Cover Page

Contents

Cover

Series

Title Page

Copyright

Foreword

Introduction

Part I: Expert Views

Chapter 1: Origins of the Crisis and Suggestions for Further Research

1.1 Introduction

1.2 The Real Economy: Actors and Markets

1.3 The Financial Techniques: Products and Methods

1.4 The Global Risk Management Challenge

1.5 Conclusion

Chapter 2: Quantitative Finance: Friend or Foe?

2.1 What Future for VaR Models?

2.2 What Future for Pricing Models?

2.3 Conclusion

Part II: Credit Derivatives: Methods

Chapter 3: An Introduction to Multiname Modeling in Credit Risk

3.1 Introduction

3.2 The Copula Model

3.3 Reduced Form Loss Models

3.4 Markovian Projection, Local Intensity, and Stochastic Local Intensity Models

3.5 Forward Loss Models

3.6 Further Issues in Credit Modeling

3.7 Acknowledgments

Chapter 4: A Simple Dynamic Model for Pricing and Hedging Heterogeneous CDOs

4.1 Introduction

4.2 Model

4.3 Semianalytic Approach

4.4 Model Calibration

4.5 Hedging a CDO Tranche

4.6 Portfolio with Heterogeneous Recovery Coefficients

4.7 Markovian Projection onto the Default Contagion Model

4.8 Stochastic Recovery Coefficients

4.9 Conclusion

Appendix 4A: Derivation of the Fokker-Planck Equation (4.3)

Appendix 4B: Markovian Projection onto the One-Dimensional Markov Chain

Appendix 4C: Self-Consistency Criterion for the Semianalytic Approximation

Acknowledgments

Chapter 5: Modeling Heterogeneity of Credit Portfolios: A Top-Down Approach

5.1 Introduction

5.2 Top-Down Default Time Matrices

5.3 Thinning by Bootstrap and Iterative Scaling

5.4 Single-Name Sensitivities

5.5 Modeling Notional Heterogeneity

5.6 Random Recovery in Top-Down Models

5.7 Conclusion

Acknowledgments

Chapter 6: Dynamic Hedging of Synthetic CDO Tranches: Bridging the Gap between Theory and Practice

6.1 Introduction

6.2 Hedging of CDO Tranches: Theoretical Issues and~Perspectives

6.3 From Theory to Hedging Effectiveness

6.4. Conclusion

Chapter 7: Filtering and Incomplete Information in Credit Risk

7.1 Introduction

7.2 A Short Introduction to Stochastic Filtering

7.3 Credit Risk Models under Incomplete Information

7.4 Structural Models I: Duffie and Lando (2001)

7.5 Structural Models II: Frey and Schmidt (2009)

7.6 Constructing Reduced-Form Credit Risk Models via Nonlinear Filtering

7.7 Numerical Case Studies

Chapter 8: Options on Credit Default Swaps and Credit Default Indexes

8.1 Introduction

8.2 Credit Default Swaps

8.3 Options on Credit Default Swaps

8.4 CIR Default Intensity Model

8.5 Options on Credit Default Indexes

8.6 Market Models for CDS Spreads

8.7 Acknowledgments

Part III: Credit Derivatives: Products

Chapter 9: Valuation of Structured Finance Products with Implied Factor Models

9.1 Introduction

9.2 Valuation of Structured Finance Instruments

9.3 Implied Factor Models and Weighted Monte Carlo

9.4 The ABX Indexes

9.5 Examples

9.6 Conclusion

Acknowledgments

Chapter 10: Toward Market-Implied Valuations of Cash-Flow CLO Structures

10.1 Introduction

10.2 Description of the Cash-Flow CLO Structure

10.3 Description of the Valuation Framework

10.4 Numerical Examples

10.5 Summary and Discussion of Open Issues

Acknowledgement

Chapter 11: Analysis of Mortgage- Backed Securities: Before and After the Credit Crisis

11.1 Market Structure

11.2 Prepayment

11.3 Yields and OAS

11.4 Selection of Calibration Instruments

11.5 Interest Rate Models

11.6 Index Projection

11.7 Monte Carlo Analysis

11.8 Parallelization

11.9 Calculating Greeks

11.10 Validation

11.11 Conclusion

Acknowledgments

Part IV: Counterparty Risk Pricing and Credit Valuation Adjustment

Chapter 12: CVA Computation for Counterparty Risk Assessment in Credit Portfolios

12.1 Introduction

12.2 General Counterparty Risk

12.3 Counterparty Credit Risk

12.4 Multivariate Markovian Default Model

12.5 Numerical Results

12.6 Acknowledgments

Chapter 13: Structural Counterparty Risk Valuation for Credit Default Swaps

13.1 Introduction

13.2 Modeling Two-Dimensional Default Risk

13.3 Counterparty Risk

13.4 First-to-Default on Two Underlyings

Appendix 13A: Theoretical Default Legs for CDS

Appendix 13B: First Hitting Time in a Polyhedral Domain

Chapter 14: Credit Calibration with Structural Models and Equity Return Swap Valuation under Counterparty Risk

