There are a lot of cliches in retirement planning. And a lot of them are cliches for a good reason – they’re good advice. But not always, and more than a few of them are right for the wrong reasons.
One of those cliches is that with investing you need to focus on the long term to reduce the risks from investing. It is absolutely true that investing is a long-term activity, but is that second part true? Do stocks (and other types of investments) get safer the longer you hold onto them?
What Does Long Term Even Mean?
Before we start looking at whether stocks are safer in the long run, it’s worth pausing to ask what does “long run” even mean? Because it’s never really addressed – it’s just sort of not the short run. Which, yes, is true. But it isn’t all that useful.
Part of the problem is the fundamentally (nearly) random and chaotic nature of the financial markets. The whole purpose of focusing on the long run is to get past all of the noise in the markets. If you simply wait long enough, all of the short-term volatility will cancel itself out, and you’ll just be left with the underlying expected return, right?
This is one of those things that is both true and false at the same exact time. The financial markets are absolutely full of short-term noise that essentially means nothing and should be ignored. But that’s not quite the full story.
As an example of this, the S&P 500 Index has gone up by about 520 points year to date (January 1, 2023 to October 17, 2023). But going by the daily closing prices, it bounced around by about 5,496 points. The sum of the daily movements has been a little more than 10.5 times the actual movement over that same time. To put this in perspective, this is the equivalent of driving from my hometown of Portland, Maine to Washington, DC, where Retirement Researcher is based (a drive of 539 miles), by way of Las Vegas with a stopover to see the Grand Canyon (5,385 miles in total1). And this is over the course of less than a year.
But even year to year returns are still pretty noisy. You still need to string together a number of years to minimize the truly random noise that is constantly present in market returns. Just like with diversification there is no bright line where you stop being undiversified and start being diversified, there is no specific point where short term turns into long
What’s Left After the Noise
However, the comparison of time horizon to investment diversification extends even further (in fact, about the idea that investing gets safer the longer you do it is often referred to as Time Diversification). Once we get rid of the random portion of the return, we are still left with the underlying risk characteristics.
Investing in the financial markets is an inherently risky activity. And not on that, this risk is exactly why we invest in the financial markets. The whole point of investing is to capture higher expected returns, which do not exist without the corresponding risk. And this risk is not the simple random noise that goes away by simply waiting it out. The risk is fundamental to the investment.
We can see this clearly when we think about buying a single stock. When you buy shares in IDEXX (one of the few publicly traded companies in Maine) you’re making a bet on that company. There is a fundamental risk as to how that company will do in the future. And that risk will never go away. You can hold those shares for 50 years and that risk will still be there. The returns we are looking for from our investment is effectively compensation for holding onto this risk.
We can get risk of a lot of the random noise by simply waiting until it cancels itself out, but we still have the risk inherent to the investment. No matter what you do, investing is risky. And we should be happy about that.
Looking at the (Simulated) Outcomes
Let’s set aside the theory for a minute though. Let’s say that time diversification does actually work and that we can cancel out the year-to-year risk. What does that actually do to our retirement plans?
One of the fundamental truths of investing is that we invest to be able to spend in the future. We put our money at risk so that we can consume more than we saved. We care about our total return – the amount that we can withdraw from our investment accounts. Well, let’s look at that.
By using a simulation technique called Monte Carlo Analysis we can run a whole bunch of simulations and see how often certain things happen. Essentially, we randomize one (or more) key variables across different simulations and see how things play out. In this case, we’ll run a thousand simulations, and we’ll just vary investment returns in each year of the different simulations.
So, what are the results?
Well, pretty much what we expected to see (running a thousand simulations does tend to round off the sharp corners in the numbers). To start off with, we assumed an average annual return of 8%, and a standard deviation of 10%. We also assumed an initial portfolio value of $100,000, and we’re going to let the simulations run for 50 years each (though we are going to be checking in on them periodically).
First things first, the observed annual returns tracked like we would expect – we looked in at 10 year intervals, and all of the average returns were right around 8%. That’s good to see, but not particularly interesting.
What is interesting is the distribution of the outcomes. Even eliminating the fundamental nature of investment risk – all of the deviation from the 8% average return is purely random noise that should cancel itself out – we still see substantial differences in outcomes even out to 50 years. And if we look at the result in dollar terms, especially at 50 years.
We can look at this distribution in a whole bunch of ways, but one of the most intuitive ways is to look at the 25th and 75th percentile results. The 25th percentile result is the one that is better than 25% of the results, and the 75th percentile result is the one that is better than 75% of the results. So, in this case, since we are working with 1,000 simulations, the 25th percentile is the 250th best result, and the 75th percentile is the 750th best result. In other words, half the time you are between the 25th and 75th percentiles and half the time you are outside of them. You can certainly do a lot of math and get a much more precise picture of the distribution, but this works for eyeballing purposes.
Let’s look at the numbers.
|10 Years||20 Years||30 Years||40 Years||50 Years|
|25th Percentile||$174,687 (5.74%)||$328,342 (6.12%)||$636,198 (6.36%)||$1,247,758 (6.51%)||$2,485,797 (6.64%)|
|75th Percentile||$251,942 (9.68%)||$581,241 (9.20%)||$1,288,568 (8.89%)||$2,756,958 (8.65%)||$5,844,599 (8.48%)|
There are a couple of things that are worth pointing out here. Focusing on the returns, the longer the simulation has run – the tighter the returns get. There was a difference of nearly four percentage points between the 25th and 75th percentile returns at 10 years, but less than a two percentage point difference between the 25th and 75th percentile returns at 50 years. The longer the simulation runs the more the good and bad returns cancel out on average.
But there’s an important effect when we start looking at the total returns (the actual dollar values) rather than just the annualized return numbers at different points in time.
|10 Years||20 Years||30 Years||40 Years||50 Years|
|Ratio between 75th and 25th percentile||1.44||1.77||2.03||2.21||2.35|
Even with just the purely random part of the risk, the difference between the good and bad outcomes actually increases through time. If stocks were truly safer in the long run, we should not be seeing this – we should be seeing these numbers converge. The fact that we don’t, even in the best-case scenario for time diversification, says that it just doesn’t work.
Stocks do not get safer the longer you hold them. But let’s turn back to the theory for a minute.
The Markets Only Look Forward
In our simulation, the returns were truly random (or as random as a computer can be). But those returns were centered on a specific number. And the longer the simulations ran the more confident you could be that your observed returns would converge on that number.
The real world does not operate like that. There is no specific number that is the “true” return for any given security or the market as a whole. Markets are constantly moving and adjusting. And that means that the expected return of every security is constantly in flux – especially over the long term.
The markets do not care about what has happened in the past. No good or bad returns are ever due. The market is only concerned about the future. What happens next – and how that squares with expectations – is what will move prices. Not how well a security has done against previous expectations.
To steal an aphorism from the traders, you shouldn’t hold something you wouldn’t buy today. It’s (very) easy to get carried away, but if you wouldn’t design the same portfolio, you have today as you did in the past, you should be thinking about why your portfolio looks the way it does.
Discipline and focusing on the long term are absolutely crucial to having a good investment experience and successful retirement. But it’s not because investing gets safer over time. It’s because, just like with traditional diversification, you can get past the noise and focus on the fundamental risk and return relationships that drive the returns to help us achieve our retirement spending goals.
- All distances from Google Maps. ↩︎