An Empirical Evaluation of Structural Credit Risk Models
Nicola A. Tarashev, BIS working papers no. 179, July 2005
This paper uses firm-level data to compare the performance of six structural credit risk models in terms of how successfully they predict default. Structural models focus on the changes in value of a corporate obligor's assets and assume that default occurs when asset values cross some lower threshold. The paper finds that the models perform well but that they don't fully capture the effects of the business and credit cycles - and so could be improved by the incorporation of macroeconomic variables. Endogenous default models, which allow borrowers to "choose" the moment of default, perform better than exogenous default models - and also produce better and quite accurate estimates of optimal capital when filtered through the Basel II Internal Rating Based approach. The analysis suggests that the most material borrower characteristics are the firm's leverage ratio, default recovery rate, and risk-free rate of return.

Decomposing Credit Spreads
Rohan Churm and Nikolaos Panigirtzoglou, Bank of England working paper no. 253, 2005
This paper uses a structural model of credit risk to try to decompose credit spreads into their main components: expected default, uncertainty about the rate of default, liquidity, regulation and tax. It looks at how well the model fits historical default frequencies and calculates an average historical compensation for credit risk that can be compared to the average observed credit spread. The results show, among other things, that a large part of the credit spread of investment grade debt is due to non-credit risk factors, while the reverse is true for high-yield debt.

Top Commercial Real Estate Trends
SVP Commentary, SRC Insights, Second Quarter 2005
This short article looks at the rise in commercial real estate lending in the banks under the supervision of the Federal Reserve Bank of Philadelphia, noting that, as the trend is set to continue, banks must make more effort to manage their concentration risks. A concentration exists when loans with similar risk characteristics add up to more than 25% of the bank's capital structure (tier 1 plus ALL). At the moment, the writers of the paper see no evidence of emerging systemic problems with CRE in the banking sector, but they say they are seeing deals that could lead to unsustainable pricing and lower portfolio quality.

Loss Given Default and Economic Capital
Jon Frye, Capital and Market Insights, Federal Reserve Bank of Chicago, December 2004
It's becoming increasingly important for banks to understand the Loss Given Default risk parameter associated with their various credit portfolios. When estimating credit risk, a complicating factor is that the Loss Given Default parameter often tends to rise at the same time that the Probability of Default risk factor rises. This article explores this problem and offers a framework for understanding the coordinated rise of default and loss rates, making clear how various key concepts such as "expected LGD" and "stressed LGD" relate to economic capital. 

The Nature of Credit Risk in Project Finance
Marco Sorge, BIS Quarterly Review, 6 December 2004
Are longer maturity loans in project financing necessarily riskier than shorter maturity loans? This paper looks at some key characteristics of project finance, such as non-recourse debt, political risk, and ex ante spreads of project finance lending to argue that long-term project financing is not necessarily perceived by lenders as riskier than short-term project finance lending. The finding contrasts with most other forms of lending, where credit risk tends to increase with maturity. Instead, the paper finds that the particular characteristics of credit risk in project finance suggest a “hump-shaped” term structure of loan spreads.

The Syndicated Loan Market: Structure, Development and Implications
Blaise Gadanecz, BIS Quarterly Review, 6 December 2004
Since the 1990s, syndications have become a main source of funding for corporate borrowers. This paper reviews the development of the syndicated loan market, looking at both its history and at market structure. It compares the structure of various syndicated loan markets around the world, and looks at the impact of the advent of secondary loan trading and “relative value” trading.

Exploring the Relationship Between Credit Spreads and Default Probabilities
Mark Manning, Bank of England working paper, 17 August 2004
It is tempting to assume that the difference in price between a credit-risk-free bond and a credit risky bond is strongly related to the expectation of default. Yet, as this paper explains, the truth is more complex. The paper applies a structural credit model to explore the problem using data from Sterling bond markets. It concludes that the variability in the spreads of high-grade A-rated credits is not strongly related to default probability. Instead, the spread is determined by other factors such as liquidity. In the case of lower quality BBB-rated bonds, default risk seems to explain about a third to a half of spread variation, depending on the modeling technique that is applied. 

Market Dynamics Associated with Credit Ratings: A Literature Review
Fernando Gonzalez et al., European Central Bank, June 2004
Agency credit ratings are an important feature of the financial markets and they play an increasingly important role in bank credit risk management and regulatory capital calculations. This paper looks at the many conflicting demands that the markets and regulators have put on agency ratings. It says that ratings have become hard-wired into the financial system in various ways (e.g., rating based trigger clauses in lending covenants) that can have a significant effect on market dynamics. It also summarises rating methodology and takes a look at the problem of credit ratings and default time horizons.

The Importance of Effective Credit Cultures at Community Banks
Federal Reserve Bank of Philidelphia, 1^st Quarter 2004
The author of this paper says that regulators should start to worry if the CEO of a bank can’t describe clearly the credit culture of the bank. But what do we mean by “credit culture”? The author says that it is an amalgam of written policies and procedures with many intangibles such as traditions, skills, philosophies and standards. The article offers an easy-to-read reminder of the importance of credit culture, and questions the tendency to appoint bank CEOs who lack a firm grounding in credit.

