Monte Carlo Simulations: What Are Their Limits?

    In the last decade or so, a “Monte Carlo simulation”(MCS) became the “gold standard” of retirement planning assistance. Named after the European gambling center, this calculation projects the odds of attaining your retirement financial goals. Unfortunately, MCSs don’t generally include scenarios like recent stock market meltdowns. Though the method runs portfolios through hundreds (or even thousands) of market situations, they assign very low odds to extreme market events. Unfortunately, these so-called “Black Swan events” appear to be happening all too frequently. So how much credence should some one planning for retirement give to MCSs?

    To answer this question let’s consider how simulations work. Though there’s no uniform approach, a good simulation typically has you enter information on: your age, retirement date, value of your investment assets, asset mix, income goals, annual contributions toward retirement, guaranteed retirement income sources, and other details. The calculating program then runs hundreds (or again, even thousands) of market scenarios based on past market performance. It incorporates assumptions on long-term expected investment returns, market volatility, inflation, and other factors. The resulting calculation provides a “rate of success,” that is, what percentage of expected market scenarios allows you to support your retirement income goals.

    It’s important to recognize that MCSs represent a significant improvement over past retirement planning approaches which simply assigned a set return to an asset class and assumed consistent performance every year. However, critics argue that MCSs are hardly a panacea because they fail to incorporate scenarios like the recent stock market swoon. In fact, William Bernstein, author of “The Four Pillars of Investing,” was quoted in the Wall Street Journal as saying, “just add 20 percentage points to the probability of failure your MCS predicts, and you’ll be O.K.” Others complain that different MCS calculators do not make the same assumptions about interest rates, inflation, and market gyrations. Also, many MCSs assume that market returns reflect a “bell-shaped” distribution. This means that the likelihood of a 6% market return (toward the middle of the curve) would be dramatically higher than the likelihood of an extreme return (e.g., a 50% loss occurring at the extreme end of the curve). MCS proponents argue that it should be that way because extreme returns are just that: extreme and infrequent.

    It’s important to recognize that some MCS models are better than others. For example, some MCS models assumes much higher odds of extreme returns than would be expected by a bell-shaped curve. Likewise, Morningstar has modified to its MCS model to allow its clients to assume returns that fall along a “bell shaped curve,” or to select a model which presumes higher likelihoods of more extreme market behavior. Other firms are considering similar modifications to their MCS models.

    How much should you rely on MSCs to help you figure out if you can afford to retire? First, recognize that a MCS is one tool among many to help you answer this question. Regardless of improvements to MCSs, no methodology will ever be able to accurately predict very infrequent events. This means that every retirement plan should have enough flexibility to respond to unpredictable occurrences that will inevitably happen. This could mean being open to working part-time, especially during your retirement’s early years. You could also mean gradually converting some or all of your assets to guaranteed life-time incomes (annuitizing). Delaying the collection of Social Security Benefits until you reach your full retirement age could be another prudent strategy. And finally, you could separate your retirement expenses into categories of “must haves” versus “would like to haves,” to identify an acceptable fall back life-style if extreme market events happen.

   As far as using MCSs, follow the recommendation of the Retirement Income Industry Association to only use MSCs that run at least tens of thousands (and preferably, hundreds of thousands) of scenarios. Currently, some MCS projections rely on as little as one thousand scenarios for their depictions. Larger numbers of simulations include more extreme events in their runs and reduce the possibility of relying on an overly rosy investment scenario for your retirement projections. Check with the MCS sponsor you use to make certain that they adhere to the Retirement Income Industry Association recommendation. As with most things, your retirement plan should be prudent and flexible to respond to “unhappy surprises” ignored by less than robust MCSs.


Filed under: Musings-Articles, Uncategorized