Skip to content

Rescue-Effect Audit

Diagnostic audit — instantiates Source–Sink Viability Management

Periodically tests whether a sink's apparent health is genuine local recovery or merely a rescue effect — persistence borrowed from a source — by asking what it would do if the support were removed.

A Rescue-Effect Audit exists to answer one dangerous question the archetype keeps raising: is this unit actually getting better, or does it just look alive because something is holding it up? Its defining move is the counterfactual-of-withdrawal — it does not read the sink's current numbers, which a rescue flow flatters, but asks what the sink's own trajectory would be if the support stopped. A unit that would keep rising is recovering; a unit that would fall back below viability is being rescued, and its healthy appearance is borrowed. This is the mechanism that catches the archetype's central self-deception — mistaking rescued persistence for recovery — and it does so as a periodic audit rather than a live gauge, because the judgment requires stepping back and probing, not just watching a dial.[1]

Example

A retailer runs a satellite store that has looked "turned around" for a year — foot traffic and sales are up, and the regional manager wants to stop the special support it receives (extended hours staffing paid from the flagship's budget, shared inventory from the flagship's overflow). A rescue-effect audit is the check before that call. Rather than trusting the store's own sales line, it profiles the dependency: how much of the satellite's traffic is walk-over from the nearby flagship's overflow, how reliable that spillover is, and what happened during a recent week when the flagship was closed for renovation.

That closed week is the natural experiment the audit seizes on — satellite sales dropped by roughly half. The finding: the satellite's recovery is largely a rescue effect; remove the flagship's spillover and it slides back below the threshold at which it covers its own costs. It has not switched from sink to self-sustaining source; it is a sink whose deficit is being quietly filled. The audit's output is a dependency profile and a role verdict — still a rescued sink, not recovered — which stops a withdrawal that would have looked safe and been a mistake.

How it works

  • Profile the dependency, not the surface health. Characterize the support the unit receives — its amount, reliability, pathway, and what interrupting it would do — so "how it's doing" is separated from "how much is being sent."
  • Ask the withdrawal counterfactual. Estimate the unit's own net trajectory with the rescue flow removed, using natural experiments (a week the source was down), staged reductions, or comparison to unsupported peers.
  • Classify: recovery vs. rescue. Recovery means the unit clears its viability threshold on its own rates; rescue means it only clears it with imports. The distinction, not the current level, is the verdict.
  • Watch for a genuine role switch. Track whether a rescued sink is actually trending toward self-sufficiency (or a source toward decline), so a real transition is caught and a false one is not declared.

Tuning parameters

  • Audit cadence — how often the test is run. Frequent audits catch a slipping unit early but cost effort and can over-react to a single bad period; rare ones let masked failure run.
  • Counterfactual strength — from a cheap peer comparison to an actual staged withdrawal. Stronger tests give a more credible verdict but risk harming the unit to learn about it.
  • Recovery threshold — how much unaided margin counts as "recovered" rather than "borderline." Set it tight and you keep support on units that could stand alone; set it loose and you cut units that then relapse.
  • Interruption-risk weighting — how heavily the profile weights the reliability of the rescue flow, not just its size, since a large but fragile subsidy is a different risk than a small dependable one.

When it helps, and when it misleads

Its strength is puncturing the most expensive illusion in the archetype: it separates units that have genuinely recovered from units whose health is on loan, so support is withdrawn from the first and maintained (or rethought) for the second — and it catches the moment a rescued unit truly does switch to standing on its own.

Its failure modes turn on the counterfactual. The withdrawal test can be too weak to bite — a gentle probe that a rescued unit rides out on momentum, yielding a false "recovered." Or it can be run backwards: an auditor who wants to cut support designs a harsh probe, or one who wants to keep it designs a toothless one, so the verdict is decided before the test. And declaring recovery off a single good period ignores that rescue effects are often seasonal. The discipline is to define the recovery threshold and the counterfactual before running the audit, prefer natural experiments over wishful reasoning, and require the unaided trajectory to hold across more than one cycle before calling a rescued sink recovered.

How it implements the components

Rescue-Effect Audit realizes the diagnosis-of-dependence slice of the archetype — the components that establish whether persistence is borrowed and whether a role is changing:

  • sink_dependency_profile — its core product: how much support the unit takes, how reliable that support is, and what its removal would do — the anatomy of the dependence.
  • role_switch_monitor — the ongoing check for whether a rescued sink is genuinely crossing toward self-sufficiency (or a source drifting toward deficit), so real transitions are recognized and false ones are not.

It diagnoses dependence but does not measure the raw flows it reasons about — those come from the Dispersal or Transfer Tracer; it does not forecast the network-level consequences of the dependence — that is the Metapopulation Model; and it does not decide what to do once a unit is judged a rescued sink — that is the Restoration Priority Matrix.

Notes

The audit produces a verdict, not a label change. Distinguishing rescue from recovery is its job; formally re-classifying a unit's status in the system's records — and everything that follows administratively — belongs to a Role Reclassification Review, which consumes this audit's finding. Keeping the diagnosis separate from the re-labeling is what lets an auditor deliver an uncomfortable "still dependent" verdict without owning the political act of demoting a unit on the books.

References

[1] The rescue effect — immigration or inflow from a source lowering a local unit's risk of falling below viability — is a named result in metapopulation ecology, and it is precisely why occupancy or surface health can hide a demographic sink: the unit persists on borrowed input, not on its own balance. ↩