FREE NEWSLETTER
IT’S HUMAN NATURE to be impressed by things that sound sophisticated or seem complex. In the world of personal finance, this certainly applies to the planning tool known as Monte Carlo analysis.
Its roots go back to the 1940s, when it was developed by Stanislaw Ulam, a physicist working on the Manhattan Project. Today, it’s a popular way to assess the strength of a proposed retirement plan. If you’ve seen presentations indicating that a financial plan has a particular probability of success, that likely came out of a Monte Carlo simulation. Because of its highly mathematical underpinnings, this type of analysis tends to be viewed as rigorous and its results reliable. It’s not perfect, though.
In the past, I’ve discussed some of the shortcomings of Monte Carlo simulations. Chief among them is the issue that retirement can’t be characterized as having a binary outcome. It’s too simplistic to say that someone’s retirement will either be a success or a failure. As researcher David Blanchett noted in a recent article, “Monte Carlo failures aren’t like plane crashes.” Retirement—thankfully—is much more nuanced.
That leaves us with a crucial question: If we don’t use Monte Carlo analysis, how do we assess the robustness of a retirement plan? How can we be sure a particular plan won’t result in a retiree outliving his or her money? Below are nine strategies to consider.
1. Asset allocation. Adherents of Monte Carlo laud its ability to protect investors from sequence-of-return risk. That’s the risk faced during the first years of retirement, when poor investment returns can be especially damaging. Because Monte Carlo analysis looks at thousands of possible sequences of returns, it can help identify this risk. That’s useful, but there’s a simpler and more intuitive way to protect against an unwelcome sequence: through portfolio structure. If you hold sufficient dollars outside of stocks and in more stable assets like bonds or cash, that can help immunize your portfolio against a bear market in stocks, even one that persists for a number of years.
2. Scenario analysis. An appeal of Monte Carlo analysis is that it effectively “rolls the dice” hundreds or thousands of times. That can help investors explore a range of potential outcomes. A downside to this approach: Each time the computer rolls the dice, it only changes one variable—typically just the projected market return. This makes Monte Carlo analyses incomplete. In the real world, many more variables are in play, and it’s important to see how those changes might also impact the sustainability of a plan.
For instance, you might look at the impact of different spending levels or different housing scenarios. What if you purchased a vacation home or you downsized? You might also look at the impact of a large charitable gift. And, of course, you might want to analyze the impact of a change in tax rates. You can’t test everything, but it’s important to look beyond the single variable that Monte Carlo simulations tend to use.
3. Tax diversification. To help shield your portfolio from ever-changing tax rules, try to strike a balance among the four main types of accounts: a cash account, a taxable investment account, traditional tax-deferred retirement accounts and tax-free Roth accounts. This sort of balance is important because it helps retirees better control their tax rate from year to year.
Already have the four types of account? You might also look into a health savings account, though eligibility will depend on the type of health insurance you carry. If you have children or grandchildren, a 529 education savings account is a great idea—especially with the new escape hatch that allows for the partial conversion of leftover 529 dollars to a Roth IRA. Contributing to some of these accounts—such as a Roth—might mean forgoing a tax benefit this year. But remember, it’s the long term that matters. Sometimes, it’s worth paying more tax this year if it means paying less down the road.
4. Guaranteed income. No doubt about it, income annuities have a bad reputation—but don’t dismiss them entirely. What most people fear is that they buy an annuity on Monday and then die on Tuesday. That’s an understandable concern, but it’s also unlikely. Locking in some guaranteed income can provide both financial security and peace of mind. At the very least, remember that Social Security represents probably the best annuity of all—with both inflation protection and a survivor benefit. I suggest doing everything you can to maximize that benefit.
5. Real estate. Rental property seems to be the asset that people either love or hate. That’s understandable. A friend described arriving at a tenant’s door to pick up an overdue rent check and being greeted by a pit bull. That’s every prospective landlord’s worst fear—but it’s also unlikely. Real estate is attractive because it delivers passive income, isn’t highly correlated with the stock market, allows for the use of leverage and provides unique tax benefits. It isn’t for everyone, but it’s worth considering.
