Gordon Aitken’s work on SJP is very important. He has discovered a difference in perception between how SJP (and maybe the wealth industry) perceives itself and how the stock market prices it. Is its book of business so fragile?

There is a number in St. James’s Place’s half-year accounts that the market has decided not to believe. The company says the business it has already written, the pensions, bonds and ISAs of its 1,064,000 clients, is worth £22.21 per share: £11.5bn in total, on the company’s own published figures. The stock market values the entire company at £6.1bn.

The market is declining to pay £5.3bn of the company’s own published valuation of business that already exists. For the shares to be fairly priced today, that published number would need to be written down by 47%. Not optimistic at the margin, not flattered by a friendly assumption or two: wrong by nearly half, in a published figure, with the methodology and sensitivities disclosed.

This piece is about that gap: what the number is, why the market stopped reading it, and what it would take for the market to be right. I have also published a 37-page research report on the company, (payhip.com/b/Xa6OE),  which covers the ground in full; this article takes one question from it.

The measure the market forgot

It is the sum of two parts. The first is the net assets the company holds now, which is cash and capital, counted rather than estimated; £1.6bn at 30 June. The second, £9.9bn of the £11.5bn, is the value of in-force business: the present value of the fees expected from every contract already sold, after costs, after tax, after allowing for the customers who will leave early. Add them together and you have what a rational buyer would pay for the company if it never wrote another policy.

I spent 30 years analysing insurers, and for much of that time embedded value was the anchor of every conversation. The sector drifted away from it over the last decade; most companies stopped publishing it, and analysts moved to accounting earnings that fit a standard spreadsheet. I wrote in March about how the industry forgot its own best valuation tool. St. James’s Place is one of the few large UK companies that still publishes the number twice a year, with its methodology and sensitivities disclosed.

Which makes the current situation odd. The one company still showing its workings is the one the market has decided to disbelieve by the widest margin in its modern history.

Half price, and what it implies

The discount has causes. An AI scare that began in February around the launch of Altruist’s adviser technology, and the commentary that followed, took the whole wealth sector down. Then, over the summer, adviser practices advising £5bn were reported to be leaving for Söderberg & Partners, and the shares fell by a fifth in three weeks. Real questions, both. The question for this piece is narrower: what would have to be true of the embedded value itself for the market to be right?

The exodus in the price

Valuation is a different question, and it is the one that matters beyond the next headline. The company does tell us something about withdrawals: its published sensitivity shows that raising withdrawal rates by 10% proportionately costs £529m of embedded value, 4.6% of the total. It cannot carry the departure question, because there is no single base rate to scale: the company models persistency, how long clients stay, by product, age and duration, and publishes only sample points on those curves in its 2025 Solvency and Financial Condition Report, published in March 2026: from 1.3% for 10-year onshore bonds held in trust to 21.25% for post-crystallisation pensions at age 80. Connecting departures to value needs one simplifying assumption: I assume the value of in-force business scales with the funds that leave, so a departing pound takes its 4.1p of in-force value with it. The net assets do not leave with the clients; they are shareholder capital, and legacy business leaving early pays exit charges into them. The company’s own disclosures support the scaling: its published stress for a 10% fall in asset values costs 9% of embedded value. On that basis, hold everything else where the company sets it and ask the only question that matters: how much client money must leave for today’s share price to be the correct embedded value?

The current share price is the correct valuation of this business only if £130bn of client money leaves and is never replaced. Since the net assets stay behind, the burden falls entirely on the in-force value, so the requirement is not 47% of the book but more than half of every pound the company manages, against the £5bn advised by the practices reported to be leaving, whose complete loss would justify a discount of 2%. Now stack that against what management says happens when practices leave. Its experience is that it retains around half the advisers in a departing multi-adviser practice and around half the client funds where an individual adviser does leave. Combining the two, management argues that only around 25% of a departing practice’s funds are ultimately at real risk. On that assumption, the entire Partnership leaving would put around a quarter of the book at risk, less than half of what the current price requires. Assume management is wrong and the true transfer rate is double their experience, 50%: the entire Partnership walking out the door still does not remove the £130bn of client money the price implies. There is no arithmetic on the company’s own numbers in which enough assets leave. I test transfer rates, including accounts far harsher than management’s, in the full report.

