There has been a dramatic increase in the popularity of drawdown products since April 2015. It is likely that many, when they opted for these choices, did not realise all the risks they faced.
A well-informed pensioner might well know about the risk of living too long or not getting a good return on his or her pension pot, but will probably not have realised that the timing of investment returns is also crucially important. Getting off to a bad start can ruin even a cautious drawdown strategy.
Current equity market volatility and the prospect of further falls should give us pause for thought and concern for the many pensioners who took advantage of the Pension Freedoms last year and opted for drawdown. As an industry, we should be asking ourselves how we can help them avoid running out of retirement savings midway through their retirement – a fate that has befallen many pensioners in the US and Australia, where pension freedom has been available for decades.
To get a true picture of the risks that pensioners face, and to enable successful ongoing management during retirement, advisers and pensioners need to use a stochastic asset model to show the range of potential outcomes. Deterministic projections are worse than useless. For the most part, the assumptions used (often a best estimate, low and high returns) are based on little evidence and, where there is some analysis, it is historical and not forward-looking. As we all know because it has been drilled into us by the FCA, “past performance is no guide to the future”.
If this were not enough to condemn deterministic projections, the final nail in their coffi – as a means of managing drawdown – is that there is no allowance whatsoever for the ups and downs of investment markets. The timing of investment returns with income drawdown is a critical risk factor. Poor returns in the early years of drawdown, when retirement savings are at their highest, can do irreparable damage.
So is the answer to use a stochastic model to manage income drawdown? Well, sadly, this is not quite so simple. The term ‘stochastic’ is a generic term, and stochastic models can differ widely in terms of how they are built and what they are suitable for. There are two basic types of model:
• Mean, variance co-variance (MVC) models, and
• Economic scenario generators (ESGs).
MVC models are very simple, and are primarily used for making short-term tactical asset allocation decisions. They use a single set of assumptions about the expected return for each asset class, its volatility (or variance) and how asset classes move relative to one another (co-variance). The set of assumptions can either be based on current market conditions or on views about long-term “normal” conditions. What an MVC model cannot do is tell you anything about how markets might progress from current conditions into the future. MVC models provide a ‘snapshot’ at a single point in time.
ESGs, on the other hand, model the economy and investment markets from today’s current conditions into the future. An ESG models the ‘investment journey’ and is, therefore, the appropriate type of model for forecasting retirement income outcomes from drawdown. The chances of low returns in the early years of drawdown can be seen, and plans can take this risk into account. MVC models provide no information about market movements from year to year.
Because they ignore the ups and downs of investment cycles, MVC models are not suitable for use with pensioners wishing to plan income withdrawals. Furthermore, MVC models can give wildly different views of the future depending on the assumption set used for projections. By contrast, because ESGs start from the same place – current market conditions – and chart a course into the future (which may be different depending on the model used), the differences in forecasts produced by different ESGs are less pronounced.
While the technical details of stochastic models are for many readers a complete turn-off, they do matter, and using the wrong model will potentially have very damaging consequences for pensioners. For an easy way to see why MVC models are not suitable for constructing drawdown plans, one simply has to look at the product bias that they can produce. All the models illustrated below are real models used in the UK. The diagrams below show the median (‘best estimate’) and the low forecasts. The optimistic outcome is not shown because to make the decision between income drawdown and annuity, the important comparisons are the best estimate (that is, what is the most likely) outcome and the worst outcome (if things go badly).
‘Normal’ conditions
What the diagrams show is that, depending on the assumptions used in the MVC model, that is, based on (a) long-term ‘normal’ conditions or (b) based on current market conditions, very different conclusions on the relative attractiveness of annuities and drawdown would be reached. In the first example, assuming long-term normal conditions prevail into the future, there appears to be considerable upside in using drawdown and little downside compared to an annuity. The bias is clearly towards using drawdown for delivering retirement income. If current market conditions are assumed in the MVC model, precisely the opposite conclusion would be reached. There is no upside on a best-estimate basis and considerable downside if the worst case scenario occurs – an income of £25,000 will exhaust pension savings by age 75.
As can be seen, the two ESGs give a more balanced picture and the differences between models tend to be less extreme. On a best estimate forecast there is some upside to balance the risk that the income of £25,000 a year can only be sustained until age 75. The choice between drawdown and annuity is therefore driven by the amount of risk the pensioner is prepared to take.
If MVC models are capable of producing such a distorted picture of the comparison between drawdown and annuity, they are equally unreliable for helping a pensioner manage drawdown in retirement. What is needed is a model that shows how investments may move from current market conditions into the future and this is precisely what an ESG does.
The main conclusion to be drawn is that due diligence needs to be undertaken on the ‘stochastic black box’ used for forecasting retirement income. The wrong model has the potential to create serious problems for pensioners. Currently, there is no transparency or requirement for the providers of stochastic models to make any information available. This makes comparisons between models almost impossible and due diligence extremely difficult.
The FCA needs to set out some disclosure standards for the producers of financial planning tools and stochastic asset models. This should be done before problems arise, not afterwards, when the claims for compensation from pensioners start to roll in.
Bruce Moss is founder and strategy director of risk profiler eValue
Key Points
Getting off to a bad start can ruin even a cautious drawdown strategy.
Because they ignore the ups and downs of investment cycles, mean co-variance models are not suitable for use for pensioners wishing to plan income withdrawals.
Due diligence needs to be undertaken on the ‘stochastic black box’ used for forecasting retirement income.
