There is a large body of academic evidence demonstrating that individual investors are subject to the “disposition effect.” It has been documented among U.S. retail stock investors, foreign retail investors, institutional investors, homeowners, corporate executives and in experimental settings.
Those suffering from this phenomenon, which was initially described by Hersh Shefrin and Meir Statman in their 1985 paper, “The Disposition to Sell Winners Too Early and Ride Losers Too Long: Theory and Evidence,” tend to sell winning investments prematurely to lock in gains and hold on to losing investments too long in the hope of breaking even.
Standard explanations for the disposition effect—such as tax considerations, portfolio rebalancing and informed trading—have been proposed and dismissed, leaving explanations that rely on investor preferences, such as prospect theory. Prospect theory implies a willingness to maintain a risky position after a loss and to liquidate a risky position after a gain. It also requires that investors derive utility as a function of gains and losses rather than the absolute level of consumption.
As Tobias Moskowitz explained in his 2010 AQR working paper, “Explanations for the Momentum Premium,” the disposition effect “creates an artificial headwind: when good news is announced, the price of an asset does not immediately rise to its value due to premature selling or lack of buying. Similarly, when bad news is announced, the price falls less because investors are reluctant to sell.” The disposition effect therefore creates predictability in stock returns and thus provides an explanation for the momentum premium.
Additional research into the disposition effect, including a 2012 study by Itzhak Ben-David and David Hirshleifer, “Are Investors Really Reluctant to Realize Their Losses? Trading Responses to Past Returns and the Disposition Effect,” has found that investors sell more when they have larger gains and losses.
Stocks with both larger unrealized gains and larger unrealized losses (in absolute value) will thus experience higher selling pressure. This temporarily pushes down current prices and leads to higher subsequent returns when future prices revert to their fundamental values.
Joseph Engelberg, Matthew Henriksson and Jared Williams provide the latest contribution to research on the disposition effect with their January 2018 study, “The Portfolio-Driven Disposition Effect.” They sought to determine whether the disposition effect operates at the individual asset level or at the portfolio level.
The authors’ data sample was the same as the one employed by Brad Barber and Terrance Odean in the study, “Trading Is Hazardous to Your Wealth: The Common Stock Investment Performance of Individual Investors,” which appeared in the December 2002 issue of The Journal of Finance. It consisted of 78,000 households with 158,000 accounts between January 1991 and November 1996.
Following is a summary of Engelberg, Henriksson and Williams’ findings:
- There is no disposition effect for a stock if the remaining portfolio is up. However, if the remaining portfolio is down, a stock with a gain is more than twice as likely to be liquidated than a stock with a loss. (Note that, at least for taxable accounts, this is exactly the opposite of what efficient tax management would require, which is the harvesting of losses). The results were significant at well below the 1% level of statistical significance (t-stats were in the 20s).
- Results were not impacted when the authors screened for the most extreme winners and losers.
- The authors reached the same results when they considered proxies for investor sophistication, such as professional jobs or high income.
- They even examined the possibility that investors were engaged in “framing tricks” to maximize their realization utility. For example, perhaps investors are more willing to recognize a loss when their portfolio is at a gain because they are able to match the losing stock with a winning stock whose gain exceeds the losing stock’s loss. By realizing both transactions simultaneously, the investor can mentally account for this dual transaction as a single realized gain. To ensure this was not driving results, they restricted their sample to days on which an investor sells just one stock. Again, they found the same results.
Engelberg, Henriksson and Williams concluded there is a large disposition effect only when the remaining portfolio is at a loss. As to the most likely explanation for the effect, they concluded that “investors derive utility from both paper gains and realized gains and that they take utility by realizing gains when they have disutility from unrealized losses.” In other words, “when their portfolio has paper losses, they compensate by realizing gains.”
They explain: “When an investor’s overall portfolio is down, the investor will receive a lot of negative utility from the paper losses, so she should be especially likely to seek a burst of positive utility from realizing a paper gain to offset some of the negative utility she has received due to the poor performance of her portfolio. This could explain why we find such a strong disposition effect when an investors’ portfolio is down.”
Another interesting finding the authors presented was that when a stock is at a gain, but the portfolio is at a loss, upon realizing the gain, investors are most likely to keep the proceeds in cash—it is important to investors that the gain “stay” realized.
The well-documented disposition effect not only provides us with explanations for behavioral-based anomalies, such as momentum, but also with an opportunity to better understand how our behaviors can negatively impact our results. We cannot learn from our behavioral mistakes unless we are aware of them.
Larry Swedroe is the director of research for The BAM Alliance, a community of more than 140 independent registered investment advisors throughout the country.