Unbanked

Unbanked means having no account at a federally insured bank or credit union. The word is a formal classification rather than a description, and the useful thing to know about it is that two federal definitions of it are in circulation and they measure different things.

Congress supplied one when it required the survey in the first place. 12 U.S.C. 1831z directs the FDIC to conduct a survey of what insured institutions are doing to bring "those individuals and families who have rarely, if ever, held a checking account, a savings account or other type of transaction or check cashing account at an insured depository institution (hereafter in this section referred to as the 'unbanked') into the conventional finance system." That is a test about individuals and families, and it looks at a person's whole history: rarely or never.

The FDIC's own survey, which produces every number anyone quotes, asks something narrower. A household counts as unbanked when the answer to whether anyone in the household has a checking or savings account now is no. That is a test about a household, in the present tense, and it is satisfied only when nobody in the house has an account.

The gap between those two tests is the single most useful fact on this page, because it explains why the headline rate is lower than most people expect and why "unbanked" should not be read as a synonym for poor. A person who has never held an account in their life is not counted if anyone they live with has one. A person who held an account for a decade and closed it last year is counted. Neither result is an error; they are answers to different questions, and knowing which question produced a number is the difference between using it well and misusing it.

Where the number comes from, and how often it moves. The survey is the FDIC National Survey of Unbanked and Underbanked Households, run since 2009 in partnership with the U.S. Census Bureau as a supplement to the Current Population Survey. It has been conducted in June of odd-numbered years since 2011, and the report follows in the fall of the year after. The 2023 edition, published in November 2024, collected responses from 29,483 households. So any figure on this page is a snapshot roughly two years old by the time it is read, and a new report is due in the fall of each even-numbered year. The survey's own name has changed and changed back: the 2019 wave was published as the FDIC Survey of Household Use of Banking and Financial Services.

The 2023 results, with the statistical caveats the FDIC states itself. In 2023, 4.2 percent of U.S. households, about 5.6 million, were unbanked. Conversely 95.8 percent, about 128.0 million households, were banked. The rate in 2021 was 4.5 percent, and the FDIC states that the difference between the two "was not statistically significant" at the 10 percent level, which is the significance standard it applies throughout the report. Read carefully, that means the honest summary of 2021 to 2023 is that the rate did not change, not that it improved. The longer comparison does support a real change: the rate was 8.2 percent in 2011, its highest since the survey began, and had fallen by almost half by 2023, corresponding to an additional 5.3 million banked households.

The mechanism behind the fall is measured rather than assumed, which is unusual for a statistic of this kind. The FDIC reports that about two-thirds of the decline between 2011 and 2023 was associated with changes in the socioeconomic circumstances of households over the period, particularly rising income and educational attainment. It also reports that after accounting for those changes, the remainder of the decline was still statistically significant. Both halves matter. Most of the improvement came from households getting better off rather than from banking becoming more accessible, and some of it did not.

Why households say they have no account, and what the answers are not. The 2023 survey asked unbanked households to select any applicable reasons and then to name the main one. Not having enough money to meet minimum balance requirements was the most cited on both measures, at 42.3 percent cited and 23.3 percent as the main reason. Distrust of banks was second on both, at 36.0 percent cited and 15.7 percent as the main reason. Avoiding a bank for privacy was cited by 33.9 percent, fees being too high by 30.7 percent, and fees being too unpredictable by 29.3 percent. Two answers sit further down and are worth naming because they describe barriers nothing about a household's budget will fix: 13.0 percent cited not having the personal identification required to open an account, and 12.3 percent cited problems with past banking or credit history. The Cash Flow guide carries the fee-and-minimum-balance share and how it has moved across waves.

Reading those answers as a single story is the commonest mistake made with them. Cost, distrust, privacy and identification are different obstacles that point at different remedies, and the survey shows they are unevenly distributed: among unbanked households with little or no interest in having an account, 37.9 percent cited privacy, against 24.2 percent of those who were very or somewhat interested, while lack of identification ran the other way. And among households that had never been banked, 20.4 percent cited lacking identification, against 4.8 percent of those who had previously had an account.

What being unbanked costs, stated in the survey's own terms. The transaction costs are the visible part, and each service has its own page: a household without an account pays for money orders, check cashing and money transfers that a bank account performs for nothing or nearly nothing. In 2023, 66.2 percent of unbanked households used neither a nonbank online payment service nor a prepaid card, which the FDIC labels "cash only". The less visible cost is credit. In 2023, 15.7 percent of all households had no mainstream credit, down from 20.0 percent in 2017. Among unbanked households the figure was 78.4 percent, against 13.0 percent of banked households. The FDIC's reading of that is that such households likely have no credit score with the nationwide credit reporting agencies at all. Having no credit record and having a record too sparse to score are different conditions, and the credit invisible and thin credit file pages draw that line and cover what it takes to leave either.

A boundary worth stating, because the two words are constantly swapped. Unbanked and underbanked are mutually exclusive FDIC categories, not degrees of the same thing. Underbanked households have an account. In 2023 the three categories accounted for the whole population: 4.2 percent unbanked, 14.2 percent underbanked, 81.6 percent fully banked.

The Census Bureau puts the FDIC's questions to a large sample of households alongside the Current Population Survey. A household that answers no to whether anyone in it currently has a checking or savings account at a bank or credit union is classified unbanked. Responses are weighted to the national population, and the FDIC publishes the rate, the estimated number of households, the reasons given, and how each figure breaks down by income, education, race and ethnicity, age, disability status, and how much a household's income varies from month to month.

A hypothetical example of why the classification does not map onto the thing most people picture. The numbers here are illustrative, not survey results.

Three households, all with a total income of about $40,000.

Household A is one person who has never held an account in her life and pays for everything in cash. She is unbanked.

Household B is four people. Three of them have never held an account. The fourth, an adult son, has a checking account he opened for a job that required direct deposit. The household is banked, and the three adults with no account of their own are not counted anywhere in the unbanked figure.

Household C is two people who held a joint checking account for eleven years and closed it last year after a run of overdraft fees. They are unbanked, which is the right answer for the survey's purposes and the wrong answer for the statute's, since they held an account for more than a decade.

Now the arithmetic that follows from those three. The published rate counts households, so household B contributes zero to it despite containing three adults who function without an account. Whatever the true number of individuals in that position is, it is larger than 5.6 million, and the survey does not measure it. That is not a criticism of the survey. It is the reason to know which unit a number counts before repeating it.

Pros (of the measure)

• The word has a definition in federal statute, which is rare for a term this widely used, and the FDIC survey's operational test is printed in the report itself rather than left implicit.
• The series runs back to 2009 on a consistent household-level question, so long-run comparisons of the unbanked rate are sound in a way that many financial statistics are not.
• The FDIC states its own uncertainty, labeling which differences are statistically significant and at what level, so a reader can tell a real change from noise.
• The survey decomposes the decline rather than just reporting it, separating the part associated with households becoming better off from the remainder.

Cons (of the measure)

• The statutory definition and the survey definition disagree, and virtually everything written about the subject quotes the second while sounding like the first.
• Counting households hides individuals. One account holder makes an entire household banked, however many adults in it have never had an account.
• The question is present tense, so a household that has just closed a long-held account is counted alongside one that has never had one.
• The survey runs every two years and reports the following fall, so the freshest available figure is usually well over a year old.
• Reasons for having no account are self-reported and multi-select, so cited shares sum to more than 100 percent and cannot be read as a ranked list of causes.