing fire-sale effects.

### International frameworks, standards, and practice notes
- NSFR final standard published in October 2014 and in observation period until 2018.
- Examples of country practices on stable funding and liquidity charges:
  - Korea: liquidity charge on daily average short-term foreign currency liabilities; rate varies from 2 to 20 basis points; adjustable based on non-core funding indicator.
  - New Zealand: core-funding ratio introduced in 2010 with minimum 65 percent of total loans and advances; increased to 70 and 75 percent in July 2011 and January 2013.
- Structural tools and identification:
  - FSB publishes annual G-SIB list using five-dimension indicators; national D-SIB identification follows similar guidelines excluding global-scope dimension.
  - Jurisdictions applying G-SIB/D-SIB surcharges include Australia; Canada; EU countries; Israel; Japan; Singapore; Switzerland; UK; US; and others.

### Selected implementation status and research agenda
- SRA and contagion modules are in mixed stages: “Implemented”, “In Development”, or “Work in progress” across authorities (BoE, ECB, BoJ, HKMA, RBI, Banco de México, BoC, U.S. Federal Reserve, IMF staff).
- Recommended structured approach:
  - Employ an EF that combines multiple models (structural bottom-up and reduced-form top-down), alternative data sets (market and supervisory), and mixed estimation and calibration methods to capture nonlinear amplification and minimize model error.
  - Continue progress on modeling direct interbank exposures, and prioritize research on fire sales, herding, information asymmetry, and integrating the nonbank financial sector.

*Source: wp18197 — Monetary and Capital Markets Department (MCM) and Systemic Risk Centre (SRC).*

### Preface.................................................................................................................

### Preface

### Collaborative program and participants
- The Monetary and Capital Markets Department (MCM) of the International Monetary Fund (IMF) and the Systemic Risk Centre (SRC) based at the London School of Economics (LSE) launched a collaborative research program into macroprudential stress testing.
- Partner institutions included: Bank of Canada (BoC), Bank of England (BoE), Bank of Japan (BoJ), Banco de México, European Central Bank (ECB), Hong Kong Monetary Authority (HKMA), Reserve Bank of India (RBI), and the U.S. Office of Financial Research (OFR).
- An Advisory Committee provided technical guidance and comments; contributors named include Alexander Brazier (BoE), Ian Christensen (BoC), Jill Cetina (OFR), Rama Cont (Imperial College London), Casper de Vries (Erasmus University of Rotterdam), Alan Elizondo (Banco de México), Itay Goldstein (Wharton School), Charles Goodhart (LSE), Cho Hoi Hui (HKMA), Malcolm Knight (SRC), Hitoshi Mio (BoJ), Deepak Mohanty (RBI), and Sergio Nicoletti-Altimari (ECB).
- Additional inputs came from Zineddine Alla (Sciences Po, Paris), José Berrospide (FRB), Jérôme Henry (ECB), Liam Girvan (BoE), Alfred Lehar (University of Calgary), Paul Nahai-Williamson (BoE), Amar Radia (BoE), and Virginie Traclet (BoC).
- Felipe Nierhoff (IMF-MCM) provided research assistance.

### Aims of the exercise
- The aim is threefold:
  - (i) present state-of-the-art approaches on macroprudential stress testing focusing on modeling and implementation challenges, including the modeling of systemic risk amplification (SRA);
  - (ii) provide a roadmap for future research and practical implementations in stress testing; and
  - (iii) discuss the potential uses of macroprudential stress tests to support macroprudential policy.

### Key messages on macroprudential stress tests (MaPSTs)
- MaPSTs are beginning to play an increasingly major role in financial sector policymaking.
- The global financial crisis in 2008 illustrated that relatively small initial losses can be magnified to systemic dimensions.
- MaPSTs attempt to quantify losses from systemic risk amplification (SRA) mechanisms, capturing losses endogenously amplified through:
  - macrofinancial feedback effects,
  - contagion across financial entities, and
  - market dynamics that can magnify moderate exogenous shocks into substantial negative financial outcomes with significant welfare losses.
- To date, MaPSTs by national authorities have mostly been diagnostic tools, used to sense sources of risk and vulnerabilities while remaining independent of the calibration of macroprudential tools.
- Leading central banks in the United States and the United Kingdom have begun to explore linking calibration of policy tools to stress testing results, elevating debate on how MaPSTs should inform macroprudential policy.
- Properly designed MaPSTs can provide quantitative, forward-looking assessments of resilience for individual banks and the financial system as a whole, supporting:
  - surveillance of macrofinancial vulnerabilities,
  - the use of macroprudential policy (MaPP) instruments,
  - risk management and decision-making in periods of financial distress, and
  - the design of recovery and resolution frameworks.
- MaPSTs also benefit financial institutions (banks, pension funds, insurance companies) concerned about tail risk.

### Challenges in modeling and implementation
- Modeling losses from SRA is challenging because:
  - amplification mechanisms are diverse and complex and can vary in structure and magnitude over time;
  - relevant data are usually scarce;
  - models constrained by available data are often subject to model error.
- Given these challenges, a single integrated model is unlikely to capture the whole range of possible amplification effects.

### Proposed approach: Encompassing Frameworks (EF)
- The paper proposes the development of “encompassing frameworks” (EF) to:
  - integrate a diverse collection of data and modeling frameworks with different characteristics;
  - maximize the information content of heterogeneous data sources; and
  - minimize potential model error.
- The paper presents various EFs being developed by authorities around the world that capture SRA mechanisms at different levels of comprehensiveness given country-specific data constraints.

### Purpose and structure of the paper
- Objective: identify major trends in theoretical and empirical modeling and provide guidance to policymakers and academics for implementation.
- The paper discusses how MaPSTs can be used for the calibration of MaPP instruments and how stress testing fits into the overall macroprudential agenda.
- The paper proceeds with:
  - Section II: interaction of stress testing and financial policy, and key principles for robust MaPSTs;
  - Section III: taxonomy to analyze systemic risk and empirical methods to measure systemic risk, with implementation challenges.

*Preface, wp18197 — Monetary and Capital Markets Department (MCM) and Systemic Risk Centre (SRC).*

### Section IV presents MaPST frameworks currently being developed. Section V examines how

### wp18197 - Section IV presents MaPST frameworks currently being developed. Section V examines how

### Stress testing: Evolution
- Stress testing emerged in the 1990s as a tool employed by financial institutions to assess their exposure to large risks.
- Supervisors rapidly adopted stress tests for microprudential regulation aimed at strictly controlling the risk of bank failure; premise: banks play a central role in assuring efficient functioning of the economy and the failure of a bank poses a significant threat to economic growth.
- Microprudential stress tests (MiPSTs) have been used to assess the risk of failure of a single institution under the dictum enshrined in the Basel capital standard: “financial stability is ensured as long as each and every institution is sound.”
- Experience from the Asian financial crisis and subsequent work led to doubt about the sufficiency of focusing exclusively on individual institutions and motivated the development of the Financial Sector Assessment Program (FSAP) in the late 1990s.
- The FSAP introduced stress testing into the policy toolkit and provided local policymakers with quantitative measures of macrofinancial vulnerabilities that complement broader financial sector assessments.
- Stress tests initially focused on resiliency to exogenous shocks for individual institutions; after the global financial crisis their use and prominence increased, evolving toward a macro perspective that incorporates interactions among financial institutions and mechanisms that can endogenously amplify shocks.

