Amateur investors fail to diversify and are better off choosing stocks at random

Whether they’re aiming to avoid high financial management fees, control their own investments, or enjoy the thrill of playing the market, more consumers are opening investment accounts and making their own stock picks.

But a new study from the UBC Sauder School of Business has found that less experienced investors are failing to diversify — and could be putting themselves at serious financial risk. The effect is so pronounced that many amateur investors would be better off choosing stocks at complete random.

For the study, researchers first asked participants to create portfolios of financial assets using tables of previous returns, and then assessed the participants’ level of financial literacy. The researchers found the investors with poor financial literacy tended to choose positively correlated assets – for example, stocks in oil companies and forestry — which tend to fluctuate in value together.

“An amateur investor might buy stocks in lumber, mining, oil and banks, and believe they are diversifying because they’re investing in different companies and sectors,” said David Hardisty, study co-author and assistant professor at UBC Sauder. “But because all of those equities tend to move in unison, it can be quite risky, because all the assets can potentially plunge at the same time.”

More experienced investors know to hedge their bets by including negatively correlated assets, which are likely to move down when others go up — or uncorrelated assets (ones that move up and down independently of the others) in order to mitigate losses.

The researchers also found that the amateur investors were actively preferring correlated assets because they seemed less complicated and more predictable.

“If it seems predictable, it seems safer and easier to track,” explains Hardisty. “Whereas if you have a combination of assets that all go in different directions, it seems chaotic, unpredictable and riskier.”

Ironically, when the study participants were encouraged to take more risk when creating a portfolio, the amateur investors ended up making safer, more diversified selections, compared to when they were encouraged to avoid risk.

“This shows that amateur investors rely on a definition of risk that greatly differs from the objective definition of portfolio risk,” said Yann Cornil, assistant professor at UBC Sauder and co-author of the study. “This can lead them to make objectively low-risk investments when they intend to take risk, or to make high-risk investments when they intend to reduce risk.”

The researchers found that when amateur investors are shown the aggregate returns of portfolios (and not merely the returns of each asset composing the portfolio), they can see that having negatively correlated or uncorrelated assets is the winning investment strategy — even if it might seem counterintuitive to play both sides.

“If you don’t diversify, when one asset does well the other ones are also going to do well. But if one does badly it’s likely the others will all do badly — and in investing, you want to avoid those worst-case scenarios,” says Hardisty, who hopes the research will encourage investors to educate themselves on investment strategies, and use the diversification tools that online investment services provide to properly balance their portfolios.

“In the best-case scenario you could make lots of money and have an extra vacation or buy a car or something like that,” he explains of the positively correlated accounts. “But if your whole portfolio crashes you could risk losing your life savings. So, the best-case scenario isn’t that much better, but the worst-case scenario is a whole lot worse.” 

UBC Sauder School of Business Assistant Professor David Hardisty co-authored Easy, breezy, risky: Lay investors fail to diversify because correlated assets feel more fluent and less risky along with UBC Sauder School of Business Assistant Professor Yann Cornil and D’Amore-McKim School of Business Associate Professor Yakov Bart. The study was recently published in Organizational Behavior and Human Decision Processes.