I just received an annual statement from my 401K retirement plan showing the costs of the various funds the plan offers.¹ If you have an employer-operated 401K or similar retirement plan, you may have wondered why you get this cost statement each year. My statement says it’s because of Department of Labor regulations, but I give the real credit to John Bogle. He invented the index fund in 1975 and founded Vanguard, an investment services company that has grown into the largest index fund provider in the world. Throughout his life, Bogle championed more transparent and lower fund costs, and as a result, he saved a whole generation of individual investors tons of money. He died at the age of 89 in January of this year.
Regarding the fees and other expenses that all investors pay to mutual and exchange-traded fund (ETF) providers, Bogle said this:
- “In investing, you get what you don’t pay for. Costs matter.”
What exactly did Bogle mean by this? After all, couldn’t it make sense to pay an expert for additional returns that exceed the fees charged? Bogle pointed out at least two important reasons why costs matter:
- Costs erode returns substantially over the long term.
- Funds with higher costs often produce lower returns.
I briefly covered these two cost issues in a previous post. But I think it’s worth commemorating the passage of John Bogle by taking a closer look at these two issues. My recent 401K statement provides a good way to examine the real-world cost choices that individual investors face.
Costs Erode Returns
My 401K statement lists the “expense ratios” of all the funds offered in the plan. An expense ratio includes all the costs the fund charges, expressed as a percentage of money invested in the fund.² The expense ratios for the funds in my plan range from 0.04% to 1.31%.
An expense ratio of even 1.3% may not sound like much. But look at this graph showing how much the 30-year growth of $10,000 is slowed by these seemingly small expenses.
This calculation uses monthly compounding of the historical average annualized return of the stock market (about 9%) and then subtracts out the expenses. In this case, the lowest expense ratio yields $45,000 more in final value as compared to the highest expense ratio. In other words, with the highest expense ratio, 37% of the potential return goes to the fund provider.
You see quite a few of these comparisons on the internet. But they’re somewhat misleading because they unrealistically assume a one-time investment contribution. A more realistic comparison is to calculate the compounding associated with monthly contributions. Here’s a similar graph but assuming monthly contributions equal to $19,000 per year, the maximum allowed under U.S. law.
Even with this more realistic scenario, the difference between the lowest and highest expense ratios is about 26% of the final value. And that equals a whopping $650,000 in lost returns over 30 years!
However, this scenario is still pretty misleading. Specifically, no 401K plan intentionally offers two identical funds with widely disparate expense ratios. My 401K plan offers a more realistic example of the kind of fund choices and range of costs that most investors encounter. But before we explore those choices, let’s look at Bogle’s second complaint about fund costs.
Higher Costs Often Produce Lower Returns
Many studies have looked at the relationship between fund costs and returns. Here’s a chart from one study by Vanguard that looked at costs and excess 10-year returns associated with small-cap blend funds. “Excess returns” refers to the amount of return above a comparable index benchmark.
Charts like this are used to argue that higher costs are correlated with lower returns. But in all fairness, the data are very scattered. Although no coefficient of determination (a measure of correlation strength) was provided for this chart, I suspect the negative correlation is quite weak. On the other hand, these data clearly show there’s no positive correlation between higher costs and higher returns. Another Vanguard study assessed a wider variety of stock funds as shown in this graph.
The picture is remarkably consistent across all types of stock funds. There’s a weak negative correlation between excess returns and costs but with a large scatter in the data.
Morningstar looked at this question another way using a “success ratio”, which counts the proportion of funds that survived (stayed in business) and achieved better returns than the average of all funds in the same category. Here’s the key graph from that study looking at 5-year returns.
They split all funds in each category into five groups (quintiles) arranged from low to high cost. The success ratio for the lowest cost quintile was in the 50% to 60% range across the categories, while the success ratio was only in the 15% to 25% range for the highest cost quintile.
Now we can start to see why Bogle said, “you get what you don’t pay for”. All other things being equal, the more you pay in fees and expenses, the more likely your returns will be diminished.
Limited Choices
This is pretty much where I’ve stopped my cost analysis in previous posts. But while looking at my 401K statement, it struck me that many investors don’t have free reign to pick any fund (or combination of funds) in the world. Given that 401K plans (and similar 403b plans) are among the most common tax-deferred retirement accounts, many folks have far fewer options.
For example, my 401K plan offers 32 fund choices, 12 of which are target-date funds, which are nearly identical except for the ratios of stocks to bonds. The rest of the funds include several bond funds as well as stock funds that focus on international stocks, different stock sizes (large cap to small cap), and different stock styles (mostly growth versus value). What’s the correlation between expense ratio and excess 10-year returns within this limited set of choices? Here’s the graph.
