There are no guarantees in life.
But that hasn’t stopped the financial planning industrial complex from trying to provide certainty to what is really hard to predict: how long your money will last in retirement.
In the old days, starting in 1994 with Bill Bengen’s seminal study, financial advisers estimated how long your portfolio might last using historical returns and a safe withdrawal rate. For those unfamiliar, Bengen’s research left us with the 4% rule, which is considered (rightly or wrongly) the holy grail of retirement planning in some circles. That rule says you can safely withdraw 4% per year from your nest egg over the course of 30 years and not run out of money.
Then, starting in 2005, investment firms and advisers were given the green light to use something called Monte Carlo to predict your portfolio’s probability of success — success being the probability that your nest egg would adequately fund your desired standard of living throughout your retirement.
This, they argued, was an improvement over Bengen’s approach because a computer was running thousands upon thousands of scenarios on your portfolio. And these scenarios would stress test your portfolio using various market returns, inflation rates, withdrawal rates and the like.
And the best-case outcome would be to assemble a portfolio and withdrawal rate that delivered a 70% or better probability of success. Not surprisingly, Monte Carlo has gained widespread adoption and advisers are now using (some might say overusing) it to provide comfort to investors worried about outliving their assets, and to demonstrate their value.
New research, however, is now challenging conventional wisdom. Today, you can settle for a 50% probability of success – provided you’re willing to delay your retirement and adjust your spending now, or in the future, according to a report by Derek Tharp, a lead researcher at Kitces.com and an assistant professor at the University of Southern Maine.
As an aside, any plan that calls for you to reduce your spending, to lower your desired standard of living, is a failure in my book. But that’s just one man’s opinion.
According to Tharp, all plans, even those with a 95% probability of success, require spending adjustments along the way. What’s more, at least compared to the other approaches, Monte Carlo offers you the opportunity to be approximately right rather than exactly wrong.
“If you’re going to be updating on an ongoing basis each year, adjusting your spending up or down as the market would allow, it’s really the market that’s driving your overall spending level and not so much the probability of success you choose,” Tharp said in an interview.
Given that, should advisers even be presenting Monte Carlo results to their clients? And, if so, how should they be using it?
Stop talking about Monte Carlo results
The arguments against using Monte Carlo — at least with investors — are many.
For one, investors “lack the numeric skills needed to accurately assess probability and because cognitive biases cause most people, including experts, to be insensitive to probabilities, neglect them completely as risk becomes more vivid or of greater magnitude, or view probability negatively,” James Sandidge, a principal of the Sandidge Group, wrote in a paper recently published in the Retirement Management Journal, an academic journal I edit.
Listen to our interview with Sandidge
Others, including John Nersesian, head of adviser education at PIMCO, also say not all investors can relate to the “odds of success.”
Investors, he said, may have difficulty in reconciling the difference between a 50% versus 70% probability of success. What’s more, he notes that investors often define success or loss in dollar terms and may not be able to process the significance of a higher or lower probability.
Another reason to stop using Monte Carlo with investors: It might work when it comes to saving for retirement, less so in the drawdown phase or what some call the decumulation phase.
“Accumulating wealth is a linear process and predictable, but in the nonlinear world of retirement income, returns and standard deviation are not predictors of success and therefore are unreliable inputs for Monte Carlo analysis,” wrote Sandidge.
Retirement income, said Sandidge, is governed by chaos theory, which makes it unpredictable.
Others agree. Generically, Monte Carlo analysis requires laying out the probability structure for the way that the future will evolve, said Michael Zwecher, author of Retirement Portfolios. “It’s not that the analysis can’t take into account rare, or non-normal, events, but that a great deal of thought needs to go into the analysis beforehand,” he said. “What isn’t thought out beforehand is omitted; but like color blindness, the analyst may not be aware of what they’re not seeing. It’s not so much the probabilities that make Monte Carlo analysis tricky to use, but the uncertainties — events unforeseeable with any useful probability — that require caution.”
Sandidge also noted in his paper that Monte Carlo takes a systematic approach to risk and cash-flow allocations, but advisers are likely to adjust those allocations, and the butterfly effect means each adjustment can cause outcomes to change significantly. Thus, flawed inputs and systematic approaches lead to flawed outputs that are likely to have little correlation to a real-world setting.”
Others agree. Monte Carlo is heavily dependent on the inputs provided, including capital market assumptions, asset class correlations and retiree longevity, said Nersesian. “A small change in any of these inputs can produce a significantly different projection,” he said. “Think ‘garbage in, garbage out.’”
Behavioral finance biases, including loss aversion, mental accounting and anchoring, may also limit the utility of Money Carlo, said Nersesian. “Successful advisers need to understand and identify the behavioral makeup of their clients in order to use Monte Carlo most effectively in guiding successful client decisions.”
