IFRS 9 aims to streamline and strengthen risk measurement and reporting of financial instruments in an efficient, forward-looking manner. This new accounting standard will have far-reaching impacts on accounting practices and performance results. This article focuses specifically on the IFRS 9 impairment model and challenges in interpreting the IFRS 9 requirements. We suggest solutions for meeting requirements in areas such as portfolio segmentation, thresholds for transitions among impairment stages, and calculating expected credit losses, leveraging Moody's Analytics expertise in credit risk modeling.
IASB published the final version of IFRS 9 in July 2014, which marked the completion of replacing IAS 39. The revised impairment model aims to provide users with more transparent and useful information regarding expected credit losses.
One of the key differences between the two standards, with large implications, is the clarification and methodology for recognizing impairment. Under the old IAS 39 "incurred loss" model, impairment depended upon how a financial instrument was classified. Under the new IFRS 9 model, impairment measurement is the same regardless of instrument type and classification. The new impairment model uses a single, forward-looking expected credit loss model that applies to all types of financial instruments within the scope of impairment accounting. The new model requires recognizing expected losses since origination or acquisition date. The major advantage of the new approach is increasing the timeliness of loss recognition and addressing the over-complexity of the multiple impairment approaches required under the IAS 39 "incurred loss" model.
IFRS 9 requires recognition of loss allowance for expected credit losses at all times. It further requires that this amount be updated at each reporting date to reflect changes in the credit risk of financial instruments in scope. IFRS 9 provides three approaches for recognizing impairment loss:
- A general "three-bucket" approach for regular financial instruments
- A simplified approach for lease receivables, trade receivables, and contract assets without a significant financing component
- A special, "credit-adjusted Effective Interest Rate (EIR)" method for purchased or originated credit-impaired financial instruments
The new impairment standard applies to all firms reporting under IFRS 9. In particular, requirements affect firms holding financial instruments such as loans, investments in debt,1 and trade and lease receivables. The revised IFRS 9 model will impact banks and insurance firms most, due to their large financial instrument holdings. Non-financial firms with portfolios including trade and lease receivables, debt securities, and intragroup loans must also revise current impairment loss calculations.
Firms must capture and collect historical data and other trend information required for building a forward-looking impairment model and for tracking credit risk migration since the origination and recognition of the financial instrument. Data will include the historical probability of defaults, ratings, loss amount, product features, and economic scenario variables. Firms may also need to develop new models and processes or upgrade existing models in order to identify an increase in credit risk and calculate one-year or lifetime expected losses. Gathering this granular data has been ranked the number one challenge by banks responding to a recent Moody's Analytics survey.2
The primary methodological and analytical challenges that firms may encounter while implementing an IFRS 9 impairment model will arise in the following areas:
- Portfolio segmentation techniques for credit risk modeling and expected credit losses calculation
- Application of different thresholds for assessing significant increases in the credit risk of financial instruments
- Enhancements required for PD/LGD/EAD and loss rate models, in order to achieve IFRS 9-compliant expected credit loss calculation.
We discuss these specific challenges in further detail next.
Firms typically segment portfolios along business lines, product types, and risk characteristics for impairment calculation. IFRS 9 requires a more granular and dynamic approach for portfolio segmentation. Firms must group financial assets based on shared credit characteristics that typically react in a similar way to the current environment and macroeconomic factors. These characteristics include instrument type, credit risk ratings, industry, geographical location, date of initial recognition, remaining term to maturity, and underlying collateral. Groupings are reevaluated and re-segmented whenever new, relevant information arises, such as a change in economic conditions, or when credit risk expectations change.
A true economic loss occurs when current expected losses exceed initial expectations. Recognizing lifetime expected credit losses after a significant risk increase reflects economic loss more accurately in the financial statements. To determine significant credit deterioration, a firm should consider reasonable and supportable information available without undue cost or effort, and then compare the following:
- Risk of default at the reporting date
- Risk of default at the date of initial
A significant increase in credit risk assessment may be done on a collective basis (for example, on a group or subgroup of financial instruments), if evidence is not yet available at the individual level. While IFRS 9 does not prescribe any specific approach for assessing changes in credit risk, it allows the following operational simplifications for assigning the instrument into different stages:
- A rebuttable presumption of a significant increase in credit risk when the borrower is 30 days past-due. This indicator is not absolute, but it is presumed to be the latest point.
- For instruments with low credit risk, firms can continue to recognize a 12-month allowance.
