Over the last few months, I’ve encountered investors, advisors and even some financial writers struggling to make or suggest good estimates of future investment portfolio returns.
No easy answer exists. But these inputs matter for planning.
And so maybe predictably, people fall into the habit of just extrapolating the past. Maybe assuming they, their clients or readers will enjoy the historical average return on stocks or on a balanced portfolio.
Starting with the historical average? That’s probably an okay place to begin. But I’m going to argue you and I want to make several adjustments to average returns to improve the usefulness of our forecasts. So let me share some thoughts and tips.
First Tip: Understand Whether Returns Adjusted for Inflation
A first quick tip: Understand whether you’re looking at real, adjusted for inflation, returns or not. Ideally you’ll work with real rates of return. (That’s best.) But if you must work with nominal, or unadjusted-for-inflation returns, be sure you know that. Because when you work with rates of returns without knowing whether the rates are real or nominal? It scrambles our ability to understand what’s going on.
A quick example to show you what I mean: Blogger Ben Carlson posted a few years ago that the worst-case all stocks return equals nearly 8%. He was talking about why you don’t need to worry about going all-in on stocks. Here’s the actual language:
The worst 30 year return — using rolling monthly performance — occurred at the height of the market just before the Great Depression and stocks still returned almost 8% per year over the ensuing three decades.
A little later in his post, Mr. Carlson acknowledged that his number ignored inflation. But that seems to me like a pretty important omission. Why? Because the adjusted for inflation return equals about 2.5%. That’s a huge difference.
If you earn an 8% return, you double your money every 9 years. If you earn a 2.5% return? You double your every 29 years.
By the way, you can tell Portfolio Visualizer, one popular tool, to adjust dollar amounts for inflation. You check a box. FireCalc and cFireSim both work in adjusted-for-inflation numbers.
But the big point here: Before you do anything else, nail down whether the numbers you’re looking are real or nominal, adjusted for inflation or not adjusted.
Second Tip: Recognize Variability in Investment Portfolio Returns
Another tip for thinking about returns and portfolio ending values. You and I aren’t guaranteed a predictable return if we invest in the stock market. Not even if we invest for a long time.
An example: If you invested $1,000,000 into your retirement account, paid a low .08% expense ratio for three decades, and invested 100% in US stocks, cFireSim says historically your results look like what show in the following table:
cFireSim Result | End. Bal. | Avg Return |
Average | $7,153,315 | 6.78% |
Median | $6,729,092 | 6.56% |
St. Dev. | $3,152,110 | N.A. |
Highest | $16,396,431 | 9.77% |
Lowest | $2,141,078 | 2.57% |
Lowest 10% | $3,561,032 | 4.32% |
Lowest 5% | $3,210,669 | 3.96% |
Just to be clear, the median adjusted for inflation return equals roughly 6.56%. That’s good. You or I can easily prepare for retirement with that real annual return.
But the worst-case result equals 2.57%. Five percent of investors earned 3.96% or less. And ten percent of investors earned 4.32% or less. Those numbers are not as good.
And the main point: Earning 3% or 4% over three decades as opposed to the median 6.56% annual return? You’re talking probably about hundreds of thousands or even millions of dollars of difference, as the table above shows.
Thus we want to plan for this variability. Not ignore it.
A quick sidebar here: If you’ve read something that suggests differently? Let me make this observation. I will guess what you’ve heard goes like this: Sure, in short term? You absolutely might be up and down, gosh, 50% in any given year? But over time the zigs and the zags even out. That’s sort of true. But not true to a degree you want to rely on. As at least a couple of Nobel Prize winners, Paul Samuelson and Robert Merton, have pointed out. (Economist Zvi Bodie has a great free discussion of this topic here: Wishful Thinking About the Risk of Stocks.)
Third Tip: Model the Right Saving Amounts and Timing
A mechanical point next. Sometimes when you estimate returns and portfolio values, you’re calculating the future vale of a single initial lump-sum investment made up front. For example, you might make calculations similar to those reflected in the preceding table. There, I calculated historical returns and values for a single $1,000,000 investment made at the very start of a 30-year time span.
The numbers look differently however if you save annually. For example, if you save $10,000 at the end of every year. Or if you start with a $25,000 lump sum investment and then save an additional $5,000 each year.
Thus, you want to work with calculators that show the same pattern of investments you plan to make. The popular FireCalc website doesn’t let you do this very easily. Neither does the Portfolio Visualizer website or the PortfolioCharts webside. The cFireSim web site does let you do this easily.
Fourth Tip: Factor in Fees and Expenses
A drum lots of people bang on. But one worth hitting at least one more time. I’m using a low .08% expense ratio for the calculations here. That’s a “Vanguard-low” level of fees on the hodge-podge of mutual funds I actually hold. And you can almost ignore fees at that level. A fee set at .08% equals $800 annually on a $1,000,000.
But many folks are paying more than that for an investment advisor. And if someone pays a larger fee like 1%? One would want to subtract that fee from all of the percentage returns shown in the tables above.
An average real return of 6.56%, for example, shrinks to 5.56% if you’re paying 1% in fees.
A tenth percentile real return of 4.32% shrinks to 3.32%.
If you have to pay 1% for an investment advisor? Okay. I get it. But make sure you account for that.
And make sure you understand that in a worst-case scenario? Your investment advisor may be capturing nearly half of the historical real return.
Fifth Tip: Consider Current Market Conditions
Let me end with something more, er, controversial.
I think we want to consider adjusting return estimates for higher stock market values in the U.S. at least. And also for the long-term trend in ever-lower interest rates.
Some investors appear to still confidently predict the future will mirror the past. That appears overly optimistic to me. Good data exists which suggest interest rates have been trending down over centuries. (See here.) That surely means it’s reasonable to think about earning less on bonds if you’re including those in your portfolio.
Low dividend rates and high valuations suggest future equity returns should be lower than over the last century.
Personally? I’m using Vanguard’s market outlook as my long-term forecast. (An example appears here.) Vanguard supplies a range of nominal (so not adjusted for inflation) returns by asset classes, which is good.
Two Final Comments So I Don’t Leave You Bummed Out
Before I end, a couple of remarks.
First, the stuff in the preceding paragraphs? Just to be clear: I don’t think it means you or I go off and do something unorthodox.
We will want to follow the prescriptions given by like David Swensen, John Bogle, Burton Malkiel, Bill Bernstein, the Bogleheads forum, and anyone else promoting low-cost passive investing that emphasizes traditional assets classes. Especially equities.
Second, if outcomes look a little less rosy once you adjust for the things mentioned above? The practical way to address them for most folks is probably to save more or work a longer.
Some Other Resources Related to Investment Portfolio Returns
A discussion of why high valuations suggest longer-run returns: CAPE Fatigue.
The first post in a series about having a backup “plan b” for retirement: Retirement Plan B: Why You Need One.