An estimated 37.5 million US households, or 27.8%, owned a Roth IRA in mid-2025, according to the Investment Company Institute (ICI). Among Roth-owning households, 42% reported contributing for tax year 2024 and 6% reported a withdrawal. Those are household survey results—not counts of individual accounts or a scorecard for how much you should have saved.
Source: ICI report and supplemental Figures A1, A14, A20 and A31. These figures do not describe 2026 activity.
September 2026 update: what does Fidelity's $144,523 average mean?
Fidelity's Q2 2026 retirement analysis, released September 3, reports an average IRA balance of $144,523 as of June 30, 2026. Its IRA analysis covers 20.3 million Fidelity IRA accounts and includes only active participants with a balance.
That is a provider's IRA average, not a Roth-only figure, a national household median, or a recommended balance. It cannot update ICI's $48,000 Roth household median. The two numbers describe different populations, account groupings, dates and measures.
| Figure | What it measures | Observation date |
|---|---|---|
| $48,000 | ICI median household assets in Roth IRAs: the middle of the distribution | Mid-2025 |
| $118,000 | ICI mean household assets in Roth IRAs: the arithmetic average | Mid-2025 |
| $144,523 | Fidelity average IRA balance: its provider analysis, not Roth-only | June 30, 2026 |
ICI Figure A14; Fidelity Q2 2026 analysis, IRA balance and footnote 1. Neither source supplies a personal retirement target.
Average versus median: a small example
Imagine five households with Roth balances of $10,000, $20,000, $30,000, $40,000 and $400,000. Their median is $30,000, the middle balance. Their average is $100,000: $500,000 divided by five.
Only one of those five households has as much as the average. That is why an average can feel surprisingly high. This is made-up arithmetic to explain the terms, not ICI's underlying sample or evidence of how many households fall below Fidelity's average.
Who owns a Roth IRA?
ICI estimates that 43.9 million households owned Traditional IRAs and 37.5 million owned Roth IRAs in mid-2025. Some owned both. You cannot add the two counts to find the number of IRA-owning households; ICI's estimate for households with any IRA type is 59.6 million.
Age groups are based on the household survey respondent, not the ages of every IRA owner in that household. Roth ownership was more common than Traditional ownership in the two groups under age 45.
| Respondent's age | Own Roth IRA | Own Traditional IRA |
|---|---|---|
| Under 35 | 28% | 17% |
| 35–44 | 30% | 22% |
| 45–54 | 29% | 28% |
| 55–64 | 30% | 39% |
| 65 or older | 24% | 48% |
ICI Figures A1 and A2. A household can appear in both ownership columns.
How does ownership vary with income?
Roth-owning households appear across the income range. Ownership rates are higher in the higher-income groups, but this is ownership, not eligibility to make a new contribution.
| Household income | Share owning Roth IRAs |
|---|---|
| Under $50,000 | 9% |
| $50,000–$74,999 | 21% |
| $75,000–$99,999 | 28% |
| $100,000–$149,999 | 37% |
| $150,000–$199,999 | 42% |
| $200,000 or more | 52% |
ICI Figure A6, including its income-year footnote and under-$50,000 memo row.
The IRS uses tax filing status and modified adjusted gross income (MAGI) for direct Roth contribution limits. Survey household income is not the same measure. Having a high income today also does not require you to close a Roth you already own. See the contribution eligibility guide and IRS Publication 590-A, chapter 2.
How has Roth ownership changed?
ICI's published estimates rise from 9.8 million Roth-owning households in 2000 to 37.5 million in 2025. The chart shows selected years, not a smooth annual growth path.
- 2000
- 9.8M
- 2010
- 19.5M
- 2019
- 24.9M
- 2025
- 37.5M
ICI Figure A1. Sampling and survey-method changes limit comparisons across this span.
Read the long history with care. ICI changed its sampling and weighting in 2014 and moved the ownership tracking survey to an online panel in 2022. The comparison spans those methodology changes; it is not a clean measure of growth caused only by new account openings.