There are at least two issues to consider when thinking about Monte Carlo simulations, said Branislav Nikolic, vice president of research at Cannex Financial Exchanges Limited. The first, to quote an aphorism generally attributed to the statistician George Box, is that “all models are wrong, but some are useful.”
In other words, Monte Carlo is just a very laborious sampling exercise, said Nikolic. “And the more you try to capture with a single model, the more you’re susceptible to a model error…and the more you try to accomplish the worse it is.”
Nikolic also takes issue with probability of success as a measurement without sufficient qualification. “What does it mean? Does it mean for instance that you have 95% of your income needs covered and only 5% is left to chance? That’s much different from having a portfolio and calculating its probability of survival with regular withdrawals.”
Plus, Monte Carlo doesn’t really address the magnitude of a portfolio failing nor the timing.
The bottom line: “Eliminating ‘Monte Carlo’ from conversations should lead to safer, simpler, and more personalized retirement income portfolios for investors,” wrote Sandidge.
How Monte Carlo can help
To be fair, experts say Monte Carlo is not without its benefits. “While Monte Carlo forecasting is far from perfect, it offers significant advantages over the linear forecasting approach that many investors rely on,” said Nersesian. “Linear forecasting assumes a fixed/constant return over a retirement horizon with no variability, an investment option that doesn’t exist in reality. In other words, it includes returns while ignoring risk in the planning process.”
Monte Carlo, he argued, also provides an advantage as it considers both the variability of returns and the random potential sequences of the returns. “As we know, averages can hide lots of sins,” he said. “An individual with their head located in an oven and their feet in an icebox has achieved an average temperature that may not be comfortable.”
Monte Carlo, Nersesian said, may be particularly helpful in the retirement distribution process, as success in retirement is primarily driven not by the average rate of return, but how the return is achieved — consistency and sequence of returns. “It provides a range of potential outcomes — in dollar terms that clients can appreciate — as opposed to a single projected outcome,” he said.
According to Nersesian, the range of outcomes is illustrated as future dollar values of the portfolio at a specific date, not the distribution amount. In other words, instead of showing a projected future value of $2 million at age 65 based on a single average return experience (linear forecasting), it shows a range of potential portfolio values, say $1.75 million to $2.25 million, based on various return and sequence assumptions.
And in the distribution phase, Nersesian said the adviser/investor can play “what-if” games where they would adjust the expected returns, retirement dates, and the like to show various potential distribution amounts.
“In other words, if an investor delayed their retirement date by two years — or saved additional funds, invested with greater return potentials, and the like — they could see the impact on the potential distribution amounts.”
And lastly, Monte Carlo provides, according to Nersesian, a personalized perspective on retirement: Two investors might both retire with the same amount of assets at the same age, but we will endure very different 30-year retirement experiences, depending on when they begin their retirement journey and the actual capital market returns earned over their retirement horizon.
So, what’s an adviser to do? What’s a retiree or pre-retiree to do when it comes to Monte Carlo?
Monte Carlo’s best use, in some circumstances, is not for client presentation at all but to provide the adviser with a tool to test their portfolio allocation recommendations and financial planning guidance, Nersesian said.
Think of it this way: Rarely do advisers talk about the science of portfolio management with their clients. Rarely do they mention things like Sharpe and Treynor ratios. But often, that’s what advisers are using to evaluate the performance of the portfolios they manage for clients. Perhaps the same should go for Monte Carlo.
Monte Carlo be used a conversation starter, Nikolic said, one that acknowledges, among things, that results will change given a different level of spending or withdrawal rate, and will even change year to year as well as after various lifecycle events (a death, a divorce, a remarriage and the like). In other words, the results — the probability of success — aren’t set in stone.
Zwecher, meanwhile, has a similar but slightly different take on how to use Monte Carlo as a conversation starter.
“In retirement income planning, a clustering of rare events can either be incredibly favorable, or incredibly disastrous for a client’s portfolio,” he said. “And given that, Monte Carlo analysis can provide a view useful for back-of-the-envelope planning, but the average path is of little use when the client only gets one shot at retirement. The worst paths ‘can’ be used for planning — less for what may happen, and more for planning how to respond to the arrival of bad outcomes.”
In an interview, Sandidge also said he thinks the average path is of little use. He recommends examining your probability of success given a worst-case scenario. And if your plan can succeed in the worst-case scenario that should provide some degree of comfort.
The good news is this. Just as Monte Carlo replaced Bengen, a new better tool may replace Monte Carlo.
The tool — value at risk or VAR — is gaining traction in planning circles, said Nersesian. VAR is for investors who relate better to financial outcomes in dollar terms versus percentages or probabilities. “It advances the risk management process by quantifying the maximum dollar loss that a portfolio may achieve during periods of market stress,” he said. “It provides advisers with a tool that can help them speak their client’s language.”