The Overruled Objection

There is a version of the Solvency II story that involves shadowy lobbying, a captured regulator, and an industry that pulled the wool over Whitehall's eyes. It is the version that practically writes itself, and it is wrong in the one way that matters. The regulator was not captured. The regulator fought, in public, with numbers, for two years. The regulator lost.

That distinction is the whole story. Regulatory capture is a failure of institutions; what happened between 2020 and 2024 was a decision — made by elected ministers, on the record, against the explicit and quantified advice of the supervisor — to thin the prudential buffer standing behind millions of guaranteed pensions, in exchange for a promise of infrastructure investment. It may even have been a defensible decision. What it was never given was a defence. It was sold as the removal of EU red tape, and the cost side of the ledger was left to a regulator's letter that almost nobody read.

This essay is about that letter, and about what it means that the system's response to a warning of this kind was to file it.

The machine being adjusted

To understand what was decided, you need one piece of machinery: the matching adjustment.

When a company offloads its pension scheme to an insurer — a "bulk purchase annuity," now running at £40–50 billion of transfers a year — the insurer takes the scheme's assets and reinvests them. The day before the transaction, the pensions were backed largely by gilts. The day after, the same promises are backed predominantly by credit: corporate bonds, and an ever-growing allocation to private, illiquid, internally-originated assets — equity release mortgages, infrastructure debt, social housing loans, urban regeneration finance. On a look-through basis, something approaching forty per cent of the assets behind a modern annuity book is private or structured credit, much of it originated, valued, and priced by the insurer's own group.

The matching adjustment is the rule that makes this profitable. It allows the insurer to discount its liabilities not at the risk-free rate but at the yield on the assets it holds, less a deduction — the fundamental spread — representing the credit risk it retains. Whatever yield survives the deduction is treated as an "illiquidity premium": a reward for locking money away, earned the moment the assets are bought, capitalised into the insurer's balance sheet on day one. The logic is not absurd. Annuity liabilities cannot run; pensioners cannot surrender. A hold-to-maturity investor genuinely does not bear mark-to-market risk. But the entire edifice rests on one number being right: the fundamental spread. If the deduction for retained credit risk is too small, the "illiquidity premium" is partly fictitious, the day-one capital is partly fictitious, and every bulk annuity in the country has been priced on it.

The fundamental spread is calibrated from long-run default data on publicly traded corporate bonds. The assets it is increasingly applied to are equity release notes, regeneration debt and private placements with no default history at all, rated in many cases by the insurers holding them. It does not move when market spreads move: in March 2020, when credit spreads doubled, the fundamental spread barely stirred — meaning the entire blow-out was classified, by construction, as illiquidity premium, and insurers' balance sheets improved in the teeth of a credit panic.

The Prudential Regulation Authority noticed all of this. That is where the story starts.

The warning

From 2020, the post-Brexit review of Solvency II offered the government a rare prize: a financial-services reform that could be framed as a Brexit dividend. The industry supplied the frame. The Association of British Insurers attached a number to it — over £100 billion of investment in productive finance, social infrastructure and green energy over a decade, conditional on "meaningful reform." Risk margin down, matching adjustment eligibility widened, capital released. Levelling up, funded by pension promises.

The PRA's position was more interesting than simple resistance. It supported most of the package — including, notably, the cut to the risk margin, which its chief executive had himself called the regime's most obvious defect. What it asked for in exchange was repair of the one component it believed was genuinely miscalibrated: the fundamental spread. In its 2022 discussion paper the PRA set out the case that the deduction failed to capture the uncertainty of credit losses — that it should include an explicit credit risk premium, that its insensitivity to market conditions and crude rating granularity actively incentivised insurers to reach for the highest-spread assets within each rating bucket, and that the matching adjustment benefit being claimed was likely too high precisely where the portfolios were drifting: into illiquid assets the insurers rated and valued themselves.

Sam Woods, the deputy governor running the PRA, put the structural point plainly in a 2022 speech: the bulk annuity market is so competitive that even well-intentioned firms are compelled to chase the most MA-productive assets available, or lose the business to someone who will. This is worth pausing on. The supervisor's argument was not that insurers were behaving badly. It was that the rule itself manufactured the race to the bottom, and would do so regardless of anyone's intentions.

The industry pushed back — and not frivolously. A market-sensitive fundamental spread would make annuity balance sheets swing with credit spreads despite the assets being held to maturity, importing exactly the procyclicality the matching adjustment exists to remove. Unexpected losses, they argued, are the job of the solvency capital requirement, not the discount rate. And the sector's record is, genuinely, unblemished: no UK annuity writer failed in 2008, in 2020, or in the gilt crisis of 2022 — when it was the pension schemes, not the insurers, who needed rescuing. These are serious arguments, and a serious account has to grant them. Though the second deserves its footnote: the solvency capital requirement, the industry's designated home for unexpected losses, is itself covered largely by own funds the matching adjustment manufactures — the buffer against the assets failing is built from the assumption that they won't. That rebuttal was available throughout. The Treasury never required anyone to answer it.

But the government did not win the argument. It declined to have it.

The decision

In November 2022, the Treasury published its consultation response. The risk margin would be cut by 65 per cent for life insurers. Matching adjustment eligibility would be widened to assets with "highly predictable" cashflows. And the fundamental spread — the one element the supervisor had said was broken — would be retained on its existing methodology and calibration, essentially untouched.