### Stress testing: Micro- and macroprudential objectives
- Microprudential objectives:
  - Aim to prevent failure of individual financial entities.
  - MiPSTs examine banks’ balance sheets, capital and regulatory ratios, and risk management practices.
  - Where deficiencies are identified, remedial efforts (including additional safety buffers in the form of bank capital) may be warranted.
- Macroprudential objectives:
  - MaPSTs assess the impact of an adverse scenario on the financial system’s capital, profitability, and ability to support economic activity as a whole.
  - By subjecting multiple institutions to the same scenario, MaPSTs assess the system after losses from systemic risk amplification (SRA) have materialized.
  - MaPSTs capture endogenous magnification of losses through macrofinancial feedback effects and contagion across entities and markets, and provide qualitative information on “reactions of the system” in periods of stress.

### Macroprudential stress testing and financial sector policies
- MaPSTs are an important element of the MaPP toolkit; they provide quantitative and qualitative information useful for calibrating MaPP instruments whose objective is to lower or contain systemic risk.
- The authoritative 2009 definition of systemic risk by BIS, FSB, and IMF: “the disruption to the flow of financial services that is caused by an impairment of all or parts of the financial system; and has the potential to have serious negative consequences for the real economy.”
- MaPP instruments attempt to curtail systemic risk amplification mechanisms and improve system resilience to shocks.
- Country experience shows increasing reliance on quantitative analysis for calibration of MaPP instruments, with supervisory judgment retaining an overriding role; appropriately designed MaPSTs can support calibration processes.
- MaPSTs inform design of recovery and resolution frameworks and systemic crisis management:
  - Microprudential policies are bottom up; macroprudential policies are top down.
  - Failure of a small institution can be tolerated if effects can be contained; failure of a systemically important financial institution (SIFI) requires special resolution regimes because ordinary bankruptcy procedures do not work.
  - MaPSTs can help address thresholds and scenarios where recovery is socially optimal relative to resolution.
- MaPSTs require simultaneous coordination with other policy tools:
  - Include microprudential supervision and regulation, monetary and fiscal policy.
  - Interactions between policies can produce complementarities and tensions that need resolution.
  - The task of attaining financial stability is too challenging to be left to one policy area alone.

### Current and future stress testing developments
- Over the past 15 years stress tests have moved from isolated risk management tools to core policy toolkit components.
- Priorities identified for improvement of stress tests include:
  - Integration of liquidity and solvency stress tests: integrating liquidity and solvency allows policymakers to model feedback from risks that can provoke significant losses in stressed situations.
  - Further development of dynamic balance sheet stress testing: models need to capture behavioral responses of banks, such as changes in business strategies and portfolio compositions directed at coping with external shocks; improved modeling of bank responses should allow more realistic assessment of impacts on solvency, liquidity, and profitability.
- Note: while these developments are useful for MiPSTs, they are not sufficient from a macroprudential point of view; however, if applied across all entities they could have implications for systemic loss amplification and thus be relevant for MaPSTs.

### Successful development of macroprudential stress testing: key elements and research agenda
- Elements necessary for successful development of MaPSTs:
  - Further incorporation of SRA mechanisms: conceptual taxonomy includes macrofinancial feedback effects and contagion from direct and indirect interconnectedness across financial entities and markets.
  - Expanding scope of MaPSTs to nonbank financial entities: MaPP aims to contain risks across the financial system as a whole; banks are typically primary credit providers, but systemic risk can build up outside the banking system.
    - Nonbank entities include pension funds, insurance companies, asset managers, hedge funds, and sovereign wealth funds; these entities can act as amplifiers or dampeners depending on cycle states.
    - Proposed approaches: Duffie (2011) suggests monitoring largest banks, largest asset classes, and largest counterparties; Cortes and others (forthcoming) propose the Systemic Risk and Interconnectedness framework (SyRIN) to measure systemic risk accounting for interconnectedness across banks and nonbanks.
- Data and model constraints are significant challenges that warrant a stimulating research and policy agenda aimed at improving understanding and quantification of SRA mechanisms to enhance the policy usefulness of MaPSTs.

*Source: wp18197 - Section IV presents MaPST frameworks currently being developed. Section V examines how*

### Box 1. Macroprudential Instruments

### Box 1. Macroprudential Instruments

### Macroprudential instrument groups
- Broad-based buffers/capital tools:
  - Dynamic provisioning requirements (DPRs)
  - Countercyclical capital buffers (CCyBs)
  - Countercyclical leverage ratio caps
  - Purpose: increase resilience of institutions to aggregate shocks and maintain credit supply through adverse conditions
  - Typically uniformly applied to all exposures
- Sectoral tools:
  - Examples: sectoral capital requirements; limits on loan-to-value (LTV), debt-service-to-income (DSTI), loan-to income (LTI) ratios; caps on the share of foreign currency loans
  - Typical application: mortgages (residential and commercial), but can also cover consumer and some corporate credit
  - Purpose: address vulnerabilities from deterioration in lending standards for loans originating from specific sectors; maintain resilience of lenders and/or borrowers
- Liquidity tools:
  - Examples: differentiated reserve requirements; liquidity coverage ratio (LCR, potentially calibrated by currency); net stable funding ratio; caps on the loan-to-deposit ratio; price-based tools (such as liquidity charges on non-core funding)
  - Purpose: contain build-up of liquidity risks associated with credit booms and prevent fire sales triggered by funding market disruptions

### Tools to contain systemic risk from contagion and resolvability
- Authorities may impose capital surcharges on institutions deemed systemically important, independent of specific circumstances
- In Europe, CRD IV allows introduction of a systemic risk buffer to complement surcharges (e.g., G-SII and O-SII buffers)
- Phasing and related measures:
  - These tools are starting to be phased in from 2016 to 2019 in equal steps of 25 percent
  - Measures to increase resolvability include total loss-absorbing capacity (TLAC)

### Calibration of Macroprudential Instruments
- Operationalizing MaPP involves:
  - Assessing systemic risks
  - Selecting and calibrating tools to target well-identified risks
- Policymakers assess, using a range of data sources:
  - (i) economy-wide vulnerabilities from excessive leverage and growth in total credit or asset prices
  - (ii) sectoral vulnerabilities arising from growing credit exposure in specific sectors
  - (iii) vulnerabilities from a build-up of maturity and foreign currency mismatches
  - (iv) the structure of the financial system (for example, concentration) and the level of inter-linkages within and across key classes of intermediaries
- Because any single indicator is often imperfect, multiple indicators and various analyses are used to assess the extent of each type of vulnerability and to choose appropriate tools
- Different policy tools and descriptions of current calibration procedures are laid out in Table 1 (as referenced in the source)

