Optimisation isn’t usually the first word that comes to mind when thinking about retirement. Some people have grand dreams, others hope to avoid poverty; many avoid thinking about it. Here at EvolveMyRetirement® we’ve recognised that optimisation is an important concept in retirement-planning models. And finances are the bedrock of a fulfilling retirement.
What is optimisation?
According to the Merriam-Webster online dictionary, optimisation is “an act, process, or methodology of making something (such as a design, system, or decision) as fully perfect, functional, or effective as possible”. The examples in this definition include decisions, and this is what we’ll focus on here. The kind of optimisation we’re talking about is the kind that helps explore which decisions may perform better under different assumptions. When a grandmaster makes a move in a chess game, his brain is attempting to make the optimal move. If he can see far enough ahead, he could theoretically make the objectively best move. There’s no random uncertainty. Financial decisions, however, involve uncertainty, so optimisation focuses on exploring which choices may perform better under different assumptions.
Let’s look at a very simple example where we optimise a financial decision. Say we have £1,000 to invest, and we know we won’t need it for another year. We have the choice of two savings accounts (assume there are no other possibilities). One pays guaranteed 3% interest and allows instant access. The other pays guaranteed 4% interest, but requires 90 days notice. We would need to decide into which of them we should place our £1,000. An optimisation model would typically identify the higher‑interest account as producing a higher expected return, assuming the saver is comfortable with the notice period.
That was an easy decision, between two choices, each of which has known consequences. Real-world decisions are not usually so simple. There are usually many more choices, and the consequences are often uncertain. Take investing in the stockmarket, for example. There are hundreds of companies to choose from. And the likely return on investment will be different for each, as will be the risk of losing money. Optimisation of investment decisions is incredibly hard, even for the most experienced experts.
Pension drawdown and optimisation
One of the classic problems in retirement planning is drawing down income from a pension pot. Let’s say you have £100,000 in your pension plan. How much could be withdrawn each year while aiming to avoid exhausting the funds during your lifetime? This is a problem in optimisation, and a hard one. It requires you to make a number of assumptions up front:
- Your life expectancy.
- The expected return on investment from your pension funds assets.
- The risk (volatility) of your pension fund investments.
- Whether or not your withdrawals are fixed, inflation-linked, or dependent on pension fund performance.
- The degree of risk you’re willing to accept for running out of money.
In the 1990s, to help solve this optimisation problem, William Bengen came up with the 4% Rule, which has since entered retirement planning folklore. This rule is often discussed as an illustration of historical research rather than a guide to what individuals should do. Bengen analysed the performance of a 50-50 mixture of equity and bonds over a 50 year period. He assumed a starting withdrawal of 4% of the initial capital. Each year, the withdrawal amount was increased in line with inflation. He found that, even during the worst market downturns, money never ran out in less than 33 years. There are, however, a number of important caveats:
- Life expectancy is limited to 33 years.
- It assumes a particular investment approach (50-50 equity and bonds).
- It relies on historical (and US) investment performance, which may not be a good indicator of the future.
- 50 years’ worth of data is a very small statistical sample.
- It’s limited to strictly inflation-linked withdrawals.
- It assumes zero tolerance for risk, and aims for a 100% success rate.
Despite its limitations, the 4% rule serves as an illustration (but nothing more) of financial optimisation in action. Only one variable was attempted to be optimised though: the drawdown rate.
Juggling competing decisions
If only retirement planning optimisation were as simple as optimising only the drawdown rate! In reality there are many different decisions that need to be made, and traded off against each other. For example:
- How much to save into a pension plan.
- How much to spend (discretionary spending, that is).
- The type of investment risk (e.g. equity versus bonds or cash).
- Whether and how to vary investment risk over time.
- Whether any assets should be converted into lifetime annuities, and if so how much.
- How much to draw down from pension funds.
- Whether to take out a pension cash-free lump sum all in one go, or to phase it.
- How much to draw down from other assets.
These cannot realistically be optimised in isolation. For example, spending and saving are on opposite sides of the coin. Increasing spending may boost enjoyment of life, whereas the intention of increasing saving is security. Increasing one necessarily means decreasing the other. As another example, consider buying annuities, which is intended to reduce risk, but can sometimes also reduce income compared to drawdown. This may mean reducing spending, unless something else is changed, for instance increasing the investment risk and potential returns of other assets.
Changing one variable can have a knock-on effect that an optimisation engine would need to compensate for by changing another. And to make it even harder, uncertainties abound, for example:
- Inflation will vary.
- Investment returns will fluctuate.
- Life expectancy is uncertain.
Juggling the optimisations of multiple variables when there’s uncertainty is an extremely hard problem. Many retirement planners consider it as much an art as a science.
Optimisation in EvolveMyRetirement®
There are numerous online retirement calculators, some more sophisticated than others. A few of the more sophisticated ones have the ability to run Monte Carlo simulations in order to test the effects of uncertainty. This involves running thousands of simulated test scenarios, and statistically analysing the results.
However, EvolveMyRetirement is a distinctive online calculator app that not only runs Monte Carlo simulations, but is also based on a Genetic Algorithm. This is an optimisation technique that uses simulated evolution to align a modelled financial strategy with objectives entered by the user. Hence the name of the calculator.
Users of EvolveMyRetirement have sometimes reported that the strategies generated in this way are surprising and counter-intuitive. Unlike a human, the app is not swayed by emotion or preconceived notions. It evaluates each potential strategy by how closely the likely outcome aligns with the user’s objectives. So for example, you might have intended to finish drawing down your pension before you touch your ISA investments. EvolveMyRetirement might determine that the other way round leads to better simulated outcomes based on the modelled inputs. The reasons might have to do with income tax, or maybe with eventual inheritance tax. Such a result is not presented as a recommendation, and it’s up to you whether you consider it appropriate for your actual circumstances.
The ‘noise’ created by uncertainty means that optimisation is not the same as perfection. For this reason, regenerating a strategy can lead to slightly different results. You may consider sharing your generated plan with a financial adviser, who can provide an expert opinion.
Without such a strategy-generation feature, based on an underlying optimisation engine, the alternative is often trial and error. This might be used to fine-tune a plan, but it’s an error-prone and time-consuming approach when starting from scratch.
Optimisation can be a useful way to explore retirement options.