Which generations own Roth IRAs?
| Respondent's generation | Own Roth IRA | Own Traditional IRA |
|---|---|---|
| Gen Z (born 1997–2012) | 22% | 13% |
| Millennials (1981–1996) | 31% | 22% |
| Gen X (1965–1980) | 30% | 31% |
| Baby Boomers (1946–1964) | 26% | 47% |
| Silent/GI generations (1904–1945) | 20% | 50% |
ICI Figure A4. All respondents were at least 18, so the Gen Z row does not represent children in that generation.
This is a snapshot of different groups. It does not tell us which IRA people opened first or predict how the same households will save as they get older.
What does the household profile look like?
| Measure | Roth-owning households | Households without IRAs |
|---|---|---|
| Median respondent age, 2025 | 51 | 49 |
| Median pretax household income, 2024 | $137,500 | $60,000 |
| Respondent has college or postgraduate degree | 65% | 26% |
| Respondent employed full- or part-time | 74% | 56% |
The rows describe separate characteristics; there is no single person who represents all of them. In particular, the employment row describes the respondent, not whether anyone in the household works.
Roth IRAs also commonly sit alongside workplace savings. In Figure A8, 87% of Roth-owning households had a defined-contribution account or defined-benefit pension coverage; 79% had a defined-contribution account. That broader category includes 401(k)-style plans, not only 401(k)s. These overlapping percentages do not establish which account should be funded first.
How often do Roth-owning households contribute?
For tax year 2024, 42% of Roth-owning households reported contributing to a Roth, compared with 23% of Traditional-owning households contributing to a Traditional IRA.
- Roth-owning households
- 42%
- Traditional-owning households
- 23%
ICI Figure A20. Each percentage uses households owning that IRA type as its denominator.
This measures whether a contribution happened, not which group saved more overall. People may save through workplace plans or other accounts, and the two IRA-owning populations overlap. Among contributing households, the median respondent age was 42 for Roth contributors and 52 for Traditional contributors (Figure A20).
Among Roth-owning households that did not contribute, 24% said they had no extra money, 18% said they were saving enough through a workplace plan, 28% said they were retired and no longer saving, and 21% reported being ineligible. Respondents could give multiple reasons. The ineligibility response does not identify a particular tax rule or establish that a backdoor Roth is available to that household (Figure A19).
How often do Roth-owning households withdraw?
- Roth-owning households
- 6%
- Traditional-owning households
- 33%
ICI Figures A29 and A31. Each percentage uses households owning that IRA type as its denominator.
In tax year 2024, 6% of Roth-owning households and 33% of Traditional-owning households reported withdrawals. Conversely, 94% and 67% reported no withdrawal. No withdrawal does not mean no trading, no contributions or a guaranteed increase in value.
Among the Traditional-owning households that withdrew, 70% said they determined the amount using the required minimum distribution (RMD) rule. That is not 70% of all Traditional IRA owners, withdrawal transactions or dollars. It also does not establish that they otherwise would have withdrawn nothing (Figures A26 and A29).
There is an important rule difference: an original Roth IRA owner has no lifetime RMDs. Inherited Roth IRAs have beneficiary distribution rules. The survey comparison does not isolate the effect of that rule from differences in age, retirement status or other circumstances. See Roth IRA RMD rules and IRS Publication 590-B, chapter 2.
Among Roth-owning households that withdrew, 32% reported using money for living expenses and 23% for home purchase, repair or remodeling. Multiple uses could be reported. A reported use is not proof that the withdrawal qualified for a tax or penalty exception (Figure A31).
How much do Roth-owning households hold?
ICI Figure A14 reports $48,000 in median household Roth IRA assets and $118,000 in mean assets. These are combined household holdings in that IRA type, not necessarily one account. The corresponding Traditional IRA median was $125,000.
