CAPITAL ONE FINANCIAL CORP

The preparation of financial statements in accordance with U.S. GAAP requires management to make a number of judgments, estimates and assumptions that affect the amount of assets, liabilities, income and expenses on the consolidated financial statements. Understanding our accounting policies and the extent to which we use management judgment and estimates in applying these policies is integral to understanding our financial statements. We provide a summary of our significant accounting policies under “Item 8. Financial Statements and Supplementary Data—Note 1—Summary of Significant Accounting Policies.”

We have identified the following accounting estimates as critical because they require significant judgments and assumptions about highly complex and inherently uncertain matters and the use of reasonably different estimates and assumptions could have a material impact on our results of operations or financial condition. Our critical accounting policies and estimates are as follows:

•Loan loss reserves

•Goodwill

•Fair value

•Customer rewards reserve

We evaluate our critical accounting estimates and judgments on an ongoing basis and update them as necessary, based on changing conditions.

Loan Loss Reserves

We maintain an allowance for credit losses that represents management’s current estimate of expected credit losses inherent in our credit card, consumer banking and commercial banking loans held for investment portfolios as of each balance sheet date. We also reserve for the uncollectible portion of finance charges and fees related to credit card loan receivables in the allowance for credit losses consistent with the methodology we use to estimate the allowance for credit losses on the principal portion of our credit card loan receivables. We also separately reserve for unfunded lending commitments that are not unconditionally cancellable. We recognize changes in our allowance for credit losses and reserve for unfunded lending commitments through the provision for credit losses. The allowance for credit losses was $16.3 billion as of December 31, 2024, compared to $15.3 billion as of December 31, 2023.

Our allowance for credit losses and reserve for unfunded lending commitments utilize models to derive a quantitative estimate of credit losses that is supplemented with additional qualitative considerations to capture risks and uncertainties not included in the quantitative result. Our estimate of expected credit losses, for all loan and unfunded lending commitments, includes a reasonable and supportable forecast period of one year and then reverts over a one-year period to historical losses at each relevant loss component of the estimate. We use externally produced consensus estimates as inputs for our forward-looking macroeconomic forecast and consider other forecasts and sources of uncertainty to develop the quantitative component. This quantitative result is then supplemented qualitatively by management for economic uncertainty, including the consideration of alternative macroeconomic scenarios, changes and trends in loan portfolios that may not be captured in the quantitative component. These adjustments represent management’s judgment of the imprecision and risks inherent in the processes and assumptions used in establishing the allowance for credit losses.

We have an established process, using analytical tools and management judgment, to determine our allowance for credit losses. Significant management judgment is required to determine the relevant information and estimation methods used to arrive at our best estimate of lifetime credit losses. Establishing the allowance on a quarterly basis involves evaluating and forecasting both credit and macroeconomic variables. The macroeconomic forecast used to inform both quantitative and qualitative components of our allowance for credit losses estimate is sensitive to certain variables, such as the U.S. Unemployment Rate, and the U.S. Real Gross Domestic Product (“U.S. Real GDP”) Growth Rate assu mptions. Our December 31, 2024 allowance assumes that the quarterly average U.S. unemployment rate is approximately 4.3% by the fourth quarter of 2025 and annual U.S. Real GDP grows by 1.8% in 2025.

In addition to macroeconomic factors, many credit factors inform our allowance for credit losses, including, but not limited to, historical loss and recovery experience, recent trends in delinquencies and charge-offs, risk ratings, the impact of bankruptcy

filings, the value of collateral underlying secured loans, account seasoning, changes in our credit evaluation, underwriting and collection management policies, seasonality, credit bureau scores, current general economic conditions, changes in the legal and regulatory environment and uncertainties in forecasting and modeling techniques used in estimating our allowance for credit losses.

