Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.
Eesti Pank – August 2016 © 2016 Kreis
Systemic Risk in a Structural Model
of Bank Default Linkages
Yvonne Kreis
Gutenberg ...
Eesti Pank – August 2016 © 2016 Kreis
2
This Paper
• Structural model of bank default in an asset-
based approach
• Focus ...
Eesti Pank – August 2016 © 2016 Kreis
3
Agenda
Motivation
Model
Correlation
Implications & Regulation
Conclusion
Eesti Pank – August 2016 © 2016 Kreis
Motivation I: Level of Interconnectedness
• Banks are interconnected on several laye...
Eesti Pank – August 2016 © 2016 Kreis
Motivation II: Impact of Systemic Risk
• Severe systemic events affect all banks.
• ...
Eesti Pank – August 2016 © 2016 Kreis
Motivation III: Default
• Default results from the balance sheet.
• Regulation focus...
Eesti Pank – August 2016 © 2016 Kreis
7
Agenda
Motivation
Model
Correlation
Implications & Regulation
Conclusion
Eesti Pank – August 2016 © 2016 Kreis
Structural Model of Default
• Individual bank defaults iff assets < debt
<
• Introdu...
Eesti Pank – August 2016 © 2016 Kreis
Normal Periods vs. Crisis Periods
• Normal periods
encompasses fundamental factors, ...
Eesti Pank – August 2016 © 2016 Kreis
Default Frequency
• Individual default probability
• Indicator variable for default ...
Eesti Pank – August 2016 © 2016 Kreis
Density of Default Frequency MN
11
parameters: p=1%; N=1,000
Eesti Pank – August 2016 © 2016 Kreis
12
Agenda
Motivation
Model
Correlation
Implications & Regulation
Conclusion
Eesti Pank – August 2016 © 2016 Kreis
Data
• Analyses for the U.S. banking sector 1980 – 2015
• Selection of “core” sector...
Eesti Pank – August 2016 © 2016 Kreis
Time Series of Average Correlation
14
Eesti Pank – August 2016 © 2016 Kreis
Time Series of Average Correlation
15
Eesti Pank – August 2016 © 2016 Kreis
Asset Correlations
• Asset correlations between banks have strongly
increased over t...
Eesti Pank – August 2016 © 2016 Kreis
17
Agenda
Motivation
Model
Correlation
Implications & Regulation
Conclusion
Eesti Pank – August 2016 © 2016 Kreis
Risk Measures
• For correlation=0, Law of Large Numbers implies
– suggests that defa...
Eesti Pank – August 2016 © 2016 Kreis
Conditional Expected Default Frequency
19
Eesti Pank – August 2016 © 2016 Kreis
U.S. Sample: Temporal Evolution of CEDF
20
Eesti Pank – August 2016 © 2016 Kreis
Outlook: Relative Systemic Capital
Supplement on Correlation
21
Eesti Pank – August 2016 © 2016 Kreis
22
Agenda
Motivation
Model
Correlation
Implications & Regulation
Conclusion
Eesti Pank – August 2016 © 2016 Kreis
Conclusion
• Approach to systemic risk based on well-known
structural model to credi...
Upcoming SlideShare
Loading in …5
×