### Practical approach to calibration: guided discretion
- Country experience:
  - Macroprudential policymaking increasingly relies on quantitative analysis
  - Judgment retains an overriding role because of difficulties and data limitations in assessing systemic risk
- Rationale for discretion:
  - Existing tools provide only partial coverage of potential risks
  - Tentative signals on likelihood of systemic risk events provide limited information about the need for macroprudential actions
  - The influence of policy actions on market participants’ behavior and expectations is an area where quantitative approaches currently offer limited guidance (CGFS 2016)

*Source: wp18197 - Box 1. Macroprudential Instruments*

### Box 2. Quantifying SRA Losses Based on a Reduced-Form Approach: An EF

### Box 2. Quantifying SRA Losses Based on a Reduced-Form Approach: An EF

### Overview
- This EF supplements loss estimates of individual entities from MiPST modules with contagion losses from SRA mechanisms that are based on publicly available and market-based information and estimated by the “systemic risk amplification loss” (SRA-loss) module put forth in Alla and others (2017).
- The quantification of SRA losses is based on the valuation of bank assets and the realization of specific events such as the failure of a financial entity, or a scenario in which a group of entities falls into distress.
- These SRA losses are conditional on stressed macrofinancial scenarios and the realization of specific events that are typically marked by a single financial entity, or a group of financial entities falling into distress.
- A figure in the source "summarizes the proposed EF."

### Encompassing Framework for Macroprudential Stress Tests — Key Steps
- 1. The framework takes a macrofinancial scenario as a starting point. Given the assumptions regarding the scenario, risk parameters (probabilities of default, loss given default, and exposures at default for different assets and entities) are individually estimated for each of the financial institutions analyzed (see left hand side of the figure).
- 2. These parameters are used as inputs to estimate losses and profitability for each entity under the MiPST framework and scenario.
- 3. These parameters are also employed as inputs to estimate the system’s multivariate density (and distress dependence structure) from which SRA losses are quantified (see right hand side of the figure). SRA losses are conditional losses, derived from the multivariate distribution representing the financial system and using an asset pricing model to compute the expected valuation of each firm’s assets under any stress event of interest. Obvious candidates for events that should be checked involve entities failing the MiPST (for example, the entities whose capital adequacy (CA) falls below a predetermined “hurdle rate” after the entity goes through the MiPST). Moreover, the impact of the default of any entity on the system (SRA losses) can also be analyzed in this framework. The approach, therefore, is stochastic since it allows analysts to estimate losses conditional on being on any of the tails of the multivariate density and the probabilities of such events. SRA losses can be added to the losses of individual entities estimated in MiPSTs.
- 4. The multivariate framework also permits an easy integration of nonbank financial intermediaries into the analysis of systemic risk; thus, interactions between banks, insurance companies, pension funds, investment funds, and hedge funds can be considered for the quantification of losses due to systemic risk amplification mechanisms (Cortes and others 2017).

### Modeling and Inputs
- Risk parameters estimated for each financial institution: probabilities of default, loss given default, and exposures at default.
- Inputs are used both for entity-level MiPST loss and profitability estimation and for estimating the system’s multivariate density and distress dependence structure to quantify SRA losses.
- SRA-loss computation uses an asset pricing model to compute expected valuation of each firm’s assets under stress events of interest.
- Conditional events to analyze include entities failing the MiPST (e.g., CA falling below a predetermined “hurdle rate”) and defaults of any entity to assess systemwide impact.

### Coverage and Extensions
- The framework accommodates multiple model choices to estimate parameters, including Merton-type approaches, Value at Risk approaches, non-arbitrage models that rely on credit default swap spreads, bond spreads, and so on.
- The multivariate approach facilitates integrating nonbank financial intermediaries—banks, insurance companies, pension funds, investment funds, and hedge funds—when quantifying losses from systemic risk amplification mechanisms.

*Developed by the IMF*

### Box 2. Quantifying SRA Losses Based on a Reduced-form Approach: An EF Developed

### Box 2. Quantifying SRA Losses Based on a Reduced-form Approach: An EF Developed

### Purpose and capabilities of the SRA-loss module
- Quantify losses due to amplification mechanisms in the bank and nonbank sectors.
- Assess whether specific entities would be able to “survive” (that is, if their CA would be above the hurdle rate) the additional losses brought by SRA.
- Calculate the contribution to SRA losses from each “connecting” entity in the system, and incorporate the decomposition of contributions into the likelihood of the event and the intensity (amount) of induced SRA losses.

### Method for estimating distress dependence across entities
- Need: estimate a system’s multivariate density to characterize the distress dependence structure across entities so that contagion losses can be quantified when one entity suffers a shock.
- Proposed statistical method (Segoviano 2006):
  - Uses observed market data on stock returns and probabilities of distress (PoDs) of individual entities.
  - Constructs a multivariate distribution of asset returns that is consistent with the default model proposed by Merton (1973) and with the observed PoDs.
  - Chooses the multivariate distribution that is closest to a prior distribution calibrated to match stock returns data.
  - The constructed multivariate density allows inference of the (unobserved) distress dependence structure — the “system’s interconnectedness structure.”
  - As PoDs change over time, the inferred distress dependence structure is updated, allowing estimation consistent with market perceptions as macrofinancial conditions evolve.

### Features and advantages of the reduced-form, market-data-based approach
- Incorporates market-updated views of risk spillovers from both direct contagion (contractual obligations) and indirect contagion (market price channels, asset fire sales, information asymmetries), including nonlinear increases in periods of high volatility.
- Key operational advantages:
  - Incorporate updates in a system’s distress dependence structure based on market perceptions of direct and indirect contagion in a timely manner.
  - Quantify SRA contagion losses without, ex-ante, needing to assume structured agent interactions and behaviors that can change in unknown manners in periods of distress.
  - Estimate SRA contagion losses from readily available market information without the need for highly detailed and granular supervisory information that is not available in many countries, nor to institutions like the IMF.
  - Compute complementary measures of systemic risk that provide supportive information to various systemic risk policy objectives (Segoviano and Espinoza 2017).
- Outcome: policymakers can obtain crisis-consistent estimates of contagion losses and complementary systemic-risk measures that are interpretable, incorporate market-perceived structural changes of agent interactions, and can change quickly and nonlinearly in periods of distress.

### Limitations and complementarities
- The EF is not complete and has intrinsic limitations common to quantitative methodologies.
- The nonstructural (reduced-form) approach:
  - Minimizes model error and simplifies implementation.
  - Does not provide much insight into specific amplification mechanisms because estimates are generated from a reduced-form statistical model.
- Despite limitations, quantifying contagion losses is necessary to assess the potential magnitude of SRA losses — crucial knowledge for policymakers.
- Insights into transmission channels can be complemented by MaPSTs and other contagion approaches discussed in the broader text.

