Practical Tips · Case Study · June 2026
When an insured asks for an expert’s other files, insurers cite one case above all others to slam the cabinet shut. Read past the headline and Tilem v. Travelers is a catalog of how to lose — not a rule that the files are off-limits.
The companion tip explains why the second Demer factor — the expert’s pattern across other insureds’ claim files — can only be proven through discovery, and why California courts grant that discovery far more often than they deny it. This is the case the industry holds up to argue the opposite. It is worth studying precisely because it lost, and because how it lost is a map of what a prepared plaintiff does differently.
Jeffrey Tilem’s home was damaged by the rainstorms of late 2016 and early 2017. Travelers paid more than $45,000 for the water that came through the roof, then denied the structural claim — the cracked walls, jammed doors, and shifting floors he valued above $500,000. Its retained firm, Coronado, attributed the movement to “slope creep,” natural instability predating the rain, and Travelers invoked the earth-movement exclusion. Tilem’s own engineer said the rain did it. The carrier kept Coronado’s answer.
Tilem’s theory was that Coronado does this for a living — that it “always favors” Travelers and “always harms” its insureds. So he asked for the proof: every report Coronado had prepared for Travelers’ insureds since 2012, and identification of every claim in which the firm had been retained — roughly sixty-eight files. Travelers answered with a number of its own: more than $80,000 and 400-plus hours to find, review, and redact them. The magistrate denied the motion on three grounds — relevance, proportionality, and the privacy of the nonparty insureds — and those are the words insurers now quote whenever an insured asks for an expert’s other files. The decision is Tilem v. Travelers Commercial Ins. Co., No. CV 17-6904-RSWL (ASx), 2018 WL 4963124 (C.D. Cal. May 22, 2018).
Tilem did not hold that the other files are off-limits. It held that this plaintiff asked for them the wrong way.
The reasoning that does the real work is borrowed from an older case, Ricotta v. Allstate: even if every Coronado report ever written favored the insurer, favorability is not inaccuracy, and to know whether any one was actually wrong you would have to re-inspect that house in its prior condition — a mini-trial for every file. Multiply by sixty-eight and the discovery looks disproportionate. It is a tidy argument, and it proves far too much. Taken seriously, no volume of one-sided opinions could ever justify looking at the pattern, because each data point would first have to be re-litigated before it could count. A reviewer who has never once found for a policyholder across dozens of files has produced a pattern that is itself evidence — exactly the kind of probability the inference-of-bias standard credits. The objection confuses admitting the pattern with proving each case within it. And often the pattern need not be re-litigated file by file at all, because the bias shows on the face of the reports — the same parsimonious, result-driven reasoning and the same procedural shortcuts, repeated across matters. (That the magistrate framed the question as “favorable isn’t the same as wrong” owes something to the posture of the case: the plaintiff never pressed the inference-of-bias standard, so the framing went unanswered — a gap a prepared plaintiff closes.)
| What sank the request | What it actually shows |
|---|---|
| Wrong claim pled. Tilem alleged individual bad faith but asked for pattern-and-practice discovery — and conceded at the hearing he could not assert a pattern-and-practice cause of action. | California allows OICF discovery to prove a pattern and practice (Colonial Life). The court declined to order pattern discovery in a case with no pattern claim — a pleading problem, not a doctrine. |
| No threshold showing. The only support offered that Coronado was biased was a single other lawsuit making the same accusation. | A court cannot take the truth of another case’s allegations as established. The theory arrived without a foundation a deposition or a narrowed request could have supplied. |
| The door was left open. The denial was without prejudice; the court pointed Tilem to a deposition and said it “would consider allowing the discovery with appropriate safeguards.” | This is the opposite of a bar. The court invited a renewed, better-founded request — and treated privacy as a reason to redact and protect, conditioned on relevance, not a wall. |
The reason insurers love Tilem is that it lets them describe a trap: you cannot have the files until you show the expert is biased, and you cannot show bias without the files. But that is not the holding. The law does not ask you to prove bias to earn discovery; it asks you to plead the pattern as a pattern and to make a threshold showing that the other files are relevant to it. Those are different burdens, and the second is one a prepared plaintiff can carry — through the claim chosen, the duties invoked (the insurer’s obligation to weigh the insured’s interests equally and to conduct a fair investigation, not merely a thorough one), and a request narrowed to the biased participant and bounded in time, geography, and scope. Tilem is not the wall. It is the instruction manual for getting over it.
Read the deep-dive → The OICF tip → The bias-evaluation service →
See the issue this case turns on in The Files the Insurer Hopes You Won’t Ask For, the standard the pattern serves in The Standard Comes Before the Factors, and the framework as a whole in Demer’s Paradigm for Assessing Biased Insurance Experts (Advocate Magazine, 2024).
Distilled from the project’s own analysis and case-wiki entry for Tilem v. Travelers Commercial Ins. Co., 2018 WL 4963124 (C.D. Cal. 2018). Holdings — Tilem; Ricotta v. Allstate Ins. Co., 211 F.R.D. 622 (S.D. Cal. 2002); Colonial Life & Accident Ins. Co. v. Superior Court, 31 Cal.3d 785 (1982) — cited for their principles. Educational and informational only; not legal advice.