Retirement, like life, is fundamentally uncertain. Not only do we not know how long we're going to live, we also don't know things like future market returns, future inflation, even when we are going to retire.
These uncertainties make retirement planning difficult; however, one way to provide context regarding these uncertainties is to run what is commonly known as a Monte Carlo projection.
The uncertainty associated with Monte Carlo is often used by financial advisors (and educators) to sell strategies that have less uncertainty (e.g., some type of insurance product with a guarantee).
For example, I recently came across someone suggesting withdrawal strategies were analogous to a plane crashing, which is an incredibly misleading analogy. In reality, people will adjust their spending in retirement as situations warrant. Additionally, no financial plan, strategy or product can guarantee a successful outcome for an investor, given the myriad of risks and uncertainties in real life.
While there are notable imperfections with outcome metrics commonly employed in Monte Carlo simulations (the probability of success) and some of the underlying assumptions (e.g., static withdrawal rates), I still believe Monte Carlo models can be useful for financial advisors and investors, but context is important, especially when some folks in the industry attempt to sell against the approach using misleading analogies.
Decision-Making Under Uncertainty
Monte Carlo modeling was developed by John von Neumann and Stanislaw Ulam during World War II to improve decision-making under uncertain conditions (related to the development of nuclear weapons) and named after the Monte Carlo Casino in Monaco. Over the last few decades, Monte Carlo projections have become an incredibly common way for financial advisors to demonstrate the uncertainty associated with financial planning projections and accomplishing various financial goals (e.g., retirement).
The key differentiator with a Monte Carlo projection versus other projection methodologies, like time value of money calculations, is the element of chance (i.e., randomness). Typically, forecasted returns are the only random (or stochastic) variable in financial planning programs that employ Monte Carlo; however other variables, like age of death or age of retirement, could be randomized as well.
The return assumptions in Monte Carlo projection can be historical, forward looking, or some combination of the two. There is no requirement that returns be normally distributed in a Monte Carlo projection; however, there may be limitations given the specific program used by an advisor (i.e., constraints are program-specific as there really aren't any limitations with Monte Carlo).
