CECL Qualitative Factors: Documentation, Governance, and Audit Findings

Under ASC 326, CECL qualitative factors are the adjustments management layers onto a quantitative loss model to reflect current conditions and reasonable, supportable forecasts that historical loss data cannot capture on its own. They are often called Q-factors. Getting them right is a balancing act: overlays that are too thin draw regulatory criticism, and overlays that are too aggressive misstate the balance sheet. Every adjustment has to be tied to observable data and documented well enough that someone else can follow the reasoning.

The Factor Categories ASC 326 Expects You to Consider

ASC 326-20-55-4 sets out a detailed list of factors an entity may consider when adjusting historical loss information. Not every factor applies to every portfolio, and institutions can consider factors beyond the list. But examiners expect management to evaluate each category and explain why a factor was or was not used. Interagency guidance from the Federal Reserve, OCC, and FDIC has confirmed that the qualitative factors from the 2006 Interagency Policy Statement remain relevant under CECL.1Federal Reserve. Frequently Asked Questions on the New Accounting Standard on Financial Instruments—Credit Losses

  • Borrower-level characteristics: financial condition, credit rating or score, asset quality, business outlook, and the borrower’s ability to make scheduled payments.
  • Asset characteristics: remaining payment terms, time to maturity, and the timing and extent of prepayments.
  • Portfolio nature and volume: shifts in the mix, size, or type of financial assets held. A move from secured to unsecured lending changes the underlying risk profile even if balances stay flat.
  • Past-due and adversely classified assets: the volume and severity of delinquencies, along with the frequency of loan modifications.
  • Collateral values: changes in the value of underlying collateral for assets where the collateral-dependent practical expedient has not been applied.
  • Lending policies and procedures: changes in underwriting standards, collection strategies, charge-off and recovery practices, and management’s knowledge of borrower operations.
  • Credit review quality: the effectiveness of the credit review and loan grading system.
  • Staff experience and depth: the experience, ability, and depth of management, lending staff, and other relevant personnel.
  • Environmental and concentration factors: exposure to changes in the regulatory, legal, or technological environment; general market conditions within a geographic area or industry; and international, national, regional, and local economic and business conditions.2Office of the Comptroller of the Currency. Allowances for Credit Losses – Comptrollers Handbook

Credit concentrations fall under that last category. Overexposure to a single industry, geography, or borrower type creates risk that a diversified historical loss rate will not reveal. NCUA guidance adds that external competitive pressures, such as a new lender entering a market with aggressive terms, can also warrant an adjustment.3National Credit Union Administration. Appendix C – Qualitative Adjustments

Data That Supports the Adjustments

Every adjustment needs a data trail. FASB has been explicit that no single data type is required, and institutions do not have to default to external consensus forecasts when internal data is sufficient and more relevant to their circumstances.4Financial Accounting Standards Board. FASB Staff Q&A – Topic 326, No. 2: Developing an Estimate of Expected Credit Losses on Financial Assets

On the internal side, loan delinquency reports, historical charge-off rates, and migration analyses show trends inside the portfolio. Loan review results and internal audit findings speak to underwriting quality and collection effectiveness. Portfolio concentration reports track exposure across sectors, geographies, and borrower types. Watch lists, risk-rating migration reports, and modification tracking flag credit quality slipping before losses hit the income statement.

External data helps calibrate how broader conditions have shifted relative to the historical look-back period. Common inputs include unemployment rates, GDP growth, interest rates, and housing price indices, often pulled from public sources such as the Bureau of Labor Statistics or the Federal Reserve’s FRED database. The important step is correlating those figures with internal loss experience to identify which macro variables actually drive losses in a given portfolio segment. The standard permits internal information, external information, or any combination, as long as it is relevant to the entity’s exposures.4Financial Accounting Standards Board. FASB Staff Q&A – Topic 326, No. 2: Developing an Estimate of Expected Credit Losses on Financial Assets

Turning the Assessment Into a Number

Once the data is gathered, each factor gets a direction and a magnitude. An upward adjustment increases the allowance when current or forecasted conditions are worse than those in the historical look-back period. A downward adjustment reduces it when the outlook is more favorable. Many institutions assign each factor a risk level (low, moderate, or high) that maps to a range of basis points, then pick a value inside that range based on the supporting evidence. The sum across factors is the total qualitative overlay for that portfolio segment.

The overlay is added to or subtracted from the quantitative historical loss rate to produce the total expected credit loss percentage. Applying that combined rate to the current amortized cost of the portfolio yields the allowance that appears on the balance sheet. The OCC’s handbook walks through this arithmetic in an illustrative example.2Office of the Comptroller of the Currency. Allowances for Credit Losses – Comptrollers Handbook

Institutions have flexibility in where the adjustment lives. Some make a standalone adjustment at the portfolio level. Others revise individual model inputs, feeding the change back into the quantitative calculation. Either approach is acceptable as long as the results are supportable and do not overlap with risks captured elsewhere in the estimate.

Forecast Horizon and Reversion

Deciding how far out to forecast and what happens after is one of the most judgment-heavy pieces of CECL. FASB does not require forecasting over the full contractual life of every loan. The forecast only needs to cover the period for which management can develop or obtain reasonable and supportable projections. That period may differ across portfolios, products, or even individual model inputs, and it must be reassessed each reporting period.4Financial Accounting Standards Board. FASB Staff Q&A – Topic 326, No. 2: Developing an Estimate of Expected Credit Losses on Financial Assets

Beyond the forecast window, the entity reverts to historical loss information that reflects expected losses over the remaining contractual life. FASB does not prescribe a single reversion method. Acceptable approaches include immediate reversion, straight-line reversion, or any other rational and systematic method, and an entity can apply different reversion approaches to different inputs and asset classes.4Financial Accounting Standards Board. FASB Staff Q&A – Topic 326, No. 2: Developing an Estimate of Expected Credit Losses on Financial Assets

Two constraints apply. When reverting, management should consider whether the historical loss information is still relevant given current asset-specific risk characteristics such as underwriting standards, portfolio mix, and loan terms. And historical loss information should not be adjusted for current economic conditions or forecasts during the reversion and post-reversion periods; that adjustment belongs only in the forecast window. Reversion is not a one-time policy election. It is part of the loss estimate and needs the same support as any other assumption.

