401(k) Monitor

Accuracy audit, round 1

We re-read 22 filings from scratch and compared 66 facts against what this site publishes. Nothing disagreed. This page is the whole worksheet: how the 22 were picked, how they were checked, all 22 rows side by side, the single row where a filing was read two defensible ways, and the three companies we could not check.

Scored 8 August 2026, against the 161 companies published at that date.

How the 22 were picked and checked

The 25 companies were picked by a fixed rule, not by hand, so the sample cannot be chosen after the fact to flatter the result: sort every published company alphabetically and take every sixth one.

Three fields were checked per company: the employer match formula, with its type and its numbers; the employer-match vesting type, with its schedule numbers; and the auto-enrollment status, with the default deferral rate where the filing states one.

Blind first, compare second. The checker read each filing and wrote down all three answers before opening what the site says, then compared them one by one. Every disagreement is recorded with the sentence from the filing behind it, and marked as a real error or as two defensible readings of the same text.

Of the 25, we held the filing text for 22, which is where the 66 comes from: 22 companies, three facts each. The other three are listed at the foot of this page.

What the check found

Round 1: no errors in 66 facts checked. One field-level disagreement, 1.5% of checks, was a convention choice about which of two disclosed formulas leads the record. Both formulas are preserved in the data.

A defensible reading is a disagreement where both readings are supported by the filing. It is counted separately from an error and never silently as agreement.
FieldExact agreementDefensible readingsExtraction errorsVerifier errors
Match formula (type and numbers)21/22 (95.5%)1 (Paychex)00
Vesting type (and schedule numbers)22/22 (100%)000
Auto-enrollment status (and default rate)22/22 (100%)000
All fields65/66 (98.5%)100

Company by company

One row per company, three facts each. The left half of every pair is what the filing says; the right half is what this site publishes, in the vocabulary the data itself uses. Each company name links to its page, so any row can be re-checked against the filing it cites.