Yvonne Kreis. Systemic Risk in a Structural Model of Bank Default Linkages

260 views

Published on

Eesti Pank – August 2016

Published in: Economy & Finance
  • Be the first to comment

  • Be the first to like this

Yvonne Kreis. Systemic Risk in a Structural Model of Bank Default Linkages

  1. 1. Eesti Pank – August 2016 © 2016 Kreis Systemic Risk in a Structural Model of Bank Default Linkages Yvonne Kreis Gutenberg University Mainz Dietmar Leisen Gutenberg University Mainz
  2. 2. Eesti Pank – August 2016 © 2016 Kreis 2 This Paper • Structural model of bank default in an asset- based approach • Focus on the impact of asset correlations • Systemic risk measure: Conditional Expected Default Frequency • Empirical analysis for the U.S. banking sector: • Macro-prudential regulation required; SIFI capital supplement too “small”
  3. 3. Eesti Pank – August 2016 © 2016 Kreis 3 Agenda Motivation Model Correlation Implications & Regulation Conclusion
  4. 4. Eesti Pank – August 2016 © 2016 Kreis Motivation I: Level of Interconnectedness • Banks are interconnected on several layers. – Direct connection (asset investments, loans) – Indirect connection via common (systemic) factors (e.g. market risk perception) • Severe systemic events affect all banks through direct and / or indirect connections. (e.g. exposure, risk perception, trust in and between banks) • Current literature often focuses on modeling direct connections. 4
  5. 5. Eesti Pank – August 2016 © 2016 Kreis Motivation II: Impact of Systemic Risk • Severe systemic events affect all banks. • Only relevant factor during a financial crises • Single Factor Approach to Systemic Risk • Foundation: Vasicek (1987) 5
  6. 6. Eesti Pank – August 2016 © 2016 Kreis Motivation III: Default • Default results from the balance sheet. • Regulation focuses on assets. • Asset Based Approach to Systemic Risk • Foundation: Structural Model by Merton (1974) 6
  7. 7. Eesti Pank – August 2016 © 2016 Kreis 7 Agenda Motivation Model Correlation Implications & Regulation Conclusion
  8. 8. Eesti Pank – August 2016 © 2016 Kreis Structural Model of Default • Individual bank defaults iff assets < debt < • Introduce default correlation through asset correlation V = + • Study a banking system with N individual banks 8
  9. 9. Eesti Pank – August 2016 © 2016 Kreis Normal Periods vs. Crisis Periods • Normal periods encompasses fundamental factors, e.g. KMV. • Crisis periods encompasses only the systemic shock. 9 = , + , , = , + , ,
  10. 10. Eesti Pank – August 2016 © 2016 Kreis Default Frequency • Individual default probability • Indicator variable for default of bank i: Xi • Default frequency: = • Conditional default probability • Conditional distribution of default frequency 10
  11. 11. Eesti Pank – August 2016 © 2016 Kreis Density of Default Frequency MN 11 parameters: p=1%; N=1,000
  12. 12. Eesti Pank – August 2016 © 2016 Kreis 12 Agenda Motivation Model Correlation Implications & Regulation Conclusion
  13. 13. Eesti Pank – August 2016 © 2016 Kreis Data • Analyses for the U.S. banking sector 1980 – 2015 • Selection of “core” sector, i.e. 15 large banks • Daily stock price returns as approximation of asset returns • Application of a principal component analysis – over a rolling 3-month window – mean (median) factor loadings of the first principal component 13
  14. 14. Eesti Pank – August 2016 © 2016 Kreis Time Series of Average Correlation 14
  15. 15. Eesti Pank – August 2016 © 2016 Kreis Time Series of Average Correlation 15
  16. 16. Eesti Pank – August 2016 © 2016 Kreis Asset Correlations • Asset correlations between banks have strongly increased over time! • They have become very high in the past decade! • This strongly influences the bank default linkages and thus systemic risk! • How to adequately introduce to risk management? 16
  17. 17. Eesti Pank – August 2016 © 2016 Kreis 17 Agenda Motivation Model Correlation Implications & Regulation Conclusion
  18. 18. Eesti Pank – August 2016 © 2016 Kreis Risk Measures • For correlation=0, Law of Large Numbers implies – suggests that default frequency is close to individual default probability (micro-prudential regulation) • Actual numbers are far too small to capture “infinity”. • Need to capture departure from p – Study Conditional Expected Default Frequency: = [ | ] – foundation for macro-prudential regulation 18
  19. 19. Eesti Pank – August 2016 © 2016 Kreis Conditional Expected Default Frequency 19
  20. 20. Eesti Pank – August 2016 © 2016 Kreis U.S. Sample: Temporal Evolution of CEDF 20
  21. 21. Eesti Pank – August 2016 © 2016 Kreis Outlook: Relative Systemic Capital Supplement on Correlation 21
  22. 22. Eesti Pank – August 2016 © 2016 Kreis 22 Agenda Motivation Model Correlation Implications & Regulation Conclusion
  23. 23. Eesti Pank – August 2016 © 2016 Kreis Conclusion • Approach to systemic risk based on well-known structural model to credit risk • Asset correlations have a strong impact (shape of the density of default frequencies, risk measure, and thus capital requirements) • “Large” correlation levels imply that stronger macro-prudential regulation is required • Next steps: stress testing & systemic capital

×