Quarter-hour rounding survives in California timekeeping systems mostly because nobody has looked at the setting since it was installed. It is the rare exposure that costs nothing to eliminate — which is exactly why leaving it in place is hard to explain later.
There is a category of compliance problem that requires money, headcount, and a year of operational change to fix. Rounding is not in it.
Rounding is a checkbox in a timekeeping system. Somebody enabled it during implementation, probably in 2014, probably because the vendor's default configuration had it on, and probably because the payroll clerk at the time had come from a shop where quarter-hour blocks were how it was done. Nobody has looked at it since. It is still on, it is still shaving minutes off punches, and it is generating a damages model out of the employer's own data — one that a plaintiff's expert can build in an afternoon because the raw punch times and the paid times are both in the export.
This post is about why that setting is now the wrong side of a bet nobody needs to be making.
For a long stretch, California employers relied on a doctrine borrowed from federal wage-and-hour practice: rounding punches to the nearest increment was permissible if the rounding was neutral on its face and neutral as applied — that is, if over time it rounded up about as often as it rounded down and did not systematically underpay.
Two things have happened to that doctrine.
The first is Donohue v. AMN Services (Cal. 2021). The California Supreme Court held that time records showing missed, short, or delayed meal periods raise a rebuttable presumption that the meal period was not provided — and, on the rounding question specifically, the Court's reasoning turned on what rounding does to the record. The employer there ran a per-pay-period electronic system that prompted employees to confirm their meal periods. The Court described that system and said that, without the rounding, it "would have ensured accurate tracking of meal period violations." The prompt was firing against numbers that had already erased the short break it existed to surface. Rounding did not obscure the evidence at the margin; it destroyed the mechanism.
The second has not happened yet. Camp v. Home Depot U.S.A. (S277518) presses the broader question — whether an employer that captures exact time may round at all. The California Supreme Court granted review in February 2023. As of this writing the case is pending and has not been argued; briefing has been complete since late 2023, and no decision has issued. While review is pending, the Court of Appeal opinion has no precedential effect (Cal. R. Ct. 8.1115(e)(1)), so nothing has been held statewide. The question has been pressed, and it is in front of the court that will answer it.
We are not going to tell you how it comes out. What we will point out is the shape of the wager. The employer that keeps rounding is betting on a favorable answer to an open question, in exchange for a benefit that consists of slightly simpler payroll arithmetic. That is a bad trade at any odds.
Rounding is rarely pled as "rounding." It is pled as unpaid wages, and it arrives with a family.
The underlying claim is unpaid minimum wage and overtime under Labor Code §§510 and 1194 — three years back, four through the unfair competition statute, with liquidated damages available under §1194.2 on the minimum wage component. Then the derivatives attach: if the hours were wrong, the wage statements that reported them were wrong, and if the hours were wrong at separation, the final pay was short. And in a PAGA action, the civil penalties run per employee per pay period.
The reason this is a favorite of the plaintiffs' bar is not the per-employee value, which is often small. It is that the proof is clean, uniform, and comes from the defendant. There is no need to depose anyone about what happened on a particular Tuesday. The system setting applied to everyone identically, which makes the class-wide model nearly self-proving and makes individualized defenses nearly unavailable.
Two places it hides where nobody thinks to look:
Foreman-entered timesheets. A construction lead who writes down "7:00 to 3:30" every day, in half-hour blocks, is rounding. There is no system setting to find, because the rounding is happening in a human being's habit before the number ever reaches a system. The paper trail looks tidy, which is the tell — real punch data is never tidy.
Shift-start queues. A warehouse where fifty people badge in over eleven minutes and every one of them shows a 6:00 start is not recording when people started; it is recording the schedule.
Almost every other exposure in California wage-and-hour requires an employer to change how work happens. This one requires an employer to change a setting and then pay people what the raw punches say.
That matters beyond the fix itself, because of how the PAGA caps are written. Labor Code §2699(g) caps PAGA civil penalties at 15 percent where the employer took all reasonable steps before receiving either a PAGA notice or a records request under §226, §432, or §1198.5; §2699(h) caps them at 30 percent for reasonable steps taken within 60 days after the notice. The statute's enumerated examples begin with "conducted periodic payroll audits and took action in response to the results of the audit" — and an employer who ran an audit, found rounding, turned it off, and trued up the affected periods has a documentary record that matches the statutory language almost word for word.
Three hedges belong immediately next to that, not in a footnote.
The caps reach civil penalties only. Not the unpaid wages, not the liquidated damages, not the fees — and in a rounding case the wages are most of the money.
They are ceilings, not floors: the statute says penalties "shall not be more than" the stated percentage, and §2699(e)(2) lets a court exceed the limitations of subdivisions (g) and (h) notwithstanding those subdivisions. No employer establishes a number; an employer establishes an argument for an upper bound.
And no California appellate court has yet construed what satisfies "all reasonable steps." Every description of that standard — ours here included — is a reading of statutory text, not a report of holdings.
What survives all three hedges is the direction, which was doing the work anyway: the employer who turned rounding off in March and documented it has an argument in November. The employer who is still rounding in November does not have that argument at all, and the difference is categorical rather than fractional.
The diagnostic is short and does not require counsel to start.
Export the timekeeping configuration, including the change history, and read what the rounding rule is and when it was last modified. Export one quarter of raw punches at native resolution alongside the paid time for the same quarter, and compute the net difference per employee. Then ask whether any grace-period or no-work-before-shift policy exists in writing, because a grace period without a no-work rule is just unrecorded work with a friendly name.
If the net effect is negative for any material group, the question of what is owed and over what period is a legal question on your facts, and it belongs with your counsel before it belongs anywhere else.
One caution on timing, from the statute rather than from us: the 15 percent window closes when a records request under §226, §432, or §1198.5 arrives — not when a lawsuit does. The records request usually comes first, and it usually does not look like the start of anything.
The whole exercise is one export and one afternoon. If you want a second pair of eyes on it, that is what the discovery call is for — and if the setting is already off and the punches reconcile, we will say so and there will be nothing to sell you. What a shortfall is worth, and what is owed, goes to your counsel.