The Great Depression as a Credit Boom Gone Wrong
BIS working papers no. 137, Barry Eichengreen and Kris Mitchener, September 2003
The 1990s boom has renewed economists’ interest in the role of credit in macroeconomic fluctuations. This paper asks how well quantitative measures of credit boom phenomena can explain the uneven expansion of the 1920s and the slump of the 1930s. The authors complement their macroeconomic analysis with sectoral studies on the property market, consumer durables industries and high-tech sectors. They conclude that the “credit boom view” provides a useful perspective on both the boom of the 1920s and the subsequent slump, especially where the organisation and history of the financial sector led intermediaries to compete aggressively in providing credit.

Trends in Households’ Aggregate Secured Debt
Rob Hamilton, Bank of England Quarterly Bulletin, Autumn 2003
The aggregate level of UK households' secured debt relative to their income has increased by about a quarter over the past five years, and has almost tripled since 1980. Using a simple model, this article concludes that most, though not all, of this increase can be accounted for by the spread of homeownership and the fall in inflation (which has reduced the rate at which households' real debt burden is eroded over time). But the model also suggests that, because only a relatively small fraction of the housing stock changes hands each year, the aggregate level of debt responds relatively slowly to changes in house prices. So the recent increases in UK house prices could lead to continuing increases in the debt-to-income ratio over the next five to ten years.

Do Australian Households Borrow Too Much?
IJ Macfarlane, Reserve Bank of Australia Bulletin, April 2003
This article says that the large increase in household debt in Australia seen in recent years is largely a function of increased borrowing for housing. This in turn is largely driven by low interest rates. The maximum risk a typical household faces during its life cycle is larger than it formerly was mainly because more households are now staying at or near their maximum risk position for a longer period. The article concludes that the rise in household debt is unlikely to lead to a major crisis in financial institutions (such as occurred in the early 1990s after the build-up in corporate debt). But household consumption will be a lot more sensitive to economic conditions than hitherto, so we’ll see a more pronounced cutback in consumption if adverse economic conditions do occur.

An Empirical Test of a Two-Factor Mortgage Valuation Model: How Much Do House Prices Matter?
Chris Downing, Richard Stanton and Nancy Wallace, Federal Reserve Board et al, April 2003
The authors of this paper point out that most of the models currently used for pricing and hedging MBS are based solely on their most important factor, interest rates.  But there’s reason to think that this omits some important pricing variables – and the leading candidate is the level of house prices. Not only do housing prices affect the level of default, but a low house price also affects the borrower’s ability to qualify for a new loan.  The authors develop and empirically test a mortgage valuation model that takes account of these issues. They say their structural model should produce sensible results even when it is applied in economic environments unlike anything seen in the past, since the underlying factors are linked to predicted behavior through the optimizing behavior of agents in the model.

Credit Card Securitization and Regulatory Arbitrage
Charles Calomiris and Joseph Mason, Columbia Business School and Drexel University, April 2003
This paper explores off-balance-sheet financing of credit card receivables by banks. The authors find that such securitizations result in some transfer of risk out of the originating bank, but that risk often remains in the securitizing bank as a result of “implicit recourse”. Securitization with implicit recourse provides an important means of avoiding minimum capital requirements. The authors also find, however, that securitizing banks set their capital relative to managed assets according to market perceptions of their risk: they do not seem to be motivated by maximizing implicit subsidies related to the government safety net. Thus, the evidence is more consistent with the “efficient contracting” view of securitization with implicit recourse than with the “safety net abuse” view. Concerns expressed by policymakers about this form of capital requirement avoidance appear to be overstated.

Economic and Regulatory Capital Allocation for Revolving Retail Exposures
Roberto Perli and William Nayda, Federal Reserve Board and Capital One, Federal Reserve discussion paper, July 2003
The authors of this paper say the latest revision of the Internal Ratings Based approach of the Basel Committee on Banking Supervision’s New Capital Accord Proposal for retail portfolios contains a significant innovation relative to previous versions: the recognition that, for revolving credits, future margin income will be available to cover losses before a bank’s capital is threatened. They assemble a mini-portfolio of revolving exposures and compare the capital charges generated by the latest Basel’s formula with the capital charges generated by two possible earnings-at-risk internal capital allocation models, finding that Basel’s capital ratios are closer to those generated by the models for the groups with lower credit risk.

Analyzing and Explaining Default Recovery Rates: A Report Submitted to ISDA
Edward Altman, Andrea Resti and Andrea Sironi, Stern School of Business NYU et al., December 2001
This report analyzes the impact of various assumptions about the association between aggregate default probabilities and the loss given default on bank loans and corporate bonds, and seeks to explain this critical relationship empirically. The analysis has implications for the results of various value-at-risk credit risk models as well as the fundamental factors that influence fixed income portfolio models and strategies. Traditionally, credit risk researchers have treated the important recovery rate variable as a function of historic average default recovery rates, conditioned perhaps on seniority and collateral factors, and in almost all cases as independent of expected or actual default rates. The authors of this paper, after reviewing the LGD problem and literature in some detail, examine corporate recovery rates on bonds in the period 1982-2000, and instead attempt to explain recovery rates as a function of supply and demand for securities. (This paper by Edward Altman written in 1998 looks at the market for distressed debt including bank loans, while this paper from April 2002 looks at the links between default and recovery rates.)