*Source: IMF — Box 2 from "Quantifying SRA Losses Based on a Reduced-form Approach: An EF Developed" (concluded).*

### Box 3. Calibrating the CCyB Rate: The BoE Approach

### Box 3. Calibrating the CCyB Rate: The BoE Approach

### Use of stress tests to inform the CCyB
- The Bank of England (BoE) treats stress tests as one of many inputs into the setting of the countercyclical capital buffer (CCyB).
- Each year the BoE subjects the seven largest U.K. banks (covering c.80 percent of lending to the real economy) to the same macroeconomic stress scenario, the annual cyclical scenario (ACS).
- The severity of the ACS varies with the FPC’s view of system-wide risk: as risks build, severity increases; as risks crystallize or abate, severity decreases.
- The BoE extracts an ACS-implied U.K. CCyB rate from the stress test results by treating the seven participants as a single bank and isolating the U.K. cyclical impact at the system-wide CET1 ratio low point.

### Data inputs and allocation judgments
- Authorities receive data from participating institutions on the U.K. element of CCyB-relevant impact items, for example:
  - firms’ estimates for U.K. income and expenses;
  - U.K. impairment charges;
  - U.K. credit risk RWA.
- For impact items that are partly idiosyncratic or difficult to allocate between U.K. and non-U.K., the BoE uses judgment to estimate U.K. impact. Examples include:
  - dividend payments;
  - available-for-sale assets;
  - impact relating to defined benefit pension schemes.
- The BoE’s dataset is sufficiently granular to enable a prudent and consistent U.K. allocation for most items.

### Calculation method (ACS-implied CCyB rate)
- The BoE applies an equation to calculate an ACS-implied CCyB rate using stress test outcomes and U.K.-relevant inputs. Key components and definitions provided:
  - Absolute change in CCyB-relevant capital resources between start and low point of the stress.
  - Absolute change in CCyB-relevant capital requirement (hurdle rate multiplied by absolute change in CCyB-relevant RWA) between start and low point of the stress.
  - Starting point CCyB-applicable RWA base.
  - Where:
    - α = the stress test hurdle rate (see section 1.4 of the October 2015 approach document for more details on the hurdle rate).
    - b = bank participating in BoE stress test.
    - RWA = UK-relevant RWAs.
- Interpretative bullets from the source:
  - The numerator represents the total absolute change in capital resources under stress plus the absolute change in capital requirements, measured from the starting point to the stress CET1 ratio low point.
  - The impact of the stress (numerator) is converted to an RWA-based measure by dividing by the starting point CCyB-applicable RWAs base as a denominator. Determining the U.K.-relevant portion of RWAs is not straightforward; in practice, a proxy (such as RWAs related to domestic credit exposures) might be used.
  - Where the stress impact exceeds the CCoB, the residual helps to inform the setting of the ACS-implied U.K. CCyB rate. The actual U.K. CCyB rate itself is set in 0.25 percent increments.

### Process timing and decision-making
- Timeline:
  - The ACS is published end-Q1 each year.
  - Firms submit projections and the BoE undertakes its analysis over Q2 and Q3.
  - Results and decisions are disclosed in Q4.
- The BoE emphasizes this calculation helps to inform FPC discussion; the FPC’s ultimate judgment can take into account a variety of other factors and indicators, including any changes to the risk outlook that could occur after the test was set.
- Between setting the ACS and taking decisions on results, risks may crystallize or abate; where the ACS results are no longer representative, the FPC and PRA use judgment to determine the appropriate and coordinated regulatory response.

### Governance and coordination considerations
- The BoE approach highlights the role of judgment and coordination between macroprudential and microprudential authorities:
  - The FPC may decide to make up any shortfall where U.K. cyclical stress is greater than the end-state CCoB by setting a positive U.K. CCyB rate (or changing the rate where already positive).
  - Stress test outputs feed into broader governance and policy-setting; they are one input among others and do not mechanically determine the CCyB.
- The approach underscores the need for granularity in data, robust allocation rules for cross-border items, and clear institutional coordination in interpreting stress-test-informed metrics.

*Source: Box 3. Calibrating the CCyB Rate: The BoE Approach (wp18197)*

### Box 4. MaPP Institutional Framework Models

### Box 4. MaPP Institutional Framework Models

### Institutional models
- Model 1. The main macroprudential mandate is assigned to the central bank, with its Board or Governor making macroprudential decisions. This model is the prevalent choice where the central bank already concentrates the relevant regulatory and supervisory powers. Systemic risk assessment can bring together macro- and microprudential expertise and fully exploit complementarities between top-down and bottom-up risk analyses, for example, in the approach to stress tests.
- Model 2. The main macroprudential mandate is assigned to a dedicated committee within the central bank. This setup creates dedicated objectives and decision-making structures for monetary and MaPP, and can help counter the potential risk for multiple mandates affecting decision-making within the central bank “(IMF 2013)”. Unlike Model 1, it can foster an open discussion of systemic risks through participation of separate supervisory agencies and external experts in the committee.
- Model 3. The main macroprudential mandate is assigned to an interagency committee outside the central bank, in order to coordinate policy action and facilitate information sharing and discussion of system-wide risk, with the central bank participating on the committee (as in France, Germany, Mexico, and the United States). Identification and mitigation of systemic risk is a multi-agency effort. This model can accommodate a stronger role of the MoF. Participation of the MoF can be useful to create political legitimacy and enable decision-makers to consider policy choices in other fields, for example, when cooperation of the fiscal authority is needed to mitigate systemic risk.

### Country examples by model
- Central Bank Model (Board or Governor) — Model 1
  - Argentina, Belgium, Brazil*, Cyprus, Czech Republic, Estonia*, Hong Kong (SAR), Hungary, Indonesia, Ireland, Israel, Italy*, Lebanon, Lithuania, Netherlands*, New Zealand, Norway,2/ Portugal*, Russia, Singapore, Slovakia, and Switzerland.2/
- Separate Committee Model (Internal Committee) — Model 2
  - Algeria, Malaysia*, Morocco, Saudi Arabia, South Africa, Thailand, and the United Kingdom
- Committee outside the central bank — Model 3
  - Austria (M), Canada and the U.S. (M), Chile (M), Denmark (C), France (M), Germany (M), Iceland (M), India (M), Korea (M), Malta (C), Mexico (M), Poland (C), Romania (C), and Turkey (M).

- Notes:
  - 1/ Countries with an “*” have an additional council including other supervisors (for example, insurance supervisory authorities and financial market authorities) that play a coordinating role.
  - 2/ The central bank mandate is confined to the CCyB.
  - 3/ “(C)” or “(M)” indicates whether the council is chaired by the central bank or by a government minister (usually the minister of finance), respectively.
- Source: IMF-BIS-FSB (2016).

