Rethinking Monte Carlo Simulations
Monte Carlo simulations are a common tool in the financial adviser’s toolkit. They can help us frame the likelihood of success of a given course of action by applying a large number of statistical simulations of future market returns to your financial plan. The results are commonly distilled down to a single number: the probability of success (i.e. in what percentage of the hundreds or thousands of simulations did you not run out of money before the end of the plan?). Monte Carlo analysis also allows us to illustrate uncertainty by showing the large range of possible outcomes of your plan.
On paper, this approach to financial planning makes a lot of sense: We can quickly compare the relative impact of changing certain inputs (e.g. savings rate, investment allocations, retirement age, spending rates, tax strategies, etc.) to illustrate the likely value of possible courses of action. For example, “Look at how much sooner you can likely afford to retire if you boost your savings rate by 5%!”
However, employing Monte Carlo analysis as an ongoing real-world decision-making tool often results in misunderstanding, anxiety, over-confidence, and/or being overly conservative.
What does success even mean?
Nobody wants to run out of money in old age. This is a real fear among retirees and near-retirees, but having a plan that addresses this valid concern is table stakes. Moreover, success can take many forms, and there are many other more meaningful markers of a life well-lived. A comprehensive financial plan should help you align your time and your money with what you value most. After all, what good is it if you reach the top of the ladder only to discover it’s leaning against the wrong wall?
When failure is not an option, we adjust
Assuming that you have the level of affluence to be seeking out the help of a financial advisor, the truth is that almost nobody wakes up one morning having run out of money. If things aren’t going your way, you will tighten your budget, work a couple more years, or decide to move to a less expensive area. The binary success or failure framing of the Monte Carlo analysis belies the reality that a reasonable person will make adjustments over time as the future unfolds, such that total failure almost never happens. Even relatively small adjustments can make a large impact when applied over long periods.
Monte Carlo simulations are backward-looking; they show only the probability that an adjustment would have been needed to keep you from running out of money once you reach the end of the plan. We don’t get the benefit of hindsight, and as traditionally employed, the standard Monte Carlo analysis doesn’t tell you when you should make an adjustment or by how much.
The traditional Monte Carlo analysis often leads to over-conservative plans
Fixating on the probability of success often results in a plan being overly conservative. From grade school, we were conditioned to think of percentages in terms of letter grades (87% is a “B+”) or that a 95% confidence level represents statistical significance. In a world where we’re going to set a plan in motion and then never change it, shooting for 95% may be a reasonable place to start. However, this almost assures that your plan will be too conservative.
If you are willing to make small manageable adjustments, you likely could have retired years earlier, you could have enjoyed yourself a little more when your kids were younger, you could have taken that sabbatical, or made an impact in someone’s life with a gift or charitable donation. Instead, those who just focus on minimizing the chance of failure will likely reach their later years with far more in the bank than they will ever need, and “you can’t take it with you.”
What retirees actually care about:
When it comes to their withdrawal plan, retirees want clear answers to two main questions:
How much can I expect to spend per month?
When would I need to adjust that spending rate, and by how much?
Say the market hasn’t done well lately and your probability of success has gone down from 90% to 75%. You are naturally concerned, and you want answers. Monte Carlo analysis doesn’t answer these essential questions and often leads to anxiety and over-correction.
This is not to say that Monte Carlo analysis is not a valid or useful exercise- far from it. Rather, we have to rethink how we communicate the results and how we employ them within the context of an ongoing, dynamic planning process.
Certainty is rarely on offer in life, whether financially or otherwise. However, having a solid plan combined with a flexible mindset to adapt and make changes along the way is the best way to find success that is meaningful to you. Having a guide who knows the terrain and who has the tools to adapt to the unexpected can be a big help too.
Dynamic Guardrails Approach
In a future article, I’ll explore the dynamic guardrails approach to retirement planning, which addresses many of the shortcomings of a stand-alone Monte Carlo analysis. The guardrails framework gives a retiree (or soon-to-be) a concrete idea of their monthly safe spending capacity based on their retirement savings and other income sources. It uses a Monte Carlo engine behind the scenes to establish those initial spending levels as well as risk-based guardrails to let you know when it’s time to adjust your spending up or down. (i.e. if your investments fall to X dollars, we will decrease spending by Y dollars). By expressing the guardrail and spending amounts in actual dollars, the client has much greater clarity and comfort that they will be able to adapt if needed.
