Gray swan neglect: Do forecasters account for low(ish) probability events?

Area of Science:

• Behavioral Economics
• Cognitive Psychology
• Financial Decision-Making

Background:

• Economic choices rely on future predictions.
• Individuals sometimes simplify predictions by focusing on the single most likely outcome, treating it as certain (digitization bias).
• This bias's occurrence in high-stakes economic contexts is not well understood.

Purpose of the Study:

• To investigate the digitization bias in predicting future asset prices.
• To determine if individuals ignore conditional probabilities of unlikely events in financial contexts.
• To assess the prevalence of this bias across different prediction types and participant groups.

Main Methods:

• Four sets of studies were conducted involving predictions of future asset prices.
• Participants were presented with conditional probability information for both likely and unlikely events.
• The studies examined binary (direction) and continuous (expected value) predictions under various conditions, including incentive-compatible tasks and financial professionals.

Main Results:

• Participants consistently ignored conditional probability information for unlikely events.
• Reliance was placed entirely on conditional probabilities associated with more likely events.
• This digitization bias persisted even when unlikely events had probabilities as high as 30% and was observed in financial professionals.

Conclusions:

• The digitization bias significantly impacts financial predictions and economic choices.
• This bias demonstrates a systematic deviation from rational probabilistic cognition in high-stakes scenarios.
• Findings have implications for understanding decision-making in behavioral economics and financial markets.