### Accountability and communication
- With responsibility clearly assigned, accountability requirements become the check and balance. A strong accountability mechanism ensures that the authority takes seriously responsibility for stress test results and impacts on policy. The mechanism should be geared toward obliging the authority to allocate sufficient resources for the conduct of risk analysis and stress tests.
- Guided discretion should be combined with a proper degree of transparency. Open communication promotes public understanding of the factors affecting systemic risk and the need for a specific tool for promoting financial stability.
- A range of communication tools is now frequently employed, such as regular testimony to the national parliament, financial stability reports, disclosure of policy statements, and meeting records. In some cases, these tools have been required by law as accountability measures (France, Germany, and the United Kingdom).
- There has been a trend toward greater transparency among central banks and regulatory authorities over the past few decades. Public communication of stress test results has gained momentum since the onset of the global financial crisis (IMF 2012).

### Benefits of publishing stress test results
- Publication of stress test results can allow external observers to judge the resilience of financial institutions and the financial system to various risks.
- Communication can make policy more effective. Disclosure can boost market discipline by providing market participants with information about risks and resilience.
- Promoting market discipline through an effective communication regime can act as an important complement to direct policy actions undertaken by authorities.
- The authority can use disclosure to shore up market confidence, promote realistic risk pricing, and (preemptively) raise additional capital from private sources if necessary, thereby reducing the probability of sudden reversals of market sentiment.
- Even when the results are weak, public communication can have a positive impact if it is accompanied by credible contingency plans for financial institutions that reflect the authorities’ commitment to financial stability.
- Transparency can boost the credibility of MaPSTs, increasing public confidence in financial stability and the stress testing authority.
- Communication can improve policymakers’ decisions by providing a credible commitment to explain stress-testing judgments publicly.
- Publication of stress test results can enhance public accountability.40

### Trade-offs and potential costs of disclosure
- Disclosure may undermine risk sharing (Allen and Gale (2000)); e.g., exposure to losses revealed by stress testing may undermine the normal provision of liquidity in interbank markets.
- Disclosure may disincentivize private efforts toward information acquisition and may undermine due diligence.
- Disclosure of stress test results may entice financial institutions to make portfolio choices to “game” the tests.
- If contingency plans or credible backstops are not in place, disclosure can undermine market confidence.
- If stress tests are not severe enough, they can provide a false sense of confidence in the resilience of the financial system.
- Information asymmetries may persist after public disclosure; announcing stress test results does not automatically place all participants on an equal level of common knowledge.41
- Preconditions for beneficial disclosure include: stress testing should target all the relevant risks and SRA mechanisms; it should assume several shocks; produce a candid assessment; and be accompanied by a convincing framework for crisis resolution and follow-up action, including government support, if needed.

### Dimensions of transparency in stress testing
- Scenarios:
  - Most authorities disclose a substantial amount of quantitative information about their chosen stress test scenario so the public can gauge severity by comparing to previous recessions or financial crises.
  - Authorities do not typically publish paths for all variables required by participants, possibly to preserve banks’ incentives to develop their own modelling capabilities.
- Methodologies and models:
  - There is typically less transparency. Disclosure of specific stress testing models may lead to a “model monoculture” (Bernanke 2013).
  - Publishing model details can signal best practice and allow external scrutiny and constructive feedback.
- Results:
  - Most authorities disclose quantitative stress test results, including for individual institutions, but disclosure is limited; those that disclose individual bank results tend to disclose only headline metrics (for example, capital ratios or shortfalls).
- Policy actions:
  - Authorities vary in how they use stress tests to inform policy actions. Disclosure of hurdle rates or qualitative information obtained and actions taken is common, but disclosing too much may reveal commercially sensitive information.

### Best practices to reinforce transparency benefits
- Publication of a policy strategy:
  - Encourage development and announcement of a preferred policy strategy based on assessments of systemic risks and deployment of specific macroprudential tools; spell out conditions under which these tools would be deployed.
- Periodic reports of risk assessments and policy actions:
  - Publish periodic, comprehensive risk assessments including macroprudential stress testing and complementary analyses.
  - Publish ex-post assessments of measures taken to gauge success and build policy credibility; these may form part of a (semi-annual) Financial Stability Report.
- Record of meetings:
  - Publication of meeting records should establish transparency on systemic risk issues discussed and clarity regarding votes cast by members on policy decisions.
  - Records should identify key decisions; publishing voting records increases accountability and requires those opposing actions to justify their positions.
- Communication of MaPST results and methodologies must be clear because MaPP aims at reducing probability of crises (low-probability events); stressed scenarios need to be credible and calculated losses should be accompanied by a convincing narrative about how losses could arise in practice.

### Conclusion (selected findings and policy implications)
- Stress testing is widely used to assess vulnerabilities of major banks and other financial institutions to a variety of risks and to provide system-wide assessments.
- The stress testing methodology has microprudential origins and needs to account for endogenous, system-wide risk amplification mechanisms.
- Recent theoretical developments capture endogenous amplification via structural (behavioural) models and reduced-form (information-structure) approaches; applications to fire sales, leverage, liquidity, and herding illustrate amplification channels.
- Empirical implementation faces data limitations (granularity, supervisory access, backward-looking accounting data, off-balance sheet items) and modelling challenges (no single workhorse model; need to capture nonlinear amplification).
- A structured yet flexible approach — an "EF" that employs multiple models (structural bottom-up and reduced-form top-down), alternative data sets (market and supervisory), and a mix of estimation and calibration methods — is recommended.
- Progress has been made on modeling direct interbank exposures; systems with current capital and liquidity standards appear capable of absorbing extreme stress in much of this work.
- There is no settled view on incorporating amplification from fire sales, herding, and information asymmetry; integrating the nonbank financial sector remains at very early stages.
- Governance to support macroprudential stress testing requires: a clear mandate to a recognized institution, accountability to broader policymaking authorities, and a clear policy on transparency weighing costs and benefits for institutions versus the system.

*Source: wp18197 - Box 4. MaPP Institutional Framework Models (IMF).*

### REFERENCES

### wp18197 - REFERENCES

### Key themes and topics covered
- Systemic risk, financial contagion, and network models (e.g., "Systemic Risk and Stability in Financial Networks"; "Financial Contagion"; "Contagion in Financial Networks").
- Liquidity risk, funding liquidity, market liquidity, and fire-sale dynamics (e.g., "Liquidity and financial contagion"; "Market liquidity and funding liquidity"; "Fire Sales in Finance and Macroeconomics").
- Procyclicality, leverage, value-at-risk, and endogenous risk (e.g., "Liquidity and Leverage"; "Procyclical leverage and value-at-risk"; "Endogenous Risk").
- Stress testing methodologies and macroprudential stress tests, including integration with macro models and systemic loss quantification (e.g., "Macrofinancial Stress Testing—Principles and Practices"; "Macro stress testing at the Bank of Japan"; "Macroprudential Stress Tests: A Reduced-Form Approach to Quantifying Systemic Risk Losses").
- Macroprudential policy design, calibration, and instruments (e.g., "Elements of Effective Macroprudential Policies"; "The ESRB Handbook on Operationalising Macroprudential Policy in the Banking Sector"; "Guidance for national authorities operating the countercyclical capital buffer").
- Bank resolution, bail-in, and loss-absorbing capacity (e.g., "A Critical Evaluation of Bail-in as a Bank Recapitalization Mechanism"; "Total Loss-Absorbing Capacity (TLAC) Principles and Term Sheet").
- Market microstructure, risk modeling, model risk, and forecasting methods (e.g., "Model risk of risk models"; "Evaluating Density Forecasts"; Markov-switching and time-varying parameter VAR methodologies).
- Interaction between monetary policy and macroprudential measures and the macroeconomic propagation of financial stress.

