Banking Friction Report 2026: Access, Overdraft Risk and Account Control

A banking percentage is only useful when the reader knows who was measured, when the data were collected, and what the question meant. This report uses five evidence types and does not blend them into one score.

FinanceFirst interpretation: account ownership is the entry point. A useful banking relationship also needs affordability, reliable timing information, accessible controls, a workable shortfall path, and a way to recover after a problem.

Primary references: Federal Reserve report, FDIC survey, CFPB fee-incidence study, and CFPB qualitative research.

The FDIC estimated that 4.2% of U.S. households, about 5.6 million households, were unbanked in 2023. Another 14.2%, about 19.0 million households, were underbanked. The remaining 81.6%, about 109.1 million households, were fully banked under the 2023 FDIC framework. These three categories are mutually exclusive within that survey.

The overall unbanked rate fell from 8.2% in 2011 to 4.2% in 2023. The FDIC estimated that the change corresponded to about 5.3 million additional banked households. Progress at the national level is real, but it does not mean the remaining gap is evenly distributed.

The underbanked definition changed for the 2023 survey. A household was underbanked if it had a bank or credit-union account and used at least one of eight specified nonbank transaction or credit services in the prior 12 months. The 14.2% result should not be placed in a trend line with older underbanked estimates without reviewing the definition change.

Source: FDIC 2023 Household Survey and FDIC release summary.

The 4.2% household average hides much larger gaps. The FDIC reported 2023 unbanked rates of 10.6% for Black households, 9.5% for Hispanic households, 12.2% for American Indian or Alaska Native households, and 1.9% for White households. These differences are descriptive. They do not establish that race or ethnicity causes a household to be unbanked.

Working-age households headed by a person with a disability also faced a larger gap. The FDIC reported an 11.2% unbanked rate for these households versus 3.7% for working-age households without a disability. Their underbanked rates were 21.2% and 14.8%, respectively. Working-age households with a disability were 8.1% of households overall but 21.5% of unbanked households.

Education also mattered descriptively. The FDIC reported that 23.1% of households without a high school diploma were underbanked, compared with 10.4% of households with a college degree. These gaps point to service-design and affordability questions, not to shortcomings in the people measured.

Source: FDIC executive summary and FDIC release summary.

One-third of unbanked households, 33.4%, named a fee or minimum-balance reason as the main reason for not having an account in 2023. The combined category includes fees that are too high, fees that are too unpredictable, and not having enough money to meet a minimum-balance requirement. The second-most cited main reason was not trusting banks, at 15.7%.

The barriers were not identical for every household. Among unbanked households that had never held an account, 20.4% cited missing the personal identification needed to open an account, compared with 4.8% of previously banked unbanked households. Among unbanked households that were very or somewhat interested in an account, 46.6% cited not having enough money to meet minimum-balance requirements. The cited-reason percentages can sum above 100% because respondents could select more than one reason.

Editorial conclusion: a product can be technically available and still fail an access test if its balance requirement is unrealistic, its fees are hard to predict, its identification process has no supported path, or its service does not earn trust.

Source: FDIC 2023 full survey findings.

The Federal Reserve estimated that 94% of adults had a bank account in 2025, leaving 6% unbanked under the survey definition. Among adults with family income below $25,000, 21% were unbanked. The rate was 1% among adults with income of $100,000 or more, a 20-percentage-point descriptive gap.

Among adults who had an account, 12% said they paid an overdraft fee in the prior 12 months. The highest published income-group estimate was not the lowest-income band. It was 20% among banked adults with family income from $25,000 to $49,999. The estimate was 17% below $25,000, 12% from $50,000 to $99,999, and 6% at $100,000 or more.

The two percentage columns use different bases. Unbanked rates are among all adults in each group. Overdraft rates are among adults in the group who had a bank account. A person cannot be counted in both columns at the same time under this table.

Source: Federal Reserve SHED 2025, table 31.

An account can be open and still operate without much room for error. In 2025, 41% of adults said they always or often had money left at the end of the month. The share was 19% among adults with family income below $25,000 and 59% among those with income of $100,000 or more, a 40-percentage-point gap.

