When betting at a casino or taking on risk in the stock market, people tend to start out with the same strategy: stop when you’re losing and keep going when you’re winning.
In reality, they do just the opposite, according to research by Arizona State’s Rawley Heimer, University of Bonn’s Zwetelina Iliewa, Chicago Booth’s Alex Imas, and University of Mannheim’s Martin Weber.
“Once they experience a win, they realize, ‘Oh, man, I want to lock that down. I want to take that to the bank,’” Imas says. So they cash out early. The opposite happens when they experience loss—not wanting to end up with less money than they started with, people tend to keep gambling, he says.
That’s why investors often leave money on the table by bailing when they’re winning and pile up losses by sticking with bad bets, the researchers find. The results throw into question a European Union policy requiring brokerages to notify clients when their investments are eroding.
Across lab and field studies, the researchers find a discrepancy between what people plan and what they actually do with their money. For the field study, the researchers examined data spanning June 2013 to August 2015 from an online brokerage with around 188,000 traders in 150 countries. The brokerage requires traders to set loss and gain limits before buying an asset, and it records revisions the traders make to initial limits.
For example, a trader might set an initial gain limit of 20 percent and a loss limit of 10 percent. If the asset loses 10 percent of its value, the position is closed. But if the trader sees that she’s at an 8 percent loss, she can choose to change the loss limit to 15 percent in hopes that extra time will let the asset turn around.
In the field study, 46 percent of traders used a “loss-exit” strategy when opening a new position, meaning that the loss limit was closer to the purchase price than the gain limit; 32 percent followed a “gain-exit” strategy; and 22 percent set a neutral strategy, in which their loss and gain limits were identical. This tracks with a theory that most people plan to pull out of a gamble as a result of loss rather than gain. But instead, the most frequent response to accumulated losses was to widen the loss limit. Traders were also likely to close the position on rising stocks before reaching their initial profit target.
A lab experiment produced similar results. Online participants were asked to choose whether to invest 50 cents per round with a maximum of 26 rounds, knowing that with each gamble the 50 cents would either double or get wiped out. Participants most often chose to stop gambling when they were winning and keep going when they were losing.
At one online brokerage, most traders facing a loss opted to increase their previously set loss limit, while traders whose positions were gaining tended to exit early.
One motivation for the research is the EU regulation aimed at improving investor welfare by requiring brokers to disclose to clients when their positions lose a certain percentage. The law assumes that giving investors this information will lead them to sell in order to avoid incurring large losses. But, says Imas, it might instead cause them to dig in and take on even more risk.
The researchers find that a minority of people realize it’s difficult to stick to a loss-exit strategy. “It turns out that 20 percent of people are wary of gambling because they feel like, ‘Look, I know myself, and I’m going to deviate and be in worse shape,’” Imas says.
For those not in that minority, “Think about whether you would take that same risk if it was offered to you just once,” he suggests. People generally feel safer taking a risk when they think they’ll have a chance to back out of it later, even though the research finds that they don’t stick to that plan. So when presented with that initial choice, Imas says, reframe the question. Then, if you wouldn’t take that gamble without an opportunity to change course, he says, “don’t take it when you can adjust your strategy.”
Rawley Heimer, Zwetelina Iliewa, Alex Imas, and Martin Weber, “Dynamic Inconsistency In Risky Choice: Evidence from the Lab and Field,” Working paper, February 2023.
New York University’s Jonathan Haidt discusses the effects of childhoods spent under the influence of smartphones and overprotective parenting.
Capitalisn’t: The Economic Costs of a Phone-Based Childhood
Analyzing both numbers and narrative context in company reports led to better price forecasts.
Large Language Models Can Improve Stock Market Forecasts
A US president’s comments tend to calm volatility around climate-related stocks and bolster asset prices by resolving uncertainty.
Climate-Policy Pronouncements Boost ‘Brown’ StocksYour Privacy
We want to demonstrate our commitment to your privacy. Please review Chicago Booth's privacy notice, which provides information explaining how and why we collect particular information when you visit our website.