6. Tuition. Most parents feel an obligation to provide their children with an education. As one parent put it, “If my kid gets into Harvard, I’ll find a way to pay that bill—no matter what.” It’s a wonderful sentiment. But with tuition so high, it’s important to think more critically about this. One way to manage risk while also supporting your children is to have them take out loans. Then, after they graduate, contribute toward paying them down, but only to the extent that it doesn’t jeopardize your own plan.
7. Mortgage. If at all possible, I recommend paying off mortgage balances before retirement. Even if the numbers indicate that you’d be able to make the payments in retirement, reducing your fixed costs can provide you with valuable flexibility to navigate life’s inevitable curveballs.
8. Estate plan. Along with tuition, another potential budget buster is long-term care. Should you or your spouse require an in-home aide or a long-term-care facility, the costs can easily run to six figures a year. While there’s no way to predict the level of care you might need—if any—you can take steps to protect your portfolio so every last dollar isn’t consumed by a nursing home. It’s worth consulting an elder-care attorney to learn what strategies exist. Those strategies differ from state to state.
9. “Right” answers. Personal finance pundits seem to enjoy beating each other up over retirement strategies. In addition to Monte Carlo analysis and the scenario approach described above, there’s the so-called 4% rule, the RMD method, the guardrails strategy and many others—and everyone has an opinion on which is best. But since none of us can see the future, the best strategy, in my view, is to consider what each of these models has to offer. Don’t let any one model leave you feeling overconfident—or overly worried. Instead, see them all as data points for consideration. Most important: Make sure that, should your plan falter, you have a backup plan.
Adam M. Grossman is the founder of Mayport, a fixed-fee wealth management firm. Sign up for Adam’s Daily Ideas email, follow him on Twitter @AdamMGrossman and check out his earlier articles.
Want to receive our weekly newsletter? Sign up now. How about our daily alert about the site's latest posts? Join the list.
There are at least 4 measures of robustness.
1. Monte Carlo. The better tools start with arithmetic means to arrive at geometrically compounded final values. It simulates volatility by randomizing rates of returns by a standard deviation. Provides a wide range of outcomes from mega-millionaire to flat broke. Estimates probability of success as % of times portfolio goes to zero. Disregards trends, momentum and the fact a portfolio may be too small to live on LONG before going to zero. Using a 15% standard deviation aim for 80% or better chance of success. More useful with 10-yrs or less to go. Cannot be used to compare trade-offs between major choices or sensitivity to changes in plan parameters.
2.Historical Backtest aka Aft-Cast. Uses actual market performance and inflation rates for rolling 1-year time periods going back 100+ years. The shortcomings are the world, regulations, investment products and markets have changed radically over just the past 30 years and there are only three 30-yr time frames between 1932 and 2022. Aim for 100% chance of success in all 1-year periods. Also cannot be used to compare trade-offs between major choices or sensitivity to changes in plan parameters.
3.Funded Ratios. These apply net-present-value calculations to a current snapshot of income, expenses and assets. Computes ratios of expected future income over expenses. Uses fixed assumptions for discount , tax and inflation rates. Aim for a ratio of 110% or higher with 5 years to go.
4. Cash Flow Projection. The most accurate method but cannot be used until actual income and expenses for first year of retirement are known +-10%. Can be used directly to create a retirement income plan aka paychecks. Works regardless of investor profile (safety-first or probabilistic) or investing strategy such as total return (selling shares) or taking income only from rents, SS, pensions, interest, dividends and normal distribution(no selling shares) . By age 72, debt coverage ratio from reliable sources of income should be 50% or higher because by then taxes, social security, RMD’s and legacy issues are well understood.
There are at least 4 measures of robustness.
1. Monte Carlo. The better tools start with arithmetic means to arrive at geometrically compounded final values. It simulates volatility by randomizing rates of returns by a standard deviation. Provides a wide range of outcomes from mega-millionaire to flat broke. Estimates probability of success as % of times portfolio goes to zero. Disregards trends, momentum and the fact a portfolio may be too small to live on LONG before going to zero. Using a 15% standard deviation aim for 80% or better chance of success. More useful with 10-yrs or less to go. Cannot be used to compare trade-offs between major choices or sensitivity to changes in plan parameters.
