Debentures and guarantees need a harder recovery test

I'm wary of lending papers that treat a debenture, a PG, and a cross-company guarantee as three independent layers of downside protection. They may all be enforceable, yet each can lead back to the same depleted pool of wealth when the borrower fails and the lender finally needs cash rather than contractual rights. More signatures don't necessarily mean more recovery.

The government's August 2026 insolvency commentary describes a broadly unchanged total in England and Wales, with fewer creditors' voluntary liquidations offset by increases in compulsory liquidations and administrations. For lenders automating origination, that distinction matters: the route into insolvency can affect enforcement timing and how much influence a security holder actually has.

The legal reference point throughout this piece is commercial lending in England and Wales. Security law and enforcement procedures aren't interchangeable across the UK, let alone across borders. A debenture's value starts with the assets and priority it actually captures, rather than the breadth of its wording, the size of the facility, or the fact that it has been registered. Timely registration matters, but it doesn't manufacture ownership or move a junior lender ahead of an existing charge holder.

The distinction between fixed and floating security is especially easy to flatten into a database field, even though the substance of control under the agreed arrangements matters more than the label. If the borrower can freely use charged account proceeds, a lender shouldn't simply assume it has effective fixed security over them. Recoveries under a floating charge are exposed to preferential claims, including certain HMRC debts. Insolvency expenses and the prescribed part reserved for unsecured creditors also reduce what remains.

Even good priority doesn't preserve sale value. Software may depend on licences that can't be transferred, while equipment may belong to an asset financier and stock may be subject to retention of title. A qualifying floating charge can give its holder a route to appoint an administrator, subject to restrictions and competing rights; that's useful control, although it doesn't create another buyer for specialised assets.

A PG adds a claim against a person, not an automatic charge over their home or a promise that apparent equity is available. Mortgage debt and joint ownership affect what can be reached, while other guarantees may compete for the same residual wealth precisely when the business has stopped paying the owner's income. Proper execution and, where appropriate, independent legal advice reduce avoidable disputes about the guarantee's enforceability. Assessing personal assets remains necessary, and that information is often harder to refresh than company bank data.

I can see the argument that a PG encourages earlier engagement when payments slip, because the director has a personal reason to negotiate before options narrow. But I'd want evidence before turning that intuition into a lower default assumption: stronger founders may reject unlimited recourse, while those most willing to sign may have little left to lose. Selection effects make the apparent comfort difficult to price.

Cross-company guarantees deserve particular scepticism when they are described as diversification. A sister company may have independent customers and real surplus assets, but a different company number isn't evidence of independent repayment capacity. Shared cash pooling and dependence on the same owner can transmit distress across the group, while an intercompany receivable may be worthless because its debtor is the very borrower that has failed.

There's also a separate legal question about the guarantor's interests: an upstream or cross-stream guarantee needs scrutiny of corporate benefit and directors' duties, with solvency relevant to a subsequent insolvency challenge. Board minutes help document the reasoning; they don't cure a transaction that is vulnerable under insolvency law. The lender can pursue different obligors, subject to the terms of the documents and applicable law. It can't recover the same debt twice, so recovery models need to recognise overlapping asset pools.

Recovery economics after default

Consider an illustrative £250,000 exposure with a PG capped at £100,000, where enforcement produces £40,000 after all associated costs, received in two years. At an assumed annual discount rate of 15%, that's about £30,200 today, before allowing for uncertainty about whether collection succeeds. A model that counts the guarantee cap as an immediate £100,000 recovery overstates protection, even in this successful enforcement scenario, because it ignores both the shortfall against the cap and the delay.

On smaller facilities, the fixed cost of getting that far can dominate the calculation, especially where the lender must investigate assets before deciding whether litigation is economic. A defensible settlement may deliver less than the contractual claim but more than contested enforcement, once management time and the probability of an unsuccessful action are included.

This is where the fintech architecture matters. A platform needs security and guarantee records linked to the correct legal entities, with priority and discharge status kept current rather than buried in origination PDFs. The same guarantor appearing across several borrowers creates concentration, not several independent recovery pots. A refinance can also change the position unless releases and replacement security are handled correctly. Recoveries should be reported by instrument and default cohort, net of costs and alongside elapsed time, with open cases retained to avoid flattering results by reporting only the quick settlements.

None of this makes security pointless. With identifiable assets and maintained priority, a debenture can materially improve the lender's position, while a credible PG may support a settlement that wouldn't otherwise happen. The case for cross-company support is strongest where the guarantor genuinely has resources outside the borrower's failure path, and where the operational cost of enforcement is proportionate to the exposure.

For investors, I'd rather see realised recovery curves and the treatment of unresolved cases than a headline percentage of the book labelled secured. If a £250,000 lending decision only clears the return hurdle because a £100,000 signature is valued at face amount, the platform is scaling an error in its recovery assumptions.