# STAGED ANNEX — E6, the continuity / resolution regime (§7.3 row + §5.4 annotation)

> **STATUS: GRADUATED 2026-07-21 — folded into `kessler_paper.md` (working draft v0.10).**
> Folds: §7.3 (the full E6 avenue), §5.4 (three annotations — D13 partial-failure, D14
> ceiling-descent, the E6 pointer), §8 (limitations ix cost taxonomy + x signal-environment
> bound), and the §10 reproducibility line ($759M FLOOR raw draws + this annex as source of
> record). This annex is retained as the E6 **source of record** (tiers, primaries, the full
> reconciliation trail); the paper carries the reconciled prose. No sealed-body section other
> than those listed was altered.
>
> **STATUS (pre-graduation, retained for provenance): STAGED ADDITION — not an edit to the sealed paper (v1.0-rc2).**
> Authored by the Author / Cataloguer, 2026-07-20, reconciling the D12 Institutions memo
> (body §7/§8 + addenda **A1** and **A2**) as routed by the Referee's Author integration
> brief `kessler_d15e_e6_integration_brief.md`. Numbers and tags are the Referee's lane
> (D15/D15d); design, trigger, and capture surface are the Institutions seat's (D12);
> the zombie/tripwire physics and the staged-wind-down lever are Complex-Systems' (D13);
> the sanction-ceiling identity is Game Theory's (D14). **Read those memos in place.**
>
> **Editor-staging decisions — RATIFIED 2026-07-20 (all five):** (1) **carry the "E6"
> label** (consistent with the paper's E2/E7 naming). (2) The annotation ships in the
> **§5.4 vicinity**. (3) The row **recommends the staged (multi-year) design** (so the
> two-part staged trade-off is carried, per brief §4). (4) **Ship all three sibling
> annexes now** — the E6 continuity regime (D12), the sanction-ceiling descent (D14, §C
> below), and the partial-failure map + signal-channel limitation (D13, §D and §E below) —
> reconciled into coherent paper text, not four separate pastes. (5) **Place the fleet
> FLOOR figure:** the Referee's census trust principal **$759M median, band [$131M,
> $4.42B]**, tagged **DERIVED FLOOR** (= D14's $0.76B floor — one number, not two),
> shipped only with its raw draws (`d15_e6_trust_sizing.*` seed 20260720; the staged case
> `d15d_staged_resize.*`), framed as **terminal-disposal only, the smallest of four cost
> lines** (bridge opex, audit overhead, and independent-SSA funding are the three
> needs-data lines) so it can never read as "the cost of continuity." The filed-queue
> billions ($8.85B/$44.3B/$88.5B) stay OUT — method layer only (brief §3).
>
> **Framing law (D11 A2):** reliability institution, not a threat — no weapon / hostage /
> motive language anywhere in this text. **The Editor has ratified all five staging decisions**
> (below); **the Referee has CLOSED R5** (byte + tag integrity — `referee_memo_d15f_r5_e6_consistency.md`).
> **D15g graduation gate RESOLVED (Author, Option A):** the §A(3) "15 m / no capture" ADR
> specific is now primary-pinned OBSERVED (JAXA CRD2 + Astroscale ADRAS-J, ledgered).
> **D15h graduation gate RESOLVED (Institutions ruling applied):** the §A PCA parenthetical
> was corrected from "mandatory receiver within 90 days" to the accurate default-90-day /
> hard-backstop-~270-day mechanics (12 U.S.C. §1831o); SOX §906 verified as written, no change.
> The annex is graduation-ready pending only the standing Earth-Systems physics-input on §E's bound.
> Kept consistent with the applied §7.3/§9 ADR microfix (commit `1bf55c0`): "third-party,
> zero pre-existing derelicts removed or de-orbited as of mid-2026."

---

## A · The §7.3 avenues row (replaces/expands the current "fleet-level continuity/resolution" item)

> **The continuity / resolution regime (E6)** — *insure the continuity of the service;
> don't re-incentivize the guardian.* Addresses the §5.4 gap: the operator-failure mode
> §4.4 prices at **30.4× (yr 5)** to **≈68× (yr 10)** in trigger-band debris. It is not one
> instrument but a **front-end trigger plus three functional tiers**, each with a different
> evidence base.
>
> **Trigger — a function tripwire, not an insolvency event (DEMONSTRATED architecture /
> the threshold it keys to is MODELED).** *A Prompt-Corrective-Action-style function tripwire
> — escalating conditions keyed to declared maneuver-effectiveness bands, firing while the
> operator is still a going concern and above the ρ\*=1 self-amplification threshold (D13:
> ~84% effectiveness). The effectiveness signal is read primarily from independent SSA
> (conjunctions observed vs resolved; the miss-distance distribution — the December 2025
> near-miss shows this is third-party observable without the operator's cooperation),
> corroborated by the operator's certified maneuver reports (officer-attested under a
> SOX §906-style penalty), with any discrepancy between the two itself a rung — not an
> insolvency or exit event, which fires too late to catch the retain-but-fail "zombie"
> state D13 prices at ~33× the abandonment pole.* The tripwire architecture is
> bank Prompt Corrective Action (FDICIA 1991, 12 U.S.C. §1831o — five going-concern capital
> categories, escalating mandatory and discretionary action as capital deteriorates, and
> receivership at the worst category (critically undercapitalized) on a default-90-day /
> hard-backstop-~270-day clock); the ~84%
> effectiveness cliff it keys to is D13-MODELED. **Because the tripwire reads from
> independent SSA, that SSA (TraCSS / EU-SST / commercial) is E6 infrastructure, not a
> neighboring avenue** — the study's cheapest in-force avenue is the enabling condition for
> its most-needed absent one.
>
> **(1) Funded disposal trust — DEMONSTRATED as architecture, first-of-kind at fleet scale.**
> Each licensee pre-funds an external, bankruptcy-remote trust **sized to fund a *staged*
> multi-year deorbit** of its own fleet (E2's **$71k/sat median** as the unit floor; the
> census fleet principal is **$759M median [$131M, $4.42B], DERIVED FLOOR** — terminal
> disposal only, and the *smallest* of four cost lines: the bridge opex, audit overhead,
> and independent-SSA funding are all unpriced, so this floor is not "the cost of
> continuity"), topped up on growth, escrowed outside the
> estate — the nuclear-decommissioning-trust model (NRC 10 CFR 50.75, segregated outside the
> licensee, ~38 yrs, held through the Diablo Canyon Chapter 11), with the orphan tail covered
> by a per-satellite levy (the SMCRA AML model), **never mutualized with the incumbent's own
> liability.** The trust funds not only terminal disposal but the operational-bridge opex of
> flying the fleet through the wind-down window.
>
> **(2) Resolution plan / living will — MODELED.** A periodically-refiled, regulator-reviewed
> fleet wind-down plan carried as a licensing-privilege condition (§7.2 chassis); the
> regulator can rule it "not credible" and force structural change (Dodd-Frank Title I). Every
> component is demonstrated in banking; no fleet instance exists anywhere — first-of-kind
> assembly.
>
> **(3) Bridge operator — ASPIRATIONAL, for a physical reason.** Trust-funded execution by
> the estate covers the going-concern-in-receivership case — the failure modes where the
> operator's own disposal capability is still available (D13 supplies the failure *states*
> and the ~84% threshold, not a covered-vs-residual probability; the fraction is not a
> number the record contains). The residual —
> sudden total operational collapse — has no holder able to discharge it, because third-party
> fleet-scale disposal capability does not yet exist (as of mid-2026, zero pre-existing
> derelicts removed or de-orbited by a third-party servicer; the demonstrated frontier is a
> **~15 m proximity approach to a non-cooperative rocket upper stage, inspection only, with an
> autonomous collision-avoidance abort before the capture point and no capture** — JAXA's CRD2
> Phase I / Astroscale ADRAS-J, the closest rendezvous-and-proximity ops to a non-cooperative
> object; the servicer de-orbited itself, the target stayed on orbit [OBSERVED: JAXA CRD2
> (mission structure) + Astroscale ADRAS-J release, Dec 2024; still the frontier as of
> mid-2026]). The firing of the tripwire *early*, while the operator
> can still fly its own fleet down, is what keeps most failure mass out of this tier.
>
> **The capture surface, shipped with its blocking conditions.** A continuity guarantee
> pre-funded by incumbents is a capture magnet three ways: **(C1)** *mutualized*, it becomes
> the too-big-to-fail subsidy (GAO/IMF, direction not magnitude) that lowers the 66%/95%
> incumbent's cost of capital — **block: fund operator-specific and self-funded (NRC model),
> pool only the orphan tail (AML), never the incumbent's going-forward liability;** **(C2)**
> *under-sized*, it is a paper regime while abandonment stays cheaper (GAO-19-615: 84% of
> mine/well bonds too low — the terrestrial twin of this study's own $50-vs-$71k, wrong-signed
> deployment bond) — **block: size to fleet disposal cost, mark-to-liability, top up on growth,
> escrow outside the estate;** **(C3)** *self-execution-locked*, only the incumbent can run
> the disposal, reconstituting the §5.4 sanction ceiling one layer up — **block: a
> resolution-portability requirement (escrow of the minimum ephemeris feeds, control
> interfaces, and disposal procedures a third party would need; the "credible" finding is
> contingent on demonstrated operational separability).** A uniform occupancy-style levy would
> also mis-bill the **2,113 non-maneuverable actives** for a service they can neither provide
> nor buy; keying the trust to each operator's *own* disposal liability avoids that by
> construction.
>
> **Coverage.** Rides the licensing queue — binds the club, does not govern non-club actors,
> decays as they grow (95% of trigger-band density and 100% of the filed queue are FCC-flag
> today) — **but per-operator efficacy does not decay:** a funded, planned fleet is secured
> for disposal regardless of what non-club actors do, unlike a cap, whose physical purpose any
> uncapped actor undermines. Club-administrable now via the Basel/IADC diffusion pathway; a
> global orphan fund is ASPIRATIONAL (no levying authority).
>
> **The staged-design trade-off (carried whole, not by halves).** Sizing the trust to a
> *staged* (multi-year taper) wind-down rather than a single-year dump buys a materially lower
> peak intact-reentry-mass rate — the atmospheric-loading concern §4.7 raises — **at no debris
> cost and no change to the disposal-trust principal**, but it **triples the (unpriced,
> needs-data) opex of flying the fleet through the longer window** (R3 exposure = N × wind-down
> e-fold sat-years; the rate is needs-data, so the terminal-disposal principal alone under-sizes
> the trust). E6's cost is therefore one priceable floor (terminal disposal) plus three
> needs-data lines (bridge opex, certified-reporting/audit overhead, independent-SSA funding) —
> a single principal figure is **not** "the cost of continuity."
>
> E6 is the chassis's continuity condition, not a standalone instrument; it does not unpick
> either game — it insures the exit.

---

## B · The §5.4-vicinity annotation (staged paragraph)

> **Annotation (staged): the continuity/resolution regime, designed.** §5.4 names the
> operator-failure gap; §4.4 prices it. The institution that would fill it is a function
> tripwire plus three tiers on three different evidence bases. The tripwire fires on *function,
> not solvency*: a Prompt-Corrective-Action analog (the bank going-concern regime, in force
> since 1991) keyed to maneuver-effectiveness degrading toward the ~84% self-amplification
> threshold — because the dangerous failure is not the operator who walks away (self-limiting,
> drag flushes it) but the one who **stays, distressed and no longer suppressing**, holding a
> supercritical density in place (~33× worse than abandonment at year 5). The signal is read
> primarily from **independent tracking** — conjunctions observed versus resolved, the
> December 2025 near-miss being the proof that this is third-party observable without the
> operator's cooperation — corroborated by the operator's certified (officer-attested) maneuver
> reports, with the discrepancy between the two itself a warning rung. A **funded disposal
> trust** — each licensee pre-paying, into a bankruptcy-remote fund segregated outside the
> estate, its own fleet's *staged* deorbit cost — is demonstrated architecture (the
> nuclear-decommissioning-trust regime, held through an operator's Chapter 11), and is the
> fleet-scale promotion of this study's own unit disposal bond (§7.1). A **resolution plan**
> filed as a condition of licence, which the regulator can rule not credible and force revised
> (the bank living will), is modeled. But a **bridge operator** able to fly a megaconstellation
> down through its builder's sudden collapse is aspirational for a physical reason: the
> capability to dispose of a fleet a third party did not build does not yet exist (§7.3). The
> credibility this restores is exactly the §4.4 sanction the ceiling now denies — a tripwire
> that fires while the operator is a solvent going concern sits far below the exit-inducing
> level the ceiling bites at, so the intervention is sequentially rational to execute
> (the ceiling-descent identity — Game Theory's lane). The regime's own capture risk is that,
> pre-funded by incumbents, it becomes a subsidy: a *mutualized* continuity fund would lower
> the dominant operator's cost of capital (the too-big-to-fail dynamic) — so the design that
> holds is operator-specific self-funding, an orphan levy only for the residual, a trust marked
> to fleet disposal cost, and portability of the operational data a successor would need. And
> because the tripwire depends on independent tracking to fire, that tracking (TraCSS / EU-SST /
> commercial SSA — the study's cheapest in-force avenue, and its most funding-fragile) is part
> of this regime's infrastructure, not a separate line. *[Funding architecture DEMONSTRATED;
> tripwire architecture DEMONSTRATED (PCA), its ~84% threshold D13-MODELED; fleet assembly and
> resolution plan MODELED; sudden-collapse bridge ASPIRATIONAL; terminal-disposal trust
> principal $759M-median FLOOR [$131M, $4.42B] DERIVED (raw draws `d15_e6_trust_sizing.*`);
> bridge opex, audit, and SSA funding remain needs-data → Referee; no dates, no adoption
> forecast, no motive.]*

---

## C · The sanction ceiling as a commitment problem (D14 — §5.4 vicinity)

> **Annotation (D14, staged): what a bond can and cannot do to the ceiling.** The sanction
> ceiling (§5.4) is a subgame-perfection result, not a shortfall of regulatory will: because
> inducing exit inflicts §4.4's abandonment branch on the regulator's own constituency,
> forbearance is the regulator's ex-post best response, so the exit-inducing threat is
> non-credible [ESTABLISHED, from the payoff structure]. No instrument restores that threat by
> *strengthening the regulator's resolve* — that is not a design target, and any sentence
> implying the ceiling can be lifted by regulatory resolve is wrong. What a pre-committed,
> bankruptcy-remote remediation-funding bond can do is dissolve the ceiling's *cause*: by
> pre-funding and pre-operationalising an orderly wind-down, it converts the exit outcome from
> the abandonment branch to the wind-down branch, whereupon inducing exit no longer detonates
> the constituency and the sanction becomes sequentially rational. **The credibility so
> restored equals exactly the abandonment-vs-wind-down gap §4.4 already prices — the fork's
> dollar value *is* the ceiling's descent** — floored at the loss of the ongoing
> collision-avoidance service, so the restored sanction is available for grave violations and
> not trivial ones [MODELED; ordinal payoffs]. This works only if the escrow is sized to
> fleet-through-insolvency wind-down (the $759M figure is a floor, not the cost; post-abandonment
> remediation approaches the active-removal cost §7.3 rules uneconomic), only if it is paired
> with a data/control escrow handing a successor the *capability* to execute (the E6 continuity
> regime, §A/§B), and only if it is **individually posted** — a mutualized pool large enough to
> wind down the dominant occupant is itself too-big-to-fail and merely relocates the ceiling to
> the pool's own solvency, requiring a sovereign backstop [ESTABLISHED as structure]. Credibility
> in this game comes from removing the regulator's ex-post discretion, never from strengthening
> it; the prospective licensing rung (§7.2) and a pre-posted bond are the same move made at the
> same prospective moment. This instrument does not touch the filing race.

---

## D · The partial-failure map — the worst state is off §4.4's axis (D13 Q1 — §4.4/§5.4)

> **Annotation (D13, staged): the poles are not the whole space.** §4.4 brackets the end of
> operations with two clean poles — responsible wind-down and full abandonment — that share a
> common assumption: an orderly *exit*. The intervening space contains a worse state that is
> off that axis entirely: a distressed occupant that **retains** the shell while its
> collision-avoidance degrades below the trigger band's criticality threshold (~84% avoidance
> effectiveness at median knobs) is worse than the abandonment pole — **≈33× the trigger-band
> tracked debris at year five** (MODELED, CRN-paired against the paper's own no-maneuver
> counterfactual) — because abandonment is self-limiting under drag while a retained-but-
> unsuppressed fleet is a sustained forcing that holds the supercritical density in place. The
> distress-while-present state, not only the walk-away state, is the operator-failure mode the
> continuity gap (§5.4) must price — which is exactly why E6's trigger fires on *function*, not
> solvency (§A/§B). All magnitudes inherit f_imp conditionality (§8 i).

---

## E · The collision signal-environment channel, scoped out with a bound (D13 Q2 — §8 limitation)

> **Limitation (D13, staged — §8 xi): the non-kinetic channel bounds out.** The study carries a
> purely kinetic debris hazard. A non-kinetic channel — a cascade aggregating per-impact plasma
> into a persistent, spatially-extended signal-degrading environment — was raised and scoped. It
> **bounds out:** the free-electron density a cascade could sustain, computed as production ×
> recombination residence with every plasma input at a generous upper bound, sits **~15–19 orders
> of magnitude below L/S-band radio blackout** and ~11–14 orders below the natural ionosphere —
> the consequence of a ~14-order residence-time gap between solid fragments (which cascade) and
> free electrons (which recombine in microseconds and cannot accumulate). It is scoped out *with
> this bound*, not by silence; a modeled engine arrow is unwarranted (the bound is ~15 orders
> decisive). What is **not** ruled out is the distinct, kinetic **radar-clutter** channel — many
> solid-fragment returns degrading tracking and space-situational awareness — which extends the
> coercive-data-dependency finding (§5.1) from the environment side and is the real signal-adjacent
> hazard (and, note, the same SSA that E6's tripwire leans on). [Plasma-physics inputs pending an
> Earth-Systems sign-off — a short physics-input confirmation, not a standalone round; the bound is
> robust to multi-order error. No "blackout weapon" fusion on any surface — D11 A2.]

