The Concessions That Defeat Privacy and Privilege

The reach-for-it version of Part 4. This is the moves, none of the citations. It assumes Parts 1 through 3 — the pleading predicate, the scope armor, and the burden discipline — because privacy fights are lost on scale, and scale is a Parts 2–3 problem wearing a Part 4 label. The organizing principle: you defeat the privacy objection by conceding away everything legitimate it protects, leaving the insurer nothing sympathetic to withhold behind.

Sort first

1. Point to the person, in your first responsive sentence. Whose privacy? Third-party insureds’ claim files carry real interests — accommodated by the concessions below. The insurer’s own payment ledgers, retention data, 1099 issuance, and selection protocols carry no third party at all — and a privacy objection over them should be named as the substitution it is. The concessions are built for the first pile and devastating to the second: once the real interests are accommodated, the insurer’s resistance can only attach to its own institutional data, where there was never a person to protect.
2. Run the four-pile audit. Relational metrics: no third party; privacy near-pretextual. Reputational metrics (other insureds’ files): the genuine case; the concessions resolve it. Procedural irregularities (insurer–vendor–expert communications): no third party; the fight is privilege. Reasonable-measures material (protocols, audits): privacy near-frivolous; privilege is the only serious contest. The insurer’s brief will blend all four under one heading; your opposition takes them apart.

Concede the protective order — and load it with the two terms

1. Stipulate the protective order before the fight, and propose it first. It is the least restrictive mechanism adequate to the interest, and negotiating it before the motion narrows the dispute to relevance and proportionality — grounds on which you are strongest. The insurer that stipulates and then asserts privacy anyway must explain why its own stipulation fails — a posture that has decided reported cases against the objector.
2. Term one — identifiers redacted in filings, no sealing. Produced materials are used only in the litigation, and anything filed with the court is filed with nonparty personal identifiers redacted rather than filed under seal. This answers the public-record concern without the friction and disfavor of sealing motions: identities off the public page, substance in the open.
3. Term two — no contact with insureds. Neither party contacts any nonparty insured whose information appears in the production, unless that identity surfaces independently (other litigation files, public records, a separate investigation) or a later court order authorizes it. This concedes away the oldest privacy fear in the area — that discovery is a solicitation drive — before the insurer can raise it.

Concede the documents unredacted — because redaction is the cost

1. Take production unredacted under the order. Mass redaction is where the expense and burden of production live — and it is the burden the insurer will invoke to resist the whole request. The protective order protects the data; redaction just makes producing it costly. Concede unredacted production under the order, and you remove the insurer’s best privacy argument and its best burden argument at once.
2. Reserve redaction for the sensitive core only. Discrete medical conditions and diagnoses; highly sensitive individual financial data such as personal income. There the marginal privacy interest justifies the marginal cost — and the redaction pathway that takes the privacy statute out of play entirely is the fallback for those categories. Calibrate to claim type: a disability or health file carries a medical core a property file does not; sensitivity moves what warrants redaction, never whether the conditional objection becomes a bar.

Concede the names you never needed

1. Never request the personal information you do not need. The pattern proof is aggregate — counts, rates, dispositions, file contents with identities protected. Outside the class-action setting, identities become relevant only when specific grounds crystallize, and the requests should say expressly that names and identifying information are not sought until then. The concession costs nothing and strips the objection of its only sympathetic fact. This part’s case study is what the opposite looks like: contact information demanded as such, for 10,082 households, in a case that needed dispositions.
2. Know why the concession matters — and the class-action boundary. A front-loaded demand for thousands of insureds’ identities invites denial, sometimes because a court reads it as a settlement lever rather than discovery — settle now or the case grows into many cases. Conceding you neither need nor seek the names removes the suspicion. The exception is the class action, where the identities of class members are the discovery: there, courts compel them — and the same stipulated protective order you led with is how they are produced.