The relationship between cost and excess returns is essentially random, with a coefficient of determination that’s less than 0.01. (A value of 0 indicates no correlation and a value of 1 indicates a perfect correlation). So, within my 401K plan, higher costs don’t predict lower returns. But more to the point, costs don’t predict anything useful about returns. Paying for expertise that has a coin flip’s chance of success is pretty irrational. You might as well flip the coin yourself and keep the fees in your pocket.
Uncertain Returns
However, you probably noticed the handful of relatively expensive funds in my 401K that had excess 10-year returns in the 2% to 3% range. Because all the returns reported for my 401K factor in costs, some of these funds have provided excess returns over the last 10 years well beyond the fees they charged. Why not pick one of those funds and reap the additional returns that aren’t available from index funds?
One of the most pervasive warnings from investment companies is: “Past performance is no guarantee of future results”. It’s such an old cliche that we can become numb to its true meaning. In this case, 10 years of excess returns are no guarantee of another 10 years of excess returns.
In fact, I’ve summarized loads data in the pages of Mindfully Investing showing that past high fliers tend to crash and burn moving forward. And despite the existence of a few superior performing funds in any given period, between 80% to 95% of all funds regularly fail to outperform their benchmarks. Picking the most recent high flier is such a common mistake that it’s pejoratively known as “chasing performance”. Investors rush to the hottest table in the casino only to find the luck has gone cold after they place their bets.
The basic choice faced by 401K investors is whether to buy the safe low-cost index fund, which guarantees middling returns (0% excess return) or pay additional fees and take a chance on the most recent hot performer. From the way I’ve set this comparison up, you already know which is the more mindful choice. But here’s some proof using my 401K as an example.
Examining returns over one or two periods (my 401K statement provides 1-year, 5-year, and 10-year statistics) only gives you a snapshot of a fund’s ongoing performance. What if we instead examined returns over many periods? There’s a lot of different ways to do this, but I chose to compare 5-year rolling annualized returns for the life of several of the funds in my 401K plan³ including the three stock index funds (large-cap, mid-cap, and small-cap funds) and the most similar (not identical) actively managed funds offered in the same plan. Here are the graphs for each of these comparisons based on fund stock size.
Years shown on the horizontal axes indicate the end of each 5-year period. Dotted lines indicate periods when excess returns for the active funds dipped into negative territory, which means that the index fund performed better than the active fund. There were two similar active fund choices for mid-caps in my plan, so I compared both active funds to the mid-cap index fund.
Three out of the four active funds shown above underperformed the index funds about half the time. For example, if in 2013 and you picked my plan’s active large-cap fund because of its recent hot streak, you would have been disappointed by negative excess returns up through today. I’ll note that the mid-cap graph shows that the active fund represented by the orange line has done pretty darn well for nearly 18 years. But even that fund started with a 5-year excess return near -3%, and there’s no guarantee that its recent superior performance will continue. Researchers have consistently failed to find ways to predict which active funds will have superior future performance.
Conclusions
My 401K plan nicely illustrates that even when you’re faced with a limited set of options that show no apparent correlation between cost and returns, the lowest cost option (index funds) is still going to be the most mindful choice.
Wider evidence also clearly shows that the same conclusion applies to the entire universe of mutual funds and ETFs. S&P Dow Jones Indices regularly tracks the ongoing performance of over 500 U.S. funds in their SPIVA Scorecard to see if superior performance “persists” over different periods. In their latest report, they found that none (zero) of the funds were able to stay in the top quartile of performance over the last five years. And given that 82% of U.S. funds they examined failed to beat a comparable index, falling out of the top quartile means falling into negative excess returns.
John Bogle knew that “costs matter” because higher costs hobble every fund’s performance over the long-term. This is true regardless of whether you have a limited set of options or can pick any fund in the world. While at times it may look like a small subset of fund managers are justifying their higher expenses, careful tracking of fund performance, individually and in aggregate, makes it clear that any superior performance is a temporary aberration.
Finally, the comparisons I made here between index funds and similar (not identical) active funds are somewhat apples-to-oranges. Ideally, funds are compared to other funds that share the exact same benchmark. A few professionals have criticized me in the past for making these more generic comparisons between funds. But here’s the problem, rarely if ever will any 401K plan give you the option of an index fund and active fund that attempt to do exactly the same thing. My comparisons accurately reflect the real-world choices available to the 401K investor. Of course, the SPIVA results show that even if you have every choice in the world, low-cost index funds are still the best option, regardless of the merits of my analysis here.
1 – I officially retired a couple of years ago, but I’m still invested in my former company’s 401K plan. I occasionally do an hour or two of billable work for my company and the 401K provider uses that as an excuse to hold onto my money.
2 – The expense ratio does not include additional costs associated with the operation of the 401K plan itself. In my statement, operation costs are described with a page of confusing, jargon-filled text.
3 – Five years seemed like a reasonable timeframe given that barely 10 years of data exist for some of these funds.
