IRS Phone Service Keeps Failing the Same Way. No One Verifies a Fix.
Summary
A June 2026 Treasury Inspector General audit sampled 200 taxpayer calls to five IRS phone lines and found representatives failed to meet basic service standards on 26% of them -- calls dropped without warning, holds stretching past half an hour, wrong answers given. Projected across the population, TIGTA estimates that is approximately 1 million taxpayers, 1,053,230 by this piece's own arithmetic on TIGTA's two published figures, who called for help between February and May 2025 and didn't get it. It is not a new problem: TIGTA read the IRS's own internal quality reviews spanning October 2020 through July 2025 and found the same five failure types as the most common repeat problems across that span -- and no IRS policy requires anyone to confirm the fixes assigned each time ever worked.
The help-line taxpayers call last failed most
drew its sample from two functions: Compliance Services, where 18% of an estimated 1.4 million callers didn't get quality service, and Accounts Management, where 34% of an estimated 2.4 million callers didn't⧉. Broken out by the five individual phone lines sampled, the failure rate ranges from 16% on the Individual Master File Balance Due line to 48.5% -- nearly one call in two -- on the National Taxpayer Advocate line, the number a taxpayer calls after other channels have already failed them to check the status of a case that's supposed to be getting special escalated attention.
View data as table
| Individual Master File Balance Due | 16% | 8 of 50 sampled calls; 1,321,714 calls in the population |
|---|---|---|
| Automated Underreporter | 20% | 10 of 50 sampled calls; 75,519 calls in the population |
| Individual Master File Accounts | 35.3% | 12 of 34 sampled calls; 1,689,709 calls in the population |
| National Taxpayer Advocate | 48.5% | 16 of 33 sampled calls; 360,916 calls in the population |
| Identity Theft (General) | 18.2% | 6 of 33 sampled calls; 307,398 calls in the population |
Dropped calls, half-hour holds, wrong answers
Of the 52 failed calls, 22 involved a call that was dropped, disconnected, or never transferred as promised⧉ -- 8 dropped before the taxpayer's issue was resolved, 5 cut off by the representative mid-call, and 9 that ended after a representative said the call would be transferred and then wasn't. Another 20 involved a hold classified as extensive: policy caps holds at 5 to 7 minutes while a representative researches an account, but the flagged holds ran from just over 7 minutes to as long as 35. Fourteen more calls gave taxpayers inaccurate information -- a representative who didn't ask enough questions to understand the issue, or who transferred someone to the wrong department entirely. Courtesy itself wasn't the main problem: taxpayers got a courteous representative in more than 96% of all 200 calls reviewed, with only 7 flagged for rudeness, jargon, or background noise.
View data as table
| Dropped, disconnected, or not transferred | 22 | 8 dropped before resolution, 5 disconnected by the rep, 9 promised transfers that never happened |
|---|---|---|
| Extended hold time | 20 | up to 35 minutes; IRS policy caps holds at 5-7 minutes |
| Inaccurate information | 14 | wrong answers or wrong-department transfers |
| Discourteous service | 7 | rude tone, jargon, or background noise |
The IRS already promised a fix for this -- eight months earlier
This isn't the 's first time hearing it. 's prior inspection, issued October 2025 on a different FY2024 call sample, found 128 of 831 calls (15%) were dropped or disconnected⧉ and recommended annual refresher training emphasizing professional, courteous service. agreed, and told its representatives completed that training in February 2026. 's new sample -- calls from February through May 2025 -- predates that training, so this audit can't yet say whether it worked. What it can say is that the same failure types (dropped calls, long holds, inaccurate answers) turned up again in a completely separate, larger, more rigorous sample.
Five years of the same list, and nothing that checks if a fix stuck
went further than its own two samples: it pulled the IRS's own internal annual quality reports from October 2020 through July 2025 and identified the same repeat issues as the most common problems across that span⧉ -- representatives not transferring calls appropriately, not completing account research, not using professional courtesy, not advising taxpayers of next steps, not explaining response timeframes. policy requires a manager to assign a corrective action, like a training video, when an employee's quality review flags a problem. It does not require anyone -- the manager, or the internal review group itself -- to confirm the corrective action was ever completed or whether it worked. management told that high employee turnover keeps contributing to the repeat findings; 's response, in effect, is that turnover doesn't explain the absence of a system to notice when a fix hasn't taken.