Force the statute’s own answers

1. Quote the whole statute — always. The privacy statute is a compliance menu: the judicial-order and permitted-by-law exceptions (a facially valid court order or subpoena is itself statutory authority to disclose), the written-authorization procedure, and the redaction insight that takes the statute out of play. The objection, as litigated, depends on omission — the prohibition quoted, the exceptions absent. And the omission can come from the bench: this part’s anchor order quoted the statute “in relevant part” and derived a consent-only rule from the truncation. Only the full text answers it.
2. Lead with the judicial-order pathway. The order granting your motion satisfies the statute by itself. Consent is a pathway, not a precondition — the courts that treat it as a precondition are reading exceptions out of the statute, and the line of authority saying so now spans two decades and includes the anchor court’s own district.
3. Demand the particularized harm showing. An unsupported privacy objection fails against available protective machinery. The insurer must show, with evidence, why the protective order does not suffice — an “unsupported assertion of practical impossibility” has been rejected in those words. Assertion is not demonstration; brief it like Part 3’s unquantified burden declaration, because it is the same defect.

Treat consent as a specialty tool

1. The consent letter is a specialty tool, not a default. It exists to unlock identified files with the insured’s authorization — appropriate for a small, bounded, relevant population whose narrative content you actually need. Aimed at thousands, it fails on scale; skipped in sequence or drafted argumentatively, it fails on mechanics. If a letter is used: neutral text, third-party administrator, letter before identities are disclosed, opt-in/opt-out default chosen by the court on the record.
2. Argue the self-selection cost of opt-in — never accept it silently. Consent-generated samples are biased by design: low, nonrandom response rates. If the insurer proposes opt-in consent as the exclusive mechanism, it is proposing to degrade the statistical evidence under privacy’s flag. Say so — and offer the unredacted-production-under-order alternative, which protects the nonparties and preserves the complete data set.

Hold privilege to its narrow empire

1. Business records are not legal communications. Compensation data, retention counts, invoices, performance metrics, ordinary-course instructions to experts — claims-administration records, not legal advice. The working question for every withheld document: generated in the course of deciding the claim, or in anticipation of defending the decision? Only the latter reaches the privilege analysis.
2. Demand the log; treat blanket assertions as forfeitures. Privilege is claimed document by document, on a log you can test. The boilerplate blanket objection is the privilege twin of the unquantified burden declaration — a conclusion where a showing is required.
3. In ERISA, deploy the fiduciary exception. The insurer administering a plan acts for the beneficiaries; pre-decisional communications about the claim — including expert selection, compensation, and evaluation — are presumptively discoverable by the claimant they were generated to evaluate. The leading case rejected the insurer carve-out and held that a claimant’s aggressive advocacy does not end the exception.
4. Name the channeling. An insurer that routes expert selection and evaluation through counsel is converting a claims-administration function into a litigation shield. In ERISA, the fiduciary exception answers it. Outside ERISA the fight is live and unsettled — argue the function (claims administration), not the conduit (counsel), and use the log discipline to force the fight document by document.

The failure modes — check against them last

1. The concessions this part’s instructive loss never made. No concession of the names — contact information demanded for 10,082 households in a case that needed dispositions (moves 8–9). No protective order offered in place of the most restrictive mechanism on the shelf (moves 3–5). A truncated statute unchallenged, because the claimant’s own motion assumed the consent-only world (moves 10–11). An untested burden declaration, and a redaction/production posture that fought the burden instead of conceding around it (moves 6, 12). The claimants in this part’s case study had won the relevance fight and held a judicial blueprint for passable scope; the absence of concessions lost what the doctrine had won.

← Part 4 overview   The case study →   The bias-evaluation service →

Related

See the doctrine behind these moves in Defeating Privacy and Privilege Objections Through Strategic Concessions, the cost of skipping them in The Motion That Asked for Ten Thousand Names, and the burden-stage predecessor in the Part 3 burden checklist.

Distilled from the project’s Part 4 doctrinal synthesis and its reading of the conditional-privacy line, California Insurance Code section 791.13 and its exceptions, and Dobro v. Allstate Ins. Co. The case anchor for each move, and the model language, are reserved for the subscriber Implementing Kit. Educational and informational only; not legal advice.