Skip to main content

Looking for Valuant? You are in the right place!

Valuant is now Abrigo, giving you a single source to Manage Risk and Drive Growth

Make yourself at home – we hope you enjoy your new web experience.

Looking for DiCOM? You are in the right place!

DiCOM Software is now part of Abrigo, giving you a single source to Manage Risk and Drive Growth. Make yourself at home – we hope you enjoy your new web experience.

Looking for TPG Software? You are in the right place!

TPG Software is now part of Abrigo. You can continue to count on the world-class Investment Accounting software and services you’ve come to expect, plus all that Abrigo has to offer.

Make yourself at home – we hope you enjoy being part of our community.

IFSLeaseWorks is now part of Abrigo.

Diversify your portfolio and earn additional interest income. End-to-end lease origination and administration automation make it possible.

Read the press announcement

CECL Methodologies: Pros and Cons for Your Portfolio

CECL Methodologies

Pros and cons of the 7 CECL methods

The Financial Accounting Standard Board’s Current Expected Credit Loss Model, or CECL, represents a major change to how financial institutions and other entities measure credit losses, and for many banks and credit unions, the transition can seem daunting.

The extra time that the FASB plans to give many financial institutions in an updated CECL implementation timeline might tempt some management teams to “stick their heads in the sand” and instead, focus attention on more pressing deadlines. However, FASB members and audit industry experts have urged bankers to use the time wisely. They have suggested obtaining better data, building more robust internal controls, learning from results of large SEC filers (who begin implementation in 2020), and perhaps approaching CECL implementation as a business solution.

An important step in CECL implementation is selecting what methodology or methodologies the institution will use for estimating credit losses. The good news is that the FASB has emphasized that the accounting standard is not prescriptive about which methods are relevant and appropriate for estimating credit losses: “The Board has permitted entities to estimate expected credit losses using various methods because the Board believes entities manage credit risk differently and should have flexibility to best report their expectations.” – ASU 2016-13 (Topic 326) Paragraph 326-20-30-3.

However, that same flexibility in methodology selection can hamstring a financial institution if staff find it difficult to narrow down their options. Some financial institutions have described how they made key decisions on loan segmentation and methodologies, and those examples can be useful.

7 CECL Methodologies

Resources to help select the best methodology for your unique institution

Given that the CECL model is non-prescriptive, banks and credit unions have flexibility in choosing the right CECL methodologies for their institution’s unique data situation. This flexibility often leads financial institutions to ask one simple question: Where do I begin? 

CECL Methodology 1

Static Pool Analysis

The static pool analysis tracks a closed pool of loans for a configurable period of time and calculates a loss ratio on only those loans in the pool at the start date – losses to new originations in the period are not included. This is a simpler rate to calculate a N-year loss rate on a pool. If the life of loan is 2-years, then a static analysis over 2-years will yield a compliant estimation under CECL. Sometimes referred to as “cumulative.”

Recommended for:

  • Institutions without large loan pools or detailed loan level risk-maintenance data such as risk rating, delinquency, etc.
  • If disaggregating a pool results in unsatisfactory counts and lack of statistical power, this approach can yield results.

Not recommended for: 

  • Institutions that have changed underwriting standards for loans resulting in a significantly different year-over-year risk portfolio.
  • Pools of long-lived loans (3 years as a rule of thumb) as the application of forecasts becomes untenable at this point.
  • Pools without at least eight more quarters of loan-level data than the life of loan estimate (5 years for a 3 year asset). Can be difficult to apply R&S Forecasts without recessionary data periods.

CECL Methodology 2

Discounted Cash Flow

A DCF model implements a PD/LGD/EAD estimation with capabilities to account for lost interest, lost principal, vintage effects (roadmap), etc. A periodic tendency to default and absolute loss given default are applied to a projective model of the loan’s cashflow, with consideration for prepayment and principal curtailment effects. This methodology is defensible and back-testable, as it produces time-bound expectations of loss and income components. This methodology can be applied with relatively little loan-level details (as little as 1 year).

Recommended for:

  • Institutions lacking sufficient historical data
  • Institutions seeking to strongly justify forecast adjustments
  • Institutions seeking to layer peer experience when no loss experience exists at the institution level.

