The Credit Complaint Index, H1 2026

Across the 3,167,874 credit reporting complaints Americans filed with the CFPB between 1 January and 30 June 2026, the three nationwide bureaus closed complaints with relief at sharply different rates: Experian 7.27%, Equifax 43.70%, TransUnion 40.96%.

Those are Credit Booster AI’s figures, derived from the CFPB’s public company response field. The CFPB does not publish them. The exact query is printed below, and anyone who runs it lands on the same numbers.

This is version 2.0 of the Credit Booster AI Credit Complaint Index. The first edition covered calendar 2025 and is still published as the State of Credit Complaints 2026 annual report.

Which credit bureau resolves the most CFPB complaints?

Equifax, narrowly. The more interesting finding is that the three bureaus converged. Here is the same measure across three matched half-year windows, so nothing is distorted by comparing a half year against a full one.

Bureau	H1 2024	H1 2025	H1 2026	Closed with relief, H1 2026
Equifax	64.99%	67.29%	43.70%	421,239 of 963,923
TransUnion	88.22%	83.22%	40.96%	461,229 of 1,126,109
Experian	31.82%	0.87%	7.27%	68,736 of 944,875
Spread, highest to lowest	56.40 pts	82.35 pts	36.43 pts

Three separate things happened in twelve months.

Experian went up, from a very low floor. Its relief count rose from 5,371 complaints to 68,736, which is 12.8 times more, while its complaint volume rose 53.1%. A rate of 7.27% still means roughly 1 complaint in 14.

TransUnion fell hardest. Its relief count actually dropped, from 567,002 to 461,229, a fall of 18.7%, while the complaints it received rose 65.3%. Fewer resolutions against many more complaints is what a 40-point rate drop looks like underneath.

Equifax slid without collapsing. Its relief count fell 11.0% while volume rose 37.1%.

The caveat travels with the number, always. The closure label is chosen by the company being complained about. The CFPB’s December 2025 report is explicit that it “uses the term ‘relief’ for actions the NCRAs report they have taken.” A falling rate can mean less relief was granted, the same actions filed under a different label, or both. Our longer treatment is in the Experian complaint relief rate study.

The exact query, so you do not have to trust us

A statistic that only traces back to the site that published it is not a statistic. Here is the complete derivation for every bureau figure above.

https://www.consumerfinance.gov/data-research/consumer-complaints/search/api/v1/
  ?date_received_min=2026-01-01
  &date_received_max=2026-06-30
  &product=Credit reporting or other personal consumer reports
  &product=Credit reporting, credit repair services, or other personal consumer reports
  &company=Experian Information Solutions Inc.
  &size=0

Read aggregations.company_response.company_response.buckets. Add Closed with monetary relief and Closed with non-monetary relief. Divide by hits.total.value minus the In progress bucket. Swap the company string for EQUIFAX, INC. or TRANSUNION INTERMEDIATE HOLDINGS, INC. for the other two rows.

Five things will change your answer, and we publish them because we got two wrong on the first pass:

1. date_received_max is inclusive. Ending a half-year window on 2026-07-01 rather than 2026-06-30 adds a day and 23,235 complaints.
2. There are two credit reporting product labels. The CFPB renamed the product mid-series, so filtering on one silently drops years of data.
3. Company strings are case-sensitive and must match exactly as shown.
4. Exclude In progress from the denominator, or an open complaint counts as a refusal.
5. The database moves. It updates roughly daily, and companies have 15 days to respond before publication, so recent windows fill in afterward.

Two checks that this holds up. Re-running the calendar-2025 method on 11 September 2026 returns 4,810,302 credit reporting complaints against the 4,810,498 we published on 3 August 2026, a gap of 0.004%. And on the same query, Experian’s calendar-2024 relief rate comes out at 18.80% and its calendar-2025 rate at 0.44%, matching the “nearly 20%” and “less than 1%” ProPublica reported before we ever published the figure.

What independent reporting found

ProPublica got there first on the relief collapse, and the credit belongs to them. Joel Jacobs reported on 10 March 2026 that Experian resolved nearly 20% of complaints in consumers’ favor in 2024 and less than 1% in 2025, that TransUnion’s relief rate began plunging in the summer of 2025 and by October it was providing relief roughly half as often, and that Equifax’s relief mostly kept up with complaints. Our calendar-year pull reproduces all three.

On 30 April 2026, four senators wrote to the bureaus about it: Elizabeth Warren, who led, with Tammy Duckworth, Andy Kim and Lisa Blunt Rochester. The letters ask for data on disputes and complaints, information on dispute handling processes and staffing, and correspondence with the CFPB. ProPublica reported on them on 4 May 2026 and published them in full.