14.1 Introduction

14.2 The Analytically Tractable First Passage (AT1P) Model

14.3 Calibration of the Structural Model to CDS Data

14.4 A Case Study with AT1P: Lehman Brothers Default History

14.5 SBTV Model (Brigo and Morini 2006)

14.6 A Case Study with SBTV: Lehman Brothers Default History

14.7 A Fundamental Example: Pricing Counterparty Risk in Equity Return Swaps

14.8 Conclusion

14.9 Appendix 14A: AT1P Model: Proof

Chapter 15: Counterparty Valuation Adjustments

15.1 Introduction

15.2 Counterparty Risk

15.3 Counterparty Valuation Adjustment (CVA)

15.4 Modeling the CVA

15.5 CVA Calculations for Bonds

15.6 CVA Calculations for Swaps

15.7 Example Calculation

15.8 Hedging

15.9 Wrong-Way Risk and Recovery Risk

15.10 Accounting Considerations

15.11 Conclusion

Chapter 16: Counterparty Risk Management and~Valuation

16.1 Introduction

16.2 Managing and Mitigating Counterparty Credit Risk

16.3 Credit Exposure

16.4 Credit Exposure under Collateralization

16.5 Pricing Counterparty Risk

16.6 Portfolio Loss and Economic Capital

Part V: Equity to Credit

Chapter 17: Pricing and Hedging with Equity-Credit Models

17.1 Introduction

17.2 Introducing the “Smile to Credit” Pricing Model

17.3 A Market Model: Fitting the S2C Model

17.4 Conclusion

Appendix: From Stochastic Volatility to Local Volatility

Chapter 18: Unified Credit-Equity Modeling

18.1 Introduction

18.2 Jump-to-Default Extended Diffusions (JDEDs)

18.3 The Jump-to-Default Extended CEV Model (JDCEV)

18.4 Introducing Jumps and Stochastic Volatility via Time Changes

18.5 Numerical Illustration

Acknowledgment

Part VI: Miscellanea: Liquidity, Ratings, Risk Contributions, and Simulation

Chapter 19: Liquidity Modeling for Credit Default Swaps: An Overview

19.1 Introduction

19.2 Liquidity as a Spread in Reduced-Form Models

19.3 Liquidity through the CAPM Framework

19.4 Regression-Based Approaches for Measuring CDS Liquidity

19.5 Discussion, Conclusions, and Further Research

Acknowledgments

Chapter 20: Stressing Rating Criteria Allowing for Default Clustering: The CPDO Case

20.1 Introduction

20.2 Ratings

20.3 CPDO

20.4 Rating Criteria: Base Case and Stressed Case

20.5 Modification of the Standard Assumptions

20.6 Numerical Results

20.7 Conclusion

Acknowledgment

Chapter 21: Interacting Path Systems for Credit Risk

21.1 Introduction

21.2 Interacting Particle Systems

21.3 IPS for Rare Event Analysis

21.4 IPaS for Multiname Credit Risk

Chapter 22: Credit Risk Contributions

22.1 Introduction

22.2 Credit Risk Model

22.3 Risk Contributions and Capital Allocation

22.4 Marginal Contributions in the Linear, Homogeneous Case

22.5 Marginal Contributions for Linear, Nonhomogeneous Functions

22.6 Marginal Contributions for Nonlinear Risk Functions

22.7 Conclusion

Appendix: Factor Models of Credit Risk

Acknowledgments

Conclusion

Further Reading

About the Contributors

Index

Wiley Series

Since 1996, Bloomberg Press has published books for financial professionals on investing, economics, and policy affecting investors. Titles are written by leading practitioners and authorities, and have been translated into more than 20 languages.

The Bloomberg Financial Series provides both core reference knowledge and actionable information for financial professionals. The books are written by experts familiar with the work flows, challenges, and demands of investment professionals who trade the markets, manage money, and analyze investments in their capacity of growing and protecting wealth, hedging risk, and generating revenue.

For a list of available titles, please visit our Web site at www.wiley.com/go/bloombergpress.

Title Page

Foreword

The current economic environment, with its unprecedented global financial crisis and slow resolution, has stimulated a wealth of innovative and constructive solutions in credit risk modeling and credit derivatives in particular. While it is true that this volume includes some of the more important research of the past year, significantly, I credit the editors for having facilitated a heightened quality within the individual contributions.