The shape of that refusal is worth seeing plainly. The PRA had proposed rebuilding the deduction in three parts: expected loss, an explicit credit risk premium for the uncertainty of those losses, and a valuation-uncertainty allowance for assets priced by internal models rather than markets. Of the three, the first already existed; the two new ones — the credit risk premium and the valuation allowance, the components that would have bitten on self-valued illiquid debt and on the correlated repricing of sectors once thought safe — were the entire point of the reform. Neither was adopted. The deduction that emerged into law was, to a first approximation, the one the supervisor had spent two years calling inadequate.

The Bank of England's reply sits in the parliamentary record, in a letter from the Governor to the Treasury Committee, and it deserves to be quoted in substance because almost nothing else in this saga is so unambiguous: implementing the government's package, the Bank estimated, would raise the annual probability of a life insurer failure from 0.5 per cent to 0.6 per cent — a relative increase of around 20 per cent. Asked about the gap between his institution and the government, Woods was a model of regulatory understatement: the government, he observed, had been clear it was not persuaded.

Note what did not happen. The Treasury did not publish an analysis rebutting the Bank's failure-probability estimate. It did not argue the fundamental spread was correctly calibrated for self-originated equity release notes. It cited, accurately, a lack of consensus — the industry disagreed with the regulator — and chose the industry's side, wrapped in the language of slashing lingering EU burdens. The PRA's risk warning was not answered. It was acknowledged, and overridden.

Governments are allowed to do this. That is, in fact, the constitutional point: risk appetite is ultimately a political choice, and an elected Treasury outranks an appointed supervisor. A government that never accepts prudential risk in pursuit of growth ends up with neither. If ministers had stood up and said — we have heard the Bank's estimate that this raises the chance of an insurer failing by a fifth; we judge the investment benefits worth that cost; the Financial Services Compensation Scheme stands behind pensioners if we are wrong — the decision would have been honest, accountable, and arguably even right.

No minister ever said it. The reform was a red-tape story on the day it was announced and has remained one since. The single most material fact about it — that the supervisor quantified the increased risk of failure and was overruled — appears in no press release, no Budget speech, no manifesto. It appears in a letter to a select committee, where warnings go to be archived.

The tell

And then there is the detail that gives the game away.

Having declined to fix the fundamental spread in the rulebook, the government agreed that the PRA could acquire new supervisory tools: powers to require add-ons, intensified stress testing — and, most revealingly, an attestation regime. From 2024, a named senior manager at each insurer must personally attest, every year, that the fundamental spread is sufficient for the specific assets the firm actually holds, and apply a voluntary top-up where it is not.

Sit with that for a moment. If the official calibration were adequate, no attestation would be needed; the rulebook number would simply be the answer. Requiring an executive to put their personal signature — and personal regulatory liability — behind the sufficiency of a figure the government had just declined to repair is the system quietly admitting that the figure cannot be relied upon. The risk the Treasury would not price into the rules was instead relocated onto individuals, one signature at a time. The hedge is the confession.

Who holds the tail

So: where does the risk actually sit?

Not, visibly, anywhere. That is the property of this architecture that ought to disturb people most. An annuity book has no depositors to run, no policyholders who can surrender, no maturity wall, no market price for a third of its assets. If the fundamental spread is too thin — if the spread on self-priced regeneration debt was never really illiquidity premium but unrecognised credit risk — the error does not surface next quarter. It compounds silently for a decade or more, capitalised into day-one profits, recycled into the pricing of the next deal, until a genuine credit cycle marks the illiquid book to reality. The sector's unblemished record, so often cited, has been earned entirely in an era of falling rates and central banks backstopping credit markets. The regime has never met its test.

If the test goes badly, the sequence is mechanical. The pensioner is protected by the Financial Services Compensation Scheme — 100 per cent, no cap, for annuities. The FSCS is funded by levies on the surviving industry, which is to say, in any system-wide stress, by a claim that ultimately lands on the state. Privatised spread in the good years; socialised tail in the bad one. The pensioner whose scheme was "secured" with an insurer, and the taxpayer standing behind the compensation scheme, are the residual risk-holders of a calibration argument they have never heard of — one the supervisor raised, quantified, and lost.

And when the reckoning comes, if it comes, there will be a particular cruelty in the record. No one will have decided to accept this risk — not officially. There will be a modernisation programme that everyone welcomed, an investment pledge that everyone applauded, and a regulator's letter, dated, filed, and answered by nobody.

The objection was overruled. It was never refuted. In British financial governance, it turns out, those are different things — and the difference is currently compounding, quietly, at gilts plus a margin.

C.J. Marsden writes on political economy at ir35andmore.com.

Sources for verification: Bank of England letter to the Treasury Committee on Solvency II reform (failure probability 0.5% → 0.6%, ~20% relative increase); HM Treasury, Review of Solvency II: Consultation — Response, 17 November 2022 (risk margin cut 65%, fundamental spread methodology retained); Sam Woods, Solvency II: Striking the balance, Bank of England speech, July 2022 (competitive compulsion toward MA-productive assets); PRA DP2/22 (credit risk premium case); ABI statement, 17 November 2022 (£100bn productive investment over ten years); PRA PS10/24 (MA attestation regime); PRA SS5/24 (funded reinsurance).