The Discovery Objections Series · Part 3 · Case Study · July 2026
An arbitrator had already rejected the insurer’s expert. The discovery court found the bias data relevant. And the motion still failed — to a declaration nobody tested and a compromise with a hole nobody saw.
Every part of the Discovery Objections Series pairs the doctrine with a case study, and each is a case the policyholder should have won. Part 3’s study is Morales v. Allstate Northbrook Indem. Co., No. 5:20-cv-02577-ODW-SHK, 2021 WL 6618614 (C.D. Cal. Dec. 20, 2021), and it is one of the more instructive losses in this series because of everything the claimant had already won. The relevance fight — the one insurers usually make everything turn on — was over, and Morales had won it. What defeated her was the material of this part: an untested burden declaration, a mini-trial rationale, a rejected compromise, and a vendor payment structure that had quietly booby-trapped the compromise she rejected. Four layers, each with a fix.
A car accident; the at-fault driver’s carrier paid its $15,000 limit; Morales pursued underinsured-motorist benefits under her own Allstate policy, with limits of $100,000. Allstate arranged what it called an “Independent Medical Examination” — retaining an otolaryngologist, Dr. Hamid Djalilian, not directly but through a third-party vendor called Litili. Hold that structure in mind; it becomes the story.
The examination went the way the complaint alleged such examinations reliably go: the report minimized Morales’s complaints and attributed them to other causes, with contentions of malingering and feigning. The UIM claim went to arbitration — where the arbitrator found Morales “a very credible witness,” rejected the malingering and feigning contentions, and awarded her $85,000, the policy maximum.
Then Morales sued for bad faith, alleging among other things that Allstate knew what it was buying when it retained Dr. Djalilian — that he was regularly retained by insurers to minimize claimants’ complaints, and that his report was the predictable product of that retention. To prove it, she served four interrogatories: identify the otolaryngologists Allstate had hired for examinations in four Southern California counties over five years; state how many times Allstate had hired Dr. Djalilian in five years; state how many times he had found malingering; state how many times he had found feigning.
Look at those requests through this series’ lenses. Temporally windowed to five years, geographically bounded to four counties, and keyed to a named expert — Part 2’s temporal, geographic, and professional-key boundaries, already on the face of the interrogatories. They sought counts — pure statistics, nothing to relitigate. This is well-bounded discovery, and the bounding is the point: a universe this narrow should have collapsed the burden claim on recomputation.
Allstate objected on every ground in the catalogue — relevance, privilege, privacy, overbreadth, burden — and the court began by handing Morales the victory that makes this case so useful. The information was relevant. Allstate had argued, on the strength of two older cases, that the “mere fact” of repeated retention cannot support a bias claim. The court distinguished those cases for exactly the reason this series has emphasized since Part 1: they address what suffices at trial, not what is discoverable — the trial-sufficiency standard is not the discovery standard. Retention-frequency and outcome-pattern statistics, the court held on the claimant’s authorities, are discoverable extrinsic evidence of expert bias.
That ruling deserves more attention than it gets. A federal court, in a first-party bad-faith case, holding squarely that the numbers this publication’s framework runs on — how often, with what findings — are proper discovery. If the order had ended there, Morales would be cited in every claimant’s brief. It did not end there.
