Information Cascades – Effects on Markets and Betting Behavior
Information Cascades – Effects on Markets and Betting Behavior
Reference:
https://www.economist.com/blogs/democracyinamerica/2011/10/mass-movements
https://www.predictit.org
In the simplest terms, Information Cascades basically occur after some time has passed, when individuals begin ignoring their own information, and base their decision off the decisions of others. To quickly illustrate, suppose someone is trying to decide between watching two different movies at a theatre that only presents 2 movies at any time. This person read a review promoting movie 1 and criticizing movie 2. He is also standing in line with 20 people ahead of himself (he is Person 21). Each person ahead of him publicly broadcasts which movie ticket they purchased. Suppose 18 of the people ahead of himself chose to watch movie 2, and only 2 people chose to watch movie 1. The idea that the the Person 21 might choose to watch movie 2, even though he read a review promoting movie 1, is an example of an information cascade. Person 21 chose to ignore his own information, and based his decision off the decisions of others.
This concept can be applied to economics as well as trading and betting. Leaders of European countries in the past have tried to “prevent a cascade imparting negative information about the value of government debt” by organizing strategic meetings. Additionally, technological advances have made it easier for people to quickly communicate their buying/selling decisions to the public. One interesting application of this concept can be shown through the example of the political prediction market (gambling site) “Predictit.org”. It is interesting to think about the payoff/reward for following information cascades, verses choosing to use one’s own information to make a prediction. One could choose to follow the crowd, where a particular market might be selling candidate A for 80 cents. This would suggest that 80% of the people playing in that market believe that candidate A will win. However, even if this prediction is correct, their profit is 20 cents since they make $1 (i.e., only 25%). The way that this site entices its users is by making riskier bets more tantalizing. Suppose instead, the same person chooses not to follow the crowd. Instead, they break the information cascade by betting on candidate B winning for 20 cents. If instead, candidate B does in fact win, their profit is 80 cents (400% profit). [My brother actually made a decent profit from following this logic — he did his own research, collected his own information, and bet against the market. His prediction turