## mcm-lse-segoviano_session4

## Source details

**Canonical URL:** [mcm-lse-segoviano_session4](https://www.imf.org/-/media/files/news/seminars/mcm-lse-segoviano_session4.pdf)

## Other formats

- [Markdown version](/-/media/files/news/seminars/mcm-lse-segoviano_session4.pdf.md)
- [Structured JSON version](/-/media/files/news/seminars/mcm-lse-segoviano_session4.pdf.json)

---

### Overview
- Presentation by Miguel Angel Segoviano, Monetary and Capital Markets Department, International Monetary Fund.
- Draws material from Alla et al. 2017. Views are those of the authors and do not represent IMF policy.
- Central problem statement (Goodhart, 1975): policy makers face information and time constraints that worsen in crises as volatility increases, history becomes a poor guide, market perceptions gain weight, and reaction times decrease.

### Challenges to quantifying Systemic Risk (SR)
- Conceptual channels and mechanisms highlighted:
  - Direct Contagion (domino effects; BIS 1994; Kaufman 1995).
  - Indirect Contagion and market amplification mechanisms: fire sales, information asymmetry, collateralized agreements, interactions across banks and non-banks, illiquidity spirals, leverage, bank runs.
  - References cited: Shleifer and Vishny (2011); Khandani and Lo (2011); Segoviano et al. (2017); Brunnermeier and Pedersen (2009); Greenwood et al. (2015); Cont and Schaanning (2016); Jacklin and Bhattacharya (1988); Kapadia et al. (2012).
- Key modeling and data challenges:
  - Generalized shocks and relationship between financial system and real economy.
  - Diverse contractual obligations, market price channels, asset fire sales, information asymmetry.
  - Data constraints, model risk, non-linear changes, structural changes, crisis-consistent estimations.
  - Need for interpretable metrics under highly restrictive data environments.

### Encompassing Frameworks (EF) — rationale and benefits
- Purpose:
  - Capture best elements of a variety of approaches because no single data set or model fully captures SRA mechanisms.
  - Provide flexible yet organized approaches to combine separate analyses.
- Cornerstone benefits:
  - Reduced risk of model error.
  - Improved assessments through complementary perspectives on risk.
  - Transferable frameworks usable across heterogeneous financial systems.
  - Cost-efficient and computationally simple, relatively light on data requirements, while preserving analytical rigor.
- Implementation principles:
  - Implemented with a combination of publicly available and supervisory-based data.
  - Embed diverse types of methods (top-down, bottom-up, microprudential stress testing (MiPST)).
  - Advance analysis cooperatively using diverse sets of data and methods.
  - Useful for parallel running with structural and reduced-form frameworks (e.g., DSGE and VAR) to improve calibrations.

### IMF Encompassing Framework (IMF-EF) — structure and characterization
- Core components:
  - SRA Losses: “Expected losses given the realization of a given event.”
  - Multivariate perspective of the financial system: time-consistent distress dependence structures and multivariate densities to derive systemic loss indicators and financial stability indicators.
  - “Crisis consistent conditional losses” based on markets’ perceptions.
  - Microprudential ST: first-order effects of adverse scenarios on individual entities; implemented by IMF (workbox), national authorities, firms, or jointly.
  - Combination of publicly available and supervisory data.
- Implementation details:
  - Uses MiPST frameworks already implemented.
  - SRA Loss based on publicly available data and market perceptions of distress dependence structures.
  - Calibration limitations: proper calibration of ex-ante modeled structures is challenging given IMF and some authorities’ data limitations.
  - Reduced-form nature: quantifies SRA Losses, identifies “connecting entities”, and estimates likelihood and intensity of contagion effects, but does NOT provide causal economic and financial mechanism explanations.
  - Parallel running: helps improve calibrations of alternative frameworks.

### SRA Loss — definitions, decomposition, and interpretation
- Definitions (formal):
  - MicroST Loss: difference between the value of bank A in normal times and its value under an adverse macroeconomic scenario.
  - SRA Loss: conditional loss assuming realization of a given financial contagion event S.
  - Total Loss: sum of MicroST Loss and SRA Loss conditional on event S.
- Venn diagram interpretation:
  - MicroST Loss: represented by adverse macroeconomic scenario area.
  - SRA Loss: additional loss corresponding to the intersection involving the realization of event S; takes into account ALL potential connections across ALL entities and their effects on the assessed bank.
  - High SRA Loss (A/B) does NOT necessarily imply a direct strong link between FE A and B; contagion can operate through intermediate FEs.
- Decomposition objectives:
  - Decompose SRA Loss to quantify likelihood and intensity of contagion events.
  - Use law of total expectations to identify connecting entities between two given entities.
- Defaulting sets and example metrics (as presented):
  - Numeric metrics (preserved exactly as presented):
    - 0.79, 0.11, 0.01, 0.09
    - R row: 0.45, 2.62, 2.30, 3.69
    - Next row: 0.35, 0.28, 0.03, 0.32

### Empirical illustration (Citi example)
- Citi (in millions of USD):
  - TARP Injection: 25,000
  - TA (Q2 2008): 2,100,385
  - TARP/TA: 1.19%
  - Loss SRA (Citi/LB)/TA: 1.14%
- Note: In December 2008, Citi received an additional 20 billion USD through TARP.

### Microprudential stress testing heterogeneity (selected statistics)
- FSAP coverage: 321 FSAPs in more than 160 Jurisdictions.
- Distribution (# of FSAPs by country): categories shown 0, 1, 2, 3, 4+ (figure referenced).
- Breakdown of micro-ST focuses (preserved labels and percentages as shown):
  - 39.39% (label context: appears associated with an item in slide)
  - Bank liquidity: 46.38%
  - Bank liquidity (alternate context): 6.25%
  - Bank solvency: 46.38% (appears in slide context)
  - Insurance solvency: 25%
  - Insurance solvency (alternate context): 13

### Conclusions (IMF-EF implications)
- EF leverages existing MiPST frameworks and publicly available data to deliver crisis-consistent conditional loss estimates based on market perceptions.
- Advantages: cost-efficient, computationally simple, useful under data constraints, and supports cooperative, diverse-method analyses.
- Limitations: reduced-form approach does not explain underlying economic/financial causes; calibration of modeled structures is challenging with limited data.
- Suggested use: run IMF-EF in parallel with alternative frameworks (structural and reduced-form) to improve calibration and robustness of systemic risk assessment.

---


_Source: https://www.imf.org/-/media/files/news/seminars/mcm-lse-segoviano_session4.pdf_