Understanding Stochastic Exposures and LGDs in Portfolio Credit Risk
Dan Rosen and Marina Sidelnikova, Algorithmics Research Quarterly, Spring 2002
This paper presents a case study on the impact of stochastic exposures and losses given default (LGD) on portfolio credit-risk estimation. The paper notes that four factors have a substantial effect on credit losses: exposure (market) volatility, credit correlations, market-credit correlations, and portfolio granularity. The authors emphasize the importance of treating stochastic exposures for economic and regulatory capital properly, discussing the limitations of recent regulatory proposals when market correlations affect exposures/LGDs and when market and credit risk are correlated. Although the examples in this paper use portfolios of derivatives, the techniques and results apply equally to other cases where LGDs, exposures and spreads are stochastic.

Recovery Rates
RiskWorX, 2002
This paper reviews the literature and software associated with the practical problem of estimating loss given default in bank portfolios. The paper summarises major empirical studies, looks at the problem of modeling recovery rates, and analyses capital relief on collateral as proposed under the New Basel Capital Accord. The paper’s conclusion tentatively suggests a number of practical steps forward for banks that are just beginning to grapple with this problem.

Depressing Recoveries
Jon Frye, Federal Reserve Bank of Chicago (from Risk Magazine, vol. 13, no. 11, November 2000)
This short but seminal article points out that if a borrower defaults on a loan, the bank’s recovery depends on the value of the collateral associated with the loan, which is itself likely to be reduced in poor economic conditions. The article points out that taking this default/LGD correlation into account complicates credit models, and goes on to offer an approximate solution that is dependent upon expected loss estimates.

Recovery Rates in the Leasing Industry
Mathias Schmit and Julien Stuyck, Leaseurope, September 2002
This paper reports on the recovery rates associated with three types of assets –  automotive, industrial and business equipment, and real estate – in the European leasing market. The analysis is based on a 37,259 defaulted lease contracts issued between 1976 and 2002 by 12 major European financial institutions. The authors compare their results with recovery rates for bonds and bank loans presented by seniority class, and test the independence of recovery rates with economic conditions. The paper comments on the current Basel Capital Proposal, which considers recovery rates as a major input in its advanced Internal Rating-Based Approach. The main conclusion is that the loss in the event of default is quite low, and that automotive leasing shows better recovery rates than the equipment segment; loss given default tends to fall when the age of the contract is far from the origination date of the lease (when recovery rates can even exceed 100%).

Secured Creditor Recovery Rates from Management Buy-outs in Distress
David Citron et al, City University Business School et al, 2002
Buy-out literature suggests that secured creditors will recoup substantial proportions of the funds they extend to finance the initial buy-out. This paper uses a unique dataset of 42 failed management buy-outs to examine the extent of credit recovery by secured lenders under UK insolvency procedures and the factors that influence the extent of this recovery. On average, secured creditors recover 62 per cent of the amount owed. The percentage of secured credit recovered rises where the distressed buy-outs is sold as a going concern and where the principal reason for failure concerns managerial factors.

LossCalc: Moody’s Model for Predicting Loss Given Default (LGD)
Greg Gupton and Roger Stein, Moody’s, February 2002
This report describes Moody’s proprietary model for predicting loss given default in bonds, loans and preferred stock. The authors point out that traditionally much more effort has been expended on calculating default risk than loss given default, though both are critical risk factors in predicting credit losses. As a result, LGD is often estimated using historical averages, an approach that offers only a very approximate rule of thumb. The report is focused on describing Moody’s model, but offers various other remarks on key issues such as defining the moment in time when any loss on a credit-risky instrument might be calculated. (Moody’s November 2000 paper on bank loan loss statistics is here, while this page offers other Moody’s papers on recovery rates.)

Default and Recovery Rates of Corporate Bond Issuers: A Statistical Review of Moody’s Ratings Performance 1970-2001
David Hamilton, Moody’s Investor Services, February 2002
This report, Moody’s 15th annual study of global corporate defaults and ratings performance, reviews the default, recovery and credit loss experience for 2001 and the historical period since 1970. It finds that default rates reached decade highs in 2001, with the default rate ending 2001 at 3.7% for all Moody’s-rated corporate bond issuers and reaching 10.2% for speculative-grade rated issuers. Despite the default of some fallen angels, rating downgrades for investment-grade issuers were not abnormally higher than average in 2001. The average recovery rate of defaulted bonds fell for the third straight year to a record low of 21% of par, although there were notably high recovery rates for a handful of fallen angel issues. The 2001 data reinforced the observation that recovery rates are inversely correlated with annual default rates, which implies that credit loss rates rise even more than default rates when default rates increase.

Suddenly Structure Mattered: Insights into Recoveries from Defaulted Debt
Karen Van de Castle, David Keisman and Ruth Yang, Fall 2001
This study analyses default loss experience for a sample of 690 bonds and 264 bank loans between 1987 and 1996. It is an update of a Standard and Poor's study that first reported that the amount of debt subordinate to an obligation has a powerful influence on that obligation's loss in default. This paper finds that the effects of such structural features as the amount of subordinated debt and the quality of collateral have grown larger since 1990.

To Provision Or Not To Provision
Claude Borio and Philip Lowe, BIS Quarterly Review, September 2001
As the authors of this paper point out, provisioning practices at banks have come under increased scrutiny over recent years from accounting and taxation authorities and from financial supervisors. This interest is evident in a number of recent policy proposals including the development of an International Accounting Standard (IAS 39) on provisioning and loan valuation; issuance of sound principles in loan accounting by banking’s Basel Committee; the proposed New Basel Accord; statistical provisioning regimes in some countries; and proposals by a Joint Working Group of standard setters to introduce fair value accounting for all financial instruments.  This article discusses the main characteristics of these recent initiatives, paying particular attention to the tensions amongst them, and lays out a simple framework within which the various proposals can be embedded. It includes worked examples of loan valuation and provisioning under various regimes.

