What is Monte Carlo and What Does it Have to Do with Your Retirement Plan?

Dana Anspach, the president and founder of Sensible Money, explains that Monte Carlo is a form of "stress-testing" your future financial wealth against various market conditions.
Publish date:

What is Monte Carlo?

How might you go about using Monte Carlo simulations for your retirement plan?

Well, first you need to understand what Monte Carlo is and what it isn’t.

According to Dana Anspach, CFP, RMA, the president and founder of Sensible Money, it’s a form of "stress-testing" your future financial wealth against various market conditions.

Monte Carlo is based on using an assumed rate of return and an assumed level of volatility as measured by standard deviation, she said. Read How to Use Monte Carlo Simulations for Retirement Planning.

So, for example, for an asset class like large-cap U.S. stocks, you may have an assumed rate of return of 10% with a standard deviation of 15%. About 68% of values would be within one standard deviation, so in any one year you may expect large cap stocks to be up by 25% or be down in value by 6% - but of course you have an additional 32% of outcomes that can fall outside those ranges, said Anspach.

With Monte Carlo, you can change certain assumptions to stress test different scenarios. For instance, the user may choose historical returns, or forward-looking returns. “If you used historical returns for the S&P 500, you'd assume an 11% average return, but if you were of the view that large-cap U.S. stocks are currently overvalued, you may assume a lower average going forward,” she said.

According to Anspach, some forms of Monte Carlo analysis test the portfolio as a whole, and others allow you to assign a return and standard deviation to each underlying asset class and test the probabilities of good or bad outcomes based on the assumed allocation.

So, what are some of the pros and cons of Monte Carlo?

Any form of stress-testing is an improvement over using only an average return or linear projection, said Anspach. “One of the biggest criticisms of Monte Carlo isn't in the testing itself, but in the way the results are framed as a probability of success or failure,” she said.

In recent years, Michael Kitces, a retirement researcher and author of the blog A Nerd's Eye View, shifted the discussion to something more meaningful by talking about the magnitude of failure and the probability of adjustment.

More recently, Derek Tharp, one of the researchers for A Nerd's Eye View, addressed another fascinating nuance - the meaning of results for a one-time test versus how to apply it if you do ongoing testing. According to Derek's research, with ongoing planning and the willingness to make adjustments, a 50% probability of success may be perfectly acceptable.

Another con of Monte Carlo, according to Anspach, is that it can test scenarios that, in reality are unlikely to materialize. “For example, suppose the Monte Carlo analysis is using forward-looking assumptions of lower-than-average interest rates, and the user wants to test their expenses against a higher-than-average inflation rate,” she said. In reality, higher inflation usually correlates to higher interest rates - we've seen that recently with bond yields rising in anticipation of higher inflation - so a real-life scenario of low interest rates and high inflation is unlikely to happen.”

In addition, Anspach noted that equity markets and asset classes tend to move in cycles. “While momentum can carry a stock or an asset class to valuations far higher than we might initially think possible, eventually, something accrues enough value that investors begin to sell,’ she said. “On the opposite side, if a selloff has gone on for many years, whether it be a stock, an entire asset class or real estate, eventually things become priced at a level that people begin to buy. In a Monte Carlo simulator, however, the random probabilities assigned can project an asset class underperforming far longer than what may actually happen in reality.”

So, should you incorporate Monte Carlo into your retirement planning?

According to Anspach, you need to answer the following questions to determine how to apply and interpret the results of your Monte Carlo simulations.

  • What are the assumptions based on? How does the Monte Carlo work?
  • Because it can be testing scenarios outside what reality is likely to deliver, rarely would you need a 100% success rate.
  • What is your willingness to make future adjustments?
  • And is this a one-time test, or will you be engaging in ongoing planning and testing?

Also, of note, there are other forms of stress testing your retirement plan, according to Anspach. Those include:

  • Fundedness. This concept used by pension plans that uses the present value of future cash flows needed.
  • Historical audit, or aftcasting. "Aftcast” is the antonym of “forecast.” Instead of making assumptions to “guess” a future outcome, an aftcast uses the actual market history to show what would have happened in the past without any predictive claim for the future," according to Jim Otar, author of Advanced Retirement Income Planning