The low credit risk exemption is often viewed as a suitable approach for wholesale and corporate exposures because firms can often map internal grades to external rating agencies. Likewise, the 30 days past-due criterion is often applied to retail portfolios because firms usually cannot map the portfolio to external ratings.
However, the Basel committee maintains higher expectations for banks implementing IFRS 9. The committee considers both the low credit risk exemption and the 30 days past-due criterion to be a "very low-quality implementation" of an expected credit loss model. The committee has strong expectations that a bank will not fall back on the 30 days past-due assumption, unless all forward-looking information has no substantive relationship with credit risk. The appropriate approach will vary by the institution's level of sophistication, the financial instrument, and data availability.
The IASB acknowledges firms may measure expected credit losses (ECL) using various techniques. While IFRS 9 does not explicitly require it, Moody's Analytics recommends that banks and insurers consider a more robust and sophisticated "expected loss approach" for most portfolios.
Many banks may leverage their existing internal credit risk management systems and expected loss calculation processes used for Basel regulatory requirements, but they will need to modify them to comply with IFRS 9 impairment requirements. Modifications include adjustments for through-the-cycle vs. point-in-time estimates and extending the Basel one-year PD/LGD/EAD to full term structures to capture the expected lifetime of financial instruments.
Other institutions may use in-house models and processes for stress testing and adjust the forecast for the forward-looking scenario rather than the stressed scenarios. Estimating "forward-looking," future economic conditions is only the first step of the adjustment process, for which institutions may need to develop single or multiple economic scenarios to calculate expected credit losses. The most challenging aspect of the change may be incorporating the macroeconomic factors forecast (interest rates, unemployment, GDP growth, etc.) into the PD/ LGD/EAD modeling and, thus, into the expected credit loss calculation. Adjusted models must reflect how such changes in factors affected defaults and losses in the past. However, it is possible that the combination of forecast factors may never have been seen historically.
Even if all the IFRS 9-compliant models for loss rate and the different components in the expected loss approach are readily available, additional issues will arise when determining the expected credit loss. Rules require discounting the expected cash shortfalls in order to obtain the current value at the reporting date. Current regulatory calculations do not discount at all or discount only from the date of the expected default point. Firms will need to modify existing systems to better capture the expected timing of credit losses and to discount future amounts to the reporting date. IFRS 9 requires the use of the effective interest rate at initial recognition when discounting the cash flows. Firms must also backfill the effective interest rate for financial instruments if this information is missing in the current accounting system. In addition, firms may need to enhance or replace a current loan loss calculation engine to accommodate the demanding computational loads of exposure level, cash flow-based, lifetime expected credit loss calculations.
Given these challenges, we next discuss potential solutions for each of the previous areas of discussion.
Implementing the IFRS 9 impairment model results in a granular and dynamic portfolio segmentation scheme. Financial instruments should be segmented based on shared credit risk characteristics. Instruments grouped together should respond to historical and current environments, as well as to forwardlooking information and macroeconomic factors in a similar way, with respect to changes in credit risk level. The grouping method should be granular enough to assess changes in credit quality leading to migration to a different credit risk rating, thus impacting the estimation of expected credit losses. Segmentation should be reevaluated and exposures re-segmented whenever there is relevant new information or whenever credit risk expectations change. Most importantly, exposures should not be grouped in such a way that the performance of the segment as a whole masks an increase in a particular exposure's credit risk. When credit risk changes after initial recognition affect only some exposures within a group, those exposures should be segmented out into appropriate subgroups.
IFRS 9 requires assessing financial instruments for significant credit risk increases since initial recognition. Firms must use change in lifetime default risk (considering quantitative and/ or qualitative information), a low credit risk exemption, and a rebuttable presumption of 30 days past-due. For instruments whose default occurrences are not concentrated at a specific point in time during the expected life, firms can use changes in one-year in default risk to approximate changes in lifetime default risk.
When using a loss rate approach to measure credit risk increases, firms should use changes in credit risk isolated from other expected loss drivers, such as collateral. Also, the loss rates should be applied to groups defined in a similar way to the groups for which the historical credit loss rates are calculated. Since loss rates should incorporate information regarding current and forward-looking economic conditions, firms should apply historical loss rates consistent with the current and expected economic conditions. If the historic economic conditions differ, an adjustment is needed. A possible approach for calculating loss rates dependent upon economic conditions is to develop a model linking loss rates with economic variables.