### Prominent authors, institutions, and publication outlets
- Repeatedly cited authors and contributors include Adrian, H.S. Shin, Battiston, Glasserman, Segoviano, Goodhart, Brunnermeier, Pedersen, Bernanke, Crockett, and others.
- Key institutional sources: Bank for International Settlements (Basel Committee on Banking Supervision), Bank of England, Financial Stability Board, International Monetary Fund, European Central Bank, Federal Reserve Board, Bank of Japan, Hong Kong Monetary Authority, Norges Bank.
- Common publication outlets: Journal of Financial Economics; Review of Financial Studies; Journal of Monetary Economics; Journal of Economic Literature; IMF Working Papers; ECB Working Paper Series; Bank of England Working Papers; FSB thematic reviews and policy papers.

### Table 1 — Macroprudential Policy Tools: selected entries and calibrations (as presented)
- Broad-based tools
  - Countercyclical capital buffer
    - Calibration note: "The credit-to-GDP gap serves as a common core indicator to guide decisions on CCyB rates, following the guidance proposed by the BCBS. In addition to the gap, policymakers draw on a number of other indicators, such as credit growth rate and asset price growth and deviations from long-term trends, and the private sector debt burden."
    - Examples of usage (Advanced economies): "Countries in European Union; Hong Kong SAR; Iceland; Norway"
    - Examples of usage (Emerging and developing economies): "Czech Republic"
  - Dynamic provisioning requirement
    - Calibration note: "The amount of loan loss provisioning is calibrated based on a pre-set formula and varies with credit cycles. By building up a countercyclical loan loss reserve in good times and then using it to cover losses in bad times, the tool smoothes provisioning costs over the cycle."
    - Examples (Advanced economies): "Spain"
    - Examples (Emerging and developing economies): "Bolivia; Chile; Colombia; Mexico; Peru; Uruguay"
  - Leverage ratio
    - Calibration note: "The ratio is simply calculated by dividing Tier 1 capital by a bank’s total exposure (both on-balance sheet and off-balance sheet). Currently, the Basel agreement stipulates it as a minimum requirement at 3 percent from 2018, but the ratio can in principle be adjusted flexibly in order to reduce incentives to adjust risk weights along credit cycles."
    - Examples (Advanced economies): "Switzerland; UK; US"
    - Examples (Emerging and developing economies): "China"
- Sectoral tools
  - Sectoral capital requirements
    - Calibration note: "Risk weights or LGD floors on a specific segment of bank loans are changed to increase resilience against unexpected credit losses, based on onsite inspection and offsite analyses including stress testing."
    - Examples (Advanced economies): "Australia; Hong Kong SAR; Ireland; Israel; Korea; Norway; Spain; Switzerland"
    - Examples (Emerging and developing economies): "Argentina; Brazil; Bulgaria; Croatia; Estonia; India; Malaysia; Nigeria; Peru; Poland; Russia; Serbia; Thailand; Turkey; Uruguay"
  - Limits on loan-to-value ratio
    - Calibration note: "Growth in mortgage loans and property prices are used jointly as core early warning indicators because they together provide powerful signals for policy actions. Some countries also use loan-level data to calculate the distribution of LTV and DSTI ratios along borrowers’ characteristics (e.g., age, income, location, number of existing mortgage loans) (e.g., Ireland, Korea, the Netherlands). The granular data makes it possible to better target the origin of systemic risks with well-tailored tools, such as differentiated limits by types of borrowers or regions."
    - Examples (Advanced economies): "Canada; Estonia; Finland; Hong Kong SAR; Ireland; Israel; Korea; Latvia; Lithuania; Netherlands; New Zealand; Norway; Singapore; Sweden"
    - Examples (Emerging and developing economies): "Brazil; Bulgaria; Chile; China; Colombia; Hungary; India; Indonesia; Lebanon; Malaysia; Poland; Romania; Thailand; Turkey"
  - Caps on debt-service-to-income ratio or loan-to-income ratio
    - Examples (Advanced economies): "Canada; Estonia; Hong Kong SAR; Ireland; Korea; Lithuania; Netherland; Norway; Singapore; UK"
    - Examples (Emerging and developing economies): "China; Colombia; Hungary; Malaysia; Poland; Romania; Thailand"
- Liquidity tools
  - Reserve requirements
    - Calibration note: "Countries change the requirement rates to alleviate liquidity pressures in the banking system and set differentiated rates by types of liabilities by examining system-wide liquidity conditions (e.g., indicators of stress in interbank markets) and maturity and currency mismatch indicators, respectively."
    - Examples (Emerging and developing economies): "Argentina; Armenia; Azerbaijan; Brazil; Bosnia and Herzegovina; Bulgaria; Cambodia; China; Colombia; Ecuador; El Salvador; Ethiopia; Fiji; Gambia; Georgia; Haiti; India; Indonesia; Kazakhstan; Lebanon; Macedonia; Moldova; Mongolia; Mozambique; Nigeria; Peru; Philippines; Romania; Russia; Saudi Arabia; Serbia; Sri Lanka; Tajikistan; Tonga; Turkey; Uruguay"
  - Liquidity buffer requirements (e.g., LCR, liquid asset ratio)
    - Calibration note: "The LCR and NSFR are calculated according to the BCBS guideline, comparing the stock of high-quality liquid assets to total net cash outflows over 30 days and the amount of a bank’s available stable funding to its required stable funding, respectively. In most countries, banks are obliged to maintain a minimum LCR of 60 percent from 2015, which will be phased in until"
    - (Table continuation beyond provided content is present in the source but not included here.)

### Additional notes from the references list
- The references comprise empirical studies, theoretical models, policy notes, working papers, speeches, and institutional guidance covering cross-cutting issues in systemic risk, stress testing, and macroprudential policy implementation.
- Several entries are noted as "forthcoming" or "n.d." indicating work in progress or forthcoming publications at the time of the source compilation.
- Footnotes in the table specify: "The list includes examples of selected macroprudential instruments, and is incomplete." and "The table includes usage of tools for purposes other than MaPP, e.g., for monetary or microprudential policy purposes." (footnote identifiers 42 and 43 retained as in source).

