I recently had the pleasure of meeting some new people over happy hour at a local meetup group. Between a few drinks on the patio on a warm early-fall afternoon, we exchanged the typical “so what do you do?” icebreakers, shared some fun stories, and had a great time.
As the introductions turned to me, I guess it’s no surprise that the topic of investing came up. I talked a little about the website I run that studies historical data to help people make wise investing decisions without having to stress about keeping up with the markets. No big deal, but I’m pretty proud of it.
I was sitting next to a smart and accomplished guy who knew a bit about investing, and his natural reaction was a perfect mix of genuine, engaged, and polite. It went something like this: “That’s really cool. But as they say, past performance is not indicative of future returns. It’s too bad that historical data not very useful for decisions today.”
Honestly, his comment didn’t surprise me at all. Not only have I heard that from plenty of investors before, but I’ve also seen the same standard disclaimer on hundreds if not thousands of investing resources over the years.
Past performance is not indicative of future results.
This straightforward statement has become so ubiquitous that many investors simply take it for granted without giving it much thought. And as a result, many intelligent people normally predisposed to evidence-based thinking in other aspects of life nonetheless dismiss the field of portfolio backtesting altogether.
But here’s the thing. That throwaway sentence is often just flat wrong.
So for my new friend asking about my investing hobby and all of my old ones here wondering how to properly apply historical data to today’s choices, let’s talk about the limits — and power — of effective portfolio backtesting.
When the past performance disclaimer is true
In order to understand how disclaimers like this don’t tell the whole story, it helps to understand why they exist in the first place.
A statement like “past performance is not indicative of future results” is actually mandated by the SEC in the marketing materials of financial institutions. And it also makes total sense in that context, as the similarly mandated reporting statistics like the 1-year, 10-year, and lifetime returns of the fund are just snapshots in time and do not communicate the underlying variability involved.
To appreciate how static performance numbers present a problem, think about how people use them. When investors shop for fund options at their favorite brokerage, it’s easy to sort them by something like the 10-year return and assume that the funds at the top are just better and will always perform like that. But in reality, that’s not how investing works.
For example, if you look at the most recent 10-year performance of a Total Stock Market fund in the US like VTI, the real return measured at the end of 2023 was about 9%. That sounds pretty great until you realize that the 10-year rolling average changes all the time. You can clearly see that in this Rolling Returns chart that shows every rolling 10-year timeframe since 1970.
Your brokerage may accurately report the most recent 10-year annualized return was 9%, but are you prepared to lose money for the next 10 years? It’s a legit possibility!
That variability presents a unique challenge when using historical data to select a fund looking forward. When you assume that your investment will return the same 9% but it actually returns something closer to -3% on your own investing timeframe, it’s normal to feel like you got duped.
I created a chart called Start Date Sensitivity to study that disconnect in rolling data snapshots. At every point in time, it looks backward 10 years and forward 10 years and measures the delta between historical expectations and future reality. Here’s what that chart looks like for the same Total Stock Market fund. The red line tracks the previous 10 years, the blue line tracks the next 10 years, and the vertical lines measure the difference between the two numbers.
In this example, the luckiest person on record was the one who started investing in 1982 and expected a -2% annualized return but received +13%. Nice! The unluckiest person was the one who started in 1999 and expected a 15.5% annualized return but received -3%. “Ouch” is putting it far too kindly. And the SDS metric is the full spread between those two extremes.
With a start date sensitivity of a whopping 33.5%, the total US stock market is pretty darned volatile. And as a result, looking at the most recent 10-year performance is simply a terrible way to estimate your future returns.
So when the site you’re reading says clearly that “past performance is not indicative of future results”, that’s where it’s coming from. Think of investing with multiple start dates like a game of darts, and it’s easy to understand just how rare it is for two darts to hit the exact same point on consecutive throws. You really should not put any faith in recent history exactly repeating.
Why the statement is misleading
While the admonishment to not base your expectations of the next throw on the results of the last one is unquestionably sound advice, the story doesn’t end there.
If a dart represents one investing timeframe, a portfolio can best be thought of as the person throwing hundreds of them. Every person has their own particular skill level and general accuracy, which becomes quite evident once you study enough data.
Just like different competitors in darts, some portfolios are more dependable at hitting the mark. My favorite visual for that is the Target Accuracy chart that measures the full range of portfolio growth paths looking at every investing start date since 1970. Here’s what that looks like for the Total Stock Market.
Look at the maximum and minimum 10-year portfolio values at the bottom of the chart, and it’s immediately evident just how wide the spread of potential outcomes has been for stocks. For a stark contrast, compare that image to the same one for simple T-Bills.
While the average 10-year real return was clearly lower for T-Bills than for stocks, bills are also a lot more predictable with a much tighter spread of outcomes.
That relative level of predictability is inherent to the assets. Stocks and bills have different natures. Stocks are wild and unpredictable, while bills are slow and steady. And barring a change in the very foundation of how the markets work, those natures are unlikely to change*. Stocks will not suddenly behave with the predictability of bills, and you will not wake up tomorrow to see that your bills gained or lost major percentages in a single day.
These differences in investing nature are highly important for investors seeking to plan for their important financial goals. And for evidence-based investors studying more complex portfolio options where the relative risk-adjusted return isn’t as obvious as comparing stocks to T-Bills, historical data is extremely important.
