Saturday, November 28, 2015

See! I was right

University of Iowa


A new study from the University of Iowa finds that once people reach a conclusion, they aren't likely to change their minds, even when new information shows their initial belief is likely wrong and clinging to that belief costs real money.

The study, co-authored by Tom Gruca, professor of marketing in the Tippie College of Business, has implications for understanding financial markets. 

He says equity analysts who issue written forecasts about stocks may be subject to this confirmation bias and do not let new data significantly revise their initial analyses.

Gruca found this confirmation bias in student traders participating in the Iowa Electronic Markets over a 10-year period during which they bought and sold real-money contracts to predict the four-week opening box office receipts for a new movie. The students analyzed markets for a total of 18 movies released between 1998 and 2008.


The research shows that even as the key first weekend box office receipts were reported, prices stayed remarkably stable as traders ignored new value-relevant information and continued to rely on their initial estimates.

Gruca says confirmation bias was induced through an assignment in which students were required to explain the thinking and analysis behind their forecast prior to trading. 

Prior studies have found that people are more apt to persist in their beliefs, despite contradictory evidence, once they've written their beliefs down, a phenomenon known as the explanation effect. This often affects a person's future actions as well.

The researchers found that prices in the IEM movie markets under-reacted upon the release of new information about box office performance when all traders were subject to confirmation bias. 

This suggests traders were reluctant to change their initial opinions, even after new information became available.

Gruca also ran several markets that included traders who did not write down their analyses -- and, thus, were not subject to the explanation effect. In those markets, prices tended to be more active as new box office information was revealed, suggesting traders who did not have to write a report more readily modified their opinions and incorporated the new information into market prices.

The success of the movies on the market also had no impact on these trading tendencies, the study found. Markets with virtually guaranteed blockbusters like Monsters Inc., Harry Potter and the Sorcerer's Stone, and Twilight exhibited the same dynamic as such quickly forgotten flops as The Fountain or Lost in Space.

"This study shows that when all traders in a market have the same bias -- in this case, confirmation bias -- market prices are not efficient and do not reflect all of the information available," says Gruca. 

"However, if some traders are not biased, then market prices efficiently reflect new, relevant information."


The study, "The Power of Priors: How Confirmation Bias Impacts Market Prices," was co-authored by Michael Cipriano, associate professor of accounting at the University of South Carolina Upstate and 1998 Tippie MBA graduate. It was published in The Journal of Prediction Markets.