How AI found the PPI’s DC Future Book (2026)

I opened my email box this morning to find this message from an old friend. If you know me or this blog you’ll guess my correspondent.

What follows is unexpurgated and I hope it is taken as constructive feedback to PPI on the DC Future Book 2026

I suspect that both my correspondent and his AI chum have been reading this blog.

I asked my biased AI to help me digest this, and it came back as follows..

---

Your overall reaction is that The DC Future Book 2026 is a useful statistical description of the DC system, but a much less convincing examination of whether that system is the right destination for UK pension policy.

Its strongest feature is that it documents some quite uncomfortable facts about DC.

Its weakness is that it tends to treat the continued expansion of DC as the background condition, and then asks how best to make that system work, rather than asking sufficiently hard whether risks are being allocated to the right people.

---

1. The most important finding is actually rather damning

The headline 90% participation rate is presented as a major success of automatic enrolment. That is fair as far as it goes. But the report itself shows that participation has been much more successful than adequacy.

The median DC pot is only £15,400, while the report says that 43% of the working-age population are under-saving against target replacement rates. It also notes that the statutory minimum has effectively become a norm rather than a floor.

That raises a fundamental question:

Has automatic enrolment solved the pension problem, or has it simply created an enormous number of people who are now technically pension savers?

The report comes close to the latter conclusion without quite saying it.

And the distribution is particularly revealing. 48% of employees earning £10,000–£20,000 are saving at the AE minimum, compared with only 12% earning £60,000–£70,000.

In other words, the people least able to compensate for inadequate contributions through other wealth are disproportionately those saving at the minimum.

---

2. The 12% recommendation deserves much more scrutiny

The report says that PPI modelling suggests that, for people earning above £18,700, 12% is the strongest lever for improving adequacy.

That is an important finding, but it is rather too easily presented as an answer.

There are at least three different questions:

1. What contribution rate produces a particular replacement rate under specified assumptions?
2. What contribution rate can people actually afford?
3. Who should bear the cost – employee, employer or taxpayer?

A modelling exercise can answer the first without answering the other two.

Indeed, the report provides evidence that employer contributions are enormously important: employers account for 68% of the £34.5bn contributed above AE minimum levels, and the additional saving is heavily concentrated among higher earners.

That suggests a much more interesting policy question than simply “should the minimum become 12%?”

It is:

Should pension adequacy be achieved principally through higher compulsory employee saving, higher employer remuneration in pension form, or some combination of the two?

That is a distributional and industrial-relations question, not merely an actuarial one.

---

3. This is where your 6% + annual 1% idea is more interesting than the report

Your earlier proposal, to start someone at 6% of pay when joining an employer and increase the contribution by one percentage point at each annual salary review, has an advantage which the report’s rather static discussion of contribution percentages misses.

It treats pension saving as a lifecycle process, rather than asking a 22-year-old to accept an apparently large permanent reduction in disposable income.

It also creates an automatic relationship between:

salary progression leading to increased pension saving leading to increasing retirement provision.

The report recognises that contribution adequacy is the central problem, but it doesn’t really explore sufficiently imaginative mechanisms for solving it.

---

4. The report is surprisingly accepting of the fragmentation it describes

One of its striking statistics is that there were 21.3 million deferred DC memberships against 11.4 million active memberships in non-micro occupational schemes in 2025. Deferred memberships represented 65% of the total.

That should perhaps be regarded as a major structural failure of the system.

Instead, the report largely treats it as a reason for better consolidation.

But consolidation doesn’t necessarily solve the underlying problem. It can simply produce:

one large pot instead of several small pots, while leaving the individual carrying exactly the same investment, longevity and decumulation risks.

Indeed, the report’s own evidence shows that convenience dominates people’s consolidation decisions: 78% cited keeping pensions together/easier access, compared with only 32% considering charges and 7% considering possible loss of guarantees or safeguarded benefits.

That ought to make policymakers extremely cautious about assuming that consolidation is automatically beneficial.

---

5. The investment discussion is one of the report’s weaker areas

The report shows that master-trust defaults have around 83% listed equities before de-risking, falling to about 30% afterwards, with bonds rising sharply.

This exposes an interesting contradiction.

During accumulation, DC savers are told to seek long-term growth. Then, perhaps 10–15 years before retirement, the system begins moving them substantially towards bonds.

But why is the member’s retirement date treated as the critical investment horizon?

If the objective is retirement income, rather than a pot that happens to be crystallised at age 67 or 68, then the relevant horizon could be another 20–30 years.

The report recognises that longevity risk continues throughout retirement, but the investment architecture remains heavily influenced by the traditional idea that someone approaching retirement needs to “de-risk” a pot aka target date lifestyle/lifestyling investment.

That is a legacy of the historic lump-sum/annuity model, which we abandoned in 2011, not necessarily a logical consequence of modern retirement-income economics.