We have a governance framework supported by processes and controls intended to ensure that our estimate of the allowance for credit losses is appropriate. Our governance framework provides for oversight of methods, models, qualitative adjustments, process controls and results. At least quarterly, representatives from the Finance and Risk Management organizations review and assess our allowance methodologies, key assumptions and the appropriateness of the allowance for credit losses. Groups independent of our estimation functions participate in the review and validation process. Tasks performed by these groups include periodic review of the rationale for and quantification of inputs requiring judgment as well as adjustments to results.

We have model policies, established by an independent Model Risk Office, which govern the validation of models and related supporting documentation to ensure the appropriate use of models for estimating credit losses. The Model Risk Office validates all models and requires ongoing monitoring of their performance.

In addition to the allowance for credit losses, on a quarterly basis, we review and assess our estimate of expected losses related to unfunded lending commitments that are not unconditionally cancellable which are generally in our Commercial Banking business. The factors impacting our assessment generally align with those considered in our evaluation of the allowance for credit losses for the Commercial Banking business. The reserve for losses on unfunded lending commitments is included in other liabilities on our consolidated balance sheets and changes to it are recorded through the provision for credit losses in our consolidated statements of income.

Although we examine a variety of externally available data, as well as our internal loan performance data, to determine our allowance for credit losses and reserve for unfunded lending commitments, our estimation process is subject to risks and uncertainties, including a reliance on historical loss and trend information that may not be representative of current conditions and indicative of future performance as well as economic forecasts that may not align with actual future economic conditions. Accordingly, our actual credit loss experience may not be in line with our expectations. We provide additional information on the methodologies and key assumptions used in determining our allowance for credit losses for each of our loan portfolio segments in “Item 8. Financial Statements and Supplementary Data—Note 1—Summary of Significant Accounting Policies.” We provide information on the components of our allowance, disaggregated by operating segment, and changes in our allowance in “Item 8. Financial Statements and Supplementary Data—Note 5—Allowance for Credit Losses and Reserve for Unfunded Lending Commitments.”

Goodwill

Goodwill represents the excess of the fair value of the consideration transferred, plus the fair value of any non-controlling interests in the acquiree, over the fair value of the assets acquired and liabilities assumed as of the acquisition date.

Goodwill totaled $15.1 billion as of both December 31, 2024 and 2023. We did not recognize any goodwill impairment in 2024 or 2023. See “Item 8. Financial Statements and Supplementary Data—Note 7—Goodwill and Other Intangible Assets” for additional information.

We perform our goodwill impairment test annually on October 1 at a reporting unit level. We are also required to test goodwill for impairment whenever events or circumstances indicate it is more-likely-than-not that an impairment may have occurred. An impairment of a reporting unit’s goodwill is determined based on the amount by which the reporting unit’s carrying amount exceeds its fair value, limited to the amount of goodwill allocated to the reporting unit. We have four reporting units: Credit Card, Auto Finance, Other Consumer Banking, and Commercial Banking.

For the purpose of our goodwill impairment testing, we calculate the carrying amount of a reporting unit using an allocated capital approach based on each reporting unit’s specific regulatory capital requirements and underlying risks. The carrying amount for a reporting unit is the sum of its respective capital requirements, goodwill and other intangibles balances. Consolidated stockholder’s equity in excess of the sum of all reporting units capital requirements that is not identified for future capital needs, such as dividends, share buybacks, or other strategic initiatives, is allocated to the reporting units and the Other category and assumed to be distributed to equity holders in future periods.

Determining the fair value of a reporting unit is a subjective process that requires the use of estimates and the exercise of significant judgment. We calculate the fair value of our reporting units using a discounted cash flow (“DCF”) calculation, a

form of the income approach. This DCF calculation uses projected cash flows based on each reporting unit’s internal forecast and the perpetuity growth method to calculate terminal values. Our DCF calculation requires management to make estimates about future loan, deposit and revenue growth, as well as credit losses and capital rates. These cash flows and terminal values are then discounted using discount rates based on our external cost of capital with adjustments for the risk inherent in each reporting unit. Discount rates used for our reporting units ranged from 8.4% to 12.7%, and we applied a terminal year long-term growth rate of 3.8% to all reporting units. The reasonableness of our DCF calculation is assessed by reference to a market-based approach using comparable market multiples and recent market transactions where available. The usefulness of market data is inherently limited due to the size and scope of our operations compared to most peer institutions and recent market transactions. The results of the 2024 annual impairment test indicated that the estimated fair values of the reporting units exceeded their carrying amounts by between 11% and 83%. We also compare the aggregate fair values of our reporting units to our market capitalization. Our assessment considers the level of premium expected to assume control of the Company in a market transaction including anticipated cost savings and other synergies that would be realized in a hypothetical transaction.

