The FCA has bowed to pressure to introduce a Consumer Redress Scheme for all members of British Steel Pension Scheme (BSPS) who were advised to transfer to personal pension alternatives when the company closed the scheme in 2017. Considering there were quite clearly many cases of incompetent and even fraudulent advice, the argument for a scheme that does the job the Financial Ombudsman Service (FOS) would otherwise have to do is compelling. And yet…

The issue raised by these extraordinary powers is whether rough justice is still justice. We are not directly affected, having not advised any BSPS members. And we have no wish to exonerate bad actors in the advice business, whose mess we are constantly being asked to clear up. But we are indirectly affected. We need to ensure that good actors do not get tarred with the same brush. And we want to minimise any unjust ‘gift’ element of redress that other firms will probably end up paying for, via the Financial Services Compensation Scheme and higher PI premiums.

This is a technical article written for people with an interest in the regulatory issues raised by Defined Benefit (DB) transfers. If you simply want to understand how to assess the merit of transferring out of a DB scheme into a personal pension arrangement, we have prepared a less technical explanation of the decision here.


An Ombudsman is a short cut intended to replicate the outcomes of a court case but without consumers incurring the uncertainties, time and costs of going to court. In granting the FCA powers to use a Consumer Redress Scheme, as a short cut on a short cut, Parliament prescribed specific conditions that must be met. These conditions imply objectivity and hence an ability to provide consistent outcomes, as a basis for fairness for both advisers and their clients. The purpose of this article is to examine whether, in the context of advice to exchange DB pension rights for a cash sum to be invested in a personal pension arrangement, any redress arrangement can meet the fairness tests of objectivity and consistency. We believe they cannot.

The apparent scale of egregious bad advice at British Steel explains the pressure on the FCA to introduce a redress scheme. It is unavoidable that this will raise the regulatory issues around general transfer advice, as BSPS can, at the same time, be both exceptional and representative.

The notable feature of pension transfer activity is that the FCA has deliberately blighted it, arguing the loss of opportunity for ‘good’ transfers is justified by preventing ‘bad’ transfers. But this assumes that it has been right to argue that there are few cases where transfers will be in a DB member’s best interests. This prior assumption, biasing everything else the FCA asserts, was always economically absurd. It depended on an unconventional logic for determining individual welfare that is not evidenced in fact. When challenged, the FCA file assessors have a number of ‘get out of gaol free’ cards that effectively deny discussion.

For the purposes of this article, we assume a BSPS Consumer Redress Scheme will make use of the same DB Transfer ‘Advice Assessment Tool’ (DBAAT) used in Past Business Reviews. We show how errors in the FCA’s logic on transfer merit generally make this assessment process unworkable in all cases where there is any room for doubt or judgement. It will not be able to generate the fair and consistent outcomes that advisers are entitled to expect. It is precisely the ‘room for doubt’ that the FCA’s false logic seeks to exclude. The implication is that a problem of legal challenge to Past Business Reviews, perhaps a judicial review, must be building that will itself cast a shadow over a Consumer Redress Scheme.

When we apply to BSPS transfers the conditions that Parliament intended the FCA to meet, we also find that a redress scheme should in this case be hard to justify. This is based on the prevalence of likely loss, and so relies on the terms offered by BSPS. We use our proprietary modelling (necessarily at this stage based on proxies rather than a range of actual cases at different dates) to estimate the scale and likelihood of any shortfall relative to the impaired pension promises of the closed scheme. We find the probability of shortfall to be very low or even negligible, even though the scheme terms were not as favourable to transfer as many.

This is based on mainstream asset classes and reasonable costs and does not exclude the possibility that actual replacement strategies were too risky, or so expensive as to be self-defeating. The worst offenders in this regard should already have been identified by the FCA, nearly five years after the event. What we should be left with is cases where there is room for doubt and therefore a requirement for objective and consistent assessment.

Fairness also assumes proportionality between the confidence in the estimate of loss and the amount of compensation. It remains to be seen how the consultation will deal with this but prima facie, without objective bases of assessment of both cause and chance of loss, the extremely onerous redress that is typically the outcome is unlikely to be proportionate.

Finally, the redress approach, be it through the Ombudsman, a Past Business Review or a Consumer Redress Scheme, is open to gaming by creating a ‘free lunch’ where ‘victims’ can keep their more flexible arrangements, add the cash compensation and augment the upside potential. The calculation of redress is based on the principle of putting them back, as close as possible, to the position they would have been in had they not transferred. It is not designed to create a gift. It cannot be justice when an adviser is found to be at fault but the finding is promptly invalidated by their client keeping the riskier position they recommended, rather than replacing it with the low-risk or guaranteed product the assessment determined to be suitable and which formed the basis of calculating the amount of redress.

Why do we care?

Why are we, a firm on the side of the angels whose own transfer cases are beyond reproach and who wrote no BSPS business, willing to stick our necks out by criticising the FCA for its treatment of other firms, some of whom were clearly rogues? The answer is simple and has nothing directly to do with BSPS or any other individual case.

  • Good regulation of investment activities requires a solid intellectual foundation in investment theory, economics and maths
  • Our business processes, which unusually in private wealth are largely ‘quantitative’, or rules based, were all carefully crafted on this foundation
  • If we cannot rely on the FCA to write or interpret rules based on the same foundation, we should worry that we too might be blighted unjustly.

Firms’ advice can look ‘beyond reproach’ on generally-accepted investment principles but not on the FCA’s. The two divergent views need aligning. The consultation required to impose a Consumer Redress Scheme under s404 of the same Act is another opportunity to try to try to bring them into better alignment.

The prompting for this article is the possibility of a scheme being introduced. But even if it isn’t, the problem I am describing exists for firms already subject to a s166 order, a Past Business Review, whose files have been internally assessed by the FCA or who find themselves defending cases brought to the Financial Ombudsman. All of these procedures are intended to replicate legal process while avoiding the costs that litigants might otherwise be unable to afford. But whereas a court might allow the specific application of the FCA’s rules and guidance on DB transfers to be challenged by a defendant firm, the same cannot be taken for granted for any of these alternatives to a legal process. Where transfers are concerned, they all promise rough justice.

The legal tests for a Consumer Redress Scheme

The main requirements of the application of the Act appear to this non-lawyer to be that:

  • the harm be applicable to the most cases covered by the scheme (on the basis, presumably, a minority of cases can be deal with cost-effectively by the Financial Ombudsman Service or by the FCA’s ability to call for Past Business Reviews from individual firms of advisers)
  • a firm subject to the redress scheme be able to determine objectively whether or not it has complied with the applicable rules and, if it has not, whether the failure has caused (or may cause) loss or damage to consumers.

In the case of British Steel, we know that a number of highly-active firms had their transfer permissions withdrawn and have been subject to review orders. The worst offenders are presumably long gone. It would be reasonable, therefore, to assume that the remaining firms that will be drawn into the scheme are those most likely to have handled nuanced cases. By ‘nuanced’ I mean that establishing fault and making the link between fault and probable loss is, for related reasons, hard to demonstrate objectively and consistently. It could of course be argued that the best way to test those cases, assuming an objective basis of testing exists, is via a redress scheme. But the nuance is not simply a question of whether the advice process complied with the applicable rules: fault is academic for the client, without loss. So determining the prevalence of probable loss is a good place to start.

The chance of loss

Though we didn’t advise any BSPS members, we have seen one example of the terms for transfer. Applying our stochastic modelling of transfers to cases with equivalent terms at the same time suggests that the probability of a shortfall against the deferred DB pension income was quite low for members close to retirement, arguably zero, and certainly zero for younger members.

This should come as no surprise. The boom in transfer activity generally arose as a logical response to pension regulation requiring the cash sum equivalent to a stream of income be calculated using low or negative discount rates (a necessary consequence of schemes having derisked), whereas the distribution of probable real equity returns is, at longer horizons, mostly or entirely higher.

Time matters because, whilst the range of uncertain outcomes keeps expanding (even with an assumption of mean reversion in equity real returns), that is offset by the assumption (which the FCA prescribed growth rates themselves require) that the trend will carry the bottom end of the possible return distribution towards and eventually above a fixed risk free rate. The driving force for transfers is not flexibility: it is higher probable outcomes. Flexibility has a value but you would not necessarily pay for it. With higher probable outcomes, particularly relative to minimum tolerable spending, there is nothing to pay for the flexibility. It is icing, not the cake.

You will not find a single acknowledgement of this text-book economic analysis of transfer merit in the FCA Handbook or guidance or in their public statements. I have no special insights why this is but I suspect it is a bunker mentality, the original position having been vocally adopted not by investment or economic experts but by supervisory staff, perhaps even senior staff opposed to the philosophy of personal responsibility that lies behind Parliament’s so-called Pension Freedoms, and nobody then wants to admit an error.

One of the ways the FCA can, if it chooses, negate this argument for transfer merit is to assert that negative real gilt yields are implicitly associated with lower future equity returns. This too is absurd. The relationship between risk free rates and expected equity returns is clearly dynamic, not fixed. Expected and actual risk premia vary with market conditions and the correlation between gilt yields and equity returns is highly unstable. If it wished to constrain advisers to using normalised return assumptions that always had a fixed relationship with real gilt yields, the FCA would have to adjust the prescribed rates for equities down, thereby removing the scope to exploit a higher expected risk premium for equities. But I doubt the Bank of England would have been amused as the logic of quantitative easing relies on encouraging risk-taking behaviour by, amongst other things expanding the risk premium, as in steepening the slope of the Capital Allocation Line.

The absurdity is not limited to the implied risk premium. The regulator would also have to intervene to prevent advisers acting as if the real gilt yield might itself normalise. Advisers will, unless actively prevented, assume that the upside for real interest rates is skewed relative to the downside, at all but short horizons. They will assume this because the market economy will not be expected to function with sustained periods of negative real yields. That means that derisking arising as draw approaches will probably be at higher real gilt yields and deferring any annuity purchase to late in retirement will probably be at higher yields.

This also undermines the FCA file assessor’s ‘get out of gaol free’ card for use when the terms were overwhelmingly favourable: an adviser should have designed a holistic plan around the retained DB pension. This is equivalent to arguing that the optimal holistic portfolio necessarily includes a large and permanent exposure to long-dated bonds with incomplete indexation to inflation. It may, but under QE it is very unlikely. Advisers do not need formal portfolio optimisation tools to know that the optimal portfolio does not now, but may at some point, include long bonds, and could even involve an annuity, but that is conditional on the normalisation of real gilt yields.

Transfer post quantitative easing is viewed as ‘arbitrage’ by advisers, and by more discerning scheme members, because two sets of actors are compelled to behave in different ways even if, in the same market conditions, they hold the same views:

  • DB pension scheme trustees have felt pressurised by regulations and accounting consequences to match both the nature and duration of their liabilities, locking in a gilt-based cost they believe will be temporary and would prefer not to lock in
  • other actors in financial markets, including DB scheme members, are under no equivalent pressure.

The arbitrage is created by some principals being able to back their beliefs and others not.

The only hint at FCA guidance consistent with the economics of transfers I have described here is not in the FCA Handbook. It is in the manual used by firms or skilled persons when assessing past business: paragraph 2.22 of the DB Transfer Advice Assessment Tool (DBAAT) Instructions. This is in the context of reliance on the DB income for meeting spending, which you might assume applies to many long-standing members of any DB scheme.

The client is reliant on income from the DB scheme in retirement if they do not have the capacity to lose it – for example because without it they would be unable to meet non-discretionary expenditure. In those circumstances, absent a suitable alternative that will provide sufficient guaranteed income to meet their needs, or evidence that the client could achieve their income objectives in retirement without a guaranteed income (for example because the client has a very short life expectancy), they should generally be advised to remain in the DB scheme.

 The client may not be reliant on income from the scheme if they can produce the same income via a suitable alternative, with or without a guarantee, and this income is able to meet their needs throughout retirement. For example, if the recommendation is to transfer to a personal pension and the cash flow modelling (in real terms) evidences that with a sustainable withdrawal rate the client will not run out of money in retirement, allowing for beyond average life expectancy and stress testing of returns.

This is a grudging acknowledgement (the second half of this paragraph having evidently been added later) that it is possible for a transfer to improve outcomes for the member simply as a function of the terms offered. It was probably forced on the FCA by those external skilled persons pointing out that a loss cannot simply be assumed without doing the sums; and the numbers are necessarily as much a function of the terms as they are of capital-market assumptions.

The BSPS terms

The one example we have seen was before the terms were improved as a result of Tata injecting additional contributions to reduce the accounting deficit. Later CETVs were probably higher. Based on that example, the terms were not particularly ‘generous’, in the sense that the implied discount rates used by trustees for calculating the CETV, implied by the multiple of income, were not as low as other schemes that were either more fully derisked (more gilts) or better funded or both. But the terms were generous enough, with CETV equivalent to 24 times the present value of the projected pension, to make improvements in welfare highly likely across the age cohorts, for reasons I will explain.

For this exercise we draw conclusions by comparison with other schemes that we did advise that offered similar terms at broadly the same time (so in similar market conditions). If adviser firms wish to provide us with BSPS examples, so we can check for consistency, we can replicate model runs using market conditions at the time. We need date of the advice, age, the CETV and the projected pension at the time (so a present value, like the CETV). We would like to include this more complete data in any response we make to the redress scheme consultation.

We extract from our reports the model-derived chance of shortfall versus the DB pension. This is calculated by assuming draw at the same real rate as the DB income, a plan end of age 100 and the 50% and 99% confidence levels for the projected replacement outcomes.

It is critical in modelling drawdown that the stress test be robust. The consequences of outliving capital are serious. Our drawdown model divides the plan into time slices of one year each and the confidence levels are for all time slices individually. Each time-slice portfolio is effectively a target-date portfolio with a stochastic return path and dynamic asset allocations. The probability of every target-date portfolio generating worst-case results, which we assume, is in practice less than for any individual one so this is a very robust test. Upside motivation is represented by the 50% confidence level or the mean sustainable real income. An explanation of our drawdown model can be found here. An interface to a simplified version of the model for planning a drawdown goal can be accessed here.

The comparisons are as follows.

  • For a person 19 years from retirement, our model showed a zero chance of a shortfall versus the DB pension The median upside (this being a number dependent on the adopted risk approach) was 133% – more than double
  • For older members very close to retirement, similar cases suggest the downside risk would have looked fairly modest and the upside much more balanced: about 15% of the outcome distribution being below the DB income at 99% confidence.

But ours is a real-return model and the shortfall risk is calculated on the basis that the DB income will not be eroded by inflation caps. Inflation matters. To an individual, a cut in real spending forced by inflation has the same impact as a cut forced by anything else. Caps do apply to both the original scheme and the BSPS replacement scheme, BSPS2, that members retaining DB benefits should (FOS thinks) have selected. Neither scheme guarantees a real income.

BSPS2 changed the inflation link in many cases from RPI to CPI and imposed a lower cap of 4% pa in deferment and 3% in payment, in the example we saw. As a reference point, the difference in cost for an RPI-indexed annuity and an annuity increasing at 3% fixed for a 55 year old is currently about 30%. Though not an exact measure of the inflation difference we want to isolate, it is notable that it is more than the likely shortfall in our comparable model runs.

In the context of stress-tested sustainable real income, only a few years of inflation above the cap (provided they are not offset by subsequent absolute falls in the price level, as distinct from falls in the rate of price change) will be sufficient to offset the shortfall risk in real income against a real-terms guarantee. Whereas the outcomes of a drawdown plan are necessarily path dependent, the inflation impact on real outcomes does not need knowledge of the entire path. Probabilities for real outcomes will change as soon as the cap has been breached. Stress tests appropriate to 2.22 of the instruction manual should assume episodes of the cap biting, as well as low real equity returns.

One of the issues the consultation will need to address is what the ‘correct’ basis is for comparison of income in the tested advice and for calculation of redress. Using BSPS2 is convenient for the benefits comparison because it removes the need to deal with another tricky problem: what were advisers saying about the PPF effects? At the time of the advice it might have been reasonable for an adviser to make comparison with the PSBS old scheme benefits and to build in the weakened inflation protection under the PPF compared with the weakened protection under PSBS2. Here too a redress scheme may seek to remove room for doubt where doubt clearly exists.

Are BSPS members different?

This approach to determining the likely prevalence of loss assumes that the 2.22 quantification of outcomes should, as a default, be available to all members of any DB scheme. In other words, there is nothing socially exclusive about rare, genuine, arbitrage opportunities. Indeed the reverse holds: they are more valuable (have greater utility) for people with less accumulated wealth and with lower incomes. In economic terms, more of their minimum needs are underpinned by State Pension and less of their upside is subject to declining marginal utility. A relatively small gain in outcomes may be transformative where a more affluent individual might barely register the lifestyle change.

In practice the FCA is probably right to treat certain types of DB scheme member as being excluded from the opportunity to improve their welfare, even when they meet the constraint of a minimum tolerable outcome in 2.22. The most obvious reason is that the member lacks comprehension: the ability to process the information required in a regulated transfer advice process or to do so in a rational way. This is naturally quite difficult to legislate for before the transfer, let alone assess after the transfer, particularly without meeting every person.

Much more subtle but also more prevalent is the disqualification factor that requires a judgement about persistence: will the member stick with the planned drawdown strategy on which the outcome probabilities depended? Persistence is needed because the plan that delivers welfare-improving outcomes also carries a high chance of delivering quite serious ‘paper’ losses at some stage along the way. Dumping may be a reasonable assumption if the member clearly lacked the comprehension and it is a possibility if their past investment experience revealed a lack of persistence. But in most cases there is genuine room for doubt. This will not be removed by asking the person, as they have a clear motive under all of these out-of-court procedures to adopt the pose of ‘buyer’s remorse’.

There has been a tendency to report the BSPS transfer frenzy as if the scheme had more than its fair share of vulnerable members, in terms of both comprehension and likely persistence. This is superficially plausible but it may not stand scrutiny (most schemes have a broad mix of members) and is condescending. Bear in mind we are not talking here about being ‘clever’, in the way that FCA staff or advisers are clever. We are talking more about being ‘smart’ and talking to other people who you think are smart. That is how the transfer opportunity generally went viral: not because we advisers are brilliant communicators.

The distinguishing feature of BSPS is probably not the profile of the members by any measure, but the fact that they were all forced by the scheme’s closure to make some difficult and stressful decisions under the pressure of deadlines. Transfer opportunities otherwise generally only apply to former employees with a deferred pension, or possibly active members very close to retiring. But there is no particular reason to assume the profile of the eligible members in an open scheme and in a scheme closed to new accruals is different. What does distinguish BSPS for the purposes of a redress scheme is that all active members coincidentally had to make choices resulting from being newly members of a DC scheme. They are getting an education of sorts.

It is realistic to assume that the relevant risk measure in the majority of cases is sustainable real income outcomes, rather than short term volatility or paper losses.

If the BSPS transfer terms and scheme benefit rules are such that transfer outcomes can be shown to be highly unlikely to lead to stress-tested shortfall for most people affected, a key condition for the scheme cannot be met. For the same reason, any presumption that transfer was likely to be unsuitable for most members has to be wrong.

How will  a redress scheme deal with nuance?

If the FCA persists with a scheme, it will spread the issues already dogging Past Business Reviews: the DBAAT process cannot deal objectively, consistently and therefore fairly with nuance. Nuance is where  the assessment of suitability cannot rely on binary tests and requires judgement or interpretation. It could choose to proceed with a different assessment basis than DBAAT but that would be unfair for firms already assessed using DBAAT . Any replacement designed to save time would need to be even more binary and less able to accommodate nuance.

There are two elements to the DBAAT process: a spreadsheet with largely yes/no questions and the Instruction manual which provides interpretative guidance. We cannot assume that the FCA’s internal assessment of transfer files always involved a manual, or the same manual. It became a requirement when the FCA’s view of transfer merit conflicted with external views based on generalised investment theory and practice. That would help explain why the FCA’s early estimates of ‘unsuitable’ DB transfer advice cases have been slashed from over half to 17% (in the press release announcing the Consumer Redress Scheme consultation). The FCA have said they cannot discuss the outcomes of Past Business Reviews, even in summary form, but I would be surprised if the dichotomy has not led to exactly the same cases being treated differently by internal and external assessors. This is also something we would like to be able to instance in our consultation submission.

What explains this dichotomy? The most obvious explanation is that the FCA’s observations are consistent with what text books refer to as a ‘safety first utility function’ whereas it is not the right utility to assume. ‘Safety first’ holds that people save without taking risk in order to meet their core liabilities, however long term in nature, and only take risk with resources in excess of that amount. They take risk having achieved sufficiency, but not in order to achieve sufficiency. The section of the manual I have quoted is an example of that thinking. To apply it in practice to retirement savings it means the FCA has to go down a particular route of requiring explicit estimates of a discretionary and non-discretionary spending, potentially years before realistic expectations can be formed for either. This is largely redundant because the safety first principle conflicts with the utility assumed in standard investment theory and is not revealed by actual investor behaviour in the UK – mercifully, or we would be even more under-provided.

Portfolio theory treats a transfer as a particular case of a general problem of choosing between a certain known outcome and a range of probable uncertain outcomes. Good financial planning makes this investment advice specific to constraints that may be derived from budgets, such as a minimum tolerable spending level, but portfolio theory does not need any constraints for it to be observable in practice. The theory needs to assume that, in any context, an appropriate risk free alternative is available to the investor on a daily basis and easily priced. It could be cash, index linked gilts or even an annuity. As most investment professionals see it, the DB pension income is just another version of a risk free real asset – or in BSPS2’s case, a semi-real asset.

The utility assumed by standard portfolio theory, when constructing rational portfolios using these assets, is not safety first. It is that people

  • are risk averse (or loss averse) but also want more
  • they recognise they need to make trade offs
  • they accept risk because there is motivation to do so.

If this were not the way most people actually make decisions, our individual balance sheets would look very different and so, more particularly, would DC members’ fund preferences and therefore the design of default funds. It is the FCA that is out on a limb by trying to attribute a different utility to DB pensions.

The DBAAT process cannot handle nuance because it does not treat the decision as a trade off, even though 2.22 introduced a constraint to the trade off, meeting core spending, that implies that value will be assigned to outcomes better than the minimum. One obvious impact of this bias is that insufficient weight will be given to upside motivation, even though this was almost certainly a driving force for both adviser and client.

My view is that this makes the DBAAT process unworkable. Advisers will try to assess their cases against standard text-book principles and the DBAAT form will not permit it. External assessors, such as accountants or compliance consultants doing the work for the adviser, may deal with this dichotomy differently, by making their answers fit the FCA’s intentions, in which case there will be a huge injustice.

It is difficult to see how this issue can be resolved other than by judicial review. Affected firms, who believe their cases are sustainable on the basis of the standard approach set out here, may be left no other option. Though individual adviser firms are the ones who face the failure risk, it is PI insurers who stand to lose the most in total. It would not therefore be surprising if a legal challenge came from the insurance industry.

Disproportionate consequences

The debate about redress cannot ignore the consequences of fault being found. A redress formula already exists, based on the principle of replicating as closely as possible the benefits ‘inappropriately’ given up. This is a very expensive consequence, because it is driven by the same low discount rates that create the opportunity for welfare-improving transfers and explain the high level of transfer activity. The practical effect for firms affected by Past Business Reviews is that their continued existence hangs on whether a very small number or a larger number of possibly indistinguishable nuanced cases, however disputed, end up requiring redress. For many this is a multi-year process. These firms are blighted: they cannot sell the firm or pay dividends and individuals cannot work elsewhere.

Very expensive consequences imply that there should be a proportionately high level of certainty about welfare being increased or reduced. This is not the case for BSPS. On the numbers alone, most transfers appear likely to be too nuanced to be that certain.

Where possible fault arises because of technical deficiencies in the advice process or documentation, redress would probably be even more disproportionate. If the deficiencies do not render the transfer unsuitable it would be more appropriate to levy fines.

For BSPS the redress formula is more complex because of the low inflation cap applicable to BSPS2. Though I have suggested the scheme proposal should fail because of it, the consultation process will still have to address this in detail and evidence its calculations.

Gaming redress

As noted, redress (whatever the route) is a free option to improve the terms of the original CETV. There is no risk or cost in presenting as a victim of bad advice. It provides a strong motivation for ‘victims’ but also for lawyers and claims management companies.

This would not be the case if redress involved sacrificing the upside and flexibility the client allegedly did not value. That would apply if they had to go back into the original DB scheme (or in this case BSPS2). It would also apply if they had to substitute all of the replacement investment strategy with the same products (pre- and at retirement) as were referenced to calculate the redress, ie those that replicate as near as possible the benefits of the scheme. Before retirement, this should be a deferred annuity or a duration-matched gilt holding. At retirement, it is an annuity.

Redress under a Past Business Review, and presumably under a Consumer Redress Scheme, does not impose this condition. Advisers are justifiably angry that cash compensation allows the beneficiary to add to, rather than alter, their existing strategy, even though, if they do, they quite clearly invalidate the conclusions about suitability, or risk preferences, that the DBAAT process led to.

In the case of the Ombudsman, it is possible to make awards that assume the member was willing and able to take risk but that the level of risk or costs in the replacement strategy were inappropriately high. This is sensible as the economics I have described are based on a capital-market context using core assets classes and without self-defeating costs. DBAAT needs to be able to make the same assessment to ensure that redress is proportionate.