DALBAR’s response to this article appears at the end of this article.
Boston-based DALBAR has published updates of its “Quantitative Analysis of Investor Behavior” study annually since 1994. The study is meant to educate investors about how the returns they earn generally lag behind the returns for market indices widely reported in the media. The study analyzes the sources of poor investor performance, finding that the bad timing behavior of buying high and selling low is the main culprit, along with fund expenses, the need for cash and a lack of cash to invest. These are all valid issues. However, my concern is that the quantitative results of the study do not properly measure the underperformance of investors. DALBAR’s method for calculating average investor returns unfairly understates these returns. There is a methodological flaw that I will explain, which is that DALBAR does not properly calculate an internal rate-of-return for an ongoing series of cash flows, which renders its results meaningless.
The DALBAR study is widely relied upon by financial services professionals. Countless advisors and investment companies cite it as an empirical demonstration of poor investor performance. I have even made such references to the study.
Yet, I have only been able to find two authors who have been critical of the study in the past. Harry Sit (The Finance Buff) was the first, and his analysis, provided in a guest post at Michael Kitces’ blog, led me onto the path of finding an even deeper problem with the study. Sit properly demonstrated the point that the market index is based on time-weighted returns (assuming the investment of a lump-sum amount at the start of the period), while investors with ongoing savings and distribution needs will experience a different money-weighted return. With ongoing contributions to investments over time, an investor will underperform the market index if returns tend to be relatively higher in the early part of the investment period when less is invested, and lower in the latter part of the investment period when more funds are invested.
That is a very important point, but it is only part of the story of what the DALBAR study is getting wrong. A key point from Sit’s article that led me on this journey was his reference to how the DALBAR study also shows results for a dollar-cost averaging investor. I’ll come back to this shortly.
The other critical piece I have found about the DALBAR study is an excerpt from the book, The Three Simple Rules of Investing: Why Everything You’ve Heard about Investing is Wrong – And What to Do Instead, by Michael Edesess, Kwok L. Tsui, Carol Fabbri, and George Peacock. That excerpt is posted here at Advisor Perspectives. The authors cited Sit’s analysis (incorrectly attributed to Kitces) about the confusion of money-weighted and time-weighted returns. They also made a new, relevant point about how if mutual fund investors (which includes many professional investors) are underperforming the market so dramatically, then who exactly is on the other side of these trades to outperform by so much? This remains as an unsolved mystery.
Let’s now dig into the problems of the DALBAR study. While the study provides analysis for rolling periods with lengths ranging from 12 months to 30 years, the 20-year investment results serve as the baseline for discussion about the study. The DALBAR study looks at investor returns for equity funds (compared to the S&P 500 as a benchmark), fixed-income funds (compared to Barclays Aggregate Bond Index) and a balanced-asset allocation fund that is not compared to a benchmark. I will base my analysis on the equity funds analysis for 20-year periods, with comparisons to the S&P 500. My analysis of S&P 500 returns is based on data from Morningstar, which leads to slightly different numbers than found in the DALBAR study reports, but this is a minor issue.
Table 1 shows the 20-year annualized returns based on monthly data for different rolling historical periods. For instance, the 2003 numbers represent the period from January 1984 to December 2003. Most recently, the row for 2016 represents January 1997 through December 2016.
By Wade D. Pfau, read the full article here.