When a ‘Win’ under Section 423 Still Leaves You with Nothing: Credit Suisse v SoftBank

by | Dec 9, 2025 | Upsettable Transactions

Every so often a case comes along that makes insolvency practitioners, funders and lawyers all sit up at once.

Credit Suisse v SoftBank is one of those cases.

On the face of it, Credit Suisse and its noteholders did something you’d normally expect to work. They used section 423 of the Insolvency Act 1986 – the ‘transactions defrauding creditors’ provision – to attack a deal that had stripped out the only real asset behind a US$440m securitisation.

They convinced the court that:

  • There was a transaction at an undervalue.
  • The debtor company did act with the prohibited purpose of prejudicing creditors.
  • They themselves were ‘victims’ of that transaction.

And yet, after a 149-page judgment, the claim was still dismissed and they walked away with nothing.

Let’s unpack what happened, in plain English, and then look at what it means for directors, funders and anyone playing with complex structures.

The set-up: Greensill, Katerra and an insolvency-remote SPV

At the heart of this case were three big players:

  • Katerra: a US-based construction group, already wobbling financially.
  • Greensill: a supply chain finance house.
  • SoftBank: a heavyweight investor, with money in both Katerra and Greensill.

The Credit Suisse fund had invested around US$440m in ‘Fairymead Notes’. Those notes sat in a securitisation structure that looked roughly like this:

  • Katerra sold receivables to an SPV called GL under a Receivables Purchase Agreement (‘the RPA’).
  • GL funded that by issuing participations up the chain, ultimately supporting the Fairymead Notes held by the Credit Suisse fund.
  • GL was an insolvency-remote shell – no staff, no other assets – its sole job was to hold that Katerra receivable and pass the cashflows through.

If you’re a secured investor, that’s the asset you’re relying on. The receivable and its security are your economic engine.

The rescue that killed the security

By late 2020, Katerra badly needed a recapitalisation. SoftBank and Greensill worked up a rescue plan that had a few moving parts:

  • The SoftBank side would provide US$440m into the Greensill group through a convertible loan note.
  • In parallel, there would be an out-of-court restructuring of Katerra.

Two agreements on 30 December 2020 – the Contribution and Exchange Agreement (CEA) and the Share Transfer Agreement (TA) – did the real damage:

  1. Under the CEA, GL released Katerra from the US$440m RPA debt and related security, in exchange for Katerra shares.
  2. Under the TA, those Katerra shares were immediately transferred to a SoftBank vehicle.

So GL went from holding a substantial secured receivable against Katerra… to holding nothing. The shares it notionally got in return went straight through to SoftBank. The securitisation chain was effectively broken.

A few months later:

  • Greensill collapsed.
  • Katerra entered US bankruptcy.
  • The Fairymead Notes defaulted.

The Credit Suisse fund and the note trustee then sued various SoftBank entities under section 423.

Section 423: what did the claimants argue?

Section 423 is a familiar tool in insolvency practitioners’ toolkit:

  • It’s about transactions at an undervalue entered into for the purpose of putting assets beyond creditors’ reach or otherwise prejudicing them.
  • It can be used even where the debtor isn’t yet in formal insolvency.
  • And it allows the court to make orders to restore the position or protect victims.

Here, the claimants said:

  • The ‘transaction’ for s.423 purposes was the CEA and TA taken together.
  • GL clearly entered into those agreements at an undervalue: it gave up a valuable secured receivable and ended up with nothing.
  • GL (through Lex Greensill) did so for the purpose of prejudicing the Fairymead noteholders, by stripping out the only asset backing the notes and shifting them into the unsecured mess of the Greensill group.
  • The SoftBank entities were beneficiaries of that deal – they ended up with shares and a de-levered Katerra – and should therefore be ordered to pay.

SoftBank, for their part, said:

  • You can’t cherry-pick the CEA and TA – they form part of a much larger ‘Greensill/Katerra transaction’ including the US$440m injection into Greensill that was supposed to fund redemption of the Fairymead Notes.
  • They thought that money would be used to repay the notes. If Greensill misused it, that wasn’t down to them.
  • GL didn’t have the necessary purpose, and in any event they didn’t end up better off because Katerra’s equity later became worthless.

What the judge actually decided

  1. What counts as the ‘transaction’?

The judge agreed with the claimants that the relevant transaction for s.423 is limited to the CEA and TA – the Impugned Transactions – because:

  • Those were the only agreements to which GL was a party.
  • GL didn’t ‘enter into’ the rest of the wider rescue package.
  • The fact that everything was part of a broader commercial deal doesn’t automatically turn it into a single statutory ‘transaction’ for s.423.

This is important: the court refused to let the wider commercial context dilute the sharp legal focus on what the debtor company itself actually did.

  1. Was there an undervalue?

Yes.

The court accepted that GL’s RPA receivable had real value at the time – roughly US$86m when you factored in expected recoveries and secured collections.

GL then released that receivable and the related security, and the shares it theoretically received were immediately diverted to SoftBank under the TA. GL was left with nothing of substance.

So:

  • The consideration GL gave (release of the Katerra debt and security) had substantial value.
  • The consideration GL received (in practical terms) was nil.

That is a textbook transaction at an undervalue under s.423(1).

  1. Did GL have the forbidden purpose?

This is where it gets more nuanced.

Everyone agreed you look at the state of mind of Lex Greensill, and attribute his purpose to GL. The judge found:

  • Lex Greensill knew exactly what GL was – an SPV whose only function was to hold the Katerra receivables for the benefit of the securitisation investors.
  • He knew that releasing the RPA would ‘kill’ the asset underpinning the Fairymead Notes and leave the investors with an unsecured claim against a struggling group.
  • He knew Credit Suisse would not consent, which is why he didn’t ask and instead concealed what was happening.
  • He was ‘between a rock and a hard place’: he wanted SoftBank to keep supporting Greensill, and that meant doing the Katerra recapitalisation, even at the expense of the securitised investors.