Importantly, chapters are split with a good balance of the theoretical versus the practical. Especially since 2007, credit markets have extraordinarily stressed both theoretical frameworks and empirical calibrations.

Similarly, and refreshingly, contributing authors are evenly split between academic versus practitioners from industry. For example, different contributions on collateralized debt obligations (CDOs) illustrate the diversity.

Parenthetically, I have been continually amazed these past two years to see that new CDO research has been the fastest growing category of research posted on DefaultRisk.com. This is quite ironic because the new issuance of CDOs has collapsed to a small fraction of 2007 levels.

I feel this volume of research is explained by not only the existing/troubling inventory of CDOs that must be managed, but also because CDOs—as a structure—offer a defined microcosm of the larger/general credit portfolio.

There are gratifyingly few redundancies across the contributing authors. Of course, it is beneficial to have a good diversity of views brought from different perspectives. For example, the different collateralized loan obligation (CLO) and residential mortgage-backed security (RMBS) chapters offer complementary discussions.

By contrast, it is not at all surprising that several contributions address the ever more liquid credit default swap (CDS) market, their options, and their liquidity. Other important and very timely contributions concern counterparty risk and credit valuation adjustment (CVA), which are here addressed in five chapters, and hybrid modeling of credit and equity, which is addressed in a novel way.

These are just a few examples of the innovation and originality in a single volume that has, among other merits, the courage to deal in a single source with several urgent topics such as the subprime crisis, pricing and hedging of credit risk, CVA, CDO, CLO, MBS, ratings, and liquidity.

Finally, this volume would not have been possible without the diligent work this past year from many people too numerous to list. Beyond the editors and contributing authors, we are grateful to all the conference attendees and panel members in September 2009 at the Université de Nice Sophia Antipolis. The interaction was spirited and the comments were invaluable. Enjoy.

GREG M. GUPTON
May 2010
DefaultRisk.com
Author of the CreditMetrics Technical Document

Introduction

The recent decade has witnessed a rapid development of more and more advanced quantitative methodologies for modeling, valuation, and risk management of credit risk, with focus on credit derivatives that constitute the vast majority of credit markets. In part, this rapid development was a response of academics and practitioners to the demands of trading and risk managing in the rapidly growing market of more and more complex credit derivative products. Even basic credit derivatives such as credit default swaps (CDSs) have witnessed considerable growth, reaching a notional value of US$45 trillion by the end of 2007, although notional amounts fell during 2008 to $38.6 trillion.1 More complex credit derivatives, such as collateralized debt obligations (CDOs), featured a global issuance of $157.8 billion in 2004, reaching $481 billion in 2007, although in 2009 this has gone down to $4.3 billion.2

The size and complexity of credit markets in general, and credit derivatives markets in particular, undoubtedly posed a challenge for (quantitative) modelers and for market practitioners. The recent turmoil in the credit markets can be attributed to many factors, but one of the factors is probably the fact that in many respects the challenge has not been properly addressed.

This volume studies aspects of modeling and analysis of credit risk that, in our opinion, have not been adequately understood in the past. This is immediately evident also from the book subtitle, in that counterparty risk, mortgage-backed securities (MBSs), liquidity modeling, ratings, and in general pricing and hedging of complex credit derivatives are all among the areas that have not been fully or adequately addressed.

An important and original feature of this book is that it gathers contributions from practitioners and academics in an equilibrated way. Whereas the practitioners' contributions are deeply grounded in concrete experience of markets and products, the contributing academics are often involved in consulting and similar activities and have therefore a real empirical knowledge of financial products as well.

We indeed found it essential, when conceiving the volume, to keep in mind two guiding principles that, according to us, have to structure the research and practice in credit risk and, more generally, in modern finance. First, research has to be rooted in experience and rely on the knowledge of the empirical behavior of markets. Losing sight of experience or disconnecting the sophisticated mathematics of modern financial theories from the day-to-day practice may be dangerous for obvious reasons, besides compromising the necessary dialogue between decision makers and quantitative analysts. Second, a high level of technicality is required to deal with current credit markets. A naive approach that would not rely on cutting-edge research would simply be condemned to fail, at least as far as derivatives or tail distributions involved in risk management are concerned. We hope the present volume will contribute to making the difficult synthesis of these two requirements (rooting in experience, technical complexity) effective.

The volume contains expert opinion articles, survey articles, as well as articles featuring the cutting-edge research regarding these aspects. This is important, as we believe that once the dust settles after the recent credit crisis, the credit markets will be facing challenges that this research addresses, so the volume may contribute to improvement of the health of the postcrisis credit universe.

The volume is directed to senior management and quants in financial institutions, such as banks and hedge funds, but also to traders and academics. In particular, academics and students in need of strengthening their understanding of how complex mathematics can be effectively used in realistic financial settings can benefit from its reading.

