Can Average Investors Really be as Bad as Studies Say?

There is no shortage of studies showing that average investors underperform the market averages, often by 2-3% per year.  However, Barry Ritholtz’s excellent book How Not to Invest says that average investors underperformed by 5% per year one decade and only 2% per year the following decade.  How could this be?  It doesn’t appear to make sense.  So, I started digging.

The Data

Here are simplified version of the chart I saw on page 382:

So, average investors trailed a basic 60/40 portfolio by 2.4 percentage points per year.  This is plausible.  Investors pay expenses, and some of them do some market timing.  Here is what I saw on the next page:

This time the behaviour gap is 3.8 percentage points per year for 20 years.  Ritholtz makes the point that “The longer the holding period, the greater the impact of errors that disrupt compounding.”  This is true as it applies to the final dollar value of your holdings.  After all, 20 years of mistakes hurts more than only 10 years of mistakes.  But these charts show annualized returns.  If investors made the same mistakes through all 20 years, we’d expect the annualized gap to stay about the same from one decade to the next.

The big increase in the behaviour gap is even more surprising given that half the 20-year returns overlap with the 10-year returns.  That first decade must have been terrible for average investors.  Using the 20-year returns and the second decade returns, we can calculate the first decade returns:

60/40 portfolio:   (1 + 7.4%) x (1 + 7.4%) / (1 + 11.1%) - 1 = 3.8%.
Average investor: (1 + 3.6%) x (1 + 3.6%) / (1 + 8.7%) - 1 = -1.3%.

Here is the first decade in chart form:

The Mystery

How could the average investor trail the basic 60/40 portfolio by 5.1 percentage points per year for a whole decade?  This is a huge gap.  The cumulative loss over the decade is 40%!  Were average investors really this dumb for the first decade and then smartened up for the second decade?

Ritholtz credits J.P. Morgan’s “Guide to the markets” for this data.  J.P Morgan puts out a quarterly report containing dozens of charts, including a chart comparing the average investor to various U.S. market averages.  The 2022 Q3 Guide to the Markets gives the 2002-2021 annualized returns, and the 2023 Q1 Guide gives the 2012-2021 annualized returns.

These guides have the same return percentages as reproduced in the book; there was no transcription error.  J.P Morgan’s data agrees that average investors trailed a 60/40 portfolio by 5.1 percentage points per year from 2002 to 2011.  If we were looking at just one investment, such as the ARK Innovation ETF, wild swings from one time period to another are certainly possible, but the J.P Morgan data represents all U.S. investors.

The Solution

The solution to our puzzle is buried in the footnotes of J.P Morgan’s charts.  The footnote to the 10-year chart includes

“Average asset allocation investor return is based on an analysis from Morningstar.”

From what I have seen, Morningstar uses a sensible methodology to calculate the average investor's behaviour gap.  However, the footnote to the 20-year chart includes

“Average asset allocation investor return is based on an analysis by Dalbar Inc., which utilizes the net of aggregate mutual fund sales, redemptions and exchanges each month as a measure of investor behavior.”  

Dalbar’s methodology for measuring the average investor’s behaviour gap is not reasonable.  Here is a quote from their 2016 “Quantitative Analysis of Investor Behavior”:

“QAIB calculates investor returns as the change in assets, after excluding sales, redemptions, and exchanges. This method of calculation captures realized and unrealized capital gains, dividends, interest, trading costs, sales charges, fees, expenses and any other costs. After calculating investor returns in dollar terms (above) two percentages are calculated:
  - Total investor return rate for the period
  - Annualized investor return rate
Total return rate is determined by calculating the investor return dollars as a percentage of the net assets, sales, redemptions and exchanges for the period.
Annualized return rate is calculated as the uniform rate that can be compounded annually for the period under consideration to produce the investor return dollars.”

Dalbar treats all fund flows as though they should have taken place at the start of their measurement period.  If you’re dollar-cost averaging from your salary, you’re treated as a fool for not having invested all the money 20 years ago, even though that makes no sense.

An example

Here’s a simple example to illustrate the problem.  Suppose you put $25,000 into an investment 20 years ago, and it has doubled three times to $200,000.  If this were the end of the story, Dalbar would say that the investment averaged 11% per year, and you averaged 11% per year for a behaviour gap of zero.

Let’s say you got a $100,000 inheritance in the past year.  You added it to your portfolio, but the return has been flat since then.  You have a total of $300,000 now.  The investments you used still produced an average return of 11% per year for 20 years.  But Dalbar treats you like a fool for not having invested the inheritance 20 years ago, even though that makes no sense.

As Dalbar sees it, you should have a million dollars now because your entire $125,000 should have been invested 20 years ago and been allowed to double three times.  Instead, you only have $300,000 now.  By Dalbar’s math, that’s an average annual return of 4.5%.  You have a behaviour gap of 6.5 percentage points per year.  Don’t you feel dumb for getting an inheritance and ruining your investment record?

Conclusion

So, the mystery is solved.  We don’t know what the actual behavior gap was for average investors from 2002 to 2011 because Dalbar’s numbers give no useful information.  Dalbar’s behaviour gap numbers get repeated all over the internet uncritically because they give financial advisors the answer they want: that average investors desperately need financial advice because they give away a huge fraction of their potential returns through their preventable mistakes.  I would be pleased if J.P. Morgan stopped quoting Dalbar behaviour gap figures, and if Ritholtz corrected his book for future printings.