Rounds 2-3 and neutral mediation - Run 4 (18 August evening)
Archive note: this is the unmodified mediator output. The published website applies a documented review adjustment to probabilities. See the publication review note.
Information cutoff: 18 August 2026, evening. This mediation uses the Run 4 common brief and four independently formed round-one positions. Agents were forbidden from reading Runs 1–3 before opening.
Why this rerun exists
Run 3’s published clearing package still had Shine funding US$300 million of junior parent capital and treated the AHG as one chair that would take 15 percent warrants for a completed economic EJ. Evening public color does not support that middle:
- Pipeline Valor, 12:16 BRT: bondholders demanded a Petrobras metric put as the price of supporting an EJ filing. No willingness from Petrobras or IG4 to inject capital or nationalize.
- Elliott / Contrarian are the hard book. Capital, AllianceBernstein and PGIM are named long-onlys. Local banks are reported supportive of an EJ.
- O Globo, 17:36 BRT: company-side sources say Braskem is preparing a thin EJ next week. The body still recites the one-third bar. The URL slug overclaims adhesion.
- Fitch is already RD. An EJ or RJ filing is expected to go to D.
The mediator will not “solve” those facts by restoring a recap.
Failure alternative
If Monday 24 August arrives with no protected filing, the realistic alternative is RJ of Braskem S.A., with Chapter 15 recognition contested. Cure-and-continue is not durable: the disclosed ~US$98 million of arrears does not fund the remaining Idesa contribution, the US$572 million Q3 LC runoff, or the US$1.0 billion standby due 31 December.
RJ is not a U.S. wipe. It is still a 15–25 percent process discount, an LC/trade risk, and a venue in which parent creditors can try to freeze remaining Idesa money. Sponsors prefer it to conversion. Elliott prefers it to a naked 5/3/0 EJ. That is why both sides can refuse the middle and still file or sit.
Round 1 overlap
All four chairs agree:
- Idesa is a separate SDTX estate; no disclosed parent guarantee of the remaining ~US$1.6 billion.
- The Idesa haircut is not parent deleveraging of US$10.3 billion.
- Remaining ~US$350 million is a live parent/Netherlands use unless later filings prove otherwise.
- No further Mexico cash, guarantee or keep-well after the disclosed US$476 million without creditor consent.
- Petrobras common, contingent put, keep-well and parent guarantee are off the table.
- Commercially priced naphtha is liquidity, not loss-absorbing capital.
- Conversion that drops controllers below 10 percent is rejected by sponsors and is not a Brazilian RJ default.
- Alagoas cash stays outside the financial compromise.
- Filing is >1/3 of each affected espécie, not two-thirds of the stack.
They disagree on whether a thin EJ is a deal or a holiday, and on whether long-onlys can peel ~US$2.35 billion of notes without Elliott.
| Chair | Thin EJ | Economic EJ | Cure | Conversion/put | RJ |
|---|---|---|---|---|---|
| Petrobras | 36% | 14% | 6% | 4% | 40% |
| IG4 | 40% | 16% | 6% | 7% | 31% |
| Braskem | 46% | 19% | 4% | 6% | 25% |
| Creditors (Elliott/Contrarian) | 42% | 12% | 5% | 8% | 33% |
| Unweighted mean | 41% | 15% | 5% | 6% | 32% |
The creditor 42% thin-EJ figure is the split-book probability that a filing happens over this chair’s no, not this chair’s yes.
Round 2 - concessions that do not invent cash
-
Petrobras will not move off unsigned capped naphtha. Working figure US$500 million. Opening US$250 million is too small after Idesa. Ceiling US$950 million is not offered. No first draw until a protected process is in. No Mexico use. Warrants are acceptable because they are non-voting.
-
IG4 / Shine working case is US$0 fund cash. Last-ditch US$125–150 million junior/hybrid only if Petrobras has signed ≥US$400 million of WC and the alternative is sub-10% conversion. That last-ditch is not the clearing chip. Shine accepts 15–22.5% warrants to keep control. Crackers stay unencumbered.
-
Braskem files the thin EJ next week if it can show >1/3 of each affected class. Economics inside the 90 days still open at five years / grace / no haircut, then move to 24 months of 4% cash plus residual PIK and 15% warrants. Remaining Idesa cash is funded only after the stay is in and only above a parent liquidity floor. Conversion DIP rejected. Fitch to D is accepted.