While academics may argue over how to best calculate the expected return and uncertainty of an investment based on asset pricing theory and underlying fundamentals, I’m generally more convinced by evidence over theory. Maybe it’s my engineering education or perhaps it’s just a core tenet of my investing personality, but regardless of the motivation you better believe I’m going to check the measurable track record of any proposed investing approach.
I honestly don’t care if the famous fund manager currently holding the darts has decades of experience and is using an advanced technique taught in elite circles. If his historical cluster is all over the map like the total stock market, there’s no way I’m wagering my own hard-earned money on his ability to hit the average bullseye on the next throw.
But with a portfolio like the Golden Butterfly? That’s a different story.
Drag the slider to compare the spread of outcomes.
Think beyond single datapoints, and past performance is absolutely indicative of future returns. But it’s not about predicting the future with precision, which is an impossible task. Instead, it’s about understanding a portfolio’s fundamental nature to match an investment to your goals. Nothing is ever guaranteed, but I believe that truly understanding the full range of past performance of a portfolio is a prerequisite to making an educated decision that stacks the odds of success in your favor.
Before we move on…
(*) One example where an investment’s fundamental nature did drastically change is when the global financial system switched suddenly from a gold standard to fiat currency in 1971. As a result, gold bullion changed overnight from a literal cash equivalent to an unpredictable investment even more volatile than stocks. You can read more about that in this article about gold, but for today’s purposes just know that the trigger was a deliberate government policy change and not some market mystery.
How to effectively study past performance
Connecting the unpredictability of the markets with the reality of a portfolio’s history is where most people get tripped up. I get it. Embracing uncertainty requires a certain mindset that takes practice.
One way to think about it is darts. While looking at the last dart to predict the next one is obviously a fool’s errand, studying the complete track record of the thrower is an excellent way to understand the likely range of outcomes so that you can plan accordingly without major surprises.
Or translated to portfolios, I think of it like this:
No portfolio is perfectly predictable, but not all portfolios are equally unpredictable.
And by finding a portfolio option where its predictability matches your expectations and can meet your needs even on the low end of potential outcomes, you can create a financial plan that allows you to save towards your goals with confidence.
To help build that confidence, I recommend these three approaches to investing research.
Ignore single timeframes
When looking at fund options at your favorite brokerage, stop paying so much attention to the default performance snapshots that they offer. And when playing with investing tools online, beware outputs that display single columns or lines. Putting too much weight into single timeframes is a sure path to disappointment.
As an alternative, take some time to tinker with the many charts I offer here at Portfolio Charts. They’re quite unique in how they’re designed to display not just one investing timeframe, but all of them simultaneously. As just one example, look at the big picture in the Heat Map and single datapoints will seem incredibly shortsighted in comparison.
Stop seeking guarantees
One common iteration of the title quote that I absolutely support with no qualifications is that “past performance is no guarantee of future returns”. I realize that when making a big money decision people want assurances. But there are simply no guarantees in investing.
Instead, look at the historical range of outcomes for a portfolio and think through how you would handle each of those situations should they happen on your own timeframe. If one of the outcomes is completely unacceptable, perhaps that’s an indicator that either the portfolio isn’t for you or you need to reevaluate your saving assumptions to aim the entire grouping higher.
There’s much more to effective investing than simply maximizing average returns. Once you find a portfolio that you’re comfortable will be close to the target, you can take it from there.
Plan for the worst, hope for the average
Instead of focusing on average returns, I would argue that a far more productive exercise is to study the times that fell well below the average. Maybe that’s the baseline return that looks at the lower-end 15th-percentile accumulation outcomes. Or perhaps it’s the safe withdrawal rate that survived even in the worst retirement timeframe on record. Almost nobody receives the exact average, and planning for the worst is a great way to mitigate any fear of an unknown future.
While some people may consider that pessimistic, I find it empowering. Create a plan that succeeded even in the worst historical case, and your chances of future success are extremely high with lots of potential upside.
It’s all about perspective
From flying darts to portfolio growth paths, uncertainty is an unavoidable part of life. But rather than shying away from past performance as an unreliable predictor, try leaning into it. There’s a very good chance that the issue is not with the data itself but simply with your perspective.
What do you really want out of investing research?
If your goal is to accurately predict future returns in order to perfectly maximize your account balances, then I have bad news for you. It’s probably not going to happen. Sure, you may get lucky for a while but it doesn’t mean that you know what you’re doing. And when your luck inevitably runs out, you’ll be left holding the bag and wondering what went wrong.
But if your goal is to become a student of history so that you can understand possible futures and choose a path with a decent chance of success, then I’m happy to report that it’s very much attainable. There are numerous ways to utilize past performance data to constructively estimate the range of future returns.
Guarantee? Of course not. But point you in the right direction with a reasonable margin of error? Absolutely.
Don’t be scared away from genuinely helpful information by dismissive legal statements. And steer clear of the cynicism that traps you in a nihilistic belief that past performance is irrelevant and nobody knows nuthin’. Historical data is my specialty, and I’m here to tell you that it can serve you very well if you just take the time to learn how to properly use it. And tools like Portfolio Charts are here to make that process easier than ever.
Past experience illuminates the path forward. Embrace knowledge to light the way.
Join the conversation
