Why the FCA is wrong about DB pension transfers
We show here our opening comments as part of a detailed response to the FCA to its proposed ban on contingent fees for advising on a transfer of a Defined Benefit (DB) pension to a personal pension.
You can read our response to CP19/25 in full here.
This follows a number of articles we have posted here on transfers (search term: 'DB pension transfers').
Fowler Drew opening observations on CP19/25
Though we are opposed to any outright ban on contingent fees, which should in principle be partly contingent, we understand the political and media context in which the FCA’s latest DB pension transfer consultation is taking place. We are more concerned about a misconception about the origin and scale of welfare benefits from transfers that is blighting a legitimate consumer opportunity and infecting the political and media context.
The misconception has an important impact on consumers by restricting access to welfare gains on a much greater scale than the FCA has claimed. It also raises doubt about any argument that the loss of those gains is justified by the mooted reduction of detriment simply by reducing the number of transfers. Neither do we think the two are necessarily mutually exclusive.
Though we believe public-policy aspects ought to dominate the debate, we also have legitimate business interests. The point at which this misconception has the greatest potential impact on firms like ours involved in transfer advice is the adjudication by the Financial Ombudsman Service (FOS) of past recommendations. If errors of interpretation by the FCA in turn infect FOS, and send the wrong signal to claims management companies, we stand to be affected if complaints are made against us, if we have to meet high Financial Services Compensation Scheme (FSCS) levies and if the reputation of the advice industry is generally damaged. These risks are vastly more important for us than the direct economic impacts of the proposals in this paper.
In our previous comments on FCA consultation papers we have pointed to the possibility of a fundamental misunderstanding at the FCA of the main economic driver for transfers, which is welfare gains from higher lifetime spending. We define this as gains in probable sustainable real income that hold even when constrained by somebody’s own tolerance of shortfall relative to the safeguarded income (whether through emotional preference or via hard economic consequences).
Though such an assertion of gain could simply be a matter of opinion, in this case it has an underlying logic that does not depend on judgement. This is the logic that links a DB scheme’s asset allocation, via TPR guidance for calculating CETVs or TVs, to the chance of exceeding the safeguarded income from drawdown, provided the risk tolerance of the scheme member (usually a deferred member) is higher than the scheme’s. The two conditions that in combination determine the scale of any gain in outcomes are i) negative real interest rates and ii) a scheme having derisked, or moved to a highly or fully risk-free (or liability-matched) asset allocation. The former being common to all schemes, in practice it is derisking that varies between schemes. When schemes held broadly the same ‘balanced’ portfolio mix as individuals, there was limited scope for a gain in outcomes and a condition other than income gains (one of those listed in the CP19/25 carve-out provision) was necessary to justify a transfer. These other conditions are indeed applicable only to a few scheme members. Though the two conditions for more general gain have been present on a growing scale since about 2012, and though the size of TV was usually above the typical threshold size for drawdown, it was the 2015 Pension Freedoms that provided the oxygen of publicity.
Where in the past we were merely suspicious this was not being understood, the arguments and data presented in this paper go a long way towards confirming our suspicions.
If, on the other hand, there is no misunderstanding, we might deduce there is instead a difference of opinion about either
• the number of schemes that have derisked sufficiently to create this opportunity (which ought to be a matter of discoverable fact) or
• the number of people with benefits from these schemes that meet a threshold requirement of both i) higher risk tolerance than the scheme’s and ii) sufficient competence or experience to evaluate the tradeoffs that risk taking invites.
It is possible that the second of these criteria is being negated at the FCA by an overriding assumption that a safety-first principle does or should apply to the funding of retirement liabilities, since this is a view explicitly expressed in the paper. As a description of an individual utility function, when applied to pensions, it means that people would not contemplate making any trade off until their core or non-discretionary spending needs had been funded with certainty. But this is clearly not the utility that is typically observed in personal finance generally or in pensions in particular, including 'default' options. It is paternalistic as well as wrong effectively to impose it on people by building it into regulatory processes. It is also directly in conflict with the public-policy objective of encouraging personal responsibility.
On the basis of about half of schemes having derisked fully or substantially and about half of those deferred or active members being excluded by competence or experience, we might expect the proportion that prima facie stands to benefit from making the trade offs involved in a transfer to be as high as 25%. The best terms, implying negligible risk of shortfall when transferring, apply to those schemes that have almost entirely derisked but even that might leave a category as large as 15%. Yet the FCA has expressed ‘surprise’ and ‘disappointment’ at a scale of actual transfers since the 2015 Pension Freedoms that we think equates to about 3% of all private-sector DB members. The difference is an order of magnitude greater than can plausibly be explained by differences of opinion about the size of the base and so has to be about what defines the base.
It is not just by trying to justify the FCA’s data interpretation that we infer a misunderstanding. CP19/25 also lists the categories of people who stand to gain. The largest category of all, those to whom the conditions described above apply, is not even mentioned.
This error is so fundamental that it can be expected to affect almost every aspect of regulation in this area, as follows.
1. Higher than expected volumes invite an alternative explanation that is neither accurate nor necessary. The ‘culprit’ then being sought needs to be one much larger and more widespread than the occurrence of exploitative or rogue advisers, for example poor standards, poor technical knowledge or perverse incentives
2 . The evaluation needed to establish the merit of an advised transfer process will be overstated, thereby excluding the use of cost-effective threshold tests of possible outcome gain by simple (and possibly remote) calculation
3. If this ‘mathematical triage’ has in fact been going on, good advisers will only advise when they pretty much know the answer, in which case a high proportion of positive transfer recommendations will be wrongly interpreted as evidence of bias
4. File reviews of suitability must be dubious if the FCA is expecting the wrong or incomplete conditions to apply to ‘good’ transfers
5. Consistency between FCA and FOS will replicate this bias, leading to unwarranted instances of redress and consequent business failures
6. Continuing high numbers of perfectly rational transfers will lead the FCA (and others) to assume incorrectly that measures already taken to improve standards are not having the desired effect
7. Estimates of detriment will be biased upwards
8. Regulatory tradeoffs will be made suboptimally, as in the Cost Benefit Analysis for CP19/25.