Don’t Rely on Monte Carlo Pension Analysis
How is your retirement plan going? Chances are, you have no idea: About a quarter of Americans saving for retirement actually have no idea how much money they’ve spent, and more than half of workers who don’t contribute to a 401(k) think they have .
Retirement is confusing and often causes feelings of anxiety and fear. Do you have enough to live on? Will you be able to cover medical expenses in a world where booking a vacation for major surgery is sometimes the best financial decision? And worst of all, will you simply run out of money and spend your last days in poverty?
Funny thing. Many people rely on financial advisors of one sort or another to reassure and guide them, and these professionals use a variety of sophisticated tools to give us some insight into how we’re doing and what our eventual retirement might look like. However, there are many variables to consider, most of which are beyond our control. As a result, if you’ve spoken to a professional about retirement planning, you’ve likely encountered what’s known as a Monte Carlo simulation —or will soon. These tools are useful, but they are also misleading and you should not rely on their predictions.
What is a “Monte Carlo simulation”?
A Monte Carlo simulation is a complex program that takes many assumptions and variables, plugs them into a model, and runs hundreds or thousands (or even hundreds of thousands) of simulations to see how changing the variables affects the outcome. It then averages the results to give you an idea of how likely different outcomes are.
In retirement planning , variables such as inflation or market returns are assigned random values for each simulation. Do enough modeling and you’ll get a general idea of how likely your strategy is to keep your money alive until a certain age. The results are typically displayed as a bell curve graph, with the most likely scenarios in the middle and outliers (unlikely events) at either end. You will also often be given a probability score that sums everything up; A Monte Carlo score of 80% means that your current plan has an 80% chance that your money will last.
Monte Carlo simulation is useful. They can test assumptions and identify potential flaws in your plans or confirm the soundness of your strategy. But Monte Carlo estimation is often oversimplified, and you should never rely solely on a Monte Carlo simulation to validate your retirement savings strategy for four main reasons.
Monte Carlo Simulation Can’t Predict the Unexpected
A Monte Carlo simulation only knows the data included in it and cannot anticipate unexpected events in its calculations. For example, financial theorist William J. Bernstein once observed that if you retired in January 1966, the first 17 years after retirement occurred during a time when market returns on the S&P 500 index were almost perfectly aligned with the rate of inflation (a period known like the Great Inflation ), essentially making zero return on your investment.
Monte Carlo simulations can’t explain things like this—they rely on probability estimates rather than analysis of once-in-a-century financial oddities or unforeseen market crashes. Even if the result is 100%, the chances of success are always lower given the possibility of surprises.
There are different models
When you receive a Monte Carlo estimate from your financial advisor, there is one more piece of software you should consider. After all, you can find free Monte Carlo simulators online, and chances are you have no idea whether your financial advisor is working with a super-sophisticated AI simulation monster or a homemade Excel spreadsheet they first created in 1999 also may not know the methodology used under the hood. If you trust your consultant, then of course it makes sense to trust their modeling, but keep in mind that if the software itself is bad, then so are its results.
Human error is a factor
Monte Carlo simulations have to make a lot of assumptions to run their models, and one of the most unreliable ones is that you – yes, you – will hold up your end of the bargain. You can tell your advisor that you plan to save a certain amount in your retirement accounts, cut back on your expenses when you retire, and be perfectly happy living a small, quiet life in a paid-for house, and then go crazy when you finally retire and spend lavishly while experiencing a sort of YOLO fever. In other words, these simulations often assume that you will automatically do what you’re supposed to do for decades and decades, despite the fact that all of human history demonstrates how sketchy we all are.
The reverse is also true. If your Monte Carlo score is low, the simulation doesn’t know that you’re actually quite adaptable, and will instantly change all your spending and saving habits as soon as you feel trouble brewing with your bills. The simulator believes that in 20 years you will run out of money, but you will find a way to survive for 40 years.
Does this mean that simulation is useless? No, it just means you can’t assume he’ll tell you everything.
Monte Carlo results can be deceiving
Finally, Monte Carlo estimates can be misleading if taken at face value. If you run 5,000 simulations and have enough money for 4,000 of them, you will get 80%. This is very good, isn’t it? But it also means that your current plan has failed 20% of the time – if you were told that you had a 20% chance of dying if you did something, you probably wouldn’t be too smug about it.
Additionally, Monte Carlo results can be changed simply by adjusting the variables. Change the inflation rate assumption or assume you put an extra 5% of your income into an IRA, and that 80% could suddenly become 90%—but it doesn’t really matter that much. All it can do is allow you to test out different scenarios and get a general idea of whether you are up to par or not.
And this is no small matter: Monte Carlo simulation has its uses. At the very least, they’ll give you some idea of how your current retirement plan will work if everything stays more or less as expected and you do exactly what you think you’ll do (including, you know, living long enough to so that it all matters). But don’t make the mistake of hearing a strong result in Monte Carlo and thinking you’re all set for the future.