FREE NEWSLETTER
MANY FINANCIAL planners say they “stress test” portfolios. That sounds like a good idea, but it isn’t well defined. I decided to do some research to see how I could apply the notion to the investments owned by my wife and me.
I came across a number of useful articles. Investopedia, one of my go-to resources for all things financial, provides this definition: “Stress testing is a computer simulation technique used to test the resilience of institutions and investment portfolios against possible future financial situations.” Forbes, meanwhile, provides a three-step process that I thought was useful.
Among financial planners, two common approaches are historical data analysis and Monte Carlo simulations. The historical approach uses past asset class returns to analyze how a portfolio would have fared historically. The idea: If your portfolio survived the past ups and downs of the financial markets, you’ll probably be okay in future.
The Monte Carlo method is similar, but takes the idea even further. It’s a mathematical method that starts with historical asset class returns for, say, each calendar year and then, over and over again, assumes these annual returns occur in a different order. This results in hundreds or even thousands of different possible scenarios, allowing you to develop a statistical distribution that shows how your portfolio might perform.
Monte Carlo simulations are especially useful for looking at whether folks will have enough money for retirement, given the amount they plan to withdraw each year from their portfolio and the range of potential returns. One big worry: Retirees get hit with atrocious returns early in retirement and those returns, coupled with their own spending, eviscerates their portfolio.
I think both the historical approach and Monte Carlo analysis are useful exercises, and they can help boost confidence in a retirement plan. But both also assume that the future will be at least somewhat like the past. What if that’s a bad assumption? That’s why I’ve added a margin of safety to our retirement plan.
How do you do that? I’m aware of three key strategies. First, you could assume your living expenses will be higher than is likely. This gives you a buffer if your portfolio doesn’t grow as expected. In fact, some planners assume spending starts high, but declines as their clients age. I’ve observed that in my parents and in-laws. But you probably shouldn’t bank on this happening. Increased medical and long-term-care expenses could kick in as your retirement progresses, so your net spending may not shrink—and it could grow.
Second, you could assume low investment returns. For instance, if your portfolio can support your lifestyle through a long retirement, even if your investment returns simply match inflation, that should make you more confident.
Finally, you might add a margin of safety by assuming you’ll need to spend down your nest egg over a very long life. I run my scenarios through to age 100. The financial planners I know worry a lot about longevity. Many folks, however, seem to be more concerned with dying young and “leaving money on the table.” This attitude often drives them to claim Social Security early. The downside: That means a permanently reduced monthly benefit and hence a weaker retirement safety net.
My engineering training tells me to do a little bit of everything—assume a higher cost of living, low investment returns and a long retirement. After all, you don’t know what or where the anomalies will show up. Can you overdo this? Sure, you could provide so much margin for safety that you live too frugally and deny yourself some well-deserved retirement fun. Still, I wouldn’t cut it too close.
Richard Connor is a semi-retired aerospace engineer with a keen interest in finance. Rick enjoys a wide variety of other interests, including chasing grandkids, space, sports, travel, winemaking and reading. His previous articles include State of Taxation, Paradise Lost and Much Appreciated. Follow Rick on Twitter @RConnor609.
Want to receive our weekly newsletter? Sign up now. How about our daily alert about the site's latest posts? Join the list.
People who demand higher Social Security COLAs have never done any of these tests.
I hope it came across that I’m no expert in predicting future markets. In fact I’m one of the worst individual stock pickers you could imagine. Given that, and the inherent uncertainty of the markets, I feel more comfortable with a conservative approach. Someone told me a long time ago the most important measure of risk tolerance is can you sleep at night.
I totally agree – assuming you mean Fidelity’s stand-alone planner. Very clearly laid out with reasonable assumptions and thoroughly documented. Uses Monte Carlo iteration of potential market returns (US stocks since 1926, bonds and foreign stocks since the 70’s).
The “significantly below average” result you refer to in Fidelity-talk means 90% of over 250 iterations of possible historical market returns provided enough income/assets throughout retirement. The lower than average and average use 75% and 50% success rates respectively.
In effect, the most optimistic scenario (average) assumes average historical returns continuing into the future and the other two scenarios assume you get stuck with an increasing amount of the bad years. I really like this.
I got very similar results with Quicken’s planner, which simply means I made similar assumptions using similar starting numbers.
What return rate would you suggest if someone uses your second strategy to assume low investment returns?
