Your Overtime Paycodes Don't Know Which Premium the FLSA Required

Starting with tax year 2026, employers have to report qualified overtime compensation separately on the W-2, in Box 12 under the new Code TT. For 2025 the IRS waived penalties and made it optional. For 2026 it is required, and the W-2s are due at the end of January.

That request lands on payroll, and payroll discovers that the number it needs was decided upstream, months ago, by a configuration nobody revisited.

What exactly does Code TT want?

Not total overtime pay. Not the 1.5x. Code TT wants the premium portion only — the extra half-rate paid above the employee’s regular rate — and only where the FLSA required it. Under the One Big Beautiful Bill Act the employee can deduct up to $12,500 of that amount ($25,000 joint), phasing out above $150,000 of modified AGI. Box 1 is unchanged and FICA still applies to all of it.

Two words do all the damage. Premium, because most reporting carries overtime as hours and a total cost, never split into straight-time and premium. And FLSA, because much of what your system calls overtime the FLSA never required.

Why is this a timekeeping question, not a payroll one?

Because payroll pays what it is told. Which hours are overtime — and on what legal basis — is decided where hours get classified, in the pay rules of your timekeeping system. By the time the week reaches payroll as “OT hours: 6.25,” the basis has been decided and, in most configurations, discarded.

To be fair to the platform: Pro WFM’s rules engine is entirely capable of allocating weekly-over-40 and state daily overtime into separate paycodes, and plenty of tenants already do, usually because a union agreement or a state audit forced the distinction years ago. There this is a reporting exercise and you are fine. The problem belongs to every configuration built before anyone needed the distinction — which is most of them, because until July 2025 nobody had a reason to care.

Where the separation actually breaks

Four shapes, in roughly the order I run into them:

1. One overtime paycode for everything. Daily, weekly, contract, and holiday OT all land in the same bucket. There is no downstream fix: you cannot recover a split from a total, and reconstructing it means re-evaluating the year against a different rule basis.

2. Daily overtime in a week that never hit 40. The expensive one. Take an employee at $22.00 base with a $1.50 shift differential, working four 10-hour days:

• Hours worked in the week: 40
• California daily overtime hours: 8 (two per day, over eight)
• FLSA regular rate: $23.50
• Premium actually paid: 8 x $11.75 = $94.00
• Qualified overtime for Code TT: $0.00

Annualize at 48 worked weeks and that is roughly $4,500 per employee of premium pay that looks like overtime premium in every report you own and contributes nothing to Code TT. For an employee near the cap it is over a third of the deduction.

Now make it a 46-hour week — four 10-hour days plus a 6-hour Friday. Daily overtime covers 8 hours, weekly-over-40 covers 6, and they overlap, because the same hour cannot draw the premium twice. The qualified amount is the premium on the FLSA-required hours: not 8, and certainly not 14. Which paycode holds which hour is decided by your allocation order — a configuration fact about your tenant, and the whole answer.

3. The regular rate is not the base rate. The FLSA regular rate includes shift differentials, nondiscretionary bonuses, and most incentive pay, so the premium is half of that rate. If your costed overtime output is built on the base wage — or if wage rates were never loaded into WFM at all, which is more common than outsiders expect — the premium dollars timekeeping can produce are not the premium the FLSA required.

4. Overtime nobody federally required. Double-time after twelve, contract overtime after eight, premium on a seventh consecutive day, holiday multipliers. Real obligations, largely not FLSA-required. And the case that catches people late: a nondiscretionary quarterly bonus raises the regular rate retroactively, which raises the FLSA premium actually owed.

The checklist to run before January

Six questions. If you cannot answer them in a week, that is itself the finding.

1. From data, not from config — list every paycode that received overtime hours in 2026, and the rule that put them there. Dormant paycodes and duplicates make the config list useless for this.
2. For each one, classify the basis: FLSA-weekly, state-daily, contractual, or blended. Blended is the answer that costs money.
3. Is the premium separable from the straight-time portion in your costed output, or do you only have the 1.5x total?
4. Does the rate behind that cost include the differentials and incentive pay the regular rate legally has to include?
5. Name your source of record — payroll register, WFM costed summary, or a reconciled join — and confirm you can reproduce the figure in March when somebody asks.
6. Write down your post-close edit policy now, before a February correction to a December timecard forces an improvised one.

A multi-state restaurant operator I worked with ran step one and found overtime hours arriving in nine paycodes, three of which blended daily and weekly bases depending on the state the location sat in. Nothing was misconfigured. It had simply never mattered which premium came from which statute, so nobody had modeled it that way. The fix was a derivation in the warehouse against punch-level data and the rule basis — not a configuration project, and the same discipline as any number that has to survive scrutiny, which I wrote about in validated versus directional labor savings.

That derivation is analytics and data platform work: the hours and the rule basis sit in your timekeeping data, and the question is whether your model carries the distinction all the way to a per-employee annual figure you would defend. Build it once, reproducibly, and next January costs an afternoon.

When NOT to build anything for this

• Single state, no daily-overtime rule, no contract overtime, no incentive pay in the regular rate. Your payroll system’s number is probably right. Validate on twenty employees and go do something useful.
• Your payroll provider already separates FLSA from non-FLSA overtime at the earnings-code level and you can prove it on a sample. Take the win.
• Wage rates do not live in WFM. Then the dollars come from payroll and timekeeping’s only job is supplying the rule basis per overtime hour. Much smaller build; confusing the two is how this becomes a six-month project.
• You want a tax answer. What qualifies is your tax advisor’s call, and the IRS has published guidance. My lane is narrower: whether your data can produce the number they tell you to produce, and produce it again in March.

The organizations that handle this well in January will not be the ones with the best payroll vendor. They will be the ones who worked out in October that “overtime” was four different legal obligations sharing a paycode.

---

If you don’t know which of your overtime paycodes the FLSA actually required, send me your paycode list and the states you operate in — that’s a one-week answer, not a project.