Sixty-three percent of adults said they would cover a hypothetical $400 emergency expense using cash or its equivalent, which includes a credit card paid in full at the next statement. Twelve percent said they could not pay the expense right now. Fifty-five percent had a dedicated three-month rainy-day fund, another 15% could cover three months using other resources, and 30% could not cover three months by any means.

Pressure was not hypothetical for many respondents. Fifty-nine percent reported at least one major unexpected expense during the prior year, and 16% did not pay all bills in the prior month. These figures do not prove that low cash-flow margin causes overdraft. They explain why a timing delay, pending hold, or recurring payment can matter even when annual income appears adequate.

Source: Federal Reserve SHED 2025.

Nearly half of banked households, 48.3%, used mobile banking as their primary account-access method in 2023. Half of all households, 49.7%, were using a nonbank online payment service such as PayPal, Venmo, or Cash App at the time of the FDIC survey. At the same time, almost all banked households used an in-person channel, either an ATM or a teller, at least once during the year.

Channel access was uneven for working-age households headed by a person with a disability. Smartphone access was 85.9% for those households and 93.0% for working-age households without a disability. Among banked households that accessed an account, mobile banking was the primary method for 47.4% of the disability group and 59.8% of the comparison group.

Inclusive design test: a bank should not treat an app as the only path to understand a fee, change an overdraft setting, verify a deadline, or resolve a negative balance. The practical standard is equivalent control across accessible digital, phone, and in-person channels.

Source: FDIC 2023 executive summary and FDIC survey release.

CFPB analysis of bank call reports found that combined overdraft and NSF revenue at reporting banks fell from $11.96 billion in 2019 to $5.83 billion in 2023. The calculated decline is 51.3%, or $6.13 billion. The CFPB estimated that the reduction saved the average household that overdrafts $185 per year.

The direction is favorable, but it does not mean overdraft friction disappeared. The dataset excludes banks with $1 billion or less in assets and every credit union. It measures aggregate revenue, not how many people were charged or the distribution of fees. Separately, the Federal Reserve found that 12% of banked adults paid an overdraft fee in 2025. The CFPB Making Ends Meet analysis found that 26.5% of consumers lived in a household charged an overdraft or NSF fee in its prior-year reference period. Those are different measures and should not be used as competing estimates.

The CFPB found that consumers in its frequent-fee group had lower average credit scores and were more likely to have subprime credit, no available credit on a credit card, and delinquent debt. The study reports association, not proof that the fees caused those credit outcomes.

Sources: CFPB fee-incidence study and CFPB bank revenue analysis.

A displayed balance is not always the amount that can safely be spent. Card authorizations can settle later, gas stations and rental companies can place temporary holds, an online purchase may not post until shipment, and an automatic payment can arrive before a deposit becomes available. Institutions also differ in posting order, cutoff times, weekend treatment, retry rules, and whether a pending item appears in the available balance.

The CFPB qualitative study repeatedly identified three themes: frustration with fee levels, uncertainty about payment timing and balance information, and a desire for clearer communication and control. The 36 participants were not a nationally representative sample, so the themes should guide questions rather than estimate prevalence.

Source: CFPB consumer-experience research.

For ATM withdrawals and one-time debit-card transactions, Regulation E generally prohibits an institution from charging an overdraft fee unless it first gives the consumer the required notice, obtains affirmative consent, and provides confirmation that includes the right to revoke consent. The default is not enrolled.

The rule does not apply in the same way to written checks, recurring debit transactions, or ACH payments. Declining debit-card overdraft coverage therefore does not eliminate every possible overdraft, NSF, merchant, or late-payment consequence. Opting in also does not require the institution to approve a transaction.

1. Ask for the current record of your ATM and one-time debit-card overdraft choice.
2. Ask separately how checks, ACH payments, and recurring debits are handled.
3. Request the per-item fee, daily limit, sustained negative-balance rule, and grace deadline in writing.
4. Compare a declined payment, standard overdraft, linked-account transfer, and credit line by total consequence, not just the headline fee.

This is an educational summary, not legal advice. Review the current account agreement and seek qualified help for a dispute. Sources: Regulation E section 1005.17 and CFPB consumer guidance.

The national sources identify where friction appears. They do not provide a current product-by-product comparison. FinanceFirst created this reproducible audit so a reader can inspect one real account using its agreement, fee schedule, app settings, recent statement, and a conversation with the institution.