PDs can also be used to identify significant credit risk increases. If using PD changes, Moody's Analytics recommends assessing the logarithmic change instead of raw changes,3 as the significance of a specific change in PD depends on the starting point.
IFRS 9 states that firms cannot simply compare the change in absolute risk over time. Instead, they should incorporate the relationship between expected life and default risk. One possible approach to doing this is to use annualized PD values instead of cumulative PD values. For instruments whose default patterns are not concentrated at a specific point in time, one can use changes in 12-month PD as an approximation of the lifetime default risk change. This approach may not be suitable for instruments with only significant payment obligations after the next 12 months, or for which changes in macroeconomic or other credit-related factors are not adequately reflected in the default risk during the next 12 months.
In addition to using PD changes, changes in PD-implied rating, expressed as notch differences, can also determine significant increases in credit risk. Ratings are sometimes preferred over PD measures, as many institutions are more familiar with internal or agency ratings. However, implied ratings have the disadvantage of being non-continuous (like PD measures). Additionally, if using an internal rating system, it must be well-designed, incorporating a reasonable number of rating categories and avoiding too many credits classified into specific categories. For IFRS 9 purposes, an internal rating system should also incorporate the relationship between expected life and default risk. The internal rating mappings, therefore, should depend on the instrument's maturity.
One challenge in calculating credit risk changes is the backfilling of credit risk assessment at origination. For this purpose, institutions must consider credit risk characteristics at initial recognition. This requires historical information such as internal ratings, external ratings, financial statements, and economic conditions statistics.
We next discuss incorporating forward-looking information into credit risk measures, as well as other challenges in calculating expected credit losses.
To overcome the expected loss calculation challenges, firms can implement different solutions to comply with IFRS 9 including existing internal models or new tools. The targeted IFRS 9 solution should possess the following characteristics:
- Applies a default definition consistent with internal credit risk practices
- Reflects an unbiased and probability-weighted amount of expected credit losses
- Is able to calculate expected losses for both one year and expected life
- Incorporates information regarding past events, current conditions, and forecasts of future economic conditions
- Discounts expected credit losses to the reporting date, using the effective interest rate as the discounting rate
- Reflects cash flows expected from collateral and other credit enhancements as part of the contractual terms
- Considers all contractual terms of the financial instrument
- Estimates the portion of the commitment to be drawn down for financial instruments that include both a loan and an undrawn commitment component
Firms can leverage Basel and stress testing models for IFRS 9 purposes. They can also utilize vendor models to help comply with IFRS 9 requirements. We recommend specific adjustments in order to comply with IFRS 9.
Most banks are subject to the Basel Capital Standards, which state three possible approaches for calculating capital requirements for credit risk: the Standardized Approach, the Foundation Internal Ratings-Based (FIRB) Approach, and the Advanced Internal Ratings-Based (AIRB) Approach. The Standardized Approach uses predefined risk weight values set by the regulator, which are not suitable for IFRS 9 requirements. However, banks estimate PD under both FIRB and AIRB, which can be used as a starting point for calculating IFRS 9-compliant PDs.
In order to use the Basel framework to obtain PDs for the IFRS 9 calculation, firms should consider the following adjustments:
I. Align the Basel definition of default and the institution's risk management practice.
IFRS 9 states that firms shall apply a definition of default consistent with the definition used for internal credit risk management purposes. However, there is a rebuttable presumption that a default does not occur later than when the instrument is 90 days past-due. The firm may rebut the presumption if it has reasonable and supportable information to determine that a more lagging criterion is more appropriate.
II. Apply adjustment for economic cycle and incorporate forward-looking information.
The desire for stable capital requirement estimates leads many banks to adopt through-the-cycle (TTC)4 PDs. Since IFRS 9 requires firms to incorporate information regarding current conditions and forecasts of future conditions, TTC PDs require a cycle adjustment incorporating forward-looking information.
In particular, firms can leverage TTC PDs and apply a cyclical adjustment. The adjustment can be based on credit cycle signals from macroeconomic variables or information from the equity or debt markets, which incorporate market participants' expectations and therefore reflect forward-looking information. Since the credit cycle affects industries in different ways, adjustments should be industry-specific. If the credit signals show an increase in risk level, PD levels should be adjusted upward. If the risk level falls, PD levels should be adjusted downward.
- One possible implementation adjustment is via a Z factor, as illustrated in Aguais, et al.,5 a single parameter that represents the credit cycle.