*Source: wp18197 - REFERENCES (IMF).*

### 2018. The NSFR final standard was published in October 2014 and is now in the observation

### wp18197 - 2018. The NSFR final standard was published in October 2014 and is now in the observation

### Stable funding requirements and liquidity charges
- NSFR final standard published in October 2014 and in the observation period until 2018.
- Korea: liquidity charge imposed on banks’ daily average balance of short-term foreign currency liabilities (so-called non-core funding).
  - Rate varies from 2 to 20 basis points.
  - Can be adjusted based on the indicator of non-core funding in the banking system.
- New Zealand: introduced the core-funding ratio to ensure banks hold sufficient retail and long-term wholesale funding.
  - In 2010, the minimum ratio was set at 65 percent of total loans and advances.
  - Increased to 70 and 75 percent in July 2011 and January 2013, taking account of funding market conditions.
- Examples of jurisdictions using stable funding requirements (e.g., NSFR, core funding ratio, LTD ratio) or liquidity charges on non-core funding:
  - Australia; Canada; countries in European Union; Hong Kong SAR; Iceland; Israel; Korea; Norway; Singapore; Switzerland; United States
  - Albania; Armenia; Argentina; Azerbaijan; Brazil; Burundi; China; Colombia; Georgia; India; Jamaica; Kosovo; Macedonia; Mexico; Morocco; Nigeria; Peru; Romania; Russia; Saudi Arabia; Serbia; Slovak; Solomon Islands; South Africa; Sri Lanka; Ukraine; Zambia
  - Ireland; Korea; New Zealand; Portugal
  - Bangladesh; Indonesia; Kuwait; Pakistan; Slovak Republic; Ukraine

### Constraints on foreign exchange positions
- Countries explicitly listed with constraints on foreign exchange positions:
  - Austria; Korea
  - Albania; Angola; Armenia; Azerbaijan; Bangladesh; Brazil; Burundi; Croatia; Democratic Republic of Congo; Gambia; Ghana; Haiti; Honduras; Kenya; Kosovo; Mauritius; Mongolia; Nigeria; Pakistan; Paraguay; Peru; Russia; Senegal; Serbia

### Structural tools: G-SIBs and D-SIBs surcharge framework
- FSB publishes annually the list of identified G-SIBs according to BCBS guideline using indicators capturing five dimensions: size, interconnectedness, lack of substitutability, complexity, and global scope of activities.
- National authorities identify D-SIBs under a similar guideline except the global-scope dimension to take into account country circumstances.
- Level of G-SIB and D-SIB surcharges can differ according to the composite score based on the indicators.
- Jurisdictions identifying G-SIBs/D-SIBs and applying surcharges include:
  - Australia; Canada; Countries in European Union; Israel; Japan; Singapore; Switzerland; UK; US
  - China; India; Indonesia; Kuwait; Nigeria; Peru; Russia; Uruguay

### International stress testing frameworks (high-level features from Table 2)
- Design thresholds and inclusion criteria noted in the table:
  - Bank inclusion threshold examples: £50bn retail deposits; $50bn total assets; €30bn total assets; varies according to country/area circumstances; varies according to country circumstances.
  - Scope/coverage examples: More than 350 banks are part of the test, including 10 major banks; Six big banks; 17 domestic banks; Seven large Norwegian banking groups; All seven D-SIBs.
- Common design elements across authorities:
  - Comprehensive macro scenario — present across multiple authorities.
  - Hurdle rate in excess of international minima — present for some authorities.
  - Full implementation of Basel III phase-in — noted as a design element.
  - Number of macro stress scenarios: ranges shown as 1-2; 2; 1; varies according to country/area circumstances; 2; 1; multiple scenarios according to country circumstances; 1; 1.
  - Frequency of exercises: Annual; Biannual; Undefined; With FSAP; Biannual; Biannual; Biannual; Annual; Annual.
- Results production modalities referenced:
  - Bottom-up (banks) and bottom-up (authorities) exercises noted.
  - Top-down (authorities) produced but not typically driver of results in some cases.
  - Hybrid approaches used in certain implementations.
- Representative counts and labels from the table:
  - BoE, US Fed, EBA/SSM, ECB, IMF, BoJ, Canada (BOC-OFSI), Bank of Korea, Norges Bank, Singapore all appear in the cross-jurisdiction comparison.

### SRA (Systemic Risk Amplification) modeling: BoE (Table 3 highlights)
- Contagion: solvency distress contagion via interbank loans (unsecured)
  - Approach description: models the revaluation of interbank debt claims (loans and debt holdings) due to changes in banks’ CA in the stress situation. Exposures are marked down in value using a structural model; resulting reductions in capital can lead to further write-downs, propagating stress through the network.
  - Data requirements: granular data on interbank exposures.
  - Challenges: on a national level, this only sees a small part of the global interbank network, and so cannot fully capture the potential impact of this channel in a global stress situation. Different plausible revaluation functions could be used, and so it is important to run sensitivity tests on modeling assumptions.
  - Reference: Bardoscia and others (2017).
- Other assets including securities, derivatives, foreign exchange: similar data needs and challenges as interbank exposures.
- Other bank nonbank assets/liabilities: noted as requiring granular data and presenting similar challenges.
- Wholesale funding cost model:
  - Uses a panel regression approach to project changes in banks’ funding costs as a function of risk-free rates and projections of banks’ solvency positions.
  - Intended to incorporate non-linear threshold and/or transition effects.
  - Data requirements:
    1. Scenario variables, bank capital projections, historic data on funding costs.
    2. Breakdown of liabilities by maturity and type.
    3. Firm projections on the impact of the stress scenario.
  - Challenges:
    1. Data limitations: regulatory data are low frequency and inconsistent through time due to accounting changes. Market data are higher frequency, but require further assumptions around the relationship between regulatory capital and market capital.
  - Reference: Dent and others (2017).
- Funding stress model:
  - Initially focusing on interbank markets, a model is being developed to assess the risks and impact of institutions withdrawing funding from each other when faced with deteriorating capital and liquidity positions.
  - Challenge: calibration is a real challenge, particularly given the significant regulatory changes since the crisis.