The judge accepted that he didn’t want to hurt the noteholders and would have preferred to pay them in full in due course. But that’s not the test. Section 423 only requires that one of his purposes in causing GL to enter the CEA/TA was to put assets beyond creditors or otherwise prejudice them. It doesn’t have to be the only or dominant purpose.

On the evidence, the judge held that this test was met.

So by this point:

  • Transaction? Yes – CEA and TA.
  • Undervalue? Yes.
  • Prohibited purpose? Yes.
  • Victims? Yes –  the Credit Suisse fund and the trustee.

If you stopped reading there, you’d assume the claimants were home and dry.  But…

Relief: the sting in the tail

But section 423 isn’t a simple ‘if you tick the boxes, defendant pays’ provision. The court has a discretion about what order to make, if any, and that discretion is restorative and protective, not punitive.

The judge made a few key points:

  1. The natural restorative order here would have been to reinstate the RPA debt and security – effectively undo the release so that GL once again had its secured receivable against Katerra.
  2. However, by the time of judgment, Katerra had gone through formal insolvency and the receivable would be worthless. Reinstating it would be a purely paper exercise.
  3. What about ordering SoftBank to pay cash? The judge treated that with caution:
    • SoftBank’s benefit from the transaction was limited to the Katerra shares.
    • At the time, those were worth around US$11.3m (based on a contemporaneous share purchase agreement).
    • After Katerra’s bankruptcy, those shares – and any economic benefit – were wiped out.
    • To require SoftBank to pay hundreds of millions, when it no longer held any benefit from the impugned transaction, would look a lot like punishment rather than restoration.

The judge emphasised that:

  • Previous cases saying ‘it will be exceptional to give no relief under s.423’ were dealing with more straightforward asset transfers.
  • Here, we were in a very unusual position: a debt-release transaction, a later third-party insolvency, and a transferee who no longer had any value to hand back.

On those facts, he held this was indeed an exceptional case where the right course was to grant no relief at all – and the claim was dismissed.

Why does this matter in the real world?

For most directors and creditors, this can feel deeply unsatisfying. You prove the bad behaviour… and still lose on remedies.

But there are some important practical lessons.

  1. Section 423 is powerful, but it isn’t a magic money tree

Section 423:

  • Can reach sophisticated structures, SPVs and securitisations.
  • Can bite on debt releases, not just straightforward asset transfers.
  • Does not require the creditors to have been identified individually at the time.

But its purpose is to restore, as far as practicable, what was lost – not to write a cheque just because someone behaved badly. Where the asset that was stripped away has genuinely gone to zero through later insolvency or market movements, the court may be very reluctant to deem someone else the insurer of that loss.

  1. Directors of SPVs still have real duties (and real exposure)

GL was just an SPV. No employees, no trading, one asset.

Yet:

  • The court pinned the purpose test squarely on Lex Greensill’s knowledge of what the SPV was there to do and who ultimately relied on it.
  • His willingness to proceed without investor consent and to conceal what had been done was central to the finding of a prohibited purpose.

If you sit on the board of a financing SPV, you aren’t simply a passenger doing what the sponsor wants. You’re expected to understand:

  • What the structure is.
  • Who the ultimate economic beneficiaries are.
  • How any restructuring will affect them.

And if you strip out the only asset without consent, you are exposing yourself – and your company – to real risk under provisions like s.423.

  1. For secured investors and funders: know your structure, and your limits

From a creditor perspective, there are some hard truths here:

  • Even when you win on a s.423 analysis, you may not recover if the underlying asset has disappeared in a subsequent insolvency.
  • In complex cross-border structures, your remedy may be practically limited by what’s left in the system, not just by how bad the behaviour was.

That doesn’t mean s.423 isn’t worth pursuing – it absolutely can be – but expectations need to be realistic. It’s a tool to restore actual value, not a general ‘anti-unfairness’ power.

  1. For anyone planning a rescue: be honest about who you’re sacrificing

Commercial rescues are messy. Sometimes, to keep a group alive, someone takes a hit.

What Credit Suisse v SoftBank shows is:

  • You can’t quietly sacrifice a particular group of creditors, especially secured ones, and hope no one notices.
  • Even if the endgame is a good-faith attempt to stabilise the wider group, if one of your conscious purposes is to strip protection from a particular constituency, s.423 is engaged.
  • Concealment – not telling the affected creditors, not seeking consent – will come back to bite you when the court reconstructs what your real purpose was.

If a rescue genuinely requires a particular creditor group to take a hit, it needs to be front-and-centre, transparent and negotiated – not done by stealth via an SPV.

Final thoughts

From an insolvency practitioner’s point of view, Credit Suisse v SoftBank is both reassuring and sobering.

Reassuring, because the court was clearly prepared to:

  • Look through complex securitisation structures.
  • Treat debt releases as transactions at an undervalue.
  • Find the necessary prejudicial purpose where a director knowingly destroys the only asset backing a creditor group.

Sobering, because, even after all of that, the brutal reality of subsequent insolvency and zero residual value meant there was simply nothing left to restore. The law can’t conjure value out of thin air.

For directors, funders and advisors, the message is simple:

  • Understand your structures.
  • Don’t treat SPVs as expendable shells.
  • Don’t assume that clever documentation will hide a prejudicial purpose.
  • And when you’re weighing up ‘who takes the pain’ in a restructuring, remember that transparency and consent are your friends. Stealth rarely ends well.