The volume provides a coherent presentation of the recent advancements in theory and practice of credit risk analysis and management with emphasis on some specific topics that are relevant to the current state and to the future of credit markets. The presented research is high-level on all the involved sides: financial, mathematical, and computational. This is the only way, we believe, that modeling should be presented and discussed so as to be meaningful, constructive, and useful. In addition, readers will also benefit from quality survey articles regarding selected topics.

The present collection of articles is one of several analytical texts that have appeared in recent months as a reaction by the quantitative community to the financial crisis that exploded in 2008. We refer, for example, to Lipton and Rennie (2007), which appeared before the crisis, and Brigo, Pallavicini, and Torresetti (2010), reporting both pre- and postcrisis research. These two books are just two examples of the necessity for the quantitative community to assess its status quo vis-à-vis financial markets in general, and credit markets in particular.

The volume opens with two expert opinion articles reflecting on the role of quantitative modeling in the past and in the future, on how it did or how it did not contribute to the current credit crisis, on what lessons modeling should incorporate from the credit crunch crisis, and on whether modeling should still be relevant.

These opening chapters form the first part of the book and are followed by articles focusing on some specific issues and areas reaching toward the frontiers of credit risk modeling, valuation, and hedging.

The second and third parts are closely related, although we found it convenient to divide their contents into two separate groups. They both deal with credit derivatives. Part II focuses on general methods in multiname credit derivatives, namely derivative products that depend on more than one credit entity at the same time. This part is meant to deal with the usual key issues of multiname credit derivatives but using revisited approaches and analysis. The topics covered include a survey of multiname credit derivatives methods and approaches, methods to deal with heterogeneity and dynamic features, analysis of hedging behavior of models, filtering and information, and the modeling of options on credit derivatives.

The focus of the third part is more oriented toward products and more specifically toward asset-backed securities (ABSs) in which the analysis of cash flows presents specific difficulties that are not present in the familiar synthetic CDO framework—although we should point out that these contributions involve general ideas and techniques that are relevant for all asset classes in the field of credit. The first chapter of this part introduces the factor models, with a particular emphasis on the ABX indexes. Topics included in the other chapters are a modeling and analysis framework for collateralized loan obligations (CLOs), a valuation and risk analysis framework for residential mortgage-backed securities (RMBSs), together with a survey of postcrisis solutions to various issues such as interest rate modeling and the handling of numerical complexity.

The fourth part is devoted to the valuation of credit valuation adjustment (CVA) and counterparty risk in the current environment. It is well-known that counterparty risk was underestimated before the subprime crisis. This contributed heavily to the crisis when many financial institutions discovered billions of dollars' worth of counterparty exposures were at risk, either directly in case the counterparty would default, or indirectly through downgrades (the downgrades of monoline insurers come to mind). Since then, counterparty risk measurement has become a key issue and has attracted a lot of attention, both from the financial industry and from academia. The first chapter of this part settles the general framework of CVA valuation, including subtle mathematical features. Topics in this part include: models and mathematical tools for CVA on credit derivatives, from both the intensity and the structural points of view; CVA for bonds and swaps; accounting issues; and advanced features related to netted positions and margin agreements.

The fifth part is devoted to equity-to-credit modeling. The idea of unifying the universes of credit and equity into a single mathematical framework is an old dream of mathematical finance. The so-called Merton model, predicting defaults by viewing the stock market value of a listed firm as a call option on its assets with a threshold computed from its debt, offers a general strategy, but the corresponding numerical results are known to be unsatisfactory: hence the need for new models incorporating advanced features such as jumps or random volatility. The two chapters introduce such models and discuss the application domain of equity-to-credit modeling that runs from joint pricing of credit and equity to relative value analysis. One of the papers in the CVA part also deals with equity-to-credit modeling but with a focus on counterparty risk for equity payoffs.

The last “Miscellanea” part gathers various contributions on important topics. They include: liquidity risk (offering a detailed survey of the existing methodologies for liquidity modeling in credit default swaps), ratings (with the case study of constant proportion debt obligations [CPDOs]), modern Monte Carlo methods (with an emphasis on interacting particle systems), and a survey of the theory of risk contributions in credit risk management.

Notes

1. International Swap and Derivatives Association, “Market Survey Year-End 2008.”

2. Securities Industry and Financial Markets Association, 2010. “Global CDO data” press release 2010-07-02.

Bibliography

Brigo, D., A. Pallavicini, and R. Torresetti. 2010. Credit models and the crisis: A journey into CDOs, copulas, correlations and dynamic models. Hoboken, NJ: John Wiley & Sons.

Lipton, A., and A. Rennie, eds. 2007. Credit correlation—Life after copulas. Singapore: World Scientific.

