FASB Accounting for Loan Losses
Carl Chatto, Managing Principal
December 8, 2014
The Financial Accounting Standards Board (FASB) has proposed significant changes not only in accounting for loan losses but also for debt securities, troubled debt restructurings and purchased loans.
The new proposal which is expected to be issued soon, puts forth a single credit loss model for all financial assets that are carried at amortized cost (AC) or at fair value through other comprehensive income (FV-OCI). It does not affect any instruments carried at fair value through net income (FV-NI).
One key element relates to removing the probability threshold for recognizing credit losses. This will result in recognizing losses sooner than under previous generally accepted accounting principles (GAAP). Other significant changes will require recognizing a credit loss allowance for some debt securities and the recognition of an allowance for loan losses when purchasing credit-impaired assets.
Under current GAAP, losses are not recognized until they are incurred or probable of a loss. FASB has essentially removed the probability threshold for credit losses, which in theory should result in higher loan loss allowances. Also, FASB now requires losses to be estimated over the life of the loan. Furthermore, these concepts are now applied to all debt instruments carried either at AC or FV-OCI; this includes loans, debt securities, accounts receivable, lease receivables and loan commitments. Equity securities are not included in the scope of the ASU.
The ASU requires immediate recognition of expected losses based on a current estimate. The current estimate should consider:
- Relevant information about expected credit losses
- Time value of money
- At least two outcomes, one of which involves a possible credit loss
Relevant information about expected credit losses should include the profile of the borrower, collateral values, economic indicators, guarantors and other information specifically related to the debt instrument. Reasonable and supportable forecasts are a key element of this, but the ASU doesn’t provide any specific guidance on forecasting economic cycles. This will be an especially subjective part of the overall loss estimate. In general, historical loss rates would provide a baseline (generally consistent with current practice) which would then be adjusted for economic and other expectations.
The true value of money would be considered either explicitly through a discounted cash flow model or implicitly through applying loss rates to amortized cost, on the basis that amortized cost represents a discounted cash flow. As a practical expedient, a collateral-dependent asset could be measured for loss by comparing its amortized cost to the fair value of the underlying collateral.
Another key input is the requirement to consider at least two outcomes including at least one with a loss. The outcomes could be probability-weighted.
There is a practical expedient for debt securities that are classified as FV-OCI that also have a fair value equal to or greater than the amortized cost and where expected losses are insignificant. For example, an entity might conclude that those investments rated AA or higher have an insignificant credit loss exposure.
The ASU required some consideration of other areas. Without going into significant detail, these considerations are:
- There is flexibility on whether to provide for the credit losses on an individual or pooled basis, if the instruments have similar risk characteristics.
- Credit losses would continue to be presented as a reduction of the amortized cost basis of loans, but also for debt securities and a separate reduction for purchased credit-impaired assets.
- Financial assets should be written off when there is no reasonable expectation of recovery of value rather than when deemed uncollectible. Since in many cases debt security values would be reduced by an allowance rather than a write-down, it is possible that the allowance could be reversed if fair values increased. This will represent a significant change for impaired debt securities. It will now be a GAAP requirement to cease interest accruals when collection of substantially all interest or principal is not probable. This will be a change in accruing income on impaired debt securities.
- There are changes to the definitions of purchased credit-impaired assets and collateral-dependent loans.
As with most accounting pronouncements, there are new footnote disclosures in the proposal. These include:
- The requirements in ASU 2010-20 (to disclose information about the credit risk in a portfolio and how it is monitored by management, aging of past due receivables and credit quality of receivables) would be expanded to include debt securities and other receivables. There will be new requirements to disclose how an entity develops its loss estimates, the factors it considers and whether those factors changed.
- Rollforward of amortized cost of loans, debt securities and lease receivables carried at AC and FV-OCI.
- Reconciliation of the difference between fair value and amortized cost for assets carried at FV-OCI.
- Reconciliation of purchase price and par value for purchased credit-impaired loans.
These changes will create implementation issues for many, if not all, institutions.
For example, the expectation is that institutions will use forecasts, a longer time horizon for historical losses, and a longer time horizon looking ahead to develop loan loss estimates. Looking back five years or more might be expected, along with a two or three year loss forecast, before reverting to historical losses as a basis for life of loan losses. Best practices have yet to be developed but this will require modifications to current ALL methods. The ability to gather data from several years back may be restricted due to system limitations, especially if institutions have changed system providers recently.
There has also been an expectation raised that loan portfolios will need to be segmented much more than they have in the past. Loan losses should, under this approach, be sorted by year of origination, average life, and risk rating to develop more refined estimates. Some observers have indicated that loan level data should be collected, not pooled. Whether this becomes a regulatory requirement has yet to be determined, but this degree of segmentation will require more robust data analysis and tracking than has been the case for most institutions and presents many time and resource implications.
By extending the forecasted loss period, there is greater emphasis on the judgments used in developing those forecasts. That will subject institutions to higher levels of documentation and review to support those forecasts. If institutions can develop strong segmented history, though, that may help support forecasted loss models.
The expectation is that this will increase the level of loan loss allowances. A representative of the OCC stated in 2013 that he expects overall allowances to increase 30% to 50% over current amounts. This would be reflected as a one-time reduction in capital rather than an expense, but the end result is essentially the same: potentially significant increases to allowances and resulting decreases in capital, at a time when Basel III is demanding higher capital ratios.
Regardless of how best practices evolve and the ultimate regulatory expectations, it is clear that implementing the new model will require more staffing and resources than it has previously, and will almost certainly result in higher loan loss allowances. Given that the proposed model to determine the level of the allowance for loan losses includes many assumptions, projections and estimates, it is questionable as to whether the proposed model would lead to improved financial reporting.
BNN can help your institution before and during the implementation. Please contact Carl Chatto or your usual BNN representative.
Disclaimer of Liability: This publication is intended to provide general information to our clients and friends. It does not constitute accounting, tax, or legal advice; nor is it intended to convey a thorough treatment of the subject matter.