The results of the 2024 annual goodwill impairment test for our Commercial Banking reporting unit concluded that, while the estimated fair value of this reporting unit exceeded its carrying amount, the percentage by which the estimated fair value of this reporting unit exceeded its carrying amount was 11%. The assumptions leveraged in the valuation of each reporting unit, including the Commercial Banking reporting unit, and the related risk of changes in those assumptions are described further below.

Assumptions used in estimating the fair value of a reporting unit are judgmental and inherently uncertain. A change in the economic conditions of a reporting unit, such as declines in business performance as a result of industry or macroeconomic trends or changes in our strategy, adverse impacts to loan or deposit growth trends, decreases in revenue, increases in expenses, deterioration in a significant loan portfolio, increases in credit losses, increases in capital requirements, deterioration of market conditions, declines in long-term growth expectations, an increase in disposition activity, adverse impacts of regulatory or legislative changes or increases in the estimated cost of capital could cause the estimated fair values of our reporting units to decline in the future, and increase the risk of a goodwill impairment in a future period. We perform sensitivity analyses around certain assumptions in order to assess the reasonableness of the assumptions and the resulting estimated fair values.

We have a governance framework supported by processes and controls intended to ensure that the accounting and disclosure for goodwill is appropriate. Our governance framework provides for oversight of assumptions, forecast inputs, methods, process controls and results.

Fair Value

Fair value, also referred to as an exit price, is defined as the price that would be received for an asset or paid to transfer a liability in an orderly transaction between market participants on the measurement date. The fair value accounting guidance provides a three-level fair value hierarchy for classifying financial instruments. This hierarchy is based on the markets in which the assets or liabilities trade and whether the inputs to the valuation techniques used to measure fair value are observable or unobservable. The fair value measurement of a financial asset or liability is assigned a level based on the lowest level of any input that is significant to the fair value measurement in its entirety. The three levels of the fair value hierarchy are described below:

Level 1: Valuation is based on quoted prices (unadjusted) in active markets for identical assets or liabilities.

Level 2: Valuation is based on observable market-based inputs other than Level 1 prices, such as quoted prices for similar assets or liabilities, quoted prices in markets that are not active, or other inputs that are observable or can be corroborated by observable market data for substantially the full term of the assets or liabilities.

Level 3: Valuation is generated from techniques that use significant assumptions not observable in the market. Valuation techniques include pricing models, DCF methodologies or similar techniques.

The degree of management judgment involved in determining the fair value of a financial instrument is dependent upon the availability of quoted prices in active markets or observable market inputs. When quoted prices and observable data in active markets are not fully available, management judgment is necessary to estimate fair value. Changes in market conditions, such as reduced liquidity in the capital markets or changes in secondary market activities, may reduce the availability and reliability of quoted prices or observable data used to determine fair value.

We have developed policies and procedures to determine when markets for our financial assets and liabilities are inactive if the level and volume of activity has declined significantly relative to normal conditions. If markets are determined to be inactive, it may be appropriate to adjust price quotes received. When significant adjustments are required to price quotes or inputs, it may be appropriate to utilize an estimate based primarily on unobservable inputs.

