POLICY ANALYSIS — APRIL 2026

The $20 Fast-Food Wage

Two years in. Six major studies. Two opposing narratives. Here's what the evidence actually supports.

Wage Floor
$20.00
Hourly minimum since Apr 1, 2024
Workers Covered
~525K
At chains with 60+ locations
Wage Jump
+25%
From the prior $16 floor
Jobs Lost (Est.)
~18K
NBER counterfactual estimate

The Law in Brief

AB 1228 — signed Sep 2023, effective Apr 2024.

AB 1228 set a $20/hour floor for limited-service restaurants in chains with 60+ locations nationwide — roughly 525,000 workers at ~30,000 establishments. A Fast Food Council can adjust the rate annually (capped at the lower of 3.5% or CPI). No 2026 increase has been enacted; a discussed 70-cent COLA bump has stalled since the council chair resigned mid-2025.

Exempt: bakeries producing on-site stand-alone bread items (as of Sep 15, 2023); restaurants in airports, hotels, event centers, theme parks, schools, museums, and gambling establishments. Cities cannot set higher fast-food-only minimums.

Employment Effect by Study

Net jobs vs. counterfactual. Same law, very different numbers.

Reich (Berkeley)Schneider (Harvard-UCSF)Clemens (NBER)EPI / BLS QCEW-20K-15K-10K-5K0-18,000-19,102

Menu Price Effect by Study

Reported price increases for fast-food menu items, percent.

Reich (Apr '26)Reich (Oct '24)Owen (UCSC)BRG / Datassential0%4%8%12%16%1.5%3.7%10%14.5%

Note: Owen (UCSC) reported a range of 8–12%; midpoint shown.

The Studies, Side by Side

Method, finding, and the most defensible critique of each.

Reich (UC Berkeley)

No job loss

Method. Glassdoor + Square payroll, Advan cell-phone proximity (4-hr stay = worker), DoorDash prices.

Finding. Wages +11%. Employment unchanged. Prices +1.5%.

Caveat. GPS proxy mistakes lingerers and delivery drivers for staff; geofencing imprecise in strip malls.

Schneider (Harvard / UCSF)

No benefit cuts

Method. Survey of 3,420 CA fast-food workers vs. 6,194 CA retail and 14,416 out-of-state controls.

Finding. Share earning <$20/hr fell 60+ pp. No reductions in hours, scheduling, or benefits.

Caveat. Captures incumbents only — by design misses workers who lost jobs and exited.

Clemens et al. (NBER)

−18,000 jobs

Method. BLS QCEW administrative payroll data (95% of US jobs). Difference-in-differences vs. rest of US.

Finding. CA fast-food employment −2.7% to −3.9% vs. counterfactual. Implied elasticity −0.09 to −0.16.

Caveat. NAICS 722513 includes restaurants not covered by AB 1228; elasticities at upper end of literature.

Employment Policies Institute

−250 hrs / worker

Method. Self-reported survey of 182 operators (Jul 2024) + CPS-ORG hours analysis + ongoing QCEW tracking.

Finding. 98% raised prices, 89% cut hours, 73% cut OT. Median worker −5 hrs/wk (~$4,000/yr).

Caveat. Industry-funded (Berman & Co.); operator survey non-random and self-selected.

Berkeley Research Group

+14.5% prices

Method. Datassential menu price data, Sep 2023 → Oct 2024. Industry consultancy, not academic.

Finding. CA fast-food prices rose ~2× the national average over the comparison window.

Caveat. Not peer-reviewed; baseline window includes pre-implementation moves and food inflation.

Owen (UC Santa Cruz)

Automation up

Method. Interviews with 100+ franchise owners/managers + financial and hiring records review.

Finding. 21% shift-work decline at one Burger King group; 12% labor-hour drop across 18 McDonald's. Prices +8–12%.

Caveat. Qualitative-heavy, no formal control group; sector automation predates AB 1228.

Why the Studies Disagree

Same law, different instruments — different answers.

THE "NO HARM" CAMP

Worker surveys, payroll-platform data, and cell-phone proximity proxies. Captures wage gains and incumbent experience well. Misses workers who exited; Advan GPS data has known imprecision in clustered retail environments.

THE "REAL HARM" CAMP

BLS QCEW administrative payroll data — gold standard for sector employment counts. Captures aggregate effects, but NAICS classification includes establishments not covered by AB 1228, blurring the law's direct effect.

Bottom Line

Both the "win-win-win" and "job-killer" framings are too clean. The defensible read: covered workers got a real 10–18% wage gain. About 16,000–19,000 net jobs were lost relative to the counterfactual, and the median retained worker lost roughly 5 hours per week. Prices rose more than Reich's 1.5% but less than industry's 14.5% — most likely 4–8% on average. Capital substitution (kiosks, AI ordering) is the longest-tail effect.

The unanswered welfare question is whether the gains to retained workers exceed the losses to displaced workers and consumers. No study has cleanly answered that.

Sources. Reich & Sosinskiy, UC Berkeley CWED (Apr 2026). Schneider, Harknett & Bruey, Harvard Kennedy / UCSF Shift Project (Oct 2024). Clemens, Edwards & Meer, NBER WP 34033 (Jul 2025). Employment Policies Institute (2024–2026). Berkeley Research Group / Datassential (Feb 2025). Owen et al., UC Santa Cruz (Mar 2026). California DIR; BLS QCEW.
Sources: EIA, CAISO, CA Energy Commission, CPUC, CA Legislative Analyst, NRC, AAA, GasBuddy, Tax Foundation, MIT/Stanford, Kpler, Bloomberg, Reuters, CalMatters, UC Davis. March 2026.
© 2026 CA Policy · About