Do not score what you cannot verify. Mark a setting as unknown until the institution provides a current answer. The audit is designed to expose uncertainty, not hide it behind a numerical rating.

The same account can work differently for people with different income schedules, accessibility needs, languages, caregiving duties, or levels of digital connectivity. These examples are prompts for testing the account, not assumptions about any group.

When an account may go negative, the useful question is not only how to avoid one bank fee. The decision should consider essential services, merchant charges, late-payment consequences, credit costs, and the time required to restore the balance.

1. Reconstruct the available balance. List the ledger balance, pending holds, scheduled ACH payments, checks, card settlements, and deposits that are not yet available.
2. Rank outgoing payments by consequence. Separate essentials and obligations with serious penalties from payments that can be moved or paused.
3. Confirm transaction-specific rules. Ask how the institution will handle each ATM, debit, ACH, check, and recurring item.
4. Use the least harmful verified path. Compare decline, merchant arrangement, grace period, linked transfer, or credit using total cost and nonfinancial consequences.
5. Capture proof. Save the alert, account screenshot, disclosure, representative name, confirmation number, and time of any change.
6. Stop repeated surprises. After the immediate issue, change the alert threshold, due date, coverage choice, or account if the current setup cannot be made predictable.

A declined essential payment can be more harmful than a fee in some circumstances. A paid overdraft can be more harmful in others. The plan does not prescribe one answer. It creates a record for a deliberate choice.

Consider an illustrative account with $20 available, a $25 grocery purchase, and a $60 scheduled utility payment. The combined outflow exceeds the available balance by $65. The result depends on settings and timing, not just arithmetic.

This is a process demonstration, not a universal cost estimate. Current terms vary by institution and account. The CSV dataset labels the scenario as a FinanceFirst framework rather than a published estimate.

The evidence suggests that meaningful access requires more than opening an account. The following design standards are an editorial synthesis of the documented problems and can be used by readers, product teams, community organizations, and financial educators.

• Predictable cost: fee amount, trigger, cap, grace period, and alternatives are stated before the decision.
• Accurate timing: available balance explains holds, pending items, deposit availability, and cutoff times.
• Real choice: a person can select which transaction types may overdraw and can revoke consent without friction.
• Accessible channels: essential controls work with assistive technology and through phone and in-person alternatives.
• Language access: critical disclosures and support are available in the language the customer understands.
• Recovery before exclusion: alerts, human support, and a clear cure path arrive before account closure becomes the only response.
• Evidence trail: the institution can show when consent was obtained, when an alert was sent, and how the balance was calculated.

No cited source publishes this exact framework. FinanceFirst built it by connecting the quantitative access and overdraft evidence with the CFPB themes of timing, communication, and control. That synthesis is the report's original contribution.

• Federal Reserve estimates are weighted self-reports from U.S. adults surveyed from October 17 to October 28, 2025.
• FDIC estimates describe U.S. households surveyed in June 2023. The underbanked definition changed from earlier waves.
• CFPB Making Ends Meet estimates describe consumers and household fee experience in a separate survey reference period.
• Call-report totals cover banks with more than $1 billion in assets. They exclude smaller banks and all credit unions.
• The CFPB qualitative study included 36 participants and is used for themes, not prevalence.
• Subgroup differences are descriptive. They do not prove that demographic identity causes a banking outcome.
• Account terms can change. Readers should verify the current agreement, fee schedule, and settings before acting.

FinanceFirst calculations: 21% minus 1% equals a 20-point low-to-high-income difference in the adult unbanked rate. 25% minus 8% equals a 17-point Black-White overdraft-incidence difference among banked adults in the 2025 Federal Reserve table. $5.83 billion minus $11.96 billion, divided by $11.96 billion, equals a 51.3% decline in reported bank overdraft and NSF revenue from 2019 to 2023. No cross-source composite score is calculated.

August 2026, version 2.0: added Federal Reserve 2025 overdraft subgroup data, resilience measures, FDIC subgroup and digital-access evidence, a full revenue series, an inclusive account-control audit, a shortfall response plan, inline source navigation, and expanded machine-readable citations.

Review the research standards, the corrections policy, or submit a data question or correction.