Practical Tips · Case Study · June 2026
A claimant had the insurer admitting it could not name a single time its expert ever sided with a policyholder — and lost on summary judgment. The full record survives. Here is what it teaches.
We almost never get to see behind a denial. A bias case that fails usually fails quietly — a trial court grants summary judgment, the appellate court affirms in a short unpublished opinion, and the record that would explain why stays in a clerk’s file no one ever pulls. We get the conclusion without the autopsy.
This case is the exception, and that is what makes it worth studying. The opinion is unpublished and disposes of the expert-bias argument in two paragraphs. But the underlying record — the briefs, the discovery responses, and a thirty-two-page declaration from a former insurance-defense lawyer — lets us reconstruct, document by document, how a claimant who had the makings of a landmark decision lost. The lesson is not really about this court. It is about architecture.
Through discovery, the insurer was made to answer two questions about the engineering firm whose report drove the denial. Asked how many times it had hired the firm, the insurer objected that the question was “overbroad and unduly burdensome” and gave no number. Asked how often that firm had ever found damage consistent with an insured’s account, it answered that it “doesn’t keep track … and therefore cannot reasonably provide an estimate.” The claimant’s expert drew the only available conclusion:
“Apart from the frequency of [the insurer’s] use of [the engineer] and the fact [the insurer] cannot even say there was a single instance in which [the engineer] agreed with the insured …” — the report bore marks of bias a competent adjuster would have caught.
An insurer that cannot identify one case, ever, in which its chosen expert took the policyholder’s side — and the claimant still lost. The facts were not the problem.
A bias case stands on three things, and they are not interchangeable. Miss any one and a sympathetic record collapses into “no triable issue.”
| Layer | What it is | What happened here |
|---|---|---|
| The standard | The legal test — an inference of bias, not proof of a corrupt mind. | Never planted. The court applied the most demanding “dishonest selection” framing because no one held it to the inference-of-bias standard. |
| The factors | The evidence that satisfies the standard — financial dependence, patterns, methodology, reasonable measures. | Half-built. Strong on the expert’s sloppy methodology; thin on the money; the pattern evidence borrowed from other lawsuits rather than built in this one. |
| The presumption | The burden-shift: once a threshold is met, the insurer must prove the expert was neutral. | Never invoked. The insurer’s “we don’t keep track” was left to count against the claimant instead of triggering the shift. |
The court framed the question as whether the insurer “dishonestly selected” its expert — language that sounds in conscious bad intent, the one thing a claimant can almost never prove directly. That framing was available because the briefing never named the governing test. The controlling standard is an inference of bias: a fair inference of financial conflict is enough to make bias cognizable, regardless of anyone’s subjective honesty. The standard you fail to assert is the standard the court gets to choose for you.
The most important factor is financial dependence, and it is the one this record most underbuilt. The claimant tried to get the numbers from the insurer — who objected and claimed not to track them — and never went to the source that had them: the expert. Compensation totals, annual assignment volume, and the share of a practice that is insurer-side live in the expert’s own billing records and tax forms. A number you can authenticate from a third party is a number the court cannot wave away.
The cross-case pattern here was real and damning — a string of look-alike denials, and an expert with no recorded instance of ever siding with an insured. But it entered through an expert’s declaration describing other lawsuits, which let the court treat this expert’s record in this case as a near-blank — “a handful of other cases.” Pattern evidence is unanswerable when it is your own record, developed in your own discovery, and merely arguable when it is borrowed.
When an insurer objects that a volume question is “unduly burdensome” and then says it “doesn’t keep track,” that is not a reason the claimant loses for want of metrics. It is the predicate for the opposite: a motion to compel, an adverse inference against a sophisticated party that could easily track its expert’s outcomes and chooses not to, and — once the threshold is met — a burden-shift the insurer cannot satisfy when it has admitted it cannot name a single pro-insured result. Here, the suppression was allowed to become the claimant’s evidentiary hole instead.
There is a further opening in this record, and it is the one most worth carrying away. The insurer told the court — repeatedly — that it had retained an “independent expert,” yet it submitted no declaration from the expert establishing that independence, resting the claim entirely on its own fraud investigator’s say-so. At the same time it professed not to “keep track” of how often it hired the expert and resisted the count as “burdensome.” But its own filings told a different story: the record showed the insurer had designated the same expert as its expert witness in other bad-faith cases and given them repeat business. A relationship the insurer’s litigation posture treated as minimal and untracked was, on the insurer’s own conduct elsewhere, substantial and ongoing.
That gap is a discovery lane, and it points straight at the factor the claimant never built — the reasonable measures an insurer must take to ensure the neutrality it asserts. An expert’s prior court declarations, and an insurer’s filings in other cases about the same expert — the designations, the declarations, the representations it makes (or pointedly avoids making) about the relationship — are discoverable and often decisive. An insurer that vouches for an expert’s independence case after case while never monitoring their compensation or their outcomes has documented, across its own filings, that it took no measures to verify the neutrality it sells to each new court. Inconsistent descriptions across cases are impeachment; a studied silence about the relationship is itself evidence.
Read the deep-dive → See the checklist → The bias-evaluation service →
The framework behind this case — the inference-of-bias standard, the rebuttable presumption, and the four factors — is laid out in Demer’s Paradigm for Assessing Biased Insurance Experts (Advocate Magazine, 2024). For why the bias fight is the one that matters, see The Loss the Insurer Is Happy to Take; for the first factor in depth, The Two Metrics That Trigger the Burden-Shift.
Quotations are from the public appellate record in Bagramyan v. Government Employees Ins. Co., No. B315018 (Cal. Ct. App. July 20, 2023) (unpublished) — the appellate briefs and the Declaration of Bruce M. Warren. The case name is omitted from the body because the opinion is unpublished and noncitable under California Rules of Court, rule 8.1115; it is identified here for source transparency only. Educational and informational only; not legal advice.