Loan Loss Provisions and Economic Slowdowns: Too Much Too Late?
Luc Laeven and Giovanni Majnoni, FRB conference paper, March 2002
This paper explains that econometric evidence shows that banks tend to postpone provisioning when faced with favorable cyclical and income conditions, until negative conditions set in. The paper suggests that risk-based regulation of loan loss provisions and reserves might lead to a beneficial dampening of the pro-cyclical effects of capital regulation. For the moment, the very different regulatory and institutional frameworks around the world lead banks in various countries to behave very differently. Bank regulators recognize that more explicit recognition should be paid to the problems associated with inadequate provisioning policies, but the solution is not easy to define due to the complicated interaction of accounting, fiscal and prudential requirements. (A comparison of different accounting standards set out as a table is available in this July 2002 paper.)

Bank Provisioning: the UK Experience
Glenn Hoggarth and Darren Pain, Financial Stability Review, Bank of England, June 2002
Bank provisions made in recognition of a deterioration in loan quality can have a significant impact on banks’ earnings and capital and are a leading indicator of a general deterioration in credit. This article examines the factors that may, in the past, have influenced provisioning by the major UK banks. It suggests that macroeconomic conditions played a particularly important role. Bank-specific factors such as the sectoral concentration of debt, especially if in risky sectors such as commercial property, were also influential.

Dynamic Provisioning: Issues and Application
Fiona Mann and Ian Michael, Financial Stability Review, Bank of England, December 2002
This article looks at the pros and cons of forward-looking ‘dynamic’ provisioning regimes in the banking industry. Under such a regime, banks make provisions based on the losses expected when loans are originated. This should deliver a rising stock of provisions when actual loan losses are unusually low, thus helping to protect banks in periods when actual losses turn out to be high. In addition, banks’ income statements would be less distorted in periods when actual losses were significantly higher or lower than the long-run expected level.

Accounting for Impaired Assets in Bank Credit Analysis
Roger B. Taillon, Standard & Poor’s Rating Direct Reprint, July 2002
Accounting for impaired assets not only differs markedly from country to country, it also offers substantial scope for management judgment. This article explains how Standard & Poor’s approaches this potentially confusing area. The author says that accounting for impaired assets can have a profound impact on bank balance sheets and income statements. To rate banks fairly, credit analysts must ask and answer the following questions: What is the definition of an impaired asset? To what extent is interest accrued on impaired assets? What is the policy for providing or reserving against losses on impaired assets? What is the policy for finally writing off impaired assets?

Loss Underreporting and the Auditing Role of Bank Exams
Jeffery W. Gunther and Robert R. Moore, Federal Reserve Bank of Dallas, March 2002
Using a unique set of banking data that contains both originally reported and subsequently revised financial variables, this article examines the incidence of adverse revisions to accounting statements. It seems that, as might be expected, banks are more likely to underreport financial losses when their financial performance is substandard. But the paper highlights the “significant auditing effect” of supervisory exams in uncovering financial problems and ensuring that bank accounting statements reflect them, even at highly rated banks. The introduction to the paper offers a clear account of the relationship between bank provisioning, bank call reports and bank exams, together with a review of research on these topics and on how they relate to share price movements in banks.

The New Capital Accord and Internal Bank Ratings
Donald van Deventer and Jaqueline Outram, Kamakura Corporation, working paper, 25 May 2002 (free registration required)
The New Basel Capital Accord has refocused attention on the quality of the credit ratings process inside banks. This paper offers a broad overview of the purpose of internal credit ratings, and relates this to other forms of credit assessment and modelling. It then offers a discussion of how internal ratings and other ratings methodologies might be tested. The paper points out that the New Accord’s testing requirement is likely to have a significant impact on the rate at which credit models are developed and applied as ratings tools.

Using Credit Risk Models for Regulatory Capital: Issues and Options
Beverley Hirtle, Mark Levonian, Marc Saidenberg et al., Federal Reserve, FRBNY Economic Policy Review, March 2001
This paper, published in 2001, describes from a regulator’s standpoint the issues associated with the development of regulatory minimum capital standards for credit risk based on banks' internal risk measurement models. The authors focus on three main areas: prudential standards that define the risk estimate to be used in the capital requirements; model standards (ie. the essential components of a comprehensive credit risk model); and validation techniques that could be used by supervisors and banks to assess model accuracy.

Revamping Your Credit Rating System
Peter Nakada, ERisk Report (free sample issue), January 2002
This article discusses six key issues as banks prepare a strategy for restructuring their internal ratings system in line with new regulatory approaches put forward by both US regulators and the Basel Committee. It looks at why banks need to act now; how to compare alternative strategies; key reasons for restructuring a ratings system; getting decisions in the right order; and how to relate business management benefits to each alternative regulatory-compliance strategy.

