Investing Blog Roundup: Monte Carlo Simulations

Monte Carlo simulations are a popular way to determine how risky a given level of spending is, for a particular set of household circumstances (i.e., assets, age, other sources of income, etc.). Different software will do such simulations differently — and provide different output as well. But the most common form out output is to show a probability of success/failure — that is, in what percentage of the simulations did the household end up depleting the portfolio before the desired length of time had elapsed.

As David Blanchett discusses in a recent article, many “failure” scenarios wouldn’t even occur in real life. In part, that’s because people tend to cut their spending, if they see that it doesn’t look like things are going according to plan. Second, Monte Carlo simulations are often run using a conservative time horizon estimate (e.g., to age 100). If a “failure” scenario shows the portfolio being depleted at, for example, age 97, there’s a good chance that the original owners of the portfolio would no longer be alive and spending from it.

In another recent article, Massimo Young and Wade Pfau point out a danger of Monte Carlo simulations, which people might not recognize. Namely, while such software randomizes the results, it does so using constraints/assumptions set by the user (or set by the software developer, if the user doesn’t have options to adjust). And the results of the simulations are very sensitive to the assumptions used.