Conflicts of interest when advising DB transfers
There was disappointment the FCA did not address ‘contingent charges’ in its DB transfer advice paper. This was never likely. It’s not specific to DB transfers and it’s already in the handbook. It’s a normal commercial source of conflict. It's far from the only one involving transfer advice.
In this post we share how we identified the conflicts and how best to manage them. We are financial advisers. We are transfer specialists. We are also wealth managers. Each poses issues and so do the connections between them. Our solutions involve the triggers for charges, the basis of calculating each charge and the design of the advice process itself in such a way as to demonstrate the absence of bias.
Why the fuss about contingent fees?
Contingent fees are only charged if the client undertakes a transaction. Any work the adviser does that does not lead to a transaction is effectively a new business acquisition cost spread across clients who do transact. Cross subsidies that would not be permitted for fiduciaries are best avoided by advice firms. But this is true of all efforts to attract new business, even in fiduciary services. It should be obvious to prospective clients of any business.
This is not the same as saying that contingent charges are inherently a source of conflict amongst the population of clients who do transact. There are cross subsidies amongst them that are introduced when charges are based on the value of the transaction whereas costs per transaction are distributed differently (or theoretically could be the same for all). But contingent fees do not have to be set by reference to the value of the transaction. Almost certainly there are costs, loosely defined, that are only crystallised when the transaction occurs and that are a function of the size of the transaction. These costs are not necessarily distributed the same way as value to the client. The client may be indifferent to them but only because they do not see the implied cross subsidies. Such size-related costs are regulatory fees, insurance and contributions to own funds required as a cushion for uninsured losses. Our internal costings led us to a proportion of 10-15% that was risk based and the balance cost or value based and so best met by a flat fee. More about that later.
Transfers: particular but not special case
This conceptual framework for charging can readily be applied to transferring out of a Defined Benefit (DB) pension scheme. But that is because it is a particular case of general principles, not a special case.
Transferring out offers the chance of a continuing fee stream for an investment manager. This has a much higher net present value than the cost and profit of the advice leading up the decision. But it has this in common with all new asset-gathering business that you either win or lose. A DB scheme member has to get advice if the transfer sum is over £30,000 but neither existing nor the proposed regulations make it a requirement the advice is provided by a party who cannot profit from the replacement investment. In fact, the new proposals, which follow the guidelines issued earlier, make it more likely the firm willing to provide advice will be willing only if they are likely to be managing the capital that has to replace (usually by drawdown) the DB income stream.
Bias dealt with in the advice process
The protection for the client does not come from the charging approach but rather the conduct of business rules around personal suitability and appropriateness that aim to prevent biased outcomes and to provide a clear audit trail that can be used to test after the event, in the case of complaints, whether the rules were applied. Opponents of contingent charges may find it difficult to grasp this as a logical if imperfect accommodation between consumer protection and normal commercial risks but they may never be satisfied by compromise.
How this suitability protection applies to DB transfers is in fact simpler to identify than many financial decisions. This is because it is a case of suitability depending on risk attitudes that can be directly discovered by testing a client’s personal indifference between a certain (ie risk free) outcome and a set of possible but uncertain (risky) outcomes. Indifference addresses the client’s ability to choose readily between the quantum of the risk free ‘hedge’ and the possible quantum of the risky ‘bets’. This is not abstract. As long as it is outcomes that are being quantified, such as real sustainable gross income or draw from capital, it is possible for the client to think directly about costs and consequences. (It’s referred to as ‘indifference’ because of the information about risk aversion contained in the shape of the payoffs to risk that the client finds to be just as valuable or beneficial as the risk free alternative. They are indifferent between the two.) A transfer makes explicit both the correct measure of the risk free alternative (the DB income and its Cash Equivalent Transfer Value) and its form or nature (in this case laying off each of capital market risk, inflation risk and longevity risk).
Both the old TVAS report format and the new required calculations make it impossible for the adviser to misconstrue the right measure for testing risk preferences. The question is, why does this regulatory requirement not apply to all decisions where risk attitudes need to be tested for by costing the risk free alternative?
Because this is about visualising consequences, it is easier to solve for the correct risk approach when the consequences are expressed as constraints on the possible alternatives. Constraints are also easier to express for a DB transfer than for many other personal goals, and typically include how prudent the assumption about how much the capital must last, the minimum tolerable draw rate or the maximum tolerance of cuts to the draw rate. The only difference the new proposals make to this exercise is that in most circumstances the planning that discovers true risk preferences will need to be holistic, not limited to the CETV, and including all the resources assigned to meeting retirement spending. Moreover, the replacement of the risk free DB income that is already implicitly part of a retirement goal should logically alter the existing risk preferences for any other capital assigned to retirement funding because that underpinning is being removed. These dependencies are one of the aspects of the proposals that make it more likely the adviser and manager will in future be the same firm.
As mentioned in our comments on the FCA paper, we find it ironic that it was the creation of a particular advice permission for transfers, designed to replicate actuarial approaches, that did in most cases make the transfer adviser and the replacement adviser or wealth manager separate entities. Ironic because, by weakening the personal suitability tests, it left consumers more rather than less exposed to exploitative, biased advice. That is an excellent reason for supporting the main thrust of the FCA’s proposals which is to bring it back into the mainstream of risk choices.
