Is “Probability of Success” Misleading Advertising?
Opinions regarding the use of Monte Carlo simulations for financial planning are divided. Personally, I try not to over-rely on it but view it as a tool for certain circumstances. Some clients may want to compare plausible long-term effects of different courses of action, or others may simply want a general sense of their long-term financial trajectory while taking into consideration a range of plausible outcomes.
If I have objections related to using Monte Carlo analysis, they have more to do with labelling and disclosure than anything. Most financial planning systems use the term “Probability of Success” (“PoS”) or “Chance of Success” as a term of art for the primary summary statistic that a Monte Carlo simulation analysis creates. PoS is the percentage of random trials that do not run out of money in a financial plan’s Monte Carlo simulation, given the plan’s assets, liabilities, and spending. The name almost trivializes the number of underlying assumptions that Monte Carlo simulation analysis makes, and as a result, it may be misleading because users can interpret it literally without thinking about some of the nuances.
Why wouldn’t a 90% PoS mean that I have a 90% chance of succeeding? Well, to start, I may not define success as having a dollar left in my portfolio when I die. Rather than thinking about this stat as my chance of success, a better interpretation might be something like the following …. this number tells me what percentage of random trials did not run out of money, taking into account planned portfolio distributions, under the assumptions made by the particular Monte Carlo engine I used including the underlying statistical distribution of investment returns, expected future returns, expected future volatility, and expected inflation.
If we want users, including financial advisors and their clients, to have a deeper appreciation for Monte Carlo’s sensitivity to inputs and its other nuances, a more descriptive name for PoS would also be helpful. My preferred term is “Portfolio Survival Rate” (“PSR”), which does not suggest predictiveness and more literally describes what the number tells us. This term is used by Portfolio Visualizer.
It would also be helpful if providers of Monte Carlo financial planning systems published out of sample results of the predictiveness of their models. Depending on the model and its assumptions, PSR may have biases in certain output ranges. And, no matter how robust a model and its assumptions are, it may turn out to be no more predictive of the future than other random number generators. Rather than relying on Monte Carlo for predictiveness, consider its output as a plausible range of outcomes – assuming the model and inputs are reasonable.
A good place to start checking into how reasonable the assumptions are is to ask your financial advisor about the return assumptions they use when providing you with a Monte Carlo simulation report. Many planning systems use historical returns as default inputs to their Monte Carlo engine. This is a textbook example of “garbage in, garbage out”. I have no idea what the next 7-10 years will bring in terms of returns, but I do know there is not much reason to believe we’ll see a repeat of the past. It is critical to use a set of reasonable forward-looking capital market assumptions as inputs.
Finally, a common shortcoming of financial planning Monte Carlo engines is that return assumptions are used for the entire plan horizon, typically lasting at least several decades. It would be an industry wide improvement to see more systems roll out the capability for regime-based return assumptions so that expected returns over 7-10 years could be modelled followed by a reversion to longer term assumptions.
Beyond labelling and model inputs, it is challenging to develop a sense of what an acceptable PoS or PSR is. There is not a true academic or industry consensus around it, perhaps because interpreting PSR is highly nuanced. Remember that most individuals and families have at least some degree of freedom to make spending adjustments. The more of a household’s non-discretionary spending that is covered by secure sources of income like pensions and social security, the greater is that household’s flexibility to adjust spending if needed. One can think of [1-PSR] as the plausibility of needing to adjust your financial plan during the plan horizon. As such, a lower PSR may be perfectly acceptable if the household’s non-discretionary spending is funded mostly with secure income.
Monte Carlo simulations are generally fairly good at illustrating a range of credible outcomes if they are based on reasonable inputs. But they are not necessarily predictive. It’s better to think of their output as illustrating a range of outcomes that may be in the ballpark rather than making predictions about what particular path your future will take.
© 2026 Philip Murphy. All rights reserved. The information presented is the opinion of the author and does not reflect the views of any other person or entity unless specified. The information provided is believed to be reliable and obtained from reliable sources, but no liability is accepted for inaccuracies. The information provided is for informational purposes and should not be construed as advice. Advisory services are only offered through IndiePlan™ LLC, an investment adviser registered with the state of New York.