ZB, N.A. v. Superior Court

(2019) 8 Cal.5th 175
Published characterizationReviewed Q3 2026Registry entry →
PAGA recovers civil penalties. Wages are not civil penalties. Overtime underpayments, meal premiums, unreimbursed expenses — none are recoverable through PAGA as penalties. The single most important analytical framework.

The 'amount sufficient to recover underpaid wages' referenced in Labor Code § 558 is not a civil penalty recoverable through PAGA. Only § 558's fixed penalty component — $50 per underpaid employee per pay period for an initial violation, $100 for subsequent violations — can be pursued in a PAGA action. The wage amount, which § 558(a)(3) directs be paid entirely to the affected employee, is compensatory relief the Labor Commissioner may recover in citation proceedings; it is not among the civil penalties PAGA deputizes private plaintiffs to collect. Justice Cuéllar wrote for a unanimous court, disapproving Thurman v. Bayshore Transit Management.

The holding rests on the taxonomy that now organizes all PAGA exposure analysis: PAGA authorizes recovery of civil penalties — law-enforcement sanctions, additional to actual losses, previously collectible only by the state — plus default penalties under § 2699(f) for violations carrying no specific penalty. Wages, premiums, and other amounts owed to employees as compensation are not civil penalties, no matter what statute mentions them.

Kalethia Lawson worked as an hourly employee for California Bank & Trust, a division of ZB, N.A. Bound by a handbook arbitration agreement with a class waiver, she filed a single-count PAGA action alleging the standard violation suite — overtime, minimum wage, breaks, wage statements, expense reimbursement — and sought 'civil penalties . . . including unpaid wages' under § 558. ZB moved to compel arbitration of the 'victim-specific' unpaid wages component, conceding under Iskanian that traditional PAGA penalties could not be arbitrated.

The motion won and backfired at once: the trial court sent the unpaid wages issue to arbitration but — reasoning that PAGA is inherently representative — ordered it arbitrated on behalf of every aggrieved employee, converting ZB's individual-arbitration strategy into a representative wage arbitration. The Court of Appeal issued a writ on a different theory: following Thurman, it held the § 558 penalty indivisible, wages included, and therefore entirely non-arbitrable under Iskanian. The Supreme Court granted review to resolve the arbitration split and resolved it by dissolving its premise.

The opinion begins from first principles. A civil penalty is a law-enforcement sanction — 'fundamentally a law enforcement action designed to protect the public and not to benefit private parties' — assessed in addition to any losses, and before PAGA collectible only by the state. Compensatory relief is different in kind: unpaid wages were always recoverable directly by employees, through their own statutory and contract claims. PAGA's innovation was to deputize employees to collect the former; it did not federate the latter into the penalty system.

Read in that light, § 558 has two distinguishable parts. The fixed per-employee, per-pay-period amounts are a classic civil penalty. The 'amount sufficient to recover underpaid wages' — which the statute directs be paid entirely to the affected employee, not split 75/25 with the state as PAGA requires of penalties — is compensation the Labor Commissioner can order in its administrative citation process. Treating it as a PAGA-recoverable penalty would let private plaintiffs route ordinary wage recovery through the penalty statute, bypassing both the wage claims the Legislature gave employees directly and the distribution rules PAGA imposes on penalties. Thurman, which had read the § 558 penalty as indivisible, was disapproved.

The arbitration question the court granted review to decide answered itself. ZB sought to compel arbitration only of the unpaid wages request — and because that request sought relief PAGA does not authorize, there was no claim to compel. The motion should simply have been denied; the trial court on remand could strike the unpaid wages allegations or permit amendment. The employer thus won the war by losing the motion: the representative wage arbitration disappeared along with the wage claim itself.

The strategic lesson outlived the procedural posture. Before ZB, employers faced PAGA complaints functioning as class actions in disguise — wage recovery for the entire workforce dressed as a penalty claim, immune from arbitration under Iskanian and from class certification requirements under Arias. After ZB, that vehicle is gone. Wage recovery at scale requires a class action, with its certification discipline, or individual claims, with their arbitration exposure. PAGA carries penalties only.

ZB converts into a mechanical audit that should be run on every PAGA demand, violation by violation. Step one: does a specific civil penalty statute cover the violation? For overtime and related wage-order violations, § 558 supplies the fixed $50/$100 amounts — and nothing more. Step two: if a specific penalty exists, is it one PAGA reaches? Step three: if no specific penalty applies, the § 2699(f) default penalty governs. At every step, any amount that is compensation to an employee — overtime shortfalls, minimum wage deficits, § 226.7 premiums at the Ferra rate, § 2802 expense reimbursement, interest — is excluded from the penalty base entirely.

Worked through a typical demand, the corrections are substantial. Meal and rest exposure is the § 2699(f) default penalty for the § 226.7 violation, not the premium amounts — Kirby's wage characterization, applied through ZB's boundary. Overtime exposure is § 558's fixed component, not the underpayment. Expense reimbursement is the default penalty, not the unreimbursed sums. Demands that load wage amounts into the penalty column overstate PAGA exposure by multiples, and the correction is not an argument about facts — it is arithmetic commanded by a unanimous Supreme Court.

The 2024 reform legislation left ZB's architecture fully intact: amended § 2699 still authorizes civil penalties only, now with caps, cure credits, and a 65/35 state-employee split — every one of which operates on the penalty base that ZB defines. If anything, the reform amplified ZB's importance, because the reasonable-steps caps are computed as percentages of properly calculated penalties; an inflated base corrupts the cap analysis too.

The wage side of the ledger did not disappear — it moved. Naranjo routes unpaid premiums into wage statement and waiting time exposure; direct and class claims remain the vehicle for wage recovery; and post-reform plaintiffs increasingly plead PAGA alongside, not instead of, those claims. The defense discipline is keeping each ledger honest: wages in the wage claims, penalties in the PAGA claim, and no amount counted twice.

Stone v. Alameda Health System (2024) 16 Cal.5th 10401 completed the boundary from the other direction. ZB answers what a PAGA plaintiff may recover; Stone answers from whom. Section 2699(b) gives 'person' the meaning it carries in section 18 and does so 'for purposes of this part,' so a public entity is not a permissible PAGA defendant at all — not for the § 2699(f) default penalty and not for a penalty specifically provided elsewhere in the Labor Code. The Court of Appeal had allowed the second while barring the first; the Supreme Court rejected the distinction and disapproved Sargent v. Board of Trustees of California State University (2021) 61 Cal.App.5th 658 as to nondefault penalties. Read together the two decisions bound the recoverable universe on both axes, and the order of operations matters: the defendant question is categorical and comes first, because it zeroes the penalty column rather than trimming it.

ZB, N.A. is the single most-used decision in PAGA exposure work because it is the one that changes the number. The wage/penalty boundary, applied category by category, routinely removes 30 to 50 percent of a demand's face value — more where premiums and expense reimbursement dominate. It is also the decision that disciplines settlement structure: allocations must respect what PAGA can and cannot recover, and LWDA review makes sloppy allocations a real risk.

Analytically, ZB completed the taxonomy Kirby and Murphy began: remedy-wages versus violation-penalties, private compensation versus state sanction. Every subsequent fight — Naranjo's derivative wages, the reform's penalty caps, even the recoverability of subsections within specific penalty statutes — is played on the board ZB drew. The Recoverability Checker tool on this site automates the category-by-category analysis.

  1. Run the three-step ZB audit on every PAGA demand before responding: specific penalty statute, PAGA recoverability, default penalty — and strip every compensatory amount from the penalty base.
  2. Demand the plaintiff's penalty math in mediation briefs and meet-and-confer: a model that includes premiums, unpaid wages, or reimbursements as penalties is wrong as a matter of law, and saying so early resets the negotiation anchor.
  3. Keep the wage and penalty ledgers separate in exposure models — wage amounts belong to direct and class claims with their own defenses, limitations periods, and arbitration posture; conflating the ledgers double-counts and misprices both.
  4. Apply the boundary to settlement architecture: PAGA allocations cover penalties subject to the statutory split and court approval, while wage consideration belongs to the individual or class components.
  5. Where a complaint seeks wages through PAGA, move against the defect directly — ZB makes the request impermissible on its face, and striking it narrows the case before any merits work.
  6. Recompute the reform's 15 and 30 percent caps from a ZB-clean penalty base — the caps compound the correction, and both steps together often reduce headline exposure by an order of magnitude.
Related Decisions
KirbyNaranjoAdolphIloffStone
On the citations
  1. 1Stone v. Alameda Health System (2024) 16 Cal.5th 1040The official citation was verified from three later published opinions, including the Supreme Court's own In re Ja.O., not from the slip header. CourtListener's cluster citations array is empty and must not be relied on.
This analysis is for informational purposes only and was last reviewed Q3 2026. Verify current status before relying on any authority.
Brinker Restaurant Corp. v. Superior CourtNaranjo v. Spectrum Security Services, Inc.