A combination of non-contingent and contingent
How then should a client expect to pay for transfer advice that is generally consistent with these principles but also specific to the nature of the transfer options? We find a degree of commonality about this amongst those transfer advisers who are motivated, like us, to provide ongoing holistic management of wealth (even when it is not organised, as in our case, to meet separate goals). We typically observe three components:
A quick and dirty test of the prima facie case for a transfer increasing personal utility
A full analysis and report
A possible transaction to be executed
We believe most firms, like us, do not charge for the first. It is a form of new business acquisition cost but not a large one. As such, it is also probably testing whether the overall management proposition (assuming a transfer) makes economic sense for both parties. Each party is qualifying the other.
The second is where the client probably assigns most of the value: getting to a good decision with clarity, composure and not too much time or effort. Unfortunately, in reality, clients will not typically assign the same value if the good decision (in spite of the first test) is to stay. They will resent paying the same as if they transferred. Though advisers must lean against this behavioural bias, there is not much they can do to prevent it. It is one of the reasons why the third charge, contingent on the transaction, is often higher than the second. The requirement that the advice be holistic will put upward pressure on this component. Indeed, it may be linked to, or replaced by, a separate charge for a holistic financial plan of which the transfer advice is only part of the deployment scope of the plan.
This runs into another constraint that may be either ‘true’ or a behavioural basis. Clients may resent paying ‘hard cash’ from a non-pension pocket as opposed to paying out of a pension pocket into which the CETV value has been paid. I’m sure the FCA understands this constraint even if opponents of contingent charges don’t.
Finally, the illogical HMRC VAT treatment between advice fees and transaction fees may be contributing to a bias to more of the fee being contingent. Clients go along with the tax saving.
The third charge is and should be a perfectly genuine contingent fee, even if that is only part of it (if the second element is also loaded onto it). As noted, advisers know full well that the liability for advice goes with execution, even if it is tested on the advice process. You are very unlikely to be sued for advising against a transfer or even for choosing to ignore an existing client’s transfer options. There is no logic to this. It is a bias both within the FCA rules and the Financial Ombudsman Service based on the old concept that financial services are less service and more product sales.
The scale of any award against you will be based on putting the client back in the position they were – or its financial equivalent since rejoining the DB scheme is not an option. Hence the size of the potential liability is a function of the amount transferred and market risks impacting it. Regulatory fees and any insurance premium loading are also a function of the size of each transfer transaction. Own funds contributions to meet claims not met by insurance, including dealing errors, are also logically size related.
In our case, because we do robustly enforce the value of the advice process independently of any transaction, as for all financial planning, we do not need to charge a high contingent fee. It is entirely justified by the contingencies it was properly designed to cover. The apparent intention of the FCA proposals to make the planning underpinning a transfer holistic may have the effect that more advisers enforce a fee for planning. But it is likely still to leave a contingent element.
The role of the decision process itself
At Fowler Drew we think there is another way in which firms can demonstrate how inherent conflicts are being managed that does not depend on the charging structure alone. We are a ‘quantitative’ firm. Quants tend to believe that an advice process built around mathematical rules instead of human judgement will make for better and more consistent decisions. But they may also recognise that it is easier to make the quantitative process appear unbiased to the client.
It is perfectly possible, and greatly to be preferred, if the decision process is derived from some optimisation algorithm, such as maximising a defined utility or benefit. The planning process uncovers the form that utility takes, which might be quite complex, and the constraints. It can (as noted earlier) discover directly the ‘true’ risk preferences. It is then a mathematical calculation what option maximises utility subject to the constraints. In the case of a DB transfer option, the answer is not open to agent bias unless the bias has been designed into the algorithm. If the same model is used for all risk decisions involving cash flow outcomes, it will quickly become apparent across much of the firm’s business that the model is biased to risk taking.
The fee model can also help deal with model bias. If the charges for wealth management are themselves fixed, and independent of the amount held in risky instead of risk free assets, there is no point the manager gaming the client (trying to draw more risky assets into the calculation) or the client gaming the manager (by withholding or concealing assets). The adviser is conspicuously indifferent (via its revenues) as to how much risk any client wants to take or how that alters over time.
Flat fees pose their own challenges, of course. It is not obvious that fees should only relate to direct costs. There are many different types of decision offering very different utility to the client, ranging from trivial to mission-critical. As a quant firm, with ongoing decision costs (including what in a conventional firm would be time costs) largely sunk in technology, it was easier to combine these two elements by creating notional licence fees for different portfolio types where the cost is a function of the model's utility but not size. In this framework, young accumulators pay less than older clients in drawdown regardless of how much has been accumulated.
Though there was a logic to each charging choice we made and evolved over the years, it is not necessarily the 'right' formula for all. Many of its features are particular to the way we organise wealth and take decisions. Philosophically, we think that idiversity is good.
I like to think the FCA takes the same view. It wants to see competing business models that encourage innovation in charging. One of the problems of the old commission regime was that it took on the characteristics of a monopoly of a single approach with rates set by colluding product manufacturers and imposed on advisers as 'distributors'. Contingent charges may look a bit like the old regime but I think that is missing important differences.