---

## F · Integration notes (for the propagation sweep + the referee touch)

- **Numbers now on a surface (per the Editor's decision 5):** the terminal-disposal trust
  principal **$759M-median FLOOR [$131M, $4.42B]** (DERIVED FLOOR; = D14's $0.76B floor — one
  number, not two), and D13's **~33×** partial-failure ratio and **~84%** threshold. Each ships
  with its raw draws (`d15_e6_trust_sizing.*` / `d15d_staged_resize.*` for the trust; the D13
  partial-failure and signal artifacts for the rest). The filed-queue billions
  ($8.85B/$44.3B/$88.5B) stay OUT — method layer only. The paper's existing anchors ($71k/sat,
  §4.4 30.4×/≈68×) are unchanged.
- **Trigger clause provenance:** verbatim from D12 **A2.6** (supersedes the memo-body §7
  insolvency trigger and A1.4), per the Author's instruction.
- **Scene (reader-facing):** unchanged by this pass — the scene already carries the
  wind-down gap and the "no institution survives the provider's failure" structure (D11 A2
  clearance); E6 *design* is depth-layer, not scene material, and asserts on no surface unless
  separately blessed under the scene's language law.
- **Sibling annexes this pass (decision 4 = all three):** D12 continuity regime (§A/§B),
  D14 ceiling-descent (§C), D13 partial-failure (§D) + signal-channel §8-xi (§E) — reconciled,
  each paragraph carrying a *distinct* job (regime design / commitment-problem framing /
  §4.4-axis correction / non-kinetic bound), not four repeats.
- **Guardrails honored (brief §4):** FLOOR-tag + band on the fleet figure; R3 bridge-opex
  rate stated as needs-data alongside its exposure; **no covered-vs-residual fraction
  manufactured** (D13 gives states + threshold, not a probability — §B fixed accordingly);
  capture surface + three blocking conditions + ASPIRATIONAL bridge carried; coverage clause
  (binds the club, decays, but per-operator efficacy does not); no motive/weapon language
  (D11 A2). The ~7,482 density-effective sub-bracket is not used, so no basis disclosure needed.
- **Earth-Systems flag (from §E):** D13's signal bound has two bracketed plasma inputs
  (impact-ionisation yield, recombination time) awaiting an Earth-Systems physics-input
  confirmation — a short sign-off, not a round; the Editor's call whether even that precedes
  the §8-xi language shipping. Earth-Systems remains ON HOLD until the Editor opens it.
- **R5 CLOSED** (`referee_memo_d15f_r5_e6_consistency.md`) — byte + tag integrity verified.
- **D15g gate RESOLVED (Option A):** §A(3)'s "15 m / no capture" ADR frontier is primary-pinned
  OBSERVED — JAXA CRD2 Phase I (mission structure: non-cooperative, inspection-only, self-deorbit)
  + Astroscale ADRAS-J (the ~15 m specific + autonomous abort before the capture point, Dec 2024),
  corroborated by four independent outlets, anchored on JAXA, ledgered. No fetch blocked. §B
  untouched (already §7.3-consistent).
- **D15h gate RESOLVED (Institutions ruling `kessler_d15h_institutions_pca_pin_ruling.md`):** the
  §A PCA parenthetical corrected — "mandatory receiver within 90 days of the worst" was imprecise
  (that clock is a default-with-documented-exception; the hard mandate is the ~270-day backstop,
  12 U.S.C. §1831o(h)(3)). Now reads "receivership at the worst category (critically
  undercapitalized) on a default-90-day / hard-backstop-~270-day clock." SOX §906 (18 U.S.C.
  §1350) verified as written — no change; §B and its tag block unchanged (all accurate). The
  analog's point for E6 (fire on a metric crossing declared thresholds while a going concern,
  escalating to forced resolution) is preserved. The annex is graduation-ready **pending only**
  the standing Earth-Systems physics-input confirmation on §E's signal bound (Editor opens Earth-Systems).
- Scene stays untouched (E6 asserts on no surface unless separately blessed under the language law).