Balances vary widely: 17% of Roth-owning households reported less than $10,000 and 14% reported $250,000 or more. The table does not establish how those larger balances were built.
A source detail worth noting: ICI Figure A10 separately reports a $48,500 overall Roth median in its conversion-status table. We use Figure A14's $48,000 for the balance distribution here; we have not blended the two or treated the difference as a change over time.
What changes with length of ownership?
- Owned less than 10 years
- $20,000
- Owned 10–19 years
- $78,000
- Owned 20 years or more
- $100,000
ICI Figure A13. Different households at one point in time, not one account's growth path.
These are different households observed in 2025, grouped by how long they had owned Roth IRAs. The chart does not follow one account from $20,000 to $100,000. Contributions, conversions, withdrawals, investments and household circumstances can differ across the groups. It cannot tell us how much of the difference time or investment returns caused.
A separate illustration of compounding
Start at $0, add an illustrative $7,000 at each year-end, and assume a constant 7% annual return after fees, with no withdrawals. The calculated balances are approximately $96,715 after 10 years, $286,968 after 20, and $661,226 after 30.
The formula is $7,000 × [(1.07n − 1) ÷ 0.07], where n is the number of years. This is nominal-dollar arithmetic, not a forecast, an actual return history, or an explanation of ICI's medians. Inflation is not deducted, returns are not guaranteed, and the annual contribution assumes eligibility; $7,000 is an example input, not a statement of the current limit.
Roth earnings are free of federal income tax when withdrawn in a qualified distribution—not simply because the owner calls it retirement. Every qualifying route requires the Roth IRA five-tax-year period and one qualifying event: reaching age 59½, qualifying disability, payment after the owner’s death, or a qualifying first-home distribution within the $10,000 lifetime limit. A penalty exception alone does not make earnings income-tax-free. See the full withdrawal rules.
What investments do households hold?
In Figure A16, 71% of Roth-owning households held mutual funds, 46% held individual stocks, 42% held exchange-traded funds (ETFs), 13% held annuities, and 11% held bank savings, money-market deposit accounts or certificates of deposit (CDs).
These are shares of households holding each investment type, not percentages of their portfolios. A household can hold several types. A Roth IRA is the account; investments inside it have their own risks. The survey does not establish that stocks are always the right choice or that holding a CD wastes a Roth. For a plain-language starting point, see what you can hold in a Roth IRA.
Where are Roth IRAs held?
ICI Figure A11 places 74% of Roth-owning households in its investment-professional channel category. Specific providers included full-service brokerages (38%), independent financial-planning firms (25%), and banks or savings institutions (16%). Direct mutual fund companies accounted for 17%, and discount brokerages for 18%.
These categories overlap because households can use more than one provider. They describe where accounts are held, not a recommendation or proof that every customer receives personal advice. The survey does not establish that younger owners opened accounts on a phone or paid lower fees.
How do households plan for retirement?
In the 2025 survey, 65% of Roth-owning households reported having a strategy for managing retirement income and assets. Among that group with a strategy, 77% consulted a professional financial adviser, and 65% identified an adviser as their primary source (Figure A34).
Also within the strategy group, 73% reviewed asset allocation, 69% developed a retirement-income plan, 66% estimated retirement expenses, 63% set aside emergency funds and 54% considered when to claim Social Security. Respondents could report multiple steps; 72% reported at least three (Figure A35). Those percentages should not be applied to all Roth-owning households.
What does the survey say about conversions?
Figure A10 reports that 21% of Roth-owning households had funded their Roth IRAs at least partly through a conversion from a Traditional IRA. Median household financial assets were $941,000 for that group and $400,000 for those without such conversions. Those assets include workplace retirement savings but exclude the primary home.
The survey associates conversion history with higher household assets; it does not show that converting caused greater wealth or was the better tax decision. A conversion moves existing retirement money and can create taxable income. A regular contribution adds new money under annual contribution rules. Our conversion tax guide explains the distinction.