- Another option for incorporating forward-looking information into an existing PD is to use a stress testing approach, where the projected PD depends upon particular economic scenarios.
- A third option is to develop a PD model that incorporates the current explanatory variables as well as forward-looking variables, such as forecasts of macroeconomic variables and/or signals from the equity market.
In addition to the aforementioned three approaches used for developing the PD model, firms can also consider simulating individual loan and collateral performance, as well as corresponding market conditions, based on historical probability distributions. With a sufficiently large number of simulation paths, the final PD becomes an unconditional risk measure, which reflects a probability-weighted outcome as required by IFRS 9.
III. Calculate lifetime PDs.
To calculate lifetime expected loss, users must construct a term structure of PDs beyond one year. Different modeling techniques include:
- Develop separate models for different time horizons and interpolate probabilities of default for intermediate maturities; developers must ensure that PDs of long horizons are higher than PDs of short horizons.
- Develop a model that uses the most up-to-date information at each point in time. The resulting PD is not time dependent, but requires forecasting the risk factors for each loan's lifetime.
- Use transition matrices, which measure the probability of moving between credit categories.
- Develop a model in which the PD is time-dependent.
Some firms also develop internal LGD models for Basel and risk management purposes, which they can leverage for IFRS 9. To use the Basel framework to obtain IFRS 9-compatible LGDs, firms should make the following adjustments:
I. Remove the downturn component.
IFRS 9 states expected loss estimations should reflect current and forward-looking expected losses, not downturn economic conditions. This method disregards the conservative approach suggested by the Banking Supervision Committee. Therefore, for IFRS 9 purposes, the downturn component should be removed.
II. Adjust the discount rate.
Basel does not specify which discount rate to use for estimating the LGD. IFRS 9 requires using the effective interest rate or an approximation thereof. Therefore, in order to use Basel models, firms should align the interest rate used or apply an adjustment. Under IFRS 9, firms must also discount expected losses to the reporting date, while Basel states discounting to the default date.
III. Incorporate forward-looking information.
Since IFRS 9 requires expected loss to be forward-looking, firms should consider building a cyclical adjustment into the LGD model. Similar to PD, the stress testing approach can also be used to produce a forward-looking LGD.
To implement the stress testing approach, firms may choose to calculate a probability-weighted, average LGD across multiple scenarios or simply use one scenario that represents the best future estimate to produce a single LGD. The implementation method should be consistent with PD and LGD.
IV. Extend the term structure.
As Basel models typically have a one-year horizon, they should be extended to provide a term structure for LGDs. Alternatively, firms can develop an LGD model that uses the most up-to-date information at each point in time. The resulting LGD requires that each explanatory variable is forecast for the entire lifetime of each loan, but it does not require an assumption on the LGD term structure.
V. Other adjustments.
Basel LGD estimations may include indirect costs related to collecting on the exposure and credit derivatives used as risk-mitigating instruments. Since IFRS 9 requirements do not include these components, they must be removed.
For financial instruments with predetermined draw and amortization terms (e.g., term loans and bonds), EAD in future periods can be calculated from known contractual terms during the cash flow generation process, taking into account probability of prepayment for prepayable loans and the probability of the call (or similar) options being exercised for bonds with contingencies.
For irrevocable loan commitment and line of credit with a loan, and an undrawn commitment component, firms may need an EAD model to estimate the instrument's exposures to credit losses. One option is to leverage the Basel EAD model, used under the AIRB approach.6 Basel defines EAD as "the expected gross exposure of the facility upon default of the obligor."7
In order to adjust the Basel EAD modeling for IFRS 9 purposes, the following modifications are needed:
i. Remove the downturn component.
If the EAD Basel estimation includes an economic downturn component, it should be removed.
ii. Extend the term structure.
For IFRS 9 purposes, the Basel EAD models should be extended beyond a one-year horizon in order to cover the expected life of the financial instrument.
Unlike the PD/LGD/EAD modeling approach discussed above, loss rate models estimate credit losses by aggregating PD, LGD, and EAD. These models are often used for short-term portfolios such as credit cards, trade and lease receivables, and some non-material exposures. In addition, medium- or small-sized firms often rely on these simple modeling approaches for loss allowance calculations.
Commonly used loss rate models include:
- Net charge-off rate model
- Roll-rate model
- Vintage loss curve model
To address the new IFRS 9 impairment model requirements, we recommend firms use a more granular and dynamic approach for portfolio segmentation by grouping financial assets based on shared credit characteristics that typically react in a similar way to the current environment and forward-looking information. Firms can implement different credit risk models for calculating the 12-month or lifetime expected losses, including the expected loss approach based on PD/LGD/EAD modeling or loss rate approach. These models can be developed internally or provided by vendors.