Significant judgment may be required to determine whether certain financial instruments measured at fair value are classified as Level 2 or Level 3. In making this determination, we consider all available information that market participants use to measure the fair value of the financial instrument, including observable market data, indications of market liquidity and orderliness, and our understanding of the valuation techniques and significant inputs used. Based upon the specific facts and circumstances of each instrument or instrument category, judgments are made regarding the significance of the Level 3 inputs to the instruments’ fair value measurement in its entirety. If Level 3 inputs are considered significant, the instrument is classified as Level 3. The process for determining fair value using unobservable inputs is generally more subjective and involves a high degree of management judgment and assumptions. We discuss changes in the valuation inputs and assumptions used in determining the fair value of our financial instruments, including the extent to which we have relied on significant unobservable inputs to estimate fair value and our process for corroborating these inputs, in “Item 8. Financial Statements and Supplementary Data—Note 17—Fair Value Measurement.”

We have a governance framework and a number of key controls that are intended to ensure that our fair value measurements are appropriate and reliable. Our governance framework provides for independent oversight and segregation of duties. Our control processes include review and approval of new transaction types, price verification, and review of valuation judgments, methods, models, process controls and results.

Groups independent of our trading and investing functions participate in the review and validation process. Tasks performed by these groups include verification of fair value measurements at least quarterly to determine if assigned fair values are reasonable, including comparing prices from vendor pricing services to other available market information.

Our Fair Value Committee (“FVC”), which includes representation from business areas, Risk Management and Finance, provides guidance and oversight to ensure an appropriate valuation control environment. The FVC reviews and approves our fair valuations at least quarterly to ensure that our valuation practices are consistent with industry standards and adhere to regulatory and accounting guidance.

We have model policies, established by an independent Model Risk Office, which govern the validation of models and related supporting documentation to ensure the appropriate use of models for pricing and fair value measurements. The Model Risk Office validates all models and requires ongoing monitoring of their performance.

The fair value governance process is set up in a manner that allows the Chairperson of the FVC to escalate valuation disputes that cannot be resolved by the FVC to a more senior committee called the Valuations Advisory Committee (“VAC”) for resolution. The VAC is chaired by the Chief Financial Officer (“CFO”) and includes other members of senior management. There were no disputes escalated to the VAC for the years ended December 31, 2024 and 2023.

Customer Rewards Reserve

We offer products, primarily credit cards, which include programs that allow members to earn rewards based on account activity that can be redeemed for cash (primarily in the form of statement credits), gift cards, travel, or covering eligible charges. The amount of rewards that a customer earns varies based on the terms and conditions of the rewards program and product. The majority of our rewards do not expire and there is no limit on the amount of rewards an eligible card member can earn. Customer rewards costs, which we generally record as an offset to interchange income, are driven by various factors such as card member purchase volume, the terms and conditions of the rewards program and rewards redemption cost. We establish a customer rewards reserve that reflects management’s judgment regarding rewards earned that are expected to be redeemed and the estimated redemption cost.

We use financial models to estimate ultimate redemption rates of rewards earned by current card members based on historical redemption trends, current enrollee redemption behavior, card product type, year of program enrollment, enrollment tenure and card spend levels. Our current assumption is that the substantial majority of all rewards earned will eventually be redeemed. We use the weighted-average redemption cost during the previous twelve months, adjusted as appropriate for recent changes in redemption costs, including changes related to the mix of rewards redeemed, to estimate future redemption costs. We continually evaluate our reserve and assumptions based on developments in redemption patterns, changes to the terms and

conditions of the rewards program and other factors. While the rewards liability is sensitive to changes in assumptions for redemption rates and costs and involves management judgment, we believe portfolio characteristics and historical performance are the best indication of future reward redemption behavior and are the primary basis for our estimate. We recognized customer rewards expense of $9.0 billion, $8.2 billion and $7.6 billion in 2024, 2023 and 2022, respectively. Our customer rewards reserve, which is included in other liabilities on our consolidated balance sheets, totaled $8.2 billion and $7.4 billion as of December 31, 2024 and 2023, respectively.

We have a governance framework supported by processes and controls that are intended to ensure that our rewards liability estimate is appropriate and reliable. Our governance framework provides for oversight of assumptions, inputs, methods, process controls and results. Additional controls are performed to ensure all underlying data used to derive the rewards liability is complete and accurate.