How should you use these numbers?
- Check who is counted. A US household, a provider's account and an individual saver are different units.
- Check what is measured. Ownership rates, contribution participation, balances and investment returns answer different questions.
- Check the year. ICI surveyed households in May–June 2025; contribution, withdrawal and income questions here concern 2024. Fidelity's separate note observes June 30, 2026.
- Keep comparisons in perspective. A balance alone omits age, other savings, pensions, spending needs and how many accounts a household has.
ICI's report draws on two weighted household surveys: a 9,021-household ownership tracking survey and a 3,214-household IRA Owners Survey. Both used Ipsos's probability-based KnowledgePanel in May–June 2025. Survey responses are estimates, not an IRS census, and small subgroup differences should not automatically be treated as meaningful.
We did not conduct either ICI survey or Fidelity's analysis. This page interprets their published results; it does not combine their underlying datasets.
Sources and data dates
- ICI, The Role of IRAs in US Households' Saving for Retirement, 2025 (PDF), Sarah Holden, Michael Bogdan and Daniel Schrass, Research Perspective 32, no. 7, June 2026. Published June 3, 2026. Methodology: printed pp. 17–18.
- ICI supplemental tables (XLSX), June 2026. Figures A1, A2, A4, A6, A8, A10, A11, A13, A14, A16, A19, A20, A26, A29, A31, A34 and A35 support the statistics above.
- Fidelity Q2 2026 retirement analysis, released September 3, 2026; IRA observation date June 30, 2026. IRA sample definition appears in footnote 1.
- IRS Publication 590-A and Publication 590-B, 2025 editions, chapter 2: contribution eligibility, conversions, qualified Roth distributions and owner/beneficiary withdrawal rules.
Sources checked September 27, 2026. Checking a source today does not change the year its data describe.
Frequently asked questions
How many US households own a Roth IRA?
ICI estimates that 37.5 million US households, or 27.8%, owned Roth IRAs in mid-2025. These are households, not individual accounts. Some households also owned Traditional IRAs.
What is the typical Roth IRA balance?
ICI Figure A14 reports median household Roth IRA assets of $48,000 and mean assets of $118,000 in 2025. These combine a household's Roth IRAs. Fidelity's $144,523 Q2 2026 average IRA balance is a different provider measure, not a Roth-only household median or a savings target.
Were Roth-owning households more likely to contribute?
Yes. For tax year 2024, 42% of Roth-owning households reported a Roth contribution, versus 23% of Traditional-owning households reporting a Traditional contribution. This measures participation, not which group saved more overall.
Does the withdrawal survey prove the effect of RMD rules?
No. In tax year 2024, 6% of Roth-owning households and 33% of Traditional-owning households reported withdrawals. Original Roth owners have no lifetime RMDs, but the comparison does not isolate that rule from differences in age and other circumstances. Inherited Roth IRAs have beneficiary distribution rules.
Corrections and updates
September 27, 2026 — Correction: Corrected the ICI report's month and source link, the income observation year, respondent-versus-household descriptions, and the Roth direct-mutual-fund provider share (17%, not 19%). The 70% RMD figure refers to withdrawing Traditional-IRA households and how they determined the amount, not 70% of transactions or dollars. Contribution participation does not measure overall saving. Added missing subgroup, overlapping-response and survey-method caveats. Removed unsupported causal, forecasting and blanket investment-location conclusions, and qualified the tax-free-earnings explanation. The compounding illustration now states timing and return assumptions and shows recalculated values. Public correction record.
September 27, 2026 — Update: Added Fidelity's September 3 Q2 2026 release as a separately labeled provider-data note. It does not replace ICI's 2025 household statistics.
Continue reading
- Roth IRA basics — how the account works.
- Roth versus Traditional IRA — the rule differences behind the labels.
- The Backdoor Roth, by the Numbers — a separate historical tax-data study.