Analysing and Combining Multiple Credit Assessments of Financial Institutions
Evangelos Tabakis and Anna Vinci, European Central Bank, working paper no 123, February 2002
The authors of this paper aim to provide a useful tool for bank credit risk managers and supervisors looking to evaluate the quality of credit assessments under the standardised and the internal ratings based approach of the New Basle Capital Accord. The paper proposes a methodology for combining external credit assessments to create benchmark ratings based on multiple sources of information. The authors model the credit assessments as products of published financial data, plus a subjective “analyst contribution”. The paper also reviews recent academic work on external credit assessments, and summarises the key points and potential impact of Basel II.

A Risk-Factor Model Foundation for Ratings Based Bank Capital Rules
Michael B. Gordy, Board of Governors of the Federal Reserve System, October 2002
The New Basel Capital Accord offers a more sophisticated approach to measuring risk than the 1988 Accord, in terms of its ability to capture the risk of individual credits using an internal ratings-based approach. But the New Accord will not allow banks to use portfolio and economic capital models that take into account diversification benefits for the purpose of calculating minimum capital requirements. This interesting paper explores the importance of this limitation of the new rules, and suggests a way in which simple ratings-based capital rules can be reconciled with the general class of credit VaR/portfolio models.

Benchmarking Default Prediction Models: Pitfalls and Remedies in Model Validation
Roger Stein, Moody’s KMV, revised 13 June 2002
This technical article discusses several strategies for testing the performance of default models. It differentiates between testing for the ‘power’ of the underlying model and the ‘calibration’ of that model. This is important so that practitioners can first identify the most powerful model available, and then calibrate this powerful model – re-calibrating it if necessary. The paper also suggests: testing models against reasonable benchmarks to understand better the relative scale of performance statistics on the data set on hand; taking care to ensure that the population from which the test sample is drawn is similar in its characteristics to the population on which it will be applied; and evaluating the variability of performance results through appropriate resampling experiments.

Methodology for Testing the Level of the EDF™ Credit Measure
Matthew Kurbat & Irina Korablev, Moody’s KMV, revised 8 August 2002
As the authors of this technical paper point out, users of default risk models need to verify that the average level of predicted defaults does indeed track the level of realized defaults over time. The paper looks at the particular problem of verifying the calibration of a set of proprietary equity-market based models offered by Moody’s KMV. The authors say that a straightforward approach to level validation – comparing mean predicted default rates with empirically observed default rates to see that they match – is made problematic by the skewed distribution of mean default rates. The paper describes how to perform level validation in the face of such skewness. It compares the levels of the models’ ‘expected default frequency’ credit measure with realized default rates from 1991-2001, and finds that the predicted and realized levels match within limits of sampling error.

Credit Risk Measurement and Pro-cyclicality
Philip Lowe, BIS working paper no. 116, September 2002
This paper tackles the thorny issue of whether, by improving credit risk measurement and linking capital more tightly to economic risk, the banking industry will deepen any credit crunches in a recession. The dangers exists because as the economy dips, default rates climb, and banks are likely to have to set aside a larger amount of risk capital against each loan. This might discourage lending at a critical point in the business cycle. The paper also offers a useful overview of bank ratings and credit modelling systems and how these all relate to one another.

A Survey of the Cyclical Effects in Credit Risk Measurement Models
Linda Allen and Anthony Saunders, Zicklin School of Business and Stern School, NYU, May 2002
One of the most contentious problems in bank credit risk management is the fear that a more precise tailoring of bank capital to credit risk exposures might worsen any credit crunches. The worry is that if banks have to set aside more risk capital as credit quality deteriorates, they will be discouraged from lending and deepen any national or global economic slowdown. The authors of this paper examine the treatment of cyclical factors in both academic and proprietary credit risk measurement models, and discuss what is meant by the so-called ‘procyclicality’ effect. They also describe models that examine the recovery rate as a function of macroeconomic factors; examine the correlation between probability of default and loss given default; and look at the relationship of all this to ‘exposure at default’ numbers.

Internal Ratings, The Business Cycle and Capital Requirements: Some Evidence From An Emerging Market Economy
Miguel Angel Segoviano and Philip Lowe, London School of Economics (FMG) and BIS, Federal Reserve Bank of Boston conference paper, April 2002
This paper uses a unique dataset from the Mexican banking industry in the later 1990s to throw some light on the question of how much regulatory capital will vary through the economic cycle under the proposed New Basel Capital Accord. It shows that regulatory capital would indeed have varied considerably over the economic cycle, particularly for banks with credits of relatively low quality; that calibration and verification of internal ratings is problematic, at least in emerging countries; and that banks might have to carry substantial ‘buffer’ capital if risk-sensitive capital rules are not to worsen the economic cycle. The authors also introduce the idea that banks might need to disclose the results of macroeconomic stress testing.

Procyclicality and the New Basle Accord – Bank’s Choice of Loan Rating System
Eva Catarineu-Rabell, Patricia Jackson, and Dimitrios P. Tsomocos, Federal Reserve Bank of Boston conference paper, April 2002
This article examines the implications of the new risk-based requirements of the proposed New Basle Capital Accord with regard to ‘procyclicality’. It also asks whether the choice of particular loan rating system by the banks would significantly increase the likelihood of sharp increases in capital requirements in recessions, creating the potential for classic credit crunches. The paper finds that rating schemes which are designed to be stable over the cycle, akin to those of the external rating agencies, would not increase procyclicality but ratings which are conditioned on the point in the cycle, akin to a Merton approach, could substantially increase procyclicality. The paper then explores whether banks are likely to choose to use a counter-cyclical, procyclical or neutral rating scheme. It finds that banks would likely not choose a stable rating approach – a finding with important implications for the new Accord.