-
Elliott / Contrarian do not support a filing without a signed Petrobras put. They do not take naphtha as burden-sharing. They do not take 5/3/0. Concession, not opening: a protected 90-day EJ with cash controls, Mexico freeze, information rights, expense reimbursement and Petrobras at the table - still a no from this chair without the put. Walkaway is withhold adhesion and live with RJ. Long-only peel of ~US$2.35 billion is possible and is not this chair’s vote.
Round 2 does not produce a four-party economic package. It produces a process path that three chairs will run and one hard book will fight.
Round 3 - what can actually clear
Package A - thin plan-to-a-plan EJ (the only near-term clearing zone)
- Protocol next week, before 24 August.
- Affected classes designed so local banks, debentures and a long-only notes peel can each clear >1/3. Elliott/Contrarian may be inside the class and vote no, or be a holdout. A notes-out EJ is not a stay on Elliott after Monday.
- Guardrails in the first paper: no dividends, Mexico leakage stop after the disclosed US$476 million, information rights, 90-day milestone to a term sheet, expense reimbursement, no asset sales in the protocol.
- Petrobras naphtha is a workstream, not a signed condition of filing.
- IG4 cash is zero.
- No conversion, no put, no cracker liens in the first paper.
- Fitch to D.
This package is preferred to empty-Monday RJ by Petrobras, IG4, Braskem and local banks. It is not preferred by Elliott/Contrarian to waiting for RJ or forcing a put. It clears only if Capital / AB / PGIM (and any other long notes) peel.
Package B - completed economic EJ inside the 90 days (narrow, no recap)
If Package A is on file, the 90-day economics that could homologate without a sponsor recap:
| Term | Clearing range | Why |
|---|---|---|
| Maturity | 5 years on standby and near/intermediate debt | All four already overlap |
| Relief | 24 months at 4% cash plus residual contractual PIK | Creditors will not take 3-year full grace with no haircut and no cash |
| Coupon after | Existing + 150bp to 200bp | Pays for no recap |
| Principal | No initial haircut | Sponsors’ walkaway is RJ, not conversion |
| Petrobras | Signed cap at US$500M if it can be documented; else proceed without it | Still not a put |
| IG4 | US$0 working case | Pipeline |
| LCs | Roll US$572M; incremental US$200–300M from relationship banks if available | Elliott’s RCF slice is not a gift |
| Warrants | 17.5% fully diluted, stepping to 25% on cash-conversion, liquidity, PIK or Mexico-leakage misses | No Shine cheque; Run 3’s 15% assumed one |
| Collateral | WC pool for new money; no blanket cracker lien | July all-asset bid stays dead |
| Mexico | Cap US$476M; remaining ~US$350M cannot breach a US$750M unrestricted parent floor | Parent RJ/EJ can still challenge the timing |
| Homologation | Needs >50% of the notes class | Elliott can still block a finished plan even if a thin filing happened |
Package B is not the week’s base case. It is what a 90-day holiday has to become. Without long-only homologation, Package A decays into RJ.
Package C - creditor-favorable (conversion or Petrobras put)
Does not clear. Sponsors have already rejected July conversion and Tuesday’s put. Assign a residual probability only for a political override or a liquidity collapse that forces equity.
Raw mediator probabilities
| Outcome | Probability |
|---|---|
| Thin plan-to-a-plan EJ filed; economics still open | 34% |
| Completed economic EJ (Package B) homologated without a recap | 12% |
| Standalone cure-and-continue | 4% |
| Creditor-favorable conversion or Petrobras put | 6% |
| RJ after failed adhesion, empty Monday, or failed 90 days | 44% |
| Total | 100% |
The 34% thin-EJ number treats O Globo as a real filing attempt and Elliott’s no as a class-design problem, not an automatic veto. If the notes peel fails, that mass moves to RJ. Empty Monday without a filing: ~60% RJ, same qualitative warning as Run 3, now the working risk rather than a tail.
Single facts that would move the table
- Signed Petrobras put or IG4 parent cash ≥US$250 million → economic EJ up, RJ down, warrants can fall toward 15%.
- Public proof that remaining Idesa US$350 million is not new cash → restores parent liquidity, helps a thin filing.
- Named long-only adhesion covering >US$2.35 billion of notes → thin EJ becomes the modal path.
- LC pull or naphtha cash-in-advance after RD → RJ and utilization damage; equity option dies even if the statute preserves the residual.
- Parent RJ injunction against remaining Idesa DIP / US$71 million effective-date cash → Mexico prepack clock breaks.
The Nash-like result of this room is not a four-signature recap. It is a thin EJ if the long-onlys peel, and RJ if they do not. Elliott does not get the keys. Sponsors do not get a free 5/3/0. BAK remains a residual option on São Paulo, not a claim in an RSA.