The Investment Decisions of Smart and Stupid People

We all know that as individual investors we're supposed to have a buy-and-hold strategy with well-diversified portfolios. But what happens when people deviate from this recommendation? It turns out the results don't depend on chance, but on how smart you are.

That's the finding of a new working paper by George M. Korniotis, a Federal Reserve economist, and Alok Kumar of the University of Texas at Austin.

Korniotis and Kumar first used data from a European survey administered to people above 50 years old in 11 countries in the region. The survey had information on things like verbal and quantitative ability, memory, age, income, wealth, education, gender. The researchers also derived a person's social network based on the level of social activities they participated in on the presumption that more participation led to a larger social network.

They found that cognitive ability, i.e. brain power, declines with age but is positively associated with education and size of social network.

They then use these findings -- with the help of some statistical gymnastics -- to look at how U.S. investors of different cognitive abilities have fared. (They took data from the 1990 Census and assumed that investors who live in more educated zip codes are likely to be more educated. The researchers also obtained investment behavior of some 62,000 retail investors.)

When there are no portfolio distortions (investors are sticking with common-sense recommendations mentioned above), smarter investors earn only one percent more per year than not-so-smart investors.

But when those distortions are higher (i.e. holdings are not diversified, trading activity is high, or holdings are titled towards local stocks), the difference widens to more than five percent. The results are surprising because these traits have been shown to lead to lower returns.

Korniotis and Kumar argue that this is a result of better information on the part of smarter investors and also the ability to use that information more efficiently through higher stock selection and timing skills.

Interestingly, wealth did not play a role in the results, but income did:

A very large proportion (about 46%) of high ability investors live in urban regions (within 100 miles of the top twenty metropolitan regions in the U.S.). In contrast, only about 17% of low cognitive ability investors are located in urban areas. Moreover, we find that high cognitive ability investors are not wealthier than low cognitive ability investors, but they earn significantly higher income in comparison to low cognitive ability investors ($126,342 versus $58,684). We also find that high cognitive ability investors exhibit a greater propensity to invest in foreign stocks, a weaker propensity to hold high dividend yield stocks, and are more likely to trade options or engage in short-selling. Collectively, these summary statistics indicate that high cognitive ability investors are more likely to adopt relatively sophisticated investment strategies.

Smart investors also hold riskier stocks while less-smart investors show a preference for "high idiosyncratic volatility" stocks which have been shown to provide low returns.

Perhaps the most interesting aspect of Korniotis and Kumar's paper is its implication for Social Security reform and the possible introduction of private accounts:

In light of our evidence, low cognitive ability investors might be better off if they did not invest directly in the stock market. Indirect investments using mutual funds and other forms of delegated investment management might be more appropriate for those investors. Similarly, while there have been attempts to privatize the social security system...under a fully privatized system, the welfare of households that do not make "wise" investment decisions could be adversely affected.