2.Historical Backtest aka Aft-Cast. Uses actual market performance and inflation rates for rolling 1-year time periods going back 100+ years. The shortcomings are the world, regulations, investment products and markets have changed radically over just the past 30 years and there are only three 30-yr time frames between 1932 and 2022. Aim for 100% chance of success in all 1-year periods. Also cannot be used to compare trade-offs between major choices or sensitivity to changes in plan parameters.
3.Funded Ratios. These apply net-present-value calculations to a current snapshot of income, expenses and assets. Computes ratios of expected future income over expenses. Uses fixed assumptions for discount , tax and inflation rates. Aim for a ratio of 110% or higher with 5 years to go.
4. Cash Flow Projection. The most accurate method but cannot be used until actual income and expenses for first year of retirement are known +-10%. Can be used directly to create a retirement income plan aka paychecks. Works regardless of investor profile (safety-first or probabilistic) or investing strategy such as total return (selling shares) or taking income only from rents, SS, pensions, interest, dividends and normal distribution(no selling shares) . By age 72, debt coverage ratio from reliable sources of income should be 50% or higher because by then taxes, social security, RMD’s and legacy issues are well understood.
Reply to tshort: 25x living expenses article referred to may be: https://www.wsj.com/articles/how-to-tell-if-your-retirement-nest-egg-is-big-enough-1421726456
It is 25x residual living expenses, to retire at 60 years
(sorry, I don’t think I’ve figures out how to reply to a specific comment)
The thing I dislike about Monte Carlo retirement success modeling is that it is basically reshuffling 70 years or so of annual market performance data into random sequences of 30 year retirements. It seems to me that this would inevitably generate impossible market performance sequences that have never occurred in history, and would likely never occur in the future.
The best analysis I’ve seen involves backtesting a given withdrawal rate against all of the actual 30 year market performance sequences we have data for going back to the 1920s and using the results to evaluate whether a given withdrawal rate (and it’s annual adjustment algorithm) “succeeds” or not.
While this doesn’t tell us much about what any given retiree is going to experience going forward, it at least helps us understand whether the approach one is using will or won’t work against every single 30 year period that has ever happened.
While a good Monte Carlo outcome is no guarantee of success, a bad outcome makes failure likely.
Bob, we all made investment mistakes. When I started my pension/profit sharing plan these two guys in raincoats wanted me to buy insurance inside the plan. Since it did not seem kosher to me, I bailed and obviously the right move. I did get suckered in to variable life and cashed it out eventually. as Bernstein says, “treat every broker, advisor, and insurance salesman as a hardened criminal.” How True!
Great points, Adam. I call our retirement plan a WAHOO strategy: we always have other options, with margin for “long tail” surprises. My only regret: I wish someone asked me earlier in life to model replacing paycheck income using savings. Our situation turned out fine in the end but it could have gone very differently.
No mention of the value of holding whole
life insurance into retirement? Why is that?
Not that I’m recommending it, but whole life worked for my wife and me. I’m 78 and bought whole life for both of us starting in my twenties, newly married, kids on the way, etc. Somewhere along the way, I realized I had made a mistake, but by that time the cash values were growing much faster than the annual premium expense, so I couldn’t bring myself to cash out this “great investment”. Fast forward to age 65 with sufficient investments not to have to worry about leaving my wife in financial difficulty and all our sons were out of the house, married and self sufficient. One day it hit me, that there was no need for any life insurance so I cashed out of all of our policies and rolled the cash over into income producing investments that provide a comfortable monthly income allowing my other investments to remain untouched. Plus, my family was insured for all of those years. Yes, I know I could have done better knowing what I know now, but I was 24. Be gentle in the comments:)
https://www.whitecoatinvestor.com/debunking-the-myths-of-whole-life-insurance/
No greater killer in pre-retirement is sequence of returns risk. Imagine your 1m portfolio where you need a 4% SWR and you get hit with a 50% loss(2008) and now your SWR is 8%. You will deplete your portfolio; the absolutely worst scenario. Wm Bernstein suggests 25X living expenses in safe investments; a bit conservative but he is a bright fellow.
Do you have a link to that 25x living expenses article? Sounds off to me.
Sequence of returns risk is inherently very risky if you’re holding a 100% stock portfolio that you rely on for living expenses.
This is exactly why you should hold a diversified portfolio (Bonds were up 5% in 2008) and enough cash to cover a couple years of expenses.