Not Recommended for: 

  • Institutions without software to implement the calculation and controls around it.
  • Institutions seeking to achieve a “benchmark” reserve level through methodological elections rather than qualitative adjustments.

CECL Methodology 3

Migration Analysis

A “Closed Cohort Migration to Loss” analysis with sub-segmentation calculations for ordinal risk characteristics (times delinquent, risk rating, ordinal risk descriptor). This methodology both describes a life-of-loan number and can accurately price the additional risk for deteriorated loans. Typically, “Pass” credits will remain with low reserve and immediately receive additional allocation on migration through the risk striations.

Recommended for:

  •  Shorter lived (<3-years) pools with risk diversity and reconciled controls around that risk description.
  • Typically sensible for renewing commercial credits.
  • Note: Requires additional statistical power and loss experience.

Not Recommended for: 

  • Pools with insufficient history as above (at least 8 quarters beyond the life of pool)
  • Longer-lived pools
  • Pools with no risk monitoring
  • Note: Can be difficult to apply R&S Forecasts
    without recessionary data periods.

CECL Methodology 4

Transition Matrix

Also known as Roll Rate Analysis, this methodology measures the tendency of a loan to transition from one state (e.g. commercial risk rating 4) to another state (e.g. commercial risk rating 5) over a period of time, and applies that tendency to project portfolio migration over a time period. As the loans migrate downward or upward adjustments in allocation are made.

Recommended for:

  • Institutions lacking a deep data history who do not wish to perform a DCF analysis
  • Institutions already leveraging this analysis

Not Recommended for: 

  • Simpler institutions without sufficient statistical power justify this form of analysis.
  • Note: Can be difficult to construct/apply framework for R&S forecasts without a deep data set.

CECL Methodology 5

Vintage Analysis

A powerful and predictive loss model for pools that are homogeneized by risk characteristics and also loan structure. The approach will calculate a loss rate that is sensitive to loan seasoning. There should be a loss curve by vintage year before applying this form of analysis. Vintage effects can be measured and included in other forms of analysis, such as a DCF model. The guidance’s requirement for vintage credit quality disclosure is not symmetrical to forward- looking vintage loss rate estimation.

Recommended for:

  • Homogeneous installment loans and mortgages.

Not Recommended for: 

  • Inappropriate for revolvers, frequently renewing credits, balloons, etc.
  • Note: Can be difficult to justify application of R&S forecasts.

CECL Methodology 6

Warm/Remaining Life

An implementation of WARM/WARL uses periodic charge-off rates and applies those rates to a projection of balances over the remaining life of an instrument. Users have latitude in the level of complexity applied in adjusting those rates for anticipated economic conditions.

Recommended for:

  • Adaptable to situations where data, or loss experience, can be lacking.
  • For shorter duration (1-year) assets, simpler methodologies may be more appropriate.
  • Note: Even when more sophisticated approaches such as DCF are elected, a remaining life measurement can provide information quickly.
    It is conceptually aligned with the standard and applicable to a variety of asset classes.

Not Recommended for: 

  • Leaning too heavily on remaining life, where loan- level data needs aren’t as high.
  • Note: May lead an institution to under invest in data capture that’s needed for other methodologies

CECL Methodology 7

Probability of Default/Loss Given Default (PD/LGD)

PD, LD, and EAD (exposure-at-default) metrics can be included in many other methodologies. For example, a DCF model applies a periodic PD to the loan’s terms over time to calculate an EAD and applies an LGD estimation. Commonly, an instutition will have 1-year PD numbers at a loan level or pool level (or seek to measure same) and leverage this in a CECL calculation. Translation from 1-year to lifetime PD numbers is not a simple or easily justifiable matter.

Recommended for:

  • Shorter-lived credits
  • High-count, low-dollar segments
  • Institutions that can separately calculate life-of-loan PDs using conditional/regression models

Not Recommended for: 

  • Pools with insufficient history as above (at least 8 quarters beyond the life of pool)
  • Longer-lived pools
  • Pools with no risk monitoring
  •  Note: Can be difficult to apply R&S Forecasts
    without recessionary data periods