I use the MAXIFI software by Larry Kotlikoff. His base assumption for a “safe” real rate of return is the 30 year Treasury. He also baselines 100 years old as the end of life. He is a big believer in safe spending rates.
Kevin
That is a great question. Years ago a fellow engineer and I did some studies with a spreadsheet tool we had created to project retirement portfolio sustainability. The conclusion we came up with was that if you beat inflation by a point or two you were probable OK. We looked at scenarios that applied to our situation – 30 year pensions, well-saved 401k, and then a fairly high social security benefit. We are part of the lucky ones. But I think the point holds – if you start with enough to retire, and your portfolio beats inflation, you shouldn’t run out of money.
I use Flexible Retirement Planner for much of my planning. I don’t use it as a Monte Carlo. Instead I use it as a deterministic model by leaving standard deviation at zero. I assume long term inflation of 3% which I thought was somewhat pessimistic, but other have said it was average at best.
Still, I’d like to know what long term return rate would qualify as “low”?
There are many variations of sensitivity analyses applied to retirement forecasting, but I prefer less automated methods that allow you to set all the variables. If errors occur in this critical planning phase of life, I want it to be my fault and not that of a software designer. The Monte Carlo simulators provide useful comparisons, but I think they make the most sense for the very wealthy (because it doesn’t really matter) or those that don’t want to put much time into the process.
Regardless of method, almost all retirement planners are disconnected from a comprehensive personal finance program, thus you have to update them independently as time passes. The further out in time we run our simulations, the greater the potential error. I use Quicken’s planner that updates every time I buy or sell in accounts I told the planner to include, update fund prices, transfer cash from bonds to my checking account, consider the purchase of a sailboat next July (including annual expenses rising at a selected rate with a final disposition after 8 years for almost nothing), etc.
As the days go by, my future approximations become less so with less effort than a planner that requires independent data entry. The probability of error decreases just like hurricane Zeta as it approaches land today. In a few hours we’ll know for sure if it’ll track as predicted or move a bit east toward Pensacola where I live. 🙂
“Still, I wouldn’t cut it too close.” I think that’s it in a nutshell. Personally, I live very far from that “cut line” and as a result don’t really think about it hardly ever. But it’s a mindset I agree. I don’t really feel I want for much and don’t feel I’m depriving myself or my family, either now or in impending retirement. My “fun” is for the most part either free or relatively inexpensive. I guess I could spend more if I wanted, but maybe it’s simply that I value the security (financially) more.
I wonder what percentage of the population gives any thought to any of this? I imagine most simply save as much as they can, maybe throw it into an index fund (if that) and hope for the best.
Good question. I’m pretty sure I think about it too much, but I’ve always enjoyed analyzing complex things. I have not met many people who enjoy discussing these kinds of topics at the level I enjoy, but that says more about me than them.
Where one invests their money, and the adjustment of one’s income stream, via the monitoring of one’s portfolio balance, if need be, can be beneficial to an investor who wants to create and control their own stream of income. Further tactical steps may be taken in order to ease the effects an unfavorable starting ” returns sequence”, through the use of a simple, arithmetic rules based process.
Research shows that a portfolio / index representative of the Large cap
“value” universe has sustained a “7%” inflation adj annual withdrawal
Research shows that a portfolio / index representative of the Large cap “value” universe has sustained a “7%” inflation adj annual withdrawal rate ( “sale of shares”, dividends reinvested ) over seventy one rolling 20 year periods since 1932 ( a decent sized sample for “stress test” )
Link https://tinyurl.com/y6key3v5 .
This has encompassed periods of Great Depression, World wars, oil price shock and high inflation, tech bubble deflation, 2008 banking crisis, pandemic?.
As there were a minimal number of periods when the income withdrawal and
negative “return sequence” depleted the portfolio to “0” ( “failure”),
when applying a simple tweak to the withdrawal rate, those periods could
be ameliorated ( see charts 1 and 2 https://tinyurl.com/y6key3v5 ).
Investment in modern, expertly managed and well diversified, large cap value index
funds ( such as the low expense Vanguard Value ( VTV ) or DFA Large Cap
), may be used for this purpose.
As the Vanguard value fund mimics the “CRSP large cap value index” / methodology as a
framework for portfolio construction, this would imply that a “passive”
value management approach is used ( based on 70 years of data ).
However, if an investor is uncomfortable with the notion of using large
cap value solely, they may employ the use of a large cap “dividend
growth” fund / universe for same income generation practice.