In 1949, Benjamin Graham famously quipped, "The investor's chief problem - and even his worst enemy - is likely to be himself." In my opinion, this statement perfectly captures the idea behind this series and the reasons for writing it. Ever since this statement was made, whether by conscious choice or random chance, academics have been gathering and analysing data that supports Graham's assertion. In particular, today's topic of "overconfidence" has been the subject of intense scrutiny by academics for years. To put it simply, overconfidence can be defined as the tendency to think you are right more than you actually are. Now, most of us are reminded of this daily (especially those of us with spouses). Most of the time, the consequences of our overconfidence are not particularly dreadful and we can skate by.
Today, however, I would like to highlight some academic research indicating that overconfidence may be the most endemic of behavioural errors that we are prone to as investors. And not only is it systematic, but it can be measured historically with astounding accuracy.
Unlike the previous biases and theories we have discussed, it is hard to attribute the definition of the concept of overconfidence to a particular party. It has been documented and researched since the early 20th century in various forms. Perhaps some of the more conclusive studies came out of the US Office of Naval Research during the 1970s-1980s with Lichtenstein and Fischoff. This duo and others conducted thought experiments with US armed forces personnel, British students, and CIA analysts. They asked simple questions-- e.g. what has a wider circulation Time magazine or Playboy?--and then asked respondents to share their level of confidence in their answer. By using the actual percentage of correct answers as the expected probability of getting a correct answer, the study revealed a tremendous bias towards overconfidence exhibited by the respondents. For example, when asked to distinguish between European and American handwriting, respondents were confident they chose correctly 70% to 80% of the time despite only succeeding half the time.
Psychologically-speaking, overconfidence results from the asymmetric weighting of supportive and unsupportive information. As an example, and without getting into too much detail, supportive or "confirming" information would be witnessing a company beating earnings expectations after having bought its shares or missing analysts' expectations after having sold their shares. Developments that favour an investor's thesis and confirm his/her belief tends to result in heightened confidence regarding their decision. When new developments do not support the investor's thesis, however, investor confidence does not tend to change much, if at all. This is called the confirmation bias--we generally seek out and hold onto information that supports our beliefs and ignore contrary information.
Over the years, many studies have confirmed the existence of overconfidence and several of its properties have been elucidated, namely: i) the less you know, the more confident you are, ii) the corollary to (i) is also true - the more you know, the less confident you are, iii) as questions increase in difficulty, you tend to be more confident, and iv) overconfidence is not correlated with age, gender, raw intellect, or race. To sum up, everyone is subject to overconfidence especially when the problem at hand is difficult and their personal knowledge limited. Sounds an awful lot like investing, doesn't it?
It turns out that many academics thought so as well. From 1991-1996, Terrance Odean and Brad Barber set upon a study of more than 66,000 households with brokerage accounts to look for a pattern of overconfidence in common stock purchases. Their hypothesis was that overconfident investors would tend to turn over their portfolios more frequently than those who were less confident. The intuition behind the idea makes good sense. For those of us who have sold stock A and bought stock B, we typically do so because we are more confident in stock B's future return prospects than stock A. The hypothesis they were looking to test was whether turning over one's portfolio more frequently led to any differences in realised returns. Their idea was that if all factors other than turnover were controlled for, then any difference in portfolio returns would be a function of the level of turnover. Using turnover as a proxy for investor confidence--with higher turnover being reflective of greater confidence and vice versa--any difference between portfolio returns would be a function of investors' confidence.
I should stress that this idea was relatively revolutionary. Prior to its publication in 2000, the forefront of academia largely believed that the frequency of portfolio turnover should have little to do with portfolio returns. In 1980, Grossman and Stieglitz set down their rational utility theorem that said investors will trade only if the marginal benefit to do so is greater than the marginal cost associated with making the trade. You can think of this as the benefit from owning stock B (e.g. higher expected returns) being greater than the cost (e.g. commissions, taxes, Etc.) associated with selling stock A and buying stock B. Therefore, on average, turnover should either be a net positive or have no impact on returns.
Odean and Barber upended this notion with their conclusive analysis that high portfolio turnover negatively effects returns in a big way. During the period 1991-1996, the value-weighted NYSE/Amex/Nasdaq total stock index returned 17.9% annualised compared to the average household's net return of 16.4% with their annual turnover equal to 75%. When divided into quintiles by turnover, however, the quintile with the highest annual turnover (roughly 200%) had an average annualised return equal to 11.2%.
I would argue a simple spread of returns between two groups isn't itself evidence of overconfidence unless other factors are controlled for. Almost anticipating these reservations, however, Odean and Barber constructed a Fama-French regression to control for the risks related to the market (i.e. beta), firm size, and firm book-to-market ratio. Furthermore, they controlled for different market environments and portfolio size. After this adjustment, the most rigorously supported conclusion was that investors hurt their gross and net performance by trading compared to a simple buy and hold strategy. Portfolios with the highest turnover underperformed the portfolios with the lowest turnover by a statistically significant margin of 46 basis points per month or 5.5% a year on a net return basis.
In summary, the data would have us believe that higher turnover (read "overconfidence in one's abilities as a stock/bond/fund-picker") tends to result in lower realised portfolio performance regardless of a portfolio's size, independent of up and down markets, and independent of equity styles or size or beta.
At this point, an astute reader will have noticed that the highest annualised return listed so far was the market return given by the value-weighted index composite of the NYSE/Amex/Nasdaq stock markets. While it is very possible to outperform indices and people do all the time, it is not probable--and that is the point. Most investors, both professional and amateur, tend to have a very hard time beating market indices consistently.
In recent years, investors have begun to realise that trading frequently hurts, higher fees and commissions hurt, and overconfidence hurts. As this type of discussion becomes more mainstream, passive investing via exchange-traded products (ETPs) or other index-trackers has begun to catch on. ETPs and other trackers offer lower turnover, lower fees, and less reliance on the selections of a manager--either professional or amateur. Keep in mind, however, that just because ETP investing lends itself to these uses doesn't mean it can't be abused. The same principles apply to ETP investing as anything else--if you turn over a portfolio of ETPs at a rate of 200% a year, don't expect the results to be any different than what we have discussed.
That concludes our series dedicated to the cross-roads of behavioural economics and investing. Hopefully, by keeping some of these core principles in mind, we can best our worst enemy in investing—ourselves.
Lee Davidson is an ETF analyst with Morningstar.
:quality(80)/cloudfront-us-east-1.images.arcpublishing.com/morningstar/5FNGF7SFGFDQVFDUMZJPITL2LM.png)
:quality(80)/cloudfront-us-east-1.images.arcpublishing.com/morningstar/EOGIPTUNFNBS3HYL7IIABFUB5Q.png)
