DC measurement for the end saver

At the investment value chain topical day around a month ago, one of the more concrete ideas discussed was the need to provide more meaningful information to the DC member. There was rapid agreement that an annual statement showing the current account balance was not helpful for members. Some even labelled this practice “positively misleading” (an increase in the account balance can be associated with a decrease in wealth if the cost of a future income stream goes up). The proposed improvement was to show members their projected income in retirement (in the UK there is statutory requirement to provide this). There was both enthusiastic support for the idea, on the grounds that this was meaningful information for individuals, as well as significant caution (how far out are we comfortable making projections? How accurate do we think this will be? Does this stray into advice territory?).

This post documents how our thinking has evolved since then, and should be considered as a strawman for knocking around and improving.

We start by assuming that the end game is to fund consumption when employment income has stopped. The individual should be able to adjust their target consumption depending on their circumstances and time preferences. We then proxy that consumption with income, which is subtly different. In rough terms we could categorise consumption as akin to DB cash flow matching, whereas providing an income is perhaps more like DB interest rate hedging. Ignoring all practicality for a moment, surely the ultimate flexibility for an individual would be for them to be able to specify their future cash flow requirements. They would then be able to tailor their different savings vehicles as they wished, perhaps using their DC pension to fund an annual holiday for the first 10 years of retirement.

The levers available to an individual are well understood, and already reflected in most DC modellers. The individual can adjust the target level of income (and potentially whether real or nominal, stable or increasing), and the preferred date of retirement. This generates the ‘DC liability’. We then turn to the asset side, where the individual can adjust the contribution rate and the level of investment risk. We plough through the devil-in-the-detail considerations, and agree on a method for producing projections of future income. DC delivery is then about asset-liability management for the individual.

This brings us to measurement for DC. Essentially the individual needs to know whether they are on target to achieve their desired future withdrawals, and at their desired time/age. The purpose of the measurement is to allow the individual to make changes to their journey as soon as possible, in order to reduce the size of the required change if it is delayed. The measure will therefore involve us in choosing which levers to hold fixed, in order to communicate progress through the variability of the remaining lever. To illustrate, we could assume the target income and contribution rate are fixed in which case the balancing item would be the age at which the individual is expected to achieved a ‘fully funded’ status. Or we could show the required contribution rate if income and age were held constant. Or show changes in the future income. In all cases a range of uncertainty should be shown around the central estimate.

Alternatively, individuals could be shown their ‘success-relative-to-target’ score (aka funded ratio – projected asset value, over projected liability value). This removes the necessity to frame the decision making in a particular way, but carries a couple of disadvantages. The first is technical in that a second number, their elapsed journey time, is necessary for informed decision making. If showing the expected age of retirement, the individual automatically knows the journey time remaining. The second is a question relating to how intuitive the measure would be. And this is likely to depend on whether an intuitive transition can be made from health and fitness apps, where the idea of taking action to hit targets is straightforward.

Whichever route is taken, we believe there is scope to increase the user-friendliness and simplicity of DC measurement, even over the modelling tools already in existence. However, there is considerable intellectual property that needs to be developed and hidden under the bonnet. Is this what we, the investment value chain, should be applying ourselves to?