The Link between Default and Recovery Rates: Implications for Credit Risk Models and Procyclicality
Edward I. Altman, Brooks Brady, Andrea Resti and Andrea Sironi, Stern School of Business, S&P Risk Solutions, et al, April 2002
This paper analyzes the impact of various assumptions about the links between aggregate default probabilities and the loss given default on bank loans and corporate bonds. It examines the literature of the last three decades with regard to how various credit models treat the recovery rate variable. It presents simulation results under three different recovery rate scenarios and examines the impact of these scenarios on the resulting risk measures, showing a significant increase in both expected and unexpected losses; and it examines the recovery rates on corporate bond defaults from 1982 to 2000. Finally, the paper analyzes how the link between default probability and recovery risk would affect the pro-cyclicality effects of the New Basel Capital Accord. It suggests that, if banks use their own estimates of LGD (as in the “advanced” IRB approach), an increase in the sensitivity of banks’ LGD due to the variation in PD over economic cycles is likely to follow.

The Dependence of Recovery Rates and Defaults
Yen-Ting Hu and William Perraudin, Birkbeck College and Bank of England, working paper, February 2002
Most banking practitioners intuitively understand that the likelihood of a default on a loan is also connected to how much the bank is likely to recover if such a default does indeed occur. But, as the authors of this paper point out, standard ratings-based models for analysing credit portfolios and pricing credit derivatives assume that defaults and recoveries are statistically independent. This paper uses Moody’s rated-bond data for January 1971 to January 2000 to show that default rates and recovery rates are, indeed, negatively correlated. Unlike earlier such analyses, the authors attempt to take into account the changing nature of the Moody’s bond universe (in terms in industry sector). They also deploy Extreme Value Theory techniques to explore the tail behaviour of total credit loss distributions in their analysis, pointing out that the degree of dependence between extreme realisations of default rates and recovery rates has a special importance for quantifying bank capital adequacy.

Financial Distress and Bank Lending Relationships
Sandeep Dahiya, Anthony Saunders and Anand Srinivasan, Georgetown University, New York University, University of Georgia, The Journal Of Finance, forthcoming 2002
How important is the news that a single borrower might default on a bank? There are reasons why it might be relatively unimportant (eg, low exposure relative to total bank exposure, relatively high recovery rates) and reasons why it might be very damaging (eg, reputational effect, implications for total bank portfolio). This paper is the first large-sample analysis of the wealth effects for lead bank shareholders when bank borrowers face financial distress. It shows that, for a lead bank, the news of default of a corporate borrower is associated with an average decline of 3.8 per cent in its stock returns over an 11-day period surrounding the date of default. News of a corporate bankruptcy is associated with a decline in bank stock returns of 1.8 per cent over a similar 11-day window. The price decline is more marked for banks with relatively high exposures. However, on average, abnormal returns to banks, on the announcement of a borrower's financial distress, are significantly and negatively related to existence of a prior past borrowing relationship with that borrower (ie, relationship banking is valuable for lenders).

Consultative Paper on Stress Testing
Monetary Authority of Singapore (MAS), January 2002
A bank's credit risk management efforts are incomplete if they do not include a comprehensive stress-testing programme, say the authors of this consultative paper. The paper defines in plain language what is meant by stress-testing in financial institutions, and offers some straightforward guidelines to a conducting stress tests: a subject that is often seen as highly quantitative. The authors point out that, although banks that possess counterparty risk rating and credit portfolio risk management tools can take better advantage of stress-testing techniques, much stress-testing best practice is equally applicable to banks that have yet to build quantitative risk management systems. The biggest handicap of stress-testing using only a qualitative risk rating system is that it is time-consuming, and that the correlations between various loans/securities in the portfolio often have to be ignored. But significant risk management benefits can still be obtained.

The United Kingdom's Small Banks' Crisis of the Early 1990s: What Were the Leading Indicators of Failure?
Andrew Logan, Bank of England, Working Paper Series No. 139, July 2001
The announcement of BCCI's closure on 5 July 1991 - see this ERisk case history - rapidly accelerated the withdrawal of wholesale funds from small and medium-sized UK banks. The authors of this study point out that within three years, a quarter of the banks in this sector (albeit a less significant sector in the UK than it is in the US) had, in some sense, failed. Their study employs a logit model to analyse the distinct characteristics of the banks that failed compared with those that survived. Perhaps not surprisingly, a number of measures of bank weakness (low loan growth, poor profitability and illiquidity) are found to be good short-term predictors of failure, as are a high dependence on net interest income and, less intuitively, low leverage. The best longer-term leading indicator of future failure, however, is rapid loan growth at the peak of the previous boom. Unlike the survivors, banks that subsequently failed thus exhibited a pronounced boom-and-bust cycle in lending growth.

Modelling Financial Instability: A Survey of the Literature
Alexandra Lai, Bank of Canada working paper, May 2002
This paper summarises the recent literature on modelling financial instability, with a special focus on what this means for the banking system. It includes a section on the literature of bank contagion, which it analyses in terms of both credit linkages and information contagion. The authors conclude that although the literature on financial crises sheds much light on the origins of financial instability, not enough work has been done to advance our understanding of financial systems as consisting of both markets and institutions. It concludes that most theories of financial markets ignore institutions, while theories of banking ignore asset markets.

