More on stochastic modelling

Our first online assembly for 2021 was kicked off when Patrick Ingram (PI) from Parmenion ‘popped in’ to answer our questions on stochastic modelling. You can watch the episode here. However, it turned out you are all massive fans of this topic and questions kept flowing, so many in fact, we ended up running out of time to answer them all. Luckily, Patrick agreed to answer them all in a follow up blog post, and that’s exactly what you will find here.


Please can you provide an example of how a stochastic model might highlight risk of the likelihood of success of a planning firm’s solutions.

PI: What a robust retirement planning stochastic model will provide, is a calculation of the combined probability of both your chances of being alive at any age up to say 110 AND your investment capital being able to fund your withdrawal plan at each stage. So, it can supply what might be described as “the gambling odds” of a plan being successful in both dimensions of time and investment outcome. This is extremely useful in proposition oversight, and advisory business risk management, at the firm level, in addition to underpinning planning for specific individuals. The caveats are that the model itself needs appropriate oversight, the assumptions set need to be credible and the plan inputs need to be coherent with the client’s goals and needs.

(About the Parmenion IMT) Is the tool best used interactively?

PI: My sense is that the Parmenion IMT can be used highly satisfactorily with clients at the start of their retirement planning journey. When it comes to the judgements needed at the moment of retirement, you would be unlikely to build and model from scratch with a client present. You would demonstrate sensitivities and options to an analytical client once you had decided on your broad recommendation. It’s not my picture of how most people want to take advice, that they want a complex proof of the adviser’s judgement calls, face to face, it’s more important to be able to reflect you understand the needs and goals and the sensitivities to risk and uncertainty. This latter point is pretty key. One of the most important inputs to a retirement plan is a view on how much “income uncertainty” an individual can (emotionally) sustain. That’s subjective and involves understanding their spending habits and mentality around money. That does not come from the model, it comes from talking things through with the client. Another point of note is that good modelling can sustain CRP design, and that is key attribute of IMT.

CAPE Shiller says markets (US) have been overvalued for 10 years+, don’t we need to be uber careful if using current forecasts?

PI: I think we are probably in agreement. JP Morgan’s end 2020 10 year forecast for US Equity is 4.3%pa against over 6.5% from the UK, in Sterling terms. You have to make some kind of forecast if you want to go into drawdown. From my personal perspective, at 60, I am looking at the central scenario for mid risk multi asset returns being close to zero in real terms (before AC) over the next 5 years. Higher and lower returns are entirely possible of course. But it’s not a sparkling picture. What it focusses attention on is strategies which will enable your drawdown portfolio to take slightly more equity risk. So the thought is, what’s better the better bet for retirement income: £100k in Risk Grade 4 or £20k in annuity and £80k in Risk Grade 6, for example? And at what combination have you the same overall level of “risk” or perhaps the right overall level of certainty?