The Accounting Puzzle in an Acquired Loan

Acquired-loan accounting under ASC Topic 326 is the subject here: why can a loan recorded at fair value also receive an allowance for credit losses? This series follows that question from the acquisition date into later reporting, so a reader can distinguish the amount embedded in the purchase measurement from the allowance, the provision expense, and the later yield adjustment. That distinction matters because the wrong boundary can make expected credit losses appear twice and can change how a loan’s performance looks after the deal closes.

A loan bought at fair value carries more than one question

When a bank acquires a loan, it does not simply carry forward the seller’s old balance. In a business combination, the acquired asset is recorded at fair value on the acquisition date. Fair value is the price that would be received to sell the asset in an orderly transaction between market participants. In a non-business-combination purchase, the initial investment generally reflects the amount paid, subject to the applicable accounting guidance. Either way, the acquirer needs a new starting point.

That starting point answers one question: what is the acquired loan worth to a market participant, or what amount did the buyer invest? It does not, by itself, answer how much of the future contractual cash flow the buyer expects to fail to collect. Topic 326 adds that second question through the ACL, an estimate of expected credit losses. The accounting problem begins when the two answers describe some of the same economics but are placed in the financial statements as if they were entirely unrelated.

A simplified thought experiment makes the overlap easier to see. Ignore interest timing, liquidity, and every other fair-value input for a moment. Suppose a loan has $100 of unpaid principal, the buyer expects to collect $95, and the acquisition measurement is therefore $95 because the expected $5 shortfall is already reflected in the price. If the non-PCD model then records a separate $5 ACL and deducts it from the asset, the reported net amount becomes $90 even though the same simplified assumptions still point to $95 of expected collection. The example is not a universal valuation formula or a journal-entry template. It isolates the relationship FASB describes: an acquisition measurement can already contain an expected-loss expectation, and a separate Day-1 allowance can repeat it.

The Financial Accounting Standards Board’s 2025-08 Update describes the concern directly. Stakeholders observed that fair-value measurement already contemplates expected credit losses. Under the non-PCD model, however, an allowance is also recognized through Day-1 credit-loss expense. The result can be a balance-sheet amount, net of the allowance, that is below the amount expected to be collected. FASB also says that the overlap can artificially inflate the subsequent yield on acquired loans.

“Double count” therefore does not mean that every purchase discount is an allowance. A loan can sell below its unpaid principal for many reasons. A buyer may price in credit risk alongside other market-participant assumptions. The important point is narrower: when the acquisition measurement already reflects expected credit losses, adding a separate Day-1 allowance without the right offset can recognize an overlapping loss twice in the accounting presentation.

This is why the issue is more consequential than a technical disagreement over labels. An acquisition-date entry affects assets, equity, and the provision line immediately. The same entry also affects the basis from which future interest income is calculated. An investor comparing two banks, an examiner reviewing an acquisition, or a manager evaluating the transaction’s performance can reach different conclusions if the same economic shortfall is presented once as part of acquisition measurement and again as an expense.

The overlap also explains why the answer cannot be found by looking only at the face amount of the discount. A discount is an observed difference between a reference amount and a purchase measurement; it is not a label that automatically tells the reader what caused every dollar of the difference. Credit expectations may be one input, while market yields, expected timing of cash flows, and other market-participant assumptions may be others. The accounting task is to assign each amount to the mechanism that the applicable model permits, not to make the balance sheet look as if it contains one neatly identified “credit discount.”

The same dollar cannot do two jobs

The cleanest way to understand the problem is to separate three ideas that can be compressed into the single word “discount.” First is the acquisition price or fair-value mark. It is the starting measurement of the asset in the acquirer’s books. Second is the ACL, which represents the expected credit loss recognized under Topic 326. Third is the noncredit discount or premium, which is a yield adjustment and is accreted or amortized into interest income over time under the applicable interest method.

Those ideas can interact, but they are not interchangeable. The gross-up approach used for purchased financial assets with credit deterioration, or PCD assets, makes the separation visible. At acquisition, the ACL is added to the purchase price or acquisition-date fair value to establish the initial amortized cost basis. The remaining difference between that basis and unpaid principal is treated as the noncredit purchase discount or premium. The acquisition-date credit expectation is therefore included in the loan’s initial accounting structure rather than being charged immediately to credit-loss expense.

In the simplified $100 example, a gross-up presentation would not simply subtract the $5 twice. It would place the acquisition-date ACL into the initial amortized-cost construction, while any remaining noncredit adjustment would follow the interest-income mechanism. That does not mean the loan is worth $100 in cash terms or that the buyer has eliminated the expected shortfall. It means the accounting model has assigned the initial credit expectation and the later yield adjustment to different places, so the same expectation is not treated as both an acquisition reduction and a new Day-1 expense.

The non-PCD route creates the concern that prompted the FASB change. The acquired loan is recorded at purchase price or fair value, and the ACL is recognized through the income statement. The accounting may be faithful to the model’s original distinction between acquired assets with and without significant deterioration, but FASB reports that stakeholders found the result uneconomic and difficult to analyze when fair value had already incorporated expected losses.

There is also a tempting but incorrect repair. A bank might try to estimate the “credit portion” of the fair-value mark and use it to reduce the ACL. The OCC’s 2026 Bank Accounting Advisory Series rejects that treatment for acquired loans that are not PCD or purchased seasoned loans. In that path, the entire fair-value mark is treated as a purchase premium or discount that will ultimately be accreted or amortized into interest income. The unaccreted or unamortized amount remains part of amortized cost, against which the ACL is evaluated, and the full expected credit loss is recognized through PCL.

That rule may seem counterintuitive until the accounting roles are kept separate. The non-PCD model does not authorize management to reverse-engineer a mark into a credit allowance after the fact. Its correction is not to make the allowance smaller; it is to decide whether the loan belongs in a gross-up category under the applicable guidance. For loans that remain outside that category, the full ACL and the purchase-price adjustment follow their specified paths.

flowchart LR A["Acquired loan"] --> B["Fair-value or purchase-price measurement"] B --> C["Separate ACL question"] C --> D{"Accounting path"} D --> E["PCD or qualifying PSL gross-up"] D --> F["Remaining non-PCD/non-PSL originated-like treatment"] E --> G["Noncredit discount or premium affects later interest income"] E --> H["Later credit changes through credit-loss expense"] F --> I["ACL recognized through credit-loss expense"] F --> J["Purchase discount or premium affects yield"] G --> K["Later reporting"] H --> K I --> K J --> K

The diagram’s two branches are not two ways to calculate one number. They are two ways of placing related economic information into the accounting model. Gross-up treats the acquisition-date allowance as part of the initial amortized-cost construction. Originated-like treatment leaves the acquisition measurement in place and recognizes the allowance through credit-loss expense. In both cases, later reporting still requires judgment about collection, credit change, and the appropriate interest-income mechanics.

That last point matters for readers who use yield as a shorthand for loan performance. If a purchase discount is accreted into interest income while an overlapping Day-1 loss has already reduced the asset and earnings, the later yield can look stronger than the underlying economics warrant. Conversely, a gross-up entry may avoid the initial expense but still produce later interest income from the noncredit portion of the adjustment. Reported income is therefore a sequence, not a single verdict on the quality of the acquisition.

The classification decision comes next

The question “Where does the ACL belong?” cannot be answered from the purchase price alone. The first important branch is whether the acquired loan has experienced more-than-insignificant deterioration in credit quality since origination. If it has, the loan is PCD and uses the existing gross-up model. A loan that was already weak when originated is not automatically PCD; the relevant comparison is its credit quality at origination with its credit quality at acquisition.

ASU 2025-08 adds a second gross-up route for qualifying purchased seasoned loans, or PSLs. At a high level, these are eligible loans that are not PCD but are economically more like loans acquired after an independent period of seasoning than loans originated by the acquirer. Eligible non-PCD loans other than credit cards acquired in a business combination are deemed seasoned. Other eligible loans must have been purchased more than 90 days after origination, and the acquirer must not have been involved in originating them. The detailed criteria and exclusions matter, but the idea is straightforward: the new route addresses a class of acquired loans for which fair value already reflects acquisition economics while the old non-PCD Day-1 expense could create the overlap.

The remaining loans stay on the originated-like path. This is a narrow amendment, not a declaration that every acquired financial asset should use gross-up. Credit cards are excluded from the new PSL definition, and PCD analysis still comes first. The rule also leaves important decisions about evidence, fair-value allocation, and the acquirer’s involvement in origination. Classification is not a substitute for analysis; it is the result of analysis.

The OCC’s BAAS is useful here because it translates the FASB change into bank-facing questions and examples. It is not an OCC rule or regulation, and its responses are tied to the facts presented. The FASB Update supplies the amendment and its rationale; the BAAS shows where a bank must make the distinction operational. Keeping those roles visible prevents a practical interpretation from being mistaken for a universal rule outside its stated scope.

The central lesson is therefore a boundary, not a formula. An acquired loan can carry a fair-value or purchase-price mark and an ACL because those measurements answer related but different questions. The double-count concern appears when the non-PCD Day-1 allowance overlaps with expected losses already reflected in acquisition measurement. Gross-up treatment is designed to place the acquisition-date allowance into the initial loan basis, while the noncredit portion remains a later yield adjustment. The next step is to see that separation in an actual journal entry and follow each amount through the first reporting period.

References