Credit risk models developed for Basel capital requirement calculation or stress testing purposes can be leveraged for IFRS 9 expected credit loss calculation as well. The forward-looking information required by IFRS 9 can be incorporated into credit risk models based on signals from macroeconomic variables or from the equity or debt markets. Possible approaches for incorporating forward-looking information include transition matrices, scenario-dependent estimations, and simulation approaches. Firms must extend the one-year PD, LGD, and EAD estimations to the instrument's lifetime, for which different statistical techniques can be used. Possible techniques include transition matrices, time-dependent models, separate models for different time horizons, and models that use the most up-to-date information at each point in time. Further, institutions will need to incorporate specific adjustments to models developed for Basel requirements.
1 Investments in equity instruments are outside the scope of the IFRS 9 impairment requirements, because they are accounted for either at Fair Value through Profit or Loss (FVTPL) or at Fair Value through Other Comprehensive Income (FVOCI), with no reclassification of any fair value gains or losses to profit or loss (i.e. the FVOCI election for equity instruments).
2 Gea-Carrasco (2015).
3 Logarithmic changes are similar to percentage changes for small fluctuations. However, logarithmic changes have more desirable properties, as they are symmetric and additive.
4 While there is no universally agreed-upon definition, the conventional view is that a rating system or a PD model with outputs that remain relatively stable across different macroeconomic conditions is a TTC system.
5 See Belkin, Suchower, & Forest (1998a, 1998b); Aguais, et al. (2004, 2006).
6 Under the Standardized and the Foundation Internal Ratings-based Approaches, firms have less flexibility with EAD calculation.
7 Basel Committee on Banking Supervision (2015).
Bao, Eric, Maria Buitrago, Jun Chen, Yanping Pan, Yashan Wang, Jing Zhang, and Janet Zhao, IFRS 9 Impairment Regulations: Implementation Challenges and Potential Solutions, Moody’s Analytics white paper, December 2015.
Basel Committee on Banking Supervision, Guidance on Accounting for Expected Credit Losses, February 2015.
Aguais, Scott, et al., M. Ong (ed.), Point-in-Time versus Through-the-Cycle Ratings, The Basel Handbook: A Guide for Financial Practitioners, London: Risk Books, 2004.
Aguais, Scott, et al., M. Ong (ed.), Designing and Implementing a Basel II Compliant PIT-TTC Ratings Framework, The Basel Handbook: A Guide for Financial Practitioners, London: Risk Books, 2006.
Belkin, Barry, Stephan Suchower, and Lawrence Forest, The Effect of Systematic Credit Risk on Loan Portfolios and Loan Pricing, Credit-Metrics Monitor, pp 17-28, 1998a.
Belkin, Barry, Stephan Suchower, and Lawrence Forest, A One-Parameter Representation of Credit Risk and Transition Matrices, Credit-Metrics Monitor, pp 17-28, 1998b.
Deloitte, Fifth Global IFRS Banking Survey: Finding Your Way, September 2015.
Ernst and Young, Facing the Challenges: Business implications of IFRS 4, 9 and Solvency II for insurers, 2015.
Gea-Carrasco, Cayetano, IFRS 9 Will Significantly Impact Banks’ Provisions and Financial Statements, Risk Perspectives Magazine, June 2015, Moody’s Analytics.
International Accounting Standards Board, International Financial Reporting Standard: IFRS 9 Financial Instruments, July 2014.
International Accounting Standards Board, Implementation Guidance: IFRS 9 Financial Instruments, July 2014.
International Accounting Standards Board, Basis for Conclusions: IFRS 9 Financial Instruments, July 2014.
International Accounting Standards Board, IFRS 9 Project Summary, July 2014.
McPhail, Joseph and Lihong McPhail, Forecasting Lifetime Credit Losses: Modeling Considerations for Complying with the New FASB and IASB Current Expected Credit Loss Models, Working paper, 2014.
Senior Director, Research
Dr. Yashan leads the research and quantitative modeling team for portfolio valuation, accounting, and balance sheet analytics. The team develops analytic and empirical models for asset valuation, IFRS 9/CECL, PPNR, and ALM.