What the CFPB itself reports

The CFPB’s annual report under FCRA Section 611(e), published December 2025, is the primary document here. Its headline figures, covering 1 January 2024 to 30 June 2025:

• More than 5.6 million complaints received in total
• Almost 4.8 million of them about credit and consumer reporting
• About 3.9 million about the three largest bureaus
• An almost 3,000% increase in complaints about those three bureaus since 1 January 2020
• Of more than 4.6 million complaints transmitted to the bureaus, about 52% were “covered complaints” under the statute

The report also measures response speed. Companies get 60 days to respond to a CFPB complaint. It finds that TransUnion consistently took about 40 days, Equifax moved around a similar length, Experian has taken nearly the full 60 days since early 2023, and in the most recent months all three averaged more than 50 days.

Keep two clocks separate. The 60 days above is the CFPB complaint response window. FCRA Section 611(a)(1)(A) sets a shorter one: a bureau must complete a reasonable reinvestigation of a disputed item within 30 days of notice, extendable by no more than 15 days under 611(a)(1)(B). Section 611(e) is what requires the CFPB to transmit these complaints to the bureaus in the first place, and requires each bureau to review them and report its determinations back.

What are Americans actually complaining about?

Every complaint carries a consumer-selected sub-issue. We counted the CFPB’s own labels, unmodified.

Sub-issue	H1 2026	Share	H1 2025	Change in count
Information belongs to someone else	1,199,390	37.86%	793,952	+51.1%
Reporting company used your report improperly	575,433	18.16%	373,651	+54.0%
Account information incorrect	388,234	12.26%	223,325	+73.8%
Investigation took more than 30 days	277,449	8.76%	103,322	+168.5%
Their investigation did not fix an error	276,350	8.72%	238,081	+16.1%
Account status incorrect	164,959	5.21%	114,149	+44.5%
Credit inquiries you don’t recognize	101,907	3.22%	136,519	-25.4%
Personal information incorrect	70,369	2.22%	51,370	+37.0%

The top complaint in America is still not a late payment or a balance. It is a file integrity problem: the wrong person’s data attached to a name, through mixed files, identity confusion or misattributed accounts. Why that category behaves differently from the rest is in the “that isn’t mine” study.

Two things moved this edition.

Timeliness complaints grew 2.69 times. “Investigation took more than 30 days” went from 4.88% of complaints to 8.76%, the largest share shift in the table. That is consumers reporting against the FCRA 611(a) clock, and it moved in the same direction as the CFPB’s own finding on response times.

Unrecognized inquiries fell. It is the only category in the top eight that dropped in absolute terms, down 25.4% while everything around it rose by half. We do not know why, and would rather say so than invent a reason.

Where do the complaints come from?

Filing rates remain wildly uneven. Each state’s H1 2026 complaints divided by its July 2025 Census population estimate gives a national rate of 9.22 per 1,000 residents over six months.

Rank	State	H1 2026 complaints	Per 1,000 residents
1	Mississippi	80,061	27.10
2	Georgia	239,205	21.16
3	Florida	456,865	19.47
4	Louisiana	82,529	17.87
5	Alabama	92,754	17.86
6	Texas	523,472	16.51
47	Maine	1,455	1.03
48	Montana	1,034	0.90
49	South Dakota	800	0.86
50	Wyoming	472	0.80
51	Vermont	377	0.58

Mississippi has overtaken Georgia at the top, having ranked second on the 2025 annual figures. A Mississippian is about 47 times more likely to file than a Vermonter. The full ranking is in the complaints by state study.

A low complaint rate is not evidence of a low error rate. It is evidence of a low complaint rate.

The surge still arrives through one door

99.87% of H1 2026 credit reporting complaints were filed through the CFPB’s web form: 3,163,821 of 3,167,874. Telephone filings numbered 2,181.

Compare that with the first half of 2020, when the same product drew 113,449 complaints. Web filings have risen 29.4 times since. Phone filings have fallen 17.4%, from 2,640 to 2,181.

Filing by phone means a person calls the CFPB and speaks to a human. It cannot be bulk-submitted, which makes it an accidental control group, and it went down while the overall series went up almost thirtyfold. The CFPB’s December 2025 report attributes part of the growth to third parties using the complaint process, including credit repair organizations and people offering credit advice on social media, and to more complaints submitted in a single session. Our channel data is consistent with that. It does not prove it.

How to read these numbers

Every figure here counts complaints filed, not people harmed or errors confirmed. One person complaining about the same account to all three bureaus produces three complaints, and nobody adjudicates whether the consumer was right. The CFPB says plainly that its database “is not a statistical sample” of consumer experience, that complaints are “not necessarily representative” of all of it, and that narratives “are not verified before publication.”

If you quote a number from this page, quote that paragraph next to it. That applies to us too.

