A Closed Claim Is Not a Closed Recovery

The closed file is a useful administrative fiction. It tells us the insured has been indemnified. The reserve is zeroed. The correspondence stops. The adjuster moves on. It doesn’t tell us the economic work has ended.

Payment ends the insured’s problem and may begin ours.

The payment may create a subrogation right.

It may start a salvage disposition.

It may reveal concurrent insurance.

In Ontario it may start a loss-transfer clock.

Those aren’t loose ends for a batch queue.

They are proceedings, with their own evidence and clocks.

Claims knows how to make an insured whole.

Recovery needs something else entirely.

It needs to know what the failed supply hose looked like before the restoration contractor took it out.

Which contract allocated the loss.

When the insured discovered the claim.

Whether the other party saw the evidence before it changed.

The first workflow is built to resolve pressure.

The second is built to preserve a right.

That is an authority boundary, not a handoff.

The evidence belongs to the first days

Recovery is flagged at first notice, not at payment.

The useful triggers are deliberately unromantic.

Another vehicle. A contractor on site last month. Water from the next unit. A young appliance on fire. A product that failed early. Each asks whether somebody other than the insured created a recoverable loss.

The adjuster still commands the proof.

The hose, the control board, the fitting: photographed in place, bagged, tagged, stored. It doesn’t stay with the restoration contractor because somebody correctly decided the kitchen has to go back together. For fire and large water losses, an origin-and-cause investigator retained within days works to NFPA 921.

The opposing expert will be measured against it.

Then the target and its insurer get a preservation notice. It names the artifact and offers a joint examination.

The public record is requested early.

Fire department and Fire Marshal reports. Police collision reports. In Ontario, the Electrical Safety Authority file.

Those records take weeks to arrive.

The insured gets written instructions too. Don’t sign a release. Don’t accept a direct repair. Don’t discuss fault.

None of this is an elaborate litigation ritual.

It is what lets a liability theory survive the repair. A file sent to recovery after the appliance went to the dump still holds immaculate PDFs. It has lost the thing that explains the fire.

The recovery screen is not a fraud score

Before any demand, run the cheap bars in order.

Direct compensation comes first.

In Ontario, Quebec, the Atlantic provinces, and Alberta since 2022, physical damage to an insured vehicle is paid by its own insurer, and statute removes the tort claim against the at-fault driver. A generic auto-subrogation queue is wrong in all of them.

Loss transfer does not rescue that queue.

Ontario loss transfer is statutory indemnification between insurers, on heavy vehicles and motorcycles.

It isn’t a tort claim against a driver.

Its two-year clock runs from refusal, not from the accident. Fault outcomes, and amounts ceded against received, are the measures.

The next screen pulls the underlying contract.

CCDC construction contracts require project insurance and waivers of subrogation among the parties. Commercial leases can put fire risk on the landlord’s insurance. An unnamed co-insured with an insurable interest may be immune. Ontario’s tenant trilogy treats tenants that way where the lease allocates fire risk.

Quebec bars subrogation within the insured’s household.

A release at the scene ends the right just as efficiently. So I would resist a “recovery likelihood” score as the first investment. A score ranks files after the route has closed. The screen answers the jurisdictional and contractual bars first, and asks about promise second.

The diary cannot start at payment

Common-law limitation periods run two years from discoverability.

Quebec generally provides three.

The decisive date is the insured’s discovery.

Not the date we issued the cheque.

Set the diary on payment and we keep a tidy timestamp and lose a valid claim.

The control is plain enough.

Diarise the limitation when the issue appears, then force review a couple of months early.

But the limitation date isn’t the whole calendar.

Ontario wants written notice within ten days for a highway or sidewalk in disrepair, and sixty days before proceeding against the Crown. Road carriage bills of lading generally require notice within nine months.

Carriage by sea has a one-year suit time.

A dashboard showing a green two-year date reports compliance on a file that died in week two. Claims has a diary discipline, and closure numbers hide its failures. Reopen rate is measured within three months, and closure ratio is claims closed over reported. Neither says whether the paid file carried forward the preservation action, the bar-screen result, the notice dates, and the limitation.

Closing restores our appetite for silence.

Recovery needs the file to stay articulate.

Payment transfers the work, it does not erase it

The subrogation right matures on payment.

Take a signed subrogation receipt with the settlement cheque, while the insured is still cooperative. In common-law provinces we sue in the insured’s name.

In Quebec we may sue in our own name.

Where an uninsured portion, a business-interruption tail, or a personal injury claim survives, the insured keeps that interest and the claims travel together.

The demand is documentation, not theatre.

Proof of loss. Invoices. Expert report. Photographs. A liability theory.

A one-month response window is the working convention.

The claim carries the insured’s deductible as well as indemnity and recoverable adjusting and expert costs.

It is not a courtesy.

Canadian practice makes the insured whole first, deductible and uninsured loss ahead of our outlay.

Report gross proceeds as recovery and we have double-counted the policyholder’s money.

Where both insurers are signatories and liability is the issue, Insurance Bureau of Canada agreements route the dispute to documents-only binding arbitration.

No costs. No oral evidence.

An uninsured or non-signatory target means litigation.

In Ontario, Small Claims Court handles claims up to $35,000, and larger claims go to Superior Court.

The route belongs in the recovery disposition.

Not in a generic status called “closed.”

Salvage and contribution need records of their own.

Salvage begins with the total-loss determination.

We take ownership, or we deduct the appraised value when the insured keeps the asset.

Ontario vehicle branding separates irreparable, salvage, and rebuilt, and the brand changes who can bid and what the asset is worth.

Contribution is a claim against another insurer on the same risk, prompted by an other-insurance clause.

Equal shares, independent liability, and maximum liability allocate identical facts differently.

None of it is subrogation because money came back.

The number that flatters is not the number that matters

Recovery reporting misleads as soon as it accepts one denominator.

Portfolio recovery ratio is recoveries over indemnity paid.

On a mixed Canadian book it sits in the low single digits, moving mostly with line mix.

Referral yield divides recoveries by paid indemnity on referred files, and often exceeds half of it.

One is low because it sees the whole book.

The other is high because somebody already picked those files.

I would not present either alone.

A team reporting sixty per cent referral yield and fifteen per cent missed opportunity is doing worse than one reporting forty and three.

The first has selected clean files well and failed to see too much of the population.

Missed opportunity comes from a blind closed-file audit.

Sample paid, closed, no-recovery files above materiality.

Then ask an examiner who never touched the file whether a third party was identifiable, evidence preserved, limitation diarised, waiver checked.

The audit is uncomfortable: it measures a decision nobody made.

Our system can’t report a right nobody identified.

It reports the closure, the payment, an optimistic yield.

It cannot report the plumber, the manufacturer, or the co-insurer who never entered a field.

Use accident year for the cohort, not calendar year.

Recoveries arrive years after payment.

Calendar-year receipts over calendar-year paid loss mix cohorts, and flatter a shrinking book.

For salvage, measure net proceeds against gross total-loss indemnity by disposal channel, then cut the result by provincial brand.

For Ontario loss transfer, measure ceded and received by accident year.

Different mechanisms need different measures.

What a control test would show

A recovery percentage will reach the board.

We should bring a control test instead.

What share of eligible paid files carries a preservation action, a bar-screen result, a limitation and notice calendar, a disposition, and an owner?

That question connects claims, legal, recovery, finance, and the system, without pretending they are one function.

It also separates the reserve view from the recovery view.

Under IFRS 17, expected recoveries belong to the group’s fulfilment cash flows where anticipated salvage and subrogation reduce the case reserve.

The recovery triangle tests whether that anticipation becomes cash.

A persistent shortfall is an aggressive reserve offset before it is a disappointing recovery result.

So build the design nobody asks for.

Make the recovery decision before closure relieves the pressure.

A referral and preservation decision at first notice. A bar screen before demand. A calendar that survives payment.

Recovery is not what happens after the claim.

It is what the claim preserved the right to do.