Post-resolution Treatment of Depositors at Failed Banks: Implications For the Severity of Banking Crises, Systemic Risk and Too Big To Fail
George Kaufman and Steven Seelig, Loyola University and IMF, Economic Perspectives (Federal Reserve Bank of Chicago), Second Quarter 2002
How insured and uninsured depositors are treated - and expect to be treated - has significant repercussions on bank stability in uncertain times. In particular, it's not always clear how quickly the financial authorities should step in to manage a crisis at a bank, and how quickly insured depositors can expect to access their money. This article reviews the problem in some depth, and argues that stepping in to resolve a crisis at a bank 'before or shortly after their net worth turns negative' and 'providing full and immediate access for insured depositors' makes the most sense from the perspective of system stability. It argues that this policy should also mean that even 'too big to fail' banks can be treated just like any other bank.

Enterprise Credit Risk Using Mark-to-Future
Algorithmics, Fall 2001
This electronic book, presented in 17 PDF chapters, offers in-depth comment on many leading-edge issues in credit risk modelling. It is introduced and partly written by researchers at Algorithmics, a software vendor with roots in the derivatives industry, with additional chapters from ratings agencies Moody’s and Standard & Poor’s. Topics include assessing creditworthiness, default correlation and loss given default; valuation of credit-risky assets; credit exposure measurement and control including derivative credit risk; and portfolio risk management techniques. The tone of the chapters varies from introductory to quite technical. As the book’s title suggests, some of the chapters are angled towards Algorithmic’s specific ‘mark-to-future’ risk philosophy, but as a whole the texts offer an excellent and unusual web resource.

Mark To Market, Oversight and Sensitivity Analysis of CDOs
Jorge Mina, RiskMetrics, December 2001 (free registration required)
A collateralised debt obligation (CDO) is a securitisation in which risky assets such as commercial loans and corporate bonds are sold to a special purpose vehicle, which in turn uses them as collateral to back the cashflows on notes sold to investors. Commercial banks increasingly use such instruments to transfer risk to investors – after chopping up the risk into different tranches to suit different investor preferences – and thus shrink their balance sheet and reduce their regulatory capital. But understanding the implications of complex CDO structures represents a challenge for investors that hold them in their portfolios. This paper introduces the key issues. It presents a detailed argument for full cash-flow simulation and scenario analysis, with typical scenarios consisting of a predetermined number of defaults and credit downgrades. The authors say this will improve valuation and risk monitoring, and help kick-start the secondary market in CDOs.

Counterparty Risk and the Pricing of Defaultable Securities
To access from the above webpage, click Papers button then Counterparty option
Robert Jarrow and Fan Yu, Cornell University and University of California, October 2001
This paper introduces a special twist on the concept of counterparty risk, which it defines as the risk that the default of a firm’s counterparty might affect the initial firm’s own default probability. In this special sense, counterparty risk is driven by the unique relationship of a firm with other firms in the economy. The authors argue that taking this risk into account is important for the pricing of defaultable bonds and credit derivatives. It’s also necessary if credit modelling is to take account of the clustering of defaults and default risks that we can see in the real world, such as the interlinked non-performing loans of certain South Korean industrial conglomerates during that country’s banking crisis. This kind of cascading of risk is more likely during a recession, but is not explained by business cycle risk alone. The authors say their discussion has important implications for the management of credit risk.

Securitization and the Efficacy of Monetary Policy
Arturo Estella, Federal Reserve Bank of New York, Economic Policy Review, Forthcoming 2002
A boom in mortgage securitisations that took hold of the markets in the 1970s is now in its mature phase, but the growth of credit risk securitisation in other markets has been as vigorous in the past five years as it was in the mortgage markets two decades ago. This paper investigates what this might mean for monetary policy. In particular, when the central bank makes a specific monetary policy move – such as increasing the overnight interbank rate by 50 basis points – is the ultimate effect on GDP different from what it would have been in the 1960s? The paper suggests that an important part of the transmission of policy through the mortgage markets occurs via liquidity and credit volumes, as opposed to the more commonly cited interest rate effects. It demonstrates ‘securitisation effects’ in the mortgage markets, and argues that these may be expected to extend to other markets as well, as new forms of securitisation allow banks to take assets off their balance sheets (and so provide banks with greater flexibility in their funding).

Houses as Collateral: Has the Link Between House Prices and Consumption in the UK Changed?
Kosuke Aoki, James Proudman and Gertjan Vlieghe, Federal Reserve Bank of New York, Economic Policy Review, Spring 2002
In the UK, in particular, there is a strong relationship between house prices, the business cycle and inflation. This is plain from the pronounced pro-cyclical pattern of house prices. This paper argues that while house prices are not a source of fundamental shocks, they are part of a wider process for transmitting shocks to the economy. The paper offers a wide-ranging discussion of this process, and also asks whether new retail credit products might have affected the link between house prices and consumption. The authors find that, in some circumstances, when presented with the same economic shocks, house prices are likely to move less than they did previously. Conversely, economists should not conclude that credit shocks have become smaller if house price movements become smaller – instead, other forms of consumption might be affected.

