IRS data study · tax year 2023
Roth IRA contribution sizes, by age
How much did people put in? Explore four annual contribution bands, one age group at a time.
This study examines the Roth IRA contribution counts published in IRS Table 6. It does not analyze the separate rollover or conversion totals. Contributions are different from the account’s balance or investment growth. IRS estimates put 9,488,414 individual taxpayers in the Roth-contribution population for 2023. About 68.4% were in the under-$6,500 band.
That broad band does not tell us whether someone made small monthly deposits or one larger payment. And this is a view of contributors—not accounts, households, all Roth owners or all eligible Americans.
Where did contributions fall?
The denominator is everyone with a Roth contribution in the selected age group. Every bar uses the same 0–100% scale.
Changes only the four bars below. The comparison and full table stay visible.
- Under $6,5006,490,152 (68.4%)
- Exactly $6,5001,853,959 (19.5%)
- Over $6,500, under $7,500214,121 (2.3%)
- Exactly $7,500930,181 (9.8%)
Counts are estimates; percentages are our calculations from the reported counts. Totals can differ slightly because of rounding.
* The exactly-$6,500 estimate has a small-sample caution; its count interval is suppressed. Read the note.
Why these dollar bands?
The general 2023 IRA contribution cap was $6,500, or $7,500 for someone age 50 or older by year-end. That helps explain the table’s cutoffs. A personal maximum also depends on eligible compensation (including qualifying spousal rules), income limits and contribution room shared with Traditional IRAs. See 2023 IRS Publication 590-A, pages 38–39.
So “exactly $6,500” is an amount band, not proof that someone used all their available room. These are historical figures; use our contribution-limit guide for current rules.
The under-$6,500 share across ages
Read each bar as the share within that age group. Rows stay in age order, with all ages as a reference.
- All ages68.4%
- Under 2077.6%
- 20–2475.3%
- 25–2970.3%
- 30–3468.2%
- 35–3973.4%
- 40–4471.5%
- 45–4976.8%
- 50–5468.2%
- 55–5959.1%
- 60–6452.6%
- 65+55.1%
Different people at different ages in one tax year: this does not track people as they grow older, explain causes or judge whether they saved enough.
Every reported group, in one table
Each band cell shows an estimated taxpayer count, then its share of that row’s total. On a narrow screen, scroll the table sideways.
| Age group | Under $6,500 | Exactly $6,500 | Over $6,500, under $7,500 | Exactly $7,500 | Total taxpayers |
|---|---|---|---|---|---|
| All ages | 6,490,15268.4% | 1,853,95919.5% | 214,1212.3% | 930,1819.8% | 9,488,414 |
| Under 20 | 174,08677.6% | 50,28322.4% | 00.0% | 00.0% | 224,370 |
| 20–24 | 707,03675.3% | 231,98624.7% | 00.0% | 00.0% | 939,021 |
| 25–29 | 975,62070.3% | 412,21329.7% | 00.0% | 00.0% | 1,387,833 |
| 30–34 | 867,24868.2% | 404,99531.8% | 00.0% | 00.0% | 1,272,243 |
| 35–39 | 791,60573.4% | 286,58226.6% | 00.0% | 00.0% | 1,078,188 |
| 40–44 | 656,32571.5% | 261,57328.5% | 00.0% | 00.0% | 917,898 |
| 45–49 | 623,06676.8% | 187,80623.2% | 00.0% | 00.0% | 810,872 |
| 50–54 | 558,76568.2% | 8,4591.0% | 53,1456.5% | 198,89724.3% | 819,265 |
| 55–59 | 449,55059.1% | 1,316 *0.2% | 59,4017.8% | 249,79032.9% | 760,057 |
| 60–64 | 371,05952.6% | 3,473 *0.5% | 64,6989.2% | 266,22437.7% | 705,454 |
| 65+ | 315,79255.1% | 5,273 *0.9% | 36,8786.4% | 215,27037.6% | 573,213 |
* Use with caution: the exactly-$6,500 estimates for ages 55–59, 60–64 and 65+ rely on small samples. Their three asterisk-marked count intervals are suppressed, not zero.
Reported counts are retained as published; rounded detail may not add to totals. Shares are rounded to one decimal place.
How we read the data
The IRS Statistics of Income study uses matched samples of tax returns and IRA information forms. These are estimates, not a census. The IRS catalog listed 2023 as the latest Table 6 year when checked October 14, 2026; the workbook identifies its release as June 2026.
The calculation: each band’s reported count divided by that age group’s reported total, multiplied by 100. Published confidence intervals concern counts; they are not intervals for these shares. We do not claim statistically significant differences. Rare missing birth years are imputed by the IRS. A new data series began in 2022, so this page makes no trend comparison with earlier years.
A limit to keep in view: dollar totals for the exact-dollar bands do not match reported counts multiplied by the band amount, beyond ordinary rounding. That discrepancy remains unresolved; count reconciliation alone cannot establish which component explains it. We retain the published counts and do not use those relationships to calculate band averages or a personal-maximum share. This does not establish that the IRS’s separate aggregate averages are wrong. These grouped data do not supply a median.
Take the data with you
Our downloads keep the reported counts, calculated shares, source references and interval flags together.
Reuse & attribution: credit RothIRAHub and link to this guide when sharing permitted excerpts of our original analysis or presentation, under our Terms of Service. We claim no ownership over the underlying U.S. government data; those data are not subject to our original-content restrictions.