Avoiding Double Counting

Double counting is where most CECL implementations get into trouble. It happens when a qualitative adjustment captures a risk the quantitative model already reflects. The OCC states plainly that management should not count the same losses in both the quantitative loss rates and the qualitative factors, because doing so overstates the allowance and violates the standard.2Office of the Comptroller of the Currency. Allowances for Credit Losses – Comptrollers Handbook

The classic example is loosened underwriting. When credit requirements are relaxed, the effect on losses has not yet appeared in the historical data, so a qualitative overlay is appropriate during the transition. As time passes and the riskier loans start charging off, the quantitative loss rate absorbs the impact. The qualitative adjustment has to decline correspondingly. Keeping the same upward overlay after actual charge-offs already reflect the loosened standards is double counting.2Office of the Comptroller of the Currency. Allowances for Credit Losses – Comptrollers Handbook

Preventing overlap requires a clear map of what the quantitative model already captures. If the model uses recent loss history that already includes recession-era charge-offs, adding a qualitative overlay for “economic downturn” risks overstating the reserve. If the model uses risk-rated loan segments and loans have already been downgraded to reflect deteriorating borrower condition, a separate add-on for “borrower financial condition” needs careful justification.

Documentation That Will Hold Up

Every adjustment needs a written record that connects the data to the decision. This usually takes the form of a formal memo covering each factor considered, the direction and size of the adjustment, and the specific evidence behind it. Regulators and auditors expect to see the actual economic reports, internal data, and analysis relied on, not just a conclusion.2Office of the Comptroller of the Currency. Allowances for Credit Losses – Comptrollers Handbook

Adjustments must be reasonable, consistently determined, and adequately supported.5SupervisionOutreach.org. Preparing for CECL Where certain factors carry more weight than others, the documentation should explain why given the current asset mix. The path from raw data to final qualitative percentage must be traceable, so a reviewer can reproduce the calculation.

Sophistication scales with the institution. Smaller or less complex institutions may rely on a narrative describing recent trends, current conditions, and management’s conclusions. Larger institutions may support their adjustments with regression analysis, sensitivity testing, or other modeling techniques. In either case, the documentation must show that each factor was evaluated and that the resulting adjustment is proportional to the identified risk.

Consistency across reporting periods matters. If a 10-basis-point overlay was applied last quarter for rising unemployment, and unemployment has risen further, the documentation should explain why the overlay moved in the expected direction, or if it did not, why. Examiners get suspicious when qualitative adjustments sit unchanged quarter after quarter despite changing conditions, because that suggests management is not actively reassessing.

Governance and Board Oversight

Because qualitative adjustments involve significant management judgment, regulators expect formal governance around them. The OCC’s Comptroller’s Handbook sets expectations at both the board and management level.2Office of the Comptroller of the Currency. Allowances for Credit Losses – Comptrollers Handbook

The board (or a designated committee) oversees the significant judgments embedded in the allowance estimate. Board-level activities include reviewing and approving the loss estimation and charge-off policies at least annually, reviewing management’s support for the estimated allowance each period, requiring management to periodically validate and revise loss estimation methods and assumptions, and approving internal and external audit plans for the allowance process while tracking resolution of any findings.

Management is responsible for establishing and maintaining governance over the loss estimation process. That includes reviewing and challenging the assumptions behind qualitative adjustments, designing effective internal controls, and periodically comparing credit loss estimates to actual charge-offs to confirm the estimate is performing as expected. Where third parties are involved, such as model vendors, management applies sound vendor risk management practices.2Office of the Comptroller of the Currency. Allowances for Credit Losses – Comptrollers Handbook The forecast period and the selection of which qualitative factors to apply are treated as assumptions that pass through governance and controls, not set-and-forget decisions.

Common Audit Findings to Avoid

Examiners have developed a consistent pattern of what they flag in CECL qualitative factor assessments.2Office of the Comptroller of the Currency. Allowances for Credit Losses – Comptrollers Handbook

  • Insufficient support: documentation that states a conclusion without showing the data and reasoning behind it fails the reasonable-support standard.
  • Double counting: adjusting for risks the quantitative loss rate already captures.
  • Failure to adjust for current conditions: relying solely on historical loss information without forward-looking data, which the standard prohibits.
  • Over-reliance on favorable historical periods: drawing loss data primarily from a period of economic growth without adjusting for the possibility that conditions may worsen.
  • Peer benchmarking as a substitute for analysis: adjusting the allowance solely to match a peer group median or target is inappropriate once an expected loss framework has been applied.
  • Static adjustments: overlays that never change despite shifting economic or portfolio conditions.
  • Data integrity problems: incomplete, inaccurate, or irrelevant data, particularly data that was not previously used for financial reporting and lacks adequate internal controls.
  • Inadequate model validation: failure to perform ongoing back-testing, benchmarking, and outcomes analysis for models used in the allowance process.

Examiners also flag weaknesses in overall governance, including lack of independent validation, ineffective model risk management, and gaps between the consolidated loss estimate and the amounts reported in regulatory filings. Material differences between the two draw immediate scrutiny and can result in matters requiring attention or enforcement actions.