---

6. The biggest missing question is risk pooling

The report describes DB, DC and CDC accurately:

• DB puts most funding, investment and longevity uncertainty on the employer.
• DC puts considerably more uncertainty on the individual.
• CDC pools some risks collectively but benefits aren’t guaranteed.

But this deserves to be the central policy question, rather than simply one classification among others.

The crucial issue isn’t:

DB versus DC.

It is:

Which risks should be borne individually, which collectively, and which by employers or the state?

DC has become dominant partly because risks have been transferred away from employers. The report could have been much more explicit about that political-economy dimension.

---

7. The treatment of DB is too passive

The report notes that only 4% of private-sector DB schemes remain open to new members, but that 8.6 million memberships remain in the system.

That is useful, but a rather poor use of statistics.  4% is misleading, 8.6 million is less so.

But the conclusion seems implicitly to be:

DB is closing leading to DC is growing leading to, “therefore, we need to make DC work better”.

There is insufficient consideration of whether some of the advantages of DB could be retained without recreating traditional final-salary schemes.

That is precisely where CDC, collective annuitisation and other forms of risk pooling become interesting.

The report does mention CDC, but it feels more like an emerging product category than a possible answer to the fundamental problem of excessive individualisation.

---

8. The “choice” argument is particularly questionable

The report repeatedly acknowledges that retirement involves an extraordinary amount of complexity.

It notes, for example, that 44% of new drawdown contracts were purchased without advice in 2025, compared with only 10% in 2014.

That is a remarkable statistic.

Yet the policy response increasingly seems to be:

more guidance + targeted support + Guided Retirement + better defaults.

But there is an uncomfortable contradiction here.

If ordinary people cannot reasonably be expected to make repeated decisions about:

• investment,
• inflation,
• longevity,
• withdrawal rates,
• sequencing risk,
• taxation,
• health,
• care costs,

then perhaps the answer isn’t better decision-making by individuals.

Perhaps the answer is designing a pension system in which fewer individually consequential decisions are required.

That is an argument for much greater collective provision.

---

9. “Guided Retirement” risks becoming a sophisticated form of managed individualism

The report says workplace schemes will have to provide one or more default pension benefit solutions, while members retain the right to choose something else.

That sounds sensible.

But it leaves the individual ultimately responsible for the consequences.

A much more radical question would be:

Why should someone who has spent 40 years saving into a pension have to become an amateur investment manager and longevity actuary at retirement?

The old DB system answered that question very simply: you received an income for life.

The challenge is to reproduce the economic advantages of that arrangement without necessarily reproducing its employer-balance-sheet structure.

---

10. The report’s sponsors are worth remembering

The report is explicitly sponsored by State Street Investment Management and Scottish Widows.

That does not invalidate the analysis. The PPI explicitly says that funding does not necessarily imply agreement with its findings.

But it does provide useful context.

The report is about how to develop and improve a large DC investment ecosystem. It is not an independent inquiry into whether the UK should have organised retirement provision differently.

That distinction matters.

---

11. The £2.2 trillion projection needs to be treated cautiously

PPI projects aggregate DC assets rising from approximately £1.3 trillion in 2025 to a median £2.2 trillion in 2046, with a range of about £2.0–£3.7 trillion.

That’s useful for understanding the potential scale of the asset pool.

But asset accumulation is not the same thing as pension adequacy.

Indeed, one of the report’s most important implicit messages is precisely that:

You can have a rapidly growing pension asset pool while millions of individuals remain inadequately provided for.

That distinction deserves much greater prominence.

My overall critique

I would characterise the report as:

excellent descriptive epidemiology of the DC patient, but insufficiently critical diagnosis of the disease.

It tells us very effectively that:

• automatic enrolment has increased participation;
• pots are growing;
• DC dominates workplace provision;
• deferred pots are proliferating;
• contributions remain inadequate for many;
• investment remains largely conventional;
• retirement is becoming increasingly complicated;
• retail and workplace pensions are becoming interconnected.

But it doesn’t ask hard enough why the system has evolved in this direction, who benefits from it, who bears the risks, and whether there are better institutional arrangements.

Most importantly, it risks conflating three quite different propositions:

1. DC is now the dominant system.
   True.

2. Therefore we need to make DC work better.
   Also reasonable.

3. Therefore the future of pensions should principally be an improved DC system.
   That does not follow.

That third proposition is the one you should challenge.

The report itself supplies much of the ammunition: a £15,400 median pot, 43% under-saving against targets, contribution rates that have barely moved, 21.3m deferred memberships, and an increasingly complex retirement process.

The really interesting pension-policy question is therefore not “How do we make DC better?”

It is:

How do we construct a pension system in which individuals can accumulate assets, obtain liquidity when they need it, participate in investment growth, and eventually obtain a reasonably secure lifetime income without having to carry all the investment and longevity risks themselves?

That  is where a “Bank of Henry” or lifetime financial instrument idea becomes particularly relevant.

The DC Future Book gets remarkably close to identifying the problem, but doesn’t really challenge the institutional architecture that creates it.