What we found
Between fiscal years 2006 and 2025, the Department of Labor recovered $2.72 billion in unpaid overtime for roughly 3.4 million workers. Over that period the number of cases fell 69% — from 8,667 a year to 2,724 — while the back wages recovered per employee rose 90%, from $650 to $1,235.
Overtime enforcement has not disappeared. It has narrowed: far fewer cases, each one substantially larger.
We reached this by downloading the Wage and Hour Division's complete compliance-action dataset and analyzing every concluded case in it. The aggregated results are published as CSV at the bottom of this page so anyone can check the arithmetic.
Overtime is where the money is
The FLSA covers both minimum wage and overtime, and public debate concentrates almost entirely on minimum wage. The enforcement record points the other way.
Across the 20-year window, overtime back wages totalled $2.72 billion against $466 million for minimum wage — a ratio of 5.8 to 1. Overtime accounts for 84.4% of all FLSA back wages recovered.
The distribution is heavily skewed. The median overtime case returns $3,639, but the 90th percentile returns $36,682. Most cases are small; a minority are very large, and those large cases increasingly drive the totals.
Where overtime violations concentrate
Restricting to the most recent decade (FY2016–FY2025) to reflect the current economy:
Two industries dominate, for opposite reasons.
- Home Health Care Services — $154.5m across 2,408 cases, at $1,551 per employee. Relatively few cases, each expensive.
- Full-Service Restaurants — $148.3m across 10,010 cases, at $997 per employee. Four times as many cases, slightly less money. This is a high-frequency, lower-value pattern.
The care sector as a whole — home health, nursing care, residential care, and services for the elderly and people with disabilities — makes up 12.6% of overtime cases but 20.9% of the dollars, at $1,121 per employee against $956 across all industries. That is consistent with the structural features of the work: live-in shifts, sleep time, travel between clients, and the long-running complexity of the companionship-services exemption. Any one of those can turn into systematic underpayment of the 40-hour threshold across an entire workforce.
The outliers are worth naming because they show what concentrated exposure looks like. Oil and Gas Extraction produced just 25 cases but $26.7m — $15,092 per employee, an order of magnitude above any other industry. Correctional Institutions ($2,369/employee across 20 cases) and State Police Protection ($2,073/employee across 26 cases) show the same shape: a handful of investigations, each covering large groups of employees with the same misapplied rule.
Geography
Texas leads on total dollars ($125.4m, FY2016–FY2025), but Texas also has more cases than any other state, and its recovery per employee is $772 — below the national average.
The more interesting figures are the states where individual recoveries run high:
- Maryland — $2,070 per employee, from only 444 cases.
- California — $1,592 per employee across $109.4m.
- Pennsylvania — $1,438 per employee across $93.7m.
A high per-employee figure with a low case count generally means investigations are finding entrenched, workforce-wide problems rather than one-off errors.
Repeat violators
Cases carrying the DOL's repeat-violator flag are 4.7% of all cases but 12.6% of overtime back wages — roughly 2.7 times overrepresented in dollar terms. Employers who get it wrong once and are investigated again tend to owe substantially more the second time.
What this means if you run payroll
The enforcement trend does not mean the risk has gone away. It means exposure has become concentrated: fewer investigations, each with a larger average liability. The practical implications:
- Regular-rate errors scale. The most common way to underpay overtime is not refusing to pay it — it is calculating the premium off base rate while excluding non-discretionary bonuses, shift differentials, and commissions. That error repeats every week for every affected employee, which is exactly how a case reaches the 90th percentile. Our FLSA regular-rate calculator isolates the correct figure.
- The 40-hour workweek is the only federal trigger. State daily overtime rules are real obligations but they are not FLSA overtime, and conflating the two causes errors in both directions. See which overtime actually qualifies.
- Reporting is now mandatory. From tax year 2026, qualified overtime must be reported separately in W-2 Box 12 under code TT. A payroll system that cannot isolate the premium portion accurately will now produce a visible, filed, wrong number — see employer reporting obligations for 2026.
Methodology, and what this data cannot tell you
Source: the Wage and Hour Division compliance action dataset, containing every concluded WHD compliance action. We used the file published 27 July 2026.
We restricted the analysis to cases with a findings end date between 1 October 2005 and 30 September 2025 — twenty complete federal fiscal years. That keeps 334,272 of 367,893 records (90.9%). We dropped 31,429 records dated before FY2006, 957 with unparseable dates, and one record dated in the year 3021. Overtime figures come from the FLSA overtime back-wage field specifically, not total case back wages.
Three limitations, stated plainly:
- This measures enforcement, not violations. A decline in cases is equally consistent with better employer compliance and with fewer investigations being conducted. The dataset contains no information that distinguishes the two, and we make no claim about which it is.
- Per-employee figures are a floor. The dataset has no overtime-specific employee count. We divided overtime back wages by the count of employees owed any FLSA back wages in the same case, which includes minimum-wage-only employees. The true per-worker overtime figure is therefore somewhat higher than what we report.
- Violation counts are not comparable across October 2025. The DOL changed its data collection procedures on 1 October 2025. We avoided violation counts entirely and used only dollar amounts and employee counts, which are unaffected.
Download the data
The aggregated tables behind every figure above, as CSV. Free to reuse with attribution.
- Overtime back wages by fiscal year, FY2006–FY2025 — cases, back wages, employees, per-employee.
- Top 40 industries, FY2016–FY2025 — by NAICS description.
- All states and territories, FY2016–FY2025.
The underlying raw file is public and can be downloaded directly from the DOL if you want to reproduce this from scratch.