Trouble With Averages, And (Deterministic) Cashflow Planning

Abraham Okusanya explains why deterministic cashflow planning models, with assumptions based on averages in financial planning, is insufficiently robust.

Never try to walk across a river just because it has an average depth of four feet.” ~ Milton Friedman

If you were to put one foot in a bucket of ice and the other foot in a bucket of boiling water, on average, your body temperature should be normal, or so goes the saying. The danger of sequencing risk and volatility drag examined above demonstrates why using deterministic cashflow planning models, with assumptions based on averages in financial planning, is insufficiently robust. Especially in retirement planning. Sadly, this approach is prevalent amongst the vast majority of UK financial planners. Most of the cashflow planning tools used by planners (such as Voyant, Prestwood/Truth) are primarily deterministic models, and even the highly regarded Certified Financial Planner accreditation is primarily assessed based on this model.

The lack of robustness and rigour aside, perhaps more worryingly, deterministic cashflow planning models risk misleading clients. Because they are often used as tools to help clients visualise likely retirement outcomes, they give clients an impression that their investments grow in a smooth, linear format over time, when in reality nothing can be further from the truth. Add to this the fact that product illustrations are also deterministic, you end up with meaningless data being served to clients, with no real grounding in reality.

Deterministic planning tools convey the illusion of certainty, where there is none. The reality is, investment outcomes are unknown, so why pretend to clients that they are? As we have shown, assumptions of long-term averages are unhelpful. They can be easily thwarted by the powerful combination of volatility drag, sequencing risk and ‘pound cost ravaging.’

These are ever-present dangers in retirement portfolios. Using tools that attempt to simplify potential outcomes, but in the process miss some of the most important factors that could impact the plan’s outcome is clearly an inadequate approach.

Some financial planners argue that ‘whatever the plan is, it would always be wrong’ or that ‘all plans are wrong, regardless of the tool you use’. This argument misses the point, as it is not about being right or wrong. The key is to ensure that the plan is rigorously tested under a range of possible but ultimately unknown outcomes. And why wouldn’t someone use the best tools available for the job? In the absence of certainty, why run oversimplified straight- line projections, when in actual fact you could run thousands of potential outcomes to estimate the probability of success or failure?

Besides investment returns, deterministic modelling tools also assume a linear path for other unknown inputs into the planning process, namely life expectancy and inflation. The concept of sequencing also applies to inflation. Higher inflation in the first decade of retirement means that clients need to increase withdrawals earlier on in retirement to maintain their lifestyle. This could wreak havoc with their retirement plan, even if inflation becomes restrained in later life.

Sequencing applies to inflation as well, because higher inflation in the first decade of retirement means that clients need to increase withdrawals earlier on in retirement to maintain their lifestyle.”

Higher inflation in the early stages of retirement results in higher levels of base portfolio withdrawals, upon which later inflation compounds. Using average inflation, without accounting for the risk of unfavourable sequencing makes deterministic models far from robust.

Some financial planners have also argued that reviewing clients’ financial plans on a regular basis enables them to overcome the serious shortcomings of deterministic models. However, the subtle yet compounding nature of sequencing risk and volatility drag means that it may be too late before it is picked up. Clients may have to rein in their portfolio withdrawals significantly to avoid running out of money. This would be unsatisfactory for clients, and uncomfortable for the planner who gave the initial recommendations.

Another scenario could be that using insufficiently robust deterministic cashflow plans may also lead planners and their clients to take a knee-jerk reaction to lower portfolio withdrawals at the first sight of negative portfolio returns.

The point here is, while regular planning meetings are a key part of helping clients stay on track, it is no proper mitigation for using tools that are inadequate for the job. Of course, some of the short-comings of deterministic models can be addressed by using a geometric mean of returns, but this doesn’t entirely overcome the fundamental flaw of implanting a visual image of a linear path in retirement. Even with very conservative assumptions, it is difficult to take account of how extreme scenarios may affect potential outcomes.