| Layer one: the untested declaration | Allstate submitted a declaration from a litigation consultant stating that answering the interrogatories would require manually reviewing “thousands of claim files” — its systems, the declaration said, offered no other way. The court relied on it, essentially unexamined: no methodology deposition, no system-specific counter-showing, no demand for the category-by-category accounting. An insurer’s own payment systems — accounts payable, vendor invoices, 1099 issuance — are precisely where retention counts live in structured form, and nobody made the court look at them. |
|---|---|
| Layer two: the mini-trial rationale | Against the malingering and feigning counts, the court reasoned that such findings “would have to be analyzed under the circumstances of each case” — the adjudicative face of burden, credited. The statistical-use answer was available: an interrogatory asking how many times an expert found malingering seeks a number, not a re-adjudication; no individual assessment need be re-decided for two hundred findings to speak to the conditions of the two-hundred-and-first. That argument, the centerpiece of this part, is absent from the record as the court describes it. |
| Layer three: the refused compromise | Allstate offered a middle path: a search of its payment records by tax identification numbers, to count payments involving a specific TIN over time. The court found the proposal reasonable “at this stage” — and noted that Morales “appears to have chosen to refuse” it without proposing any alternative. Part 2 taught what that posture costs: the all-or-nothing claimant inherits the burden of the whole dispute. The denial cites the refusal. |
| Layer four: the hole in the compromise | The layer that makes Morales a permanent teaching case. Allstate did not retain Dr. Djalilian; Litili did. The payments ran Allstate → Litili → Djalilian. A search of Allstate’s payment records for Djalilian’s tax identification number would therefore have returned little or nothing. The offered compromise, “reasonable” on its face, was aimed at a payment path that did not exist. The fix costs one sentence: accepted, provided the search covers both Dr. Djalilian’s TIN and Litili’s TIN. Nobody surfaced the hole — not the claimant, not the court — and the case now stands as the canonical illustration of the vendor-intermediated payment shield. |
Stack the fixes and the motion wins. Test the declaration — methodology deposition, alternative sources, the self-induced-burden argument for whatever incapacity survives. Answer the mini-trial rationale with the statistical-use point the interrogatories embodied. Accept the compromise, re-specified to trace the actual payment path — and note what that re-specified search really is: not a concession, but the burden objection’s answer. A payment-records query keyed to a tax identification number is the manageable, pre-aggregated form the bounding always pointed toward — a search of accounts-payable, not a manual crawl through thousands of files — once both Dr. Djalilian’s and Litili’s identifiers are named. And sequence the case the way Part 2’s case study taught: the vendor-relationship discovery first — the Allstate–Litili agreements, the reporting and access terms — which would have shown where the retention data actually lived and who had the contractual right to pull it. Armas won those agreements as “clearly relevant”; Morales never sought them. The two cases are one lesson viewed from its two sides: structure first, statistics second.
And well-bounded requests are not self-executing. Morales windowed her interrogatories to five years, held them to four counties, and keyed three of the four to a named expert — textbook bounding — and lost anyway, because the bounding was never converted into a recomputation and the residual “incapacity” Allstate claimed was never attacked as self-induced. Bounding shrinks the universe; it does not excuse the follow-through. Relevance can be won and the motion still lost — the burden fight is a separate fight, with its own record to build. Burden declarations are examined or they are conceded; there is no third option. A compromise is refused, accepted, or re-specified — and re-specification is usually the winning move, because the counter-offer both preserves the negotiating record and closes the traps. And whenever a vendor stands between the insurer and the expert, trace the money’s actual path before agreeing to search anywhere — the shield only works in the dark.
Free edition · The story
The full narrative: the arbitration Morales won, the relevance ruling in her favor, the four layers of the loss, and the one-sentence fix to the vendor-routed compromise that would have changed the outcome.
Paid edition · Ground by ground
The practitioner’s dissection: all four interrogatories verbatim, each objection and its disposition, the authorities on both sides of the relevance ruling, the format caveat that confines the holding, and the full layered rewrite.
← Part 3 overview The checklist → The bias-evaluation service →
The doctrine this case tests is laid out in The Two Faces of Burden — and the Trial That Never Happens; the Part 2 counterpart that supplies the missing move is The Court That Wrote the Winning Motion; and the pattern evidence at issue is mapped in The Files the Insurer Hopes You Won’t Ask For.
Distilled from the project’s own reading of Morales v. Allstate Northbrook Indem. Co., No. 5:20-cv-02577-ODW-SHK, 2021 WL 6618614 (C.D. Cal. Dec. 20, 2021). Source text in sources_md_conversions/cases/; case-wiki entry in wiki/cases/. The ground-by-ground rewrite is reserved for the subscriber edition. Educational and informational only; not legal advice.