Butler v. Jimmy John's Franchise, LLC
3:18-cv-00133
| S.D. Ill. | Feb 24, 2021Background
- Plaintiffs (putative class of Jimmy John’s employees, 2014–2018) allege a franchise "No‑Poach Provision" restrained competition for labor, suppressed wages, and created classwide antitrust injury under Section 1 of the Sherman Act.
- Plaintiffs relied on Dr. Hal Singer, who used wage‑regression models on Weekly Sales Reports (WSRs) covering ~615,000 employees to estimate classwide wage suppression and aggregate damages.
- Jimmy John’s attacked Singer and offered two rebuttal experts: Dr. Janusz Ordover (antitrust economist) and Dr. Justin McCrary (labor/franchise economist), who identified conceptual, data, and enforcement problems and advanced procompetitive justifications.
- Key dispute at class‑certification stage: whether plaintiffs’ expert methodology can establish predominance/common impact under Rule 23(b)(3) and whether contested expert testimony is admissible under Daubert/Rule 702.
- The court excluded Dr. Singer’s testimony because his regressions suffered a material methodological flaw (random measurement error from mixing per‑hour and per‑shift pay records that inflated wage estimates) and denied plaintiffs’ motions to exclude Ordover and McCrary, finding their methodologies reliable (challenges go to weight).
Issues
| Issue | Plaintiff's Argument | Defendant's Argument | Held |
|---|---|---|---|
| Admissibility of Dr. Singer’s wage‑regression evidence to prove classwide impact | Singer’s regressions directly measure wage effects before/after cessation of No‑Poach; fixed effects absorb measurement issues | WSRs mix per‑hour and per‑shift pay; Singer misclassified some salaried managers as hourly, producing random measurement error that inflated results | Excluded — court found Singer’s methodological error (mixed pay units not fixed) materially inflated impact and rendered testimony unreliable under Daubert/Rule 702 |
| Admissibility of Dr. Ordover’s critiques and alternate regressions | Plaintiffs: Ordover’s unpooling and state/year regressions create endogeneity/sample‑selection and false positives | Ordover: unpooling corrects data errors, tests geographic/year heterogeneity, and shows many employees unlikely injured; any data choices affect weight, not admissibility | Admitted — court found Ordover’s methods reliable; criticisms go to weight and cross‑examination |
| Admissibility of Dr. McCrary’s franchise/economic analysis (procompetitive justifications, enforcement evidence) | Plaintiffs: McCrary’s opinions lack documentary proof that procompetitive benefits caused the clause or changed when it ceased; some factual inputs unreliable | McCrary: franchising economics, training/encroachment/freeriding theory, document review (release requests ~88% granted), and regression on manager training/support his views; challenges affect weight | Admitted — court found McCrary’s reasoning grounded in literature and record; attacks on data or completeness go to weight |
| Use of expert evidence to satisfy Rule 23 predominance (impact element) | Singer: regression shows 87–91% injured and broader structure implies classwide injury | Defendants: methodological flaws, heterogeneous markets, selective/enforced clause, alternative labor markets and procompetitive purposes defeat common proof | Court: Because Singer’s model was unreliable it cannot support predominance; remaining experts survive Daubert but ultimate Rule 23 analysis requires further assessment at certification stage |
Key Cases Cited
- Daubert v. Merrell Dow Pharms., 509 U.S. 579 (1993) (trial judge gatekeeper for expert reliability under Rule 702)
- Gen. Elec. Co. v. Joiner, 522 U.S. 136 (1997) (court may exclude expert opinion unsupported by data)
- Kumho Tire Co. v. Carmichael, 526 U.S. 137 (1999) (Daubert principles apply to all expert testimony)
- Manpower, Inc. v. Ins. Co. of Pennsylvania, 732 F.3d 796 (7th Cir. 2013) (challenges to expert data frequently go to weight, not admissibility)
- Messner v. Northshore Univ. HealthSystem, 669 F.3d 802 (7th Cir. 2012) (Daubert analysis may be required at class‑certification when expert is critical)
- Comcast Corp. v. Behrend, 569 U.S. 27 (2013) (model used for class certification must line up with liability theory)
- Wal‑Mart Stores, Inc. v. Dukes, 564 U.S. 338 (2011) (predominance inquiry may overlap with merits where common proof is contested)
- Tyson Foods, Inc. v. Bouaphakeo, 136 S. Ct. 1036 (2016) (representative/statistical evidence can be appropriate in class actions)
- In re Hydrogen Peroxide Antitrust Litig., 552 F.3d 305 (3d Cir. 2008) (impact is an element requiring proof that may be common or individual)
- Bazemore v. Friday, 478 U.S. 385 (1986) (regression analysis admissible only if major factors are included)