Could the assumptions be wrong instead?

The embedded value reporting marks its own homework: every period it discloses how actual experience compared with what the assumptions expected. In 2021 and 2022 experience beat the assumptions and the company strengthened them, by more than £500m over the two years. In 2023 and 2024 experience turned worse than the strengthened assumptions, by £11m and then £136m, the 2024 miss explicitly driven by persistency. That is what assumption error looks like. The experience variance exposes the miss each period, while a separate assumption change shows whether the company has concluded that the difference should also be reflected in its expectations for future years. Since then the misses have shrunk: £53m across 2025, and just £10.2m in the first half of 2026, on a number of £11.5bn, with no assumption change made. The assumptions are currently missing reality by less than 0.1% per half.

The persistency curves are calibrated on decades of client behaviour in a world with exit charges; the August 2025 structure removed them for new business, and the Consumer Duty, in force since 31 July 2023, has lowered exit friction across the industry. Whether behaviour learned in a high-friction world survives a low-friction one is not yet knowable: at the year end the company had four months of experience under the new structure. What can be said is that nothing has shown up yet. Client fund retention improved to 95.4% in the half in which the first two departing practices actually left the register, the aggregate surrender rate, clients cashing in, ran at 4.6% annualised, and the one product line running hot, unit trusts and ISAs at 7.6% withdrawals against assumptions that appear to have been set when the rate ran nearer 5%, is not yet visible as an aggregate miss. The figure could deteriorate from here if practices leave and their clients follow, and I will be watching it. The £51bn of client funds already invested but not yet paying charges remains the least assumption-dependent part of the whole number: the money is in, and the charges begin on a timetable running to 2032.

Embedded value moves with markets, and that is the fair criticism: it rose 12% in six months partly because investment returns ran at an annualised 16% of opening funds, and a bear market would take some of that back. That is a caveat about the level, not about the discount, because a market fall drags the share price down alongside the embedded value and the ratio survives.

Buying back at half price

St. James’s Place is running a £128.1m share buy-back, retiring its own shares at half their published embedded value. If the embedded value is broadly right, buying shares at 0.53x embedded value is strongly accretive to embedded value per share: the company is retiring claims on substantially more published embedded value than the cash it spends. If the embedded value is overstated by half, the buy-back is merely neutral. The board, which sees the persistency data, the fee flows and the actuarial work behind the published number, has chosen to buy at scale. Boards do not do that with numbers they privately doubt.

The buy-back has a cost inside the business. Within the Partnership it is read by some as money that could be repairing the service and administration advisers complain of, and that criticism has force; a business built on its advisers cannot fund shareholders indefinitely at the expense of the people who gather the assets. The board is buying anyway. As evidence of what management believes about the published number, the internal unpopularity makes the signal stronger, not weaker.

What I think, and what would change my mind

What would change my mind is evidence, not argument: client retention breaking materially below the company’s 95% ambition, persistency assumptions weakening in the year-end accounts, or the buy-back quietly stopping. None of those has happened. Until one does, the simplest description of St. James’s Place is a company selling at half its published value of the business it has already written, with the future thrown in for nothing.

Gordon Aitken and Aitken Advisory

Gordon Aitken runs Aitken Advisory, providing strategic advice on life insurers and pension funds. He works with investors, insurers and advisers on transactions, capital strategy and market positioning. As he did throughout his sell-side career, he meets fund managers and investors for one-to-one briefings on the sector. If you would like to arrange a briefing or discuss a potential engagement, click the button below to email him.

The full report. Priced for an Exodus That the Numbers Do Not Support: St. James’s Place after the H1 2026 results. 37 pages: the departures sized and stress-tested against private accounts from inside the Partnership, the adviser economics, the flows, Healey’s first Budget, the AI question, forecasts to 2027 and a dated watch-list. Licences from £295 at payhip.com/b/Xa6OE.