22 companies, 66 facts, checked 8 August 2026. Scroll sideways to read the full width.
CompanyText sourceEmployer matchVesting of the matchAuto-enrollment
Filing saysPublishedVerdictFiling saysPublishedVerdictFiling saysPublishedVerdict
AllstateBatch 1tiered 100% of 1st 2% + 50% of next 4%, max 4%tiered [100/2, 50/4], max 4Agreecliff 2 yrcliff 2Agreeyes, 6%, +1%/yr to 10%, 45-day windowpresent, 6, esc 1 to 10, 45 daysAgree
AT&TFeasibility samplematch exists, no formula disclosednot_disclosed / unknownAgreenot disclosednot_disclosedAgreenot mentionednot_mentionedAgree
BoeingBatch 1match exists, no formula disclosednot_disclosed / unknownAgreeimmediateimmediateAgreeyes, default % not statedpresent, default nullAgree
CBRE GroupBatch 266-2/3% of 1st 6%, cap $6,000, max 4%single_tier [66.67/6], max 4, cap in conditionsAgreegraded 33-1/3%/yr, 100% at 3 yrgraded 33.33/66.67/100 at 1/2/3Agreeyes, 3%, 45-day windowpresent, 3Agree
CintasBatch 2discretionary, rate set by Board yearly, no formuladiscretionary_undisclosed / unknownAgreegraded 0/20/40/60/100 at <2/2/3/4/5 yrgraded 20/40/60/100 at 2/3/4/5Agreeyes, 3% after 3 monthspresent, 3Agree
CortevaBatch 2100% up to 6%, max 6%single_tier [100/6], max 6AgreeimmediateimmediateAgreeyes, 6%, +1%/yr to 15%present, 6, esc 1 to 15Agree
Darden RestaurantsBatch 1discretionary quarterly 0 to 120% of 1st 6%; actuals 65/25/80/65, then 80partially_disclosed single_tier [null/6]; actuals + 0-120% range captured in actual_rate_for_plan_year and conditionsAgreegraded 5%/quarter from 5th quarter, 100% at 24 quartersgraded, same scheduleAgreenot mentionednot_mentionedAgree
DowBatch 2tiered 100% of 1st 4% + 50% of next 2%, max 5%tiered [100/4, 50/2], max 5AgreeimmediateimmediateAgreeyes, 6%, +1%/yr to 15%present, 6, esc 1 to 15Agree
Elevance HealthBatch 1tiered 100% of 1st 4% + 50% of next 2%, max 5% (PR plan)tiered [100/4, 50/2], max 5Agreecliff 2 yrcliff 2Agreeyes, 6%, +1%/yr to 15%present, 6, esc 1 to 15Agree
FastenalBatch 2discretionary, no formuladiscretionary_undisclosed / unknownAgreeimmediate (100% vested when made)immediateAgreeyes, 5%present, 5Agree
General MillsBatch 2new hires: 100% of 1st 4% + 50% of next 4%, max 6%; legacy: 50% up to 6%tiered [100/4, 50/4], max 6; legacy cohort in conditionsAgreegraded 20/40/60/80/100 at 1-5 yrgraded, same scheduleAgreeyes, 8%, +1%/yr to 10%present, 8, esc 1 to 10Agree
Home DepotBatch 1tiered 150% of 1st 1% + 50% of next 4%, max 3.5% (Puerto Rico plan; the mainland FutureBuilder states the same formula)tiered [150/1, 50/4], max 3.5Agreecliff 3 yr (1,000-hour years)cliff 3Agreenot mentionednot_mentionedAgree
Huntington BancsharesBatch 2tiered 150% of 1st 2% + 100% of next 1%, max 4%tiered [150/2, 100/1], max 4Agreecliff 2 yrcliff 2Agreeyes, 4%, +1%/yr to 10%present, 4, esc 1 to 10Agree
International PaperBatch 2tiered 70% of 1st 4% + 50% of next 4%, max 4.8%tiered [70/4, 50/4], max 4.8Agreecliff 3 yrcliff 3Agreeyes, 8%, +1%/yr to 10%present, 8, esc 1 to 10Agree
Kimberly-ClarkBatch 2100% of 1st 5% (Mobile union: 1st 4%), max 5%single_tier [100/5], max 5, union exception in conditionsAgreeimmediateimmediateAgreeyes, 8%present, 8Agree
MascoBatch 2matching and/or profit sharing, vary by subsidiary, no formulanot_disclosed / unknown, vary-by-subsidiary capturedAgreematch immediateimmediateAgreenot mentionednot_mentionedAgree
M&T BankBatch 2100% up to 5%, max 5%single_tier [100/5], max 5Agreeimmediate (DEC/RAA separate, 5-yr)immediateAgreeyes, 1% after 6 months, +1%/yr to 10%present, 1, esc 1 to 10Agree
O'Reilly AutomotiveBatch 1tiered 100% of 1st 2% + 25% of next 4%, max 3%tiered [100/2, 25/4], max 3Agreegraded 33/66/100 at 2/3/4 yrgraded, same scheduleAgreeyes, 2%present, 2Agree
PaychexBatch 22025 plan year: 100% of 1st 3% + 50% of next 2%, max 4%; effective 2026-01-01: 100% of 1st 4%single_tier [100/4], max 4, effective_date 2026-01-01, 2025 formula in superseded_formula_noteDiffers, defensibleimmediateimmediateAgreenot mentionednot_mentionedAgree
PPG IndustriesBatch 2rate discretionary yearly, 100% in 2025 and 2024, on 1st 6%, max 6%partially_disclosed [100/6], discretion in actual_rate_for_plan_year and conditionsAgreeimmediateimmediateAgreeyes, 4%present, 4Agree
Regions FinancialBatch 2100% up to 5%, max 5%single_tier [100/5], max 5AgreeimmediateimmediateAgreeyes, 2%, +1% esc for 2-9% deferrerspresent, 2, esc 1 to 10Agree
Trane TechnologiesBatch 2100% up to 6%, max 6%single_tier [100/6], max 6AgreeimmediateimmediateAgreeyes, 4%, 35-day windowpresent, 4Agree

The one row that differed: Paychex

The Paychex filing states two match formulas, one for the plan year reported and one taking effect after it ends.

The Company provides a matching contribution in the amount of 100% of the first 3% and 50% on the next 2% of eligible pay that an employee contributes to the Plan for a total matching contribution of 4%. Effective January 1, 2026, the Company's matching contribution is in the amount of 100% of the first 4% of eligible pay that an employee contributes to the Plan.

The published record leads with the post-2026 formula, 100% of the first 4%, and keeps the 2025 plan-year formula, 100% of the first 3% plus 50% of the next 2%, in the superseded-formula note, with an effective date of 1 January 2026 and a flag saying the formula changed during the year. The checker led with the plan-year formula instead.

That is the convention every record follows: the formula shown is the current one, the effective date says when it took effect, and the superseded-formula note preserves the prior one, so the company page shows both. Both formulas produce the same maximum match, 4% of pay, and every number in the filing is present in the record. Two defensible readings of the same filing, not an error. No data was changed.

The three we could not check

Three of the 25 sampled companies had no stored filing text at the time of the audit, so there was nothing to read blind, so nothing was checked for them. They are named here rather than dropped, because a sample that quietly loses its awkward members is not a sample. They are first in line for round 2.

Where this fits

This audit measures one thing: whether the numbers published here match the filings they were read from. The pipelines that produce them, the rules they obey, and the lag built into the source documents are described on the methodology page. Found a row you read differently? Write to corrections@401kmonitor.com with the company and the filing.