A major economic assumption in most theories in finance is that individuals are rational. Yet, reality seems to suggest that people are normal not rational meaning they do not always pick the best choices available to them. Last week, at the CFA Society Sweden annual forecast dinner, BlackRock Vice President Benjamin Kelly, PhD, reminded us of
this as he talked on Cognitive Biases in Forecasting: Detection and Management. He focused on four most common investment biases, a brief summary of which I include here.
First, overconfidence. He gave an example of a test he hands out often that consists of 10 random questions on topics like how many islands there are in the Archipelago near Stockholm. The respondents are asked to give a 90% confidence interval within which a particular the answers to particular random questions lie. Most of the respondents always give a narrow band within which the answer should lie reflective of their overconfidence, perhaps between 50 and 55. The answer is actually in the thousands. Only one person had ever gotten all ten questions correctly and that was because for reach question, he gave the interval of zero to infinity interval (Smart, uh!). Notably, most forecasters, investors and investment managers tend to be overly confident in their estimates to the point of not considering the chance that they might be wrong. It is likely that the investors then seek information that affirm their predictions and avoid anything that contradict it. Essentially it is myopic in the sense that the investor does not want to face up to any contradictory evidence.
Secondly, anchoring. He gives the example of a team of analysts who met with their manager often to make predictions of the outlook for the future together. The manager would be the first to go and write his predictions in an open excel sheet before the rest of the team wrote up their predictions. Notably, their forecasts were very similar and based just near the manager´s predictions. While the manager saw this as a sign of consensus in outlook, this is actually a sign of the other employees confirming what the boss had already thought about. Anchoring is basically where the other people base their predictions not on their own judgment but on others´ predictions. Essentially, nobody wants to stand out from the rest. This bias is very evident in consensus forecasts which normally are near a mean. It is kind of a regression to the mean. The best way to deal with this is to ask each person to independently generate their estimates and send them to a central person for compilation before the meeting.
Regret bias happens where investors sell off winners too early and refuse to cede losing stocks because they want to lock in gains given their past experience of losing when they held on too long to a winning stock and of gaining when they held onto a losing Investment. Investors continue in wrong investments where they should cut their losses and exit.
A final bias he explored is sunk costs. An example would suffice: If there was a snowstorm and you had a ticket to a football game, are you more likely to buy it when you have bought it yourself or when someone else bought it for you? Most people would only go when they had bought it themselves not realizing that the money had already been paid in either case and is a sunk cost. Sunk costs should actually be irrelevant in decision making. Very often when considering investments, the investor is often worried that he has spent so much money already and is unwilling to let go of the investment even when it’s likely to be costing him.
There are several other biases as can be read here and here but these were his point of focus. When Daniel Kahneman, Nobel Prize winner and best-selling author of Thinking Fast and Slow was asked which behavioral bias he would wipe out with a magic wand if he had one (indicative of which one he considers critical), he says overconfidence (Richard Thaler, author of Misbehaving: The Making of Behavioral Economics agrees). When I asked Benjamin Kelly a similar question on which he considers most critical, he promptly answered sunk costs as. In sum, we should scrupulously re-examine our investment decisions for biases.
Erick Mokaya's Blog 