The deal a judge must bless.

A PAGA case cannot settle quietly.#

A private lawsuit usually ends when the parties agree. A PAGA case does not. Because the claim is a representative actionRepresentative actionA suit one employee brings on the State's behalf to recover penalties for an entire workforce — which, unlike a class action, needs no class certification.Lab. Code § 2699(a) — one employee suing to collect the penalties the State could have collected — the parties cannot privately dispose of it. The penalty belongs to the State; the employee collects it only as the State’s proxy. So two things are required, and the statute is short and exact about both.

Two commands sit in those two sentences. A judge must review and approve any settlement — there is no quiet, off-the-record resolution of a PAGA claim. And the LWDALWDAThe Labor and Workforce Development Agency — the state agency that receives the pre-suit notice, may choose to investigate, and collects 65% of any penalties recovered.Lab. Code § 2699(m) must receive the proposed settlement at the same time the court does, so the State is never the last to know about a deal struck in its name.

The State is a party to every PAGA settlement.#

The reason approval is mandatory is structural: in a PAGA case the State is the real party in interest, and the plaintiff is its proxy. The engine of the whole statute says so.

The employee sues to recover the civil penalties the State itself could have assessed and collected. That premise — a private plaintiff suing on the State’s behalf — was settled long before the 2024 reform. The California Supreme Court put it plainly in Arias²2Arias v. Superior Court, 46 Cal.4th 969 (2009). A PAGA action is a representative action brought on the State's behalf and need not satisfy class-action certification requirements.: a PAGA action is representative, brought on the State’s behalf, and needs no class certification. It is the State’s penalty being compromised — which is why the State must be served and a court must approve.

That the plaintiff is a proxy, not the owner of the claim, also explains a quieter point about who keeps the right to settle. In Kim³3Kim v. Reins International California, Inc., 9 Cal.5th 73 (2020). An employee who settles and dismisses individual Labor Code claims still qualifies as an 'aggrieved employee' with standing to pursue PAGA penalties; PAGA standing does not depend on maintaining an unredressed individual injury. the Court held that settling and dismissing the individual Labor Code claims does not cost a plaintiff PAGA standing — the representative penalty claim is a thing apart. The agency receives the deal and may weigh in on it; in practice it frequently does not appear, but the right to be heard is built into the statute, not left to the parties.

Where the money goes: 65 to the State, 35 to the workers.#

When a PAGA penalty is recovered — by settlement or by judgment — the statute, not the parties, decides how it is divided. The reform raised the workers’ share, but the State still keeps the majority.

Sixty-five percent goes to the LWDA; thirty-five percent is shared among the aggrieved employees — up from the old 25%. One narrow case runs differently: where the person had no employees, the flat penalty goes wholly to the agency.⁴4§ 2699(n). The no-employee flat penalty (§ 2699(f)(1)) is distributed entirely to the LWDA — there are no aggrieved employees among whom to divide a 35% share. For the ordinary case, the division looks like this:

The picture corrects the most common misreading of a PAGA number: the figure on the notice is not what lands in any worker’s pocket. The majority is the State’s, and the workers’ 35% is then split among everyone affected.

What approval scrutiny looks like.#

Here is the part it is easy to overstate, so it is worth drawing the line carefully. The statute fixes the requirement of approval — and that is nearly all it fixes. § 2699(s)(2) says the superior court must review and approve the settlement, and that the agency is served at the same time. It does not, in that subdivision, set out a multi-factor test for what makes a settlement approvable.

The statute sets the requirement of approval; the standards a court applies in deciding whether to approve are developed in the case law. Those standards — the kind of fairness-and-adequacy review a court brings to a representative settlement — are not part of this repository’s verified data, so this page does not cite a specific approval-standard decision or state its holding. What follows is a careful, general description of the kinds of things a reviewing court attends to, not a test the statute fixes.⁵5A deliberate boundary. Because no approval-standard case is in the verified set here, naming one — or reciting a multi-factor “fair, reasonable, and adequate” formulation as though § 2699(s)(2) supplied it — would assert more than the source material supports. The items below are framed as general considerations, not as a holding or a statutory checklist.

Described at that careful, general level, a court reviewing a PAGA settlement is the kind of body that tends to look at whether the deal honors what the statute does fix and does not quietly give away more than it should. In broad terms, that attention often falls on a handful of things:

• whether the allocation honors the 65/35 split the statute sets (§ 2699(m));
• whether the release is appropriately tied to the claims that were actually noticed, rather than reaching further;
• whether the attorney’s fees requested are reasonable; and
• whether the agency had its contemporaneous chance to object (§ 2699(s)(2)).

Stated that way, the division of labor is clean: the statute supplies the gate, and the case law supplies the measure applied at the gate. This page stays on the first half of that sentence and flags the second as developed elsewhere — there is no controlling decision named here because none is in the verified set.

What we still don’t know.#

A treatise earns trust by marking its own edges. The approval mechanism is young, and several questions around it are genuinely open as of July 13, 2026.

The first is the scope of the release a PAGA settlement may carry. The statute requires approval but does not, in § 2699(s)(2), define how broadly the claims released may sweep — how far beyond the specifically noticed violations a release may reach is worked out case by case, not fixed in the text.

The second is the weight of the agency’s frequently-silent participation. The LWDA must be served, and may object — but how a reviewing court should treat the agency’s common non-appearance, and whether silence carries any inference at all, is not something the statute answers.

The third is who is even in the settling group when arbitration has carved the case apart. A plaintiff compelled to arbitrate the individual claim keeps the representative claim in court — the Supreme Court so held in Adolph⁶6Adolph v. Uber Technologies, Inc., 14 Cal.5th 1104 (2023). Under California law a plaintiff compelled to arbitrate the individual PAGA claim retains standing to pursue the non-individual (representative) claims in court, rejecting the contrary suggestion in Viking River. Rested on the standing concept from Kim v. Reins. — so the boundary of the group bound by a settlement can be contested: which aggrieved employees are inside the representative claim being compromised, and which were peeled off into an individual arbitration, is a question the interaction of arbitration and the 2024 standing rule has not yet settled. These are seams to flag, not to paper over with a standard the statute does not supply.

For the architecture of the penalty being divided — the tiers, the caps, and the cure-to-zero path — see The Penalty & Exposure; for how the claim reaches this point, see Life of a Claim.

What to do now.#

Diagnosis is only half of it. Where you stand depends on who you are.

Common questions.#

The authorities.#