*Italic: Source: wp18197 - 2018. The NSFR final standard was published in October 2014 and is now in the observation*

### 3. In addition, banks project the impact of the stress

### 3. In addition, banks project the impact of the stress

### Macro-financial feedback / second-round effects
- Approaches described:
  - DSGE: closed-economy model for the euro area with financially-constrained households and firms and an oligopolistic banking sector; features frictions for credit demand and supply.
    - Data requirements: Macro data and bank level data aggregated at country level (ECB uses data from the stress test templates, PD, LGD + historical data on capital ratio and other key balance sheet variables).
    - Challenge: Calibration of the steady state.
    - Reference: Darracq Paries and others 2010.
  - MCS-GVAR: Mixed-Cross-Section Global Vector Autoregressive model for the 28 EU economies and a sample of individual banking groups to assess propagation of bank capital shocks to the economy; assesses how capital ratio shocks influence bank credit supply and aggregate demand conditional on different bank adjustment strategies (asset-side deleveraging vs capital raising vs mixture).
    - Data requirements: Macrofinancial time series across EU28 plus US, consolidated banking data from SNL and other public data vendors, IBSI/IMIR data.
    - Reference: Gross and others 2016.
  - IDHBS model: Household balance sheet model using HFCS household survey data; used for scenario and sensitivity analysis of household balance sheets and assessing borrower-based macroprudential instruments (LTV and DSTI caps).
    - Data requirements: Macro data, bank level data, household budget, and wealth survey data.
    - Challenge: Employment-related statistics missing for many countries, in particular duration of unemployment parameters.
    - Reference: Gross and others 2016.
  - Financial Macroeconometric Model (FMM) — Bank of Japan: medium-sized macro model with two sectors (financial and macroeconomic); financial institutions modeled by panel data and estimated equation-by-equation using least squares.
    - Data requirements: Detailed balance sheet data and time series of economic (nominal and real) and financial variables from the 80s.
    - Reference: Kitamura and others 2014.
  - HKMA frameworks: empirical default-rate models with Monte Carlo simulation; VAR macro model linked to a banking model with sequential MC updating (10,000 trials) to capture macrofinancial feedback.
    - Empirical finding: Using a stress-testing framework without macrofinancial feedback could lead to an under-estimation of risks.
    - Reference: Hong Kong Monetary Authority, 2016; Virolainen, 2004.

### Liquidity and leverage interactions
- Bank of Canada (BoC) modules:
  - Leverage module: banks deleverage by selling market securities (not loans) when leverage-constrained; price curves for securities are calibrated.
  - Liquidity risk module: funding withdrawal decisions modeled via global game approach; liquidity risk can materialize endogenously from creditor behavior and information contagion.
  - Data requirements: Detailed information on liquid and illiquid assets and liabilities susceptible to runs; detailed information on banks’ securities holdings.
- U.S. Federal Reserve:
  - Current CCAR/DFAST focus on capital adequacy (CA); scenarios reflect environments with liquidity strains.
  - Plans for research to incorporate funding shocks into capital stress tests and to model system-wide funding-cost effects tied to system solvency.
  - Implementation challenges: gathering information on on- and off-balance-sheet exposures, counterparties, and institution behaviors; integrated capital and liquidity stress testing may be required.
  - References: Brazier, 2015; Tarullo, 2014; Tarullo, 2016.
- Bank of Japan:
  - October 2016 FSR macro stress testing analyzed fire sales of cross-border loans and nonlinear effects from deeper disposal discounts when many institutions sell simultaneously.
  - Data requirements: funding profile and cross-border loan book of individual banks.
  - Reference: Bank of Japan 2016 Financial System Report, October 2016.

### Contagion: direct interbank exposures and network models
- Interbank network models (ECB, BoC, Banco de México, RBI, HKMA, Banco de México):
  - Purpose: study transmission of solvency shocks via bilateral bank-by-bank exposures; simulate loss cascades after assumed defaults.
  - BoC uses Eisenberg and Noe (2001) clearing algorithm to proportionally repay interbank obligations.
  - Data requirements: bilateral bank-by-bank interbank exposures; detailed interbank exposures at loan-bank level for some central banks.
  - Challenges: calibration on real bank-by-bank exposures in a dynamic manner; modeling endogenous network responses and behavioral/dynamic aspects of interbank participation.
  - References: Hałaj and Kok 2013; Martínez-Jaramillo and others 2010; Poledna and others 2015; Solorzano-Margain and others 2013.
- Banco de México:
  - Uses extremely granular interbank exposure data to construct multi-layer network models (derivatives; loans and deposits; foreign exchange; securities) and finds loans and deposits account for roughly 10% of expected systemic risk in one application.
  - Data requirements: detailed information on interbank exposures at loan-bank level.
  - Reference: Poledna and others 2015.

### Indirect contagion — fire sales and price-mediated mechanisms
- Price-mediated contagion model (described in context prior to Table 4):
  - Used to assess contagion risks from forced selling of tradable assets when banks face capital/funding constraints; forced sales reduce prices, causing mark-to-market losses on other banks’ balance sheets and potential further rounds of sales.
  - Data requirements: granular data on banks’ tradable assets; historic market data on asset prices and trading volumes to calibrate price impact of asset sales.
  - Modeling challenges: institutions’ multiple response options; choosing granularity of asset holdings; calibrating market liquidity and price-impact functions; assumptions about which assets banks sell and sales timescales.
- Agent-based and network-based fire-sale models:
  - ABM of systemic liquidity risk (ECB): captures amplification of funding shocks via fire sales, interbank linkages, funding conditions as function of solvency, and cross-holding of debt channels.
    - Data requirements: detailed asset and liability structure including maturity and encumbrance, expected return, banks capitalization; prior on probability of linkages between banks.
    - Reference: Hałaj 2016.
  - Agent-based model of bank–shadow bank system: endogenous price formation and clearing mechanism producing fire sales after exogenous liquidity shock; banks subject to liquidity requirements sell illiquid securities and endogenous price declines trigger cascades.
    - Data requirements: macro statistics on assets and liabilities of asset fund managers and banks; redemption rates.
    - Challenges: calibrating system to simulate fire-sale phenomenon to specific market structure and accounting for market depth.
    - Reference: Calimani and others mimeo.
  - Network-based model for systemic risk (26 largest euro area banking groups): spans asset commonality, concentration, and interconnectedness to produce systemic measures combining interconnectedness, asset commonality, and leverage.
    - Uses links between overlaps and capital figures under stress.
    - Reference: Battiston and others mimeo.
- Implementation status across authorities:
  - Some models are marked as "In Development" or "Work in progress"; others are "Implemented" depending on jurisdiction and model maturity (sources: BoE, ECB, BoJ, HKMA, RBI, Banco de México, BoC, U.S. Federal Reserve, IMF staff).

### Data requirements and common modeling challenges
- Common data requirements across approaches:
  - Granular bank-level balance sheet data (assets, liabilities, maturities, encumbrance).
  - Bilateral interbank exposure data (who-to-whom).
  - Historic market data on asset prices and trading volumes.
  - Household survey data for micro-macro balance sheet models (e.g., HFCS).
  - Macrofinancial time series across jurisdictions for GVAR-style models.
- Recurrent challenges:
  - Calibration of steady states, market liquidity, and price-impact functions.
  - Missing employment or duration-of-unemployment parameters in household models.
  - Estimating bilateral interconnections when data are missing — use of statistical reconstruction (e.g., maximum entropy).
  - Modeling institutions’ behavioral responses, endogenous network formation, and dynamic feedback loops.
  - Integrating capital and liquidity stress testing and endogenizing fire-sale effects remain areas for further research and implementation.

*Sources: BoE, ECB, BoJ, HKMA, RBI, Banco de México, BoC, U.S. Federal Reserve, and IMF staff.*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2018/wp18197.pdf_