PART I

Expert Views

Chapter 1

Origins of the Crisis and Suggestions for Further Research

Jean-Pierre Lardy

JPLC

We review several of the factual failures that the 2008 subprime crisis has revealed and analyze the root causes for these. Credit rating, regulation, models, accounting, leverage, risk management, and other aspects are reviewed. In each case, we survey solutions proposed as well as suggest directions for further research.

1.1 Introduction

The many roots of the 2008 financial crisis have been well covered in several publications. The aim of this review is to provide a short list of the ones most frequently raised and in each case try to distill one important aspect of the problem, the current proposals, and, possibly, what could be a direction of research to better understand the issue. A lesson from past decades is certainly that crises are not easy to forecast in terms of timing and magnitude, and when they occur (we can only forecast that they will occur), it is not always easy to separate, to paraphrase a famous quote from financier J.P. Morgan,1 what was wrong as a matter of judgment from what was wrong as a matter of principle. These same questions apply in today's modern finance of sophisticated markets, products, and models, with the additional complexity of separating, when something went wrong, a technical failure of the “machine” (or of the principles on which it is built) from a failure of the “user” (or its judgment). To use an analogy (I find it useful)—investing is like riding a bicycle, and there is always a trade-off between performance and risk and improvements from better machines or better driving.

After working 20 years in the financial markets, including roles at two investment banks in equity and credit derivatives,2 I have witnessed several stages of their development. I was lucky enough to reach levels of responsibility giving me a view on how decisions are made, good and bad, individually or collectively. Being mostly in the “engines room” kept me in the front lines of crises and allowed me to see how things work in practice on investment banks' trading floors. Last, having been present at early stages of the developments of these markets helped me to keep a healthy sense of pragmatism about them: The following paragraphs are personal reflections on the drivers of the crisis.3

In the remainder of this article, the various topics are organized into three sections: actors and markets, methods and products, and finally a last section on global risk management. To use an analogy with transportation, the first section would be about geography and population preferences; the second section about engineering of roads, airplanes, railways, and so on; and the third section about the rules of the road, safety procedures, and so forth. The choice of these three sections helps to distinguish the different natures of the topics, but the topics are greatly interrelated and overlap the sections in several ways.

1.2 The Real Economy: Actors and Markets

In this section, I review the issues in the 2008 financial crisis that are more closely related to the natural needs and organization of the real economy. This may be where the most important roots of the crisis lie, but also where alternatives are not easy to propose or to achieve quickly, or even possible to do so, especially when it comes to human behavior.

1.2.1 Loan Origination

With regard to the subprime crisis, it's legitimate to start with loan origination. Although no one yet knows what the full extent of the damage will be, the gradual deterioration of U.S. retail loan quality standards over the years is a fact. The negative incentives of securitization markets (originate to distribute), the flaws (and fraud) on documentation and appraisal values, the political environment supportive to increase home ownership, the lack of intervention by federal regulatory authorities despite several local whistle-blower cases all played a role (Berner and Grow 2008). The irony is that the United States was by far the most advanced country in terms of retail credit scoring (FICO scores, etc.).

The new regulatory proposals will force loan originators to keep more “skin in the game, ” with a vertical slice of any securitization (not cherry-picking a part of the origination).4 Further research could also explore what is the right balance between statistical credit scoring and proximity and human judgment, with all its diversity, and for which there is no substitute, in credit decisions.

1.2.2 Macroeconomic Imbalance

The increased Asian savings following the 1997 crisis, compounded with the surplus of China and oil-exporting countries, created a large supply of liquidity and a formidable demand for (apparently) high-quality fixed-income assets. Despite the large supply of notes and bonds from Western government deficits, the low-interest-rate environment fueled a demand for higher-yielding fixed-income assets. Wall Street engineered the products that met such demand, which was broadly characterized by a risk aversion for idiosyncratic risk (first-loss or nonrated products), but generally complacent for systemic risk, favoring highly rated products (especially AAA), albeit from complex structures and rating techniques.

Low interest rates also favored the emergence of the financial bubble in real estate prices, construction, and infrastructure, boosting growth and job creation—all welcomed by politicians and their communities.

The new regulatory proposals favor the creation of a systemic regulator5 to monitor these imbalances, and to raise concerns with persuasive (yet nonbinding) powers.

Further research could now explore what anticyclical macro policies can be global, targeting all countries at once, to avoid diplomatic crises.6

1.2.3 Rating Agencies

Rating agencies regularly and successfully improved their methodologies to take advantage of the increase in computing power and the increased availability of financial and market data. The wider availability of external ratings became a key component of regulation with Basel II, increasing furthermore the need for ratings. The irony is that the rating agencies' worst failures relate to credit products that were, by design, built on credit ratings, such as collateralized debt obligations (CDOs) of mezzanine asset-backed securities (ABSs).