The short version

• 3,167,874 credit reporting complaints were filed with the CFPB in H1 2026, up 49.7% year over year and 27.9 times the H1 2020 level.
• Relief rates converged: Equifax 43.70%, TransUnion 40.96%, Experian 7.27%. The highest-to-lowest gap narrowed from 82.35 points to 36.43.
• Experian rose from 0.87% to 7.27%. TransUnion fell from 83.22% to 40.96% and its relief count dropped 18.7% in absolute terms.
• The top complaint is “this isn’t mine” at 37.86%. The fastest growing of the top eight is “investigation took more than 30 days”, up 2.69 times.
• Mississippi now files the most per person at 27.10 per 1,000 in six months. Vermont files the least at 0.58.
• 99.87% arrived by web form. Phone filings fell 17.4% since H1 2020.
• The relief rate is our derivation from a public field, and the closure label is chosen by the company being complained about.

Methodology

Source. CFPB Consumer Complaint Database, search API v1, pulled 2026-09-11. The database held 17,683,411 complaints at pull time, most recent row dated 2026-09-11. The data is public domain. Rate limiting can occur at the CDN edge; the bulk CSV at files.consumerfinance.gov/ccdb/complaints.csv.zip holds the same records.

Window. date_received 2026-01-01 to 2026-06-30 inclusive. Comparison windows are the matched halves of 2020, 2024 and 2025. Calendar-year figures use 01-01 to 12-31.

Product filter. Both CFPB credit reporting product labels: “Credit reporting or other personal consumer reports” and “Credit reporting, credit repair services, or other personal consumer reports”.

Relief rate. Within-company: monetary plus non-monetary relief divided by that company’s closed complaints, where closed equals received minus In progress. Bureau matching uses the CFPB’s own normalized company strings. Reproduced twice on separate pulls with identical results.

Sub-issues. The CFPB’s Sub-issue field, unmodified, summed across parent issue labels where the same sub-issue appears under more than one. It is a consumer-selected menu item and inherits whatever the CFPB’s form nudges people toward. That is a real limitation we cannot correct for.

Population. US Census Bureau, Vintage 2025 State Population Estimates (NST-EST2025), POPESTIMATE2025, the July 1 2025 estimate, pulled 2026-09-11. Validated: the 50 states plus DC sum to the published US total of 341,784,857 exactly. National rate is 3,152,813 complaints over 341,784,857 residents. Territories and military ZIPs are excluded from both sides.

Our own record. Credit Booster AI appears zero times in the company field. We checked ourselves before writing about anyone else.

Version history

• v2.0, 12 September 2026. H1 2026 edition. Window moved to 1 January to 30 June 2026. Added the matched H1-over-H1 bureau relief series and the H1 2025 comparison column on sub-issues. Documented the inclusive date_received_max boundary after it produced a 23,235-complaint overcount in drafting. Three new CSVs published.
• v1.0, 3 August 2026. Annual edition on calendar 2025. Still published, with its own CSVs, as the State of Credit Complaints 2026.
• Next edition: Q3 2026 data, November 2026. Quarterly from there. The date changes only when the figures do.

Download the data

Free to reuse, including commercially, with attribution to Credit Booster AI and a link back to this page. Each file carries its own source header, window and pull date.

• Bureau relief rates, H1 2024 to H1 2026 (CSV, 9 rows, counts and rates)
• Complaints by sub-issue, H1 2026 versus H1 2025 (CSV, counts and shares)
• Complaints per 1,000 residents by state, H1 2026 (CSV, 51 jurisdictions)

Prior edition, calendar 2025: per-capita by state, sub-issues, submission channel 2019 to 2025.

The CFPB file is public domain. The aggregation, the per-capita normalisation and the rate derivations are ours.

Sources

• CFPB Consumer Complaint Database, search API v1, pulled 2026-09-11, per the query above
• CFPB, Annual report of credit and consumer reporting complaints, December 2025, the FCRA Section 611(e) report, for the 5.6 million, 4.8 million, 3.9 million, almost 3,000% and response-time figures
• Fair Credit Reporting Act Section 611, 15 U.S.C. 1681i, for the 30-day reinvestigation period at 611(a)(1)(A), the 15-day extension at 611(a)(1)(B), and the complaint transmission duties at 611(e)
• Joel Jacobs, “Credit Bureaus Are Leaving More Mistakes on Frustrated Consumers’ Reports Under Trump’s CFPB”, ProPublica, 10 March 2026
• Joel Jacobs, “Lawmakers Demand Answers About Growing Number of Unfixed Mistakes on Credit Reports”, ProPublica, 4 May 2026
• Letters from Senators Warren, Duckworth, Kim and Blunt Rochester to the credit bureaus, dated 30 April 2026, published by ProPublica
• US Census Bureau, Vintage 2025 State Population Estimates (NST-EST2025), July 1 2025 estimates, pulled 2026-09-11

Nothing on this page changes your credit file. The first step is free: pull all three reports at AnnualCreditReport.com and read them line by line. For the mechanics, start with how to dispute a credit report or filing a CFPB complaint against a bureau. Credit Booster AI reads all three reports and tracks what changes between pulls, on a 7-day free trial on Plus at $9.99 a month. It cannot remove information that is accurate, and no app can.