Modeling Sovereign Yield Spreads: A Case Study of Russian Debt
Darrell Duffie, Lasse Pedersen, Kenneth Singleton, revised September 2001
The default of a sovereign country is a political decision, in which governments balance the cost of making debt payments against the reputational cost of a failing to honour their promises. Unlike corporate bankruptcies, the willingness to pay is more important than the ability to pay. Section One of this academic paper succinctly reviews the problems this throws up for risk modellers, discusses the thorny problem of internal versus external debt, and surveys the relevant literature. The rest of the paper is a technical attempt to model the term structure of credit spreads for bonds of the Russian Ministry of Finance, in relation to the Russia default crisis of 1998. The aim is a pragmatic model that takes into account the default, restructuring and illiquidity risks of sovereign debt. Other Darrell Duffie articles and working papers here, some with downloads.

Rating Agency Actions and the Pricing of Debt and Equity of European Banks: What Can We Infer About Private Sector Monitoring of Bank Soundness?
Reint Gropp and Anthony J. Richards, ECB, August 2001
Market discipline is to be a significant component of the regulation of bank behaviour, according to the Basle II regulatory reforms. But that means regulators and bankers must better understand the relationship between the information about banks that is made available to the markets, and movements in bank equity and bond prices. This European Central bank working paper reviews the relevant literature, and asks whether changes in the ratings of European banks by rating agencies can at present be shown to exert any ‘market discipline’ effects. It uses some limited new empirical research to suggest that ratings announcements are regarded as having some informational value in the stock markets, but that they don’t have much effect on bond prices – possibly because rated banks are regarded as ‘too big to fail’.

Fear by Default
Robert Matthews, Royal & SunAlliance, July 2001
This short article looks at the maturing European high-yield debt market as one example of the more general relationship between debt and equity capital. From a risk-taker’s perspective, it explains why companies such as BT – once rated AAA – have been content to allow their credit ratings drop away in the pursuit of shareholder value, and argues that different definitions of default mean that the risk of investing in high-yield debt is easily misinterpreted. It also asks whether equity investment can any longer support investors hungry for income as bondholders continue to ‘extract the marrow from the best of the company market’.

Modelling Default Risk
Peter Crosbie, KMV, June 2001
This paper uses equity market examples from 2001 to reprise the influential approach to default modelling developed by California consultancy KMV in the 1990s. This approach determines default probability using information from the market price of a firm’s equity. The paper explains that the default point, the asset value at which the firm will default, generally lies somewhere between total liabilities and current, or short-term, liabilities. Using specific examples, it shows how the different default probabilities of two firms are driven by the difference in the risks of their businesses rather than by their respective asset values or leverages. It offers an unusually clear explanation of the fundamental relationship between equity volatility, asset volatility, leverage and business risk and explains how this relationship can be traced across sectors as different as banking and computer software. Underlying the paper is the insight that debt and equity “are both really derivative securities on the underlying assets of the firm”. Other KMV papers here.

Credit Ratings and the BIS Reform Agenda
Edward Altman and Anthony Saunders, Stern School of Business, NYU, March 2001
This working paper by two leading credit risk academics looks at the BIS proposals for reforming bank capital adequacy. It focuses on the mechanism underlying the regulators’ ‘standardised’ model and offers a critique of the logic behind the risk weighting of the mechanism. It also offers general insights into the relative riskiness of different grades of credit ratings. More Altman papers on ‘z’ scores, bonds, distressed debt and ratings migration here.

Benchmarking Quantitative Default Risk Models: A Validation Methodology
Moody’s, March 2000
The problem of how to verify a credit rating model is something that banks have become more interested in following the Basle II capital adequacy reform proposals. But it’s a problem that’s always been an issue for the public rating agencies. This article presents a summary of the approach Moody's uses to validate and benchmark a series of popular quantitative default risk models, including its own model for public companies. The paper discusses performance measurement and sampling techniques, as well as other practical considerations associated with performance evaluation for quantitative credit risk models. Other recent Moody’s quantitative risk analytics papers here.

Bank Loan Loss Given Default
Moody’s, November 2000
When a company defaults on its bank loan it doesn’t mean that the bank loses all its money. History tells us that, after 18 months or so, there’s usually a resolution that gives back to the bank a substantial fraction of its original capital. But how substantial, and how reliable, is this recovery rate? This paper from public rating agency Moody’s uses 181 loans defaulted loans to look at the secondary market price quotes for these assets. The study finds that the mean bank loan value in default is 69.5 per cent for senior secured debt, and 52.1 per cent for senior unsecured. This is an average, however, and the lowest 10th percentiles of recoveries are 39.2 per cent and 5.8 per cent respectively. Other more recent Moody’s quantitative risk analytics papers here.

Feedback on Survey of Banks' Preparation for the New Basle Capital Accord
Domenic Carratu, September 2001
This informal survey of some 25 UK and European banks says the more advanced will begin detailed implementation of Basle II between early 2002 and end 2003, though many smaller banks have not developed a plan and few banks have worked out budgets. It sees data gathering and management as the largest single element of work, and says demand for the Advanced IRB approach will be strong.

Making Internal Ratings Work
Michel Crouhy, Robert Mark, January 2001
The regulators are offering banks both carrots and sticks in their proposals for reforming regulatory capital. In this article, Dr Bob Mark and Dr.Michel Crouhy of CIBC discuss the implications for banks of the internal ratings based approach described in the latest proposals released on January 16.

Supervisory Standards for Internal Rating Systems
William Treacy, August 2000
This paper discusses the emerging framework of standards that will be used as critical reference points by Federal Reserve examiners in evaluating the strength of banks’ internal rating systems.

©2008 Sungard. All rights reserved. Legal Information