In fact, the rating agencies have been hurt by the consequences of the weak parts of their business models: Who pays obviously makes a difference, sophisticated quantitative methodologies should not be pushed beyond their limits, and critical human judgment must always remain (McCreevy 2009).

As concerns further research, one wonders whether perhaps ratings should incorporate some external or open-source elements (academics' and practitioners' contributions, etc.) to their methodologies or reports to keep pace with innovation and information (in particular for complex or new structures).

1.2.4 Hedge Funds

After the precedent of Long-Term Capital Management (LTCM) in 1998, there had been growing fears in the years before 2007 about the growth of the hedge fund industry, but hedge funds were not at the origin of the 2008 crisis (High-Level Group on Financial Supervision in the EU 2009). A few hedge funds failed (Amaranth, etc.), and many had to trigger gates, causing damage to their investors, but all of these were idiosyncratic events. Obviously, withdrawing excess liquidity from Bear Stearns or Lehman Brothers added to the runs on these banks, but hedge funds were no different in this respect than corporations or mutual funds, and they also withdrew excess liquidity from Goldman Sachs, Morgan Stanley, and others. The irony is that prime brokerage contracts are all about haircuts, independent amounts, stress tests, daily margining, and so on, designed to protect the banks from the hedge fund risk, and suddenly these contracts backfired against the banks' liquidity, as hedge funds were scared about the risk of their prime brokers and were left little choice (their deposits are not guaranteed like retail depositors' are). The other irony from the Lehman bankruptcy is that hedge funds active in derivatives ended up better than funds invested in cash securities (in particular in jurisdictions with no segregations of securities).

Regulators should enforce that hedge funds' play by the technical rules of the markets. Beyond that, encouraging through regulatory best practices for collateral handling seems the best direction in order to limit the systemic impact of hedge fund failure.7

1.2.5 Remunerations

For many thinkers, the remuneration of human capital is the main engine of progress.8 It is also a source of excess and injustice. However, the irony is that Wall Street, in the language of economists, is one of a very few examples of Marxist industries, where the greatest share of added value is kept by workers, instead of the capitalists' “surplus-value.” In this field, a delicate balance must be found: a better alignment of remuneration horizons in order not to give up-front rewards for long-term risks, while the virtue of division-level cash compensation budgets necessarily moderates payoffs and therefore the moral hazard that can be associated with corporate-wide stock options plans.

Further research should explore whether the remuneration problem is an extension of the shareholder versus bondholder governance issue (Billet, Mauer, and Zheng 2006). For example, should bondholders of highly levered entities have a vote in the top remunerations schemes?

1.2.6 Leverage

The social benefit of the financial system is to transform information into liquidity for the rest of the world: Assets that are illiquid but good risks are transformed by banks and markets into liquid liabilities (that are liquid assets for the rest of the world). Yet the 2008 crisis is also a consequence of an excessive leverage from banks and from the shadow banking system of banks' vehicles, money market funds, and hedge fund financing: Basel I created an incentive for banks to use off-balance-sheet vehicles for certain (apparently) low-risk assets; money market funds were also competing for the same assets with lower regulatory and capital constraints; and providing leverage to hedge funds is a highly collateralized lending and fee-generating business.

As banks' regulatory capital ratios are risk weighted and do not directly control leverage, the current discussions revolve around the accounting of a universally accepted leverage ratio (as is currently the case in the United States and Switzerland).

Further research could be conducted as to whether, for that purpose, full debt consolidation would be desirable, with the necessary accounting precaution to differentiate the part of consolidated debt that is associated with the minority interests in the equity (or minority debt).

1.3 The Financial Techniques: Products and Methods

In this section, I review the issues in the subprime crisis that are more related to the technical choices that have been made historically or more recently by the financial world to provide or facilitate its business. This may be the part where correcting mistakes is more a matter of time and experience. In financial markets, like everywhere else, progress is not linear, and knowledge is built on both successes and mistakes of prior periods.

1.3.1 Mathematics

Whether the representation of the real world by mathematics is adequate must partly involve the mathematicians' responsibility, especially as they indirectly expect to get a higher demand for their services. Initially, there is a virtuous circle where practitioners' rules of thumb can be rebuilt by more formal approaches that confirm one other and allow further innovations. After a while, innovations can go too far; naive assumptions taken for granted are no longer valid but no longer questioned; teachers and students place theory before practice; and so on. The irony is the parallel with rating agencies: a possible excess of sophistication that is not counterbalanced by experience.

There is ground for further research on what assumptions in financial mathematics should not be made by convenience. Shouldn't there be more academic refutations, counter-examples according to their consequences if they aren't verified?