- sampled 200 of 3.76 million calls to five phone lines from Feb. 15 to May 15, 2025 and found 52 -- 26% -- failed basic service standards, projecting to an estimated 1,053,230 taxpayers ('s own rounded figure: 'approximately 1 million') across Compliance Services (18% failure) and Accounts Management (34% failure)⧉.
- The National Taxpayer Advocate line -- the number taxpayers call when other channels have already failed them -- had the worst failure rate of the five lines sampled, at 48.5%.
- already promised a fix for a nearly identical finding eight months earlier -- 15% of a separate FY2024 sample was dropped or disconnected, agreed to add refresher training, and says representatives completed it in February 2026. 's new sample predates that training, so this audit doesn't yet test whether it worked; it does show the same failure types recurring in a fresh, larger sample regardless.
- read the 's own internal quality reports spanning October 2020 through July 2025 and identified the same repeat problems as the most common issues across that span, with no policy requiring anyone to confirm a corrective action -- once assigned -- was ever completed or effective. 's three new recommendations ask the to document dropped calls, remind representatives of existing procedures, and evaluate recurring issues; agreed to all three but committed only to an employee alert and reminders, not a new verification requirement.
This piece draws its figures from TIGTA's June 10, 2026 audit, Report No. 2026-100-030⧉, fetched directly from tigta.gov and read in full. The prior October 2025 inspection (Report No. 2026-IE-R001) is cited only as it appears inside this June 2026 report's own footnotes and text -- it was not independently re-fetched this iteration. The 1,053,230 combined-taxpayer figure is this piece's own arithmetic summing 's two separately published point estimates (251,502 and 801,728); 's own text rounds the combined figure to 'approximately 1 million' and does not publish a combined confidence interval. This is a service-quality audit, not a spending audit -- no dollar figure is at stake in the underlying report.
Sources(2) ▾
- Treasury Inspector General for Tax Administration, Taxpayers Continue to Experience Customer Service Issues When Calling Certain IRS Telephone Lines (Report No. 2026-100-030) (2026-06-10) — The full 18-page final audit report assessing the quality and accuracy of telephone assistance representatives gave taxpayers on five phone lines across two functions (Compliance Services and Accounts Management) from Feb. 15 through May 15, 2025. Fetched directly from tigta.gov and converted with pdftotext -layout. Used for every figure in this piece: the statistically valid sample results (Results of Review, printed pp. 3-5), the dropped-call and hold-time detail (printed pp. 4-5), the courtesy/accuracy findings (printed p. 5), the five-year repeat-issue review (printed pp. 6-7), the outcome-measure projections and confidence intervals (Appendix II, printed pp. 10-11), and the scope/methodology and per-line sample table (Appendix I, printed pp. 8-9). tigta.gov · original document
- Treasury Inspector General for Tax Administration, Limited Testing Shows Taxpayers Generally Received Courteous and Professional IRS Telephone Service (Report No. 2026-IE-R001) (2025-10-01) — 's prior Office of Inspections and Evaluations review, cited repeatedly in the June 2026 audit (footnote 3 and printed pp. 4, 8) as the immediately preceding look at telephone courtesy: it found 128 of 831 sampled FY2024 calls (15%) were dropped or disconnected, and led to a recommendation -- which agreed to and says it completed in February 2026 -- for annual refresher training emphasizing professional, courteous service. Not independently re-fetched this iteration; every figure drawn from it here is a direct citation appearing inside the June 2026 report (doc-tigta-2026-100-030), which is the load-bearing primary source for this piece. Cited here only as the report identity being referenced, not as an independently reviewed document. tigta.gov
Comments
Always open. Logged-in readers can annotate paragraphs in place.
TIGTA's June 10, 2026 audit⧉ listened to a statistically valid sample of 200 recorded calls to five telephone lines -- lines taxpayers reach for balance-due questions, identity-theft help, and underreported-income disputes -- and found 52 of them, 26%, failed the Taxpayer Bill of Rights' own standard for prompt, courteous, accurate service. Projected across the roughly 3.76 million calls those lines received from Feb. 15 through May 15, 2025, estimates approximately 1 million taxpayers didn't get the help they called for. The finding itself isn't the surprise. also read the 's own internal quality reviews spanning October 2020 through July 2025 and found the identical list of problems named as the most common repeat issues across that span -- with no policy that requires anyone to check whether an assigned fix ever closed the gap.