Analyzes IFRS 9, delves into its effects on future impairment calculations, and provides recommendations on how financial institutions can implement and leverage forward-looking credit loss models.
Previous ArticleIFRS 9 Survey Results
In this article, we use historical data to calculate and compare loan- and portfolio-level loss allowances under the incurred loss model and CECL.
July 2017 WebPage Dr. Yanping Pan, Dr. Yashan Wang
This paper investigates the impact of using EDF9 instead of EDF8 values as inputs for estimating credit portfolio risk measures within Moodys Analytics RiskFrontier®. The recent EDF9 enhancements affect portfolio risk analysis via various channels — due not only to new values for default probabilities, but also because the market Sharpe ratio (i.e. market-level risk premium) and asset return-based correlations for corporate exposures depend on time series of EDF measures. In this paper, we focus on the question of how using the new EDF9 default probabilities alter patterns in portfolio risk measures.
June 2017 Pdf Noelle Hong, Jimmy Huang, Libor Pospisil, Albert Lee, Marc Mitrovic, Sunny Kanugo, Tiago Pinheiro, Andriy Protsyk, Dr. Yashan Wang
IFRS 9 materially changes how institutions set aside loss allowance. With allowances flowing into earnings, the new rules can have dramatic effects on earnings volatility. In this paper, we propose general methodologies to measure and manage credit earnings volatility of a loan portfolio under IFRS 9. We walk through IFRS 9 rules and the different mechanisms that it interacts with which flow into earnings dynamics. We demonstrate that earnings will be impacted significantly by credit migration under IFRS 9. In addition, the increased sensitivity to migration will be further compounded by the impact of correlation and concentration. We propose a modeling framework that measures portfolio credit earnings volatility and discuss several metrics that can be used to better manage earnings risk.
This paper introduces a framework for stress testing portfolios of credit risk sensitive securities. Specifically, the framework uses a macroeconomic scenario to project stressed expected losses (EL) on the securities by accounting for credit quality changes, recovery risk effects, fluctuations in market price of risk, and interest rates paths. The calculations are carried out analytically over multiple periods.
April 2016 Pdf Sunny Kanugo, Vishal Mangla, Libor Pospisil, Dr. Yashan Wang, Kevin Yang, Ian Ward, Jay Harvey
Modeling the Joint Credit-Interest Rate Dynamics on a Multi-Dimensional Lattice Platform: Model Validation and Applications in Risk Integration
This document presents validation results for the credit-interest lattice or the multi-dimensional lattice (MDL) valuation model within Moody's Analytics RiskFrontier™.
June 2015 Pdf Dr. Yanping Pan, Dr. Yashan Wang, Sunny Kanugo, Rama Sankisa
We validate the performance of the Moody's Analytics Public EDF™ (Expected Default Frequency) model for Asian-Pacific and Japanese corporate firms during the last decade, including the recent credit crisis and its recovery period. We divide the decade into two sub-periods: an early period, 2001–2007, and a later one, 2008–2010, and then compare the model's performance during these two periods. We focus on the model's ability to prospectively differentiate between defaulters and non-defaulters, the timeliness of its default prediction, and its accuracy of levels.
October 2011 Pdf Christopher Crossen, Xu Zhang
We validate the performance of the Moody's Analytics Public EDF™ (Expected Default Frequency) model for European corporate firms during the last decade, including the recent credit crisis and its recovery period. We divide the decade into two sub-periods: an early period, 2001–2007, and a later one, 2008–2010, and then compare the model's performance during these two periods. We focus on the model's ability to prospectively differentiate between defaulters and non-defaulters, the timeliness of its default prediction, and its accuracy of levels.
October 2011 Pdf Christopher Crossen, Xu Zhang
In this paper, we validate the performance of the Moody's Analytics Public EDF™ (Expected Default Frequency) model for global financial firms during the last decade, including the recent credit crisis and its recovery period. We divide the decade into two sub-periods: an early period, 2001–2007, and a later one, 2008–2010, and then compare the model's performance during these two periods. We focus on the model's ability to prospectively differentiate between defaulters and non-defaulters, its comparison to agency credit ratings, the timeliness of its default prediction, and its accuracy of levels.
October 2011 Pdf Christopher Crossen, Xu Zhang
In this paper, we validate the performance of the Moody's Analytics Public EDF™ (Expected Default Frequency) model for North American corporate firms during the last decade, including the recent credit crisis and its recovery period.
May 2011 Pdf Christopher Crossen, Shisheng Qu, Xu Zhang