1.3.2 Models

Models generally do not take well enough into account the potential for markets to deviate far from equilibrium, especially illiquid assets. In this case, third-party models based on reasonable assumptions (such as rating agency models) usually underestimated tail risks, which were envisioned only by internal stress tests, and the later ones were often judged as lacking plausibility. Proprietary models used for their own accounts generally performed better as long as they were nimble enough to allow the user's critical eye and the ability to continually correct deficiencies. It can be preferable to have several (albeit simpler) competing models that can bring different inputs to the same risk, instead of an ultrasophisticated monolithic approach that might miss the point. The irony is again a possible excess of sophistication, crowding out caution.

Further research: From the past experiences of macroeconomic and financial models, what is the right level of complexity not to be fooled by equations?9

1.3.3 Derivative Products

The social benefit of a derivative product should be the same as that of any other financial market instrument: allowing the transfer of risk and reward from a willing seller to a willing buyer, and providing information to the rest of the world about such transfer to help further decisions to invest, produce, consume, and so on. Even with turbulence, market risks are preferable to Knightian uncertainty (Knight 1921). The successful product innovation of derivatives growth is twofold: more customized products to fit transfer preferences, more vanilla products to build liquidity and market information. The industry must balance both aspects for success.

Derivative structures are also part of the tool kit used by services offered by the financial industry. It is in such services that possible conflicts of interest are more likely.10 Last, although the management of counterparty and legal risks in derivative transactions has made tremendous progress, it is still an area of concern due to the size of these markets.

Further research: Exchanges and central clearing can improve liquid derivatives. What about public initiatives in favor of third-party collateral schemes11 to address the broader needs of bilateral contracts?

1.3.4 Funding

Historically, funding was somewhat an exception to the risk management trend to push the responsibilities of all risks as closely as possible to their originators. Trading desks have usually few options in their funding policy. The bank or institution treasury takes care of long-term cash provided by certain activities on the bid, while funding the debits of other desks at the offer: Term funding is typically not the problem of trading desks.

Moreover, financial models are doing the same by discounting risk-free future cash flows at short-term interbank “XIBOR” rates and using swap rates for medium and long-term maturities.

The global crisis of 2008 demonstrated how funding risk can derail the normal arbitrage relationship between cash and related unfunded products: The short-term London Interbank Offered Rate (LIBOR) is a real loan (and can incorporate a funding risk premium), while swap rates, which are unfunded, can significantly miss the point of prices driven by fire sales or unwinds of riskless but cash-funded products.

Further research should quantify how a global systemic short-term funding squeeze translates not only into temporary negative interest margins, but also fire-sale transactions on the basis of cash capital as term funding of last resort, prompting negative long-term swap spreads, large negative basis on cash bonds versus credit default swaps (CDSs), and so on.

1.3.5 Short Sales

In the 1990s crises (both Latin America 1994 and Asia 1997), short sales were blamed for declining markets, and authorities hastily changed the rules of equity lending and borrowing and short sales (Malaysia, Taiwan, etc.). Even though G-7 countries' markets have constant surveillance against price manipulation (as they should), similar moves happened in G-7 countries in the autumn of 2008: This is more surprising but obviously understandable. At the same time, the worst daily market moves (such as limit down days) occur when the panic of long investors finds no bid from a lack of short interest. Only the shorts are buying during crashes. Also, markets with no ability to sell short are more prone to the creation of bubbles and subsequent disasters (real estate being a prime example). In summary, the irony is that past short sales are the most natural financial market contracyclical mechanism.

In the future, we could see an interesting duality from regulators toward short sales: While market regulators continue to actively regulate the appropriate execution and reporting of short sales, shouldn't newly established systemic regulators want to encourage more efficient frameworks for term borrowing? And why not encourage a sufficient amount of long-term short interest?

1.3.6 Accounting

Accrual accounting was at the root of many disasters where banks had negative economic net worth while remaining liquid in the 1980s: Accrual accounting can allow poor management for too long. Fair value accounting was brought in to provide investors (and management) financial results where the flexibility of accrual accounting is replaced by binding market levels (directly or indirectly through predefined valuation models). Bank managers should have known that markets can be brutal; the rules applying to individual traders were suddenly applied at an entirely different scale, leading to severe systemic consequences. In particular, illiquid markets with granular positions are inefficient, and the unwinding of one losing position creates further losses and further forced sales.

Proposals seem to go in the direction of a middle ground: a simplification of doubtful refinements around available for sale (AFS), held to maturity (HTM), and so on, with the possibility of some management judgment to overrule aberrant market prices (either too high or too low), whenever necessary to reflect a reality based on prudent accounting, and not misleading automatic rules (IASB Exposure Draft 2009).

Further research could explore whether taxes can also play a role to promote prudent accounting, and also potentially reduce the volatility of tax revenues.

1.3.7 Legal

The proper handling of legal risks is critical for the financial industry where so much of the business relates to promises of future performance. To limit the temptation to walk away from wrong past decisions requires a strong legal framework. The capital markets are also very technical, and precise rules of compliance must be followed in order to prevent manipulations, trading on inside information, false rumors, and so on. The markets' legal environment has made great progress on all of this. The crisis pointed out a few important problems: strictly national bankruptcy laws where assets can be frozen in one country against the legitimate rights of alien owners (collateral transfers and rehypothecation) (King 2009). Also, certain derivatives terminations or interpretations by the nondefaulting counterparts of Lehman Brothers are controversial and being disputed (Ng and Spector 2009).

Immediate proposals call for broader use of transaction documents where the legalities are closer to the economic intent, and based on electronic format (FpML) instead of paper.

However, further research should review whether an international bankruptcy standard could be enforceable for asset conveyance—for example, by transferring the disputed asset in a third-party jurisdiction.

1.4 The Global Risk Management Challenge

In this last section are grouped issues that relate to the organization and control of the interactions or the communication of information between all the moving parts. Although they do not belong—strictly speaking—to the previous two sections, they participate directly or have great influence indirectly on the real world itself and the corresponding financial techniques.

1.4.1 Regulation

Basel I allowed a better capitalization of the banking systems following the crisis of the 1980s. Basel II was designed to correct Basel I undifferentiated risk weights, which created incentives for banks to take assets off balance sheets. Basel II greatly reduced these regulatory arbitrages but potentially increased systemic risks with the reliance on external ratings and models. The irony is that the subprime crisis—and the collapse of many off-balance-sheet structures12 inspired by Basel I—struck at the time Basel II was coming into effect.

It is critical that regulations anticipate and be aware of unintended consequences: Many of the toxic off-balance-sheet vehicles were a consequence of Basel I, and much of the demand for toxic highly rated fixed-income products was a consequence of Basel II.

Further research could explore how to address quickly regulatory weaknesses, which otherwise are systemic and procyclical by nature. A practical solution could be through fast-track specific additional disclosures required under Basel II's Pillar 3.

1.4.2 Diversification

“Since we have not more power of knowing the future than any other men, we have made many mistakes (who has not during the past five years?), but our mistakes have been errors of judgment and not of principle.”—J. P. Morgan Jr., excerpt from statement made before the U.S. Senate, 1933.

2. Managing Director, JPMorgan and Société Générale.

3. These are not reflective of views of former colleagues, clients of JPLC, or partners of the CRIS consortium.

4. The 5 percent rule of Article 122 of the EU Capital Requirement Directive (2009).

5. European Systemic Risk Board (ERSB) of the European Union.

6. It is legitimate to assume that systemic regulation is subordinated to diplomacy.

7. Such rules could also apply to limit the systemic risk of central clearing in periods of crisis.

8. Jean Bodin: Il n'est de richesses que d'hommes.

9. Improving bikes takes as much from driving experience as from pure engineering.

10. Derivative structures—with a dedicated derivative contractual setup—are opposed here to derivative products whose contractual setup is standardized.

11. Collateral of counterparts is held by an appropriate third party.

12. Structured investment vehicles (SIVs), conduits, and so on.

References

Acharya V., I. Hasan, and A. Saunders. 2002. Should banks be diversified? Evidence from individual bank loan portfolios. BIS Working Papers, September.

Berner, R., and B. Grow. 2008. They warned us about the mortgage crisis. BusinessWeek, October.

Billet, M., D. Mauer, and Y. Zhang. 2006. Stockholder and bondholder wealth effect of CEO incentives grants. University of Iowa, August.

Dimon, J. 2008. III—Fundamental causes and contributions to the financial crisis. JPMorgan Chase Annual Report, 14.

Duffie, D., and H. Zhu. 2009. Does a central clearing counterparty reduce counterparty risk? Stanford University, March.

The High-Level Group on Financial Supervision in the EU. 2009. Chaired by Jacques de Larosiére. Report (February): 24.

International Accounting Standards Board (IASB) Exposure Draft. 2009. Financial instruments: Classification and measurement (July).

King, M. 2009. Global banks are global in life, but national in death, [by] Mervyn King, Governor of the Bank of England. Financial Stability Report (June).

Knight, F. H. 1921.Risk, uncertainty and profit. Boston: Houghton Mifflin.

McCreevy, C. 2009. The credit crisis—Looking ahead. Institute of European Affairs, February.

Ng, S., and M. Spector. 2009. The specter of Lehman shadows trade partners. Wall Street Journal, September 17.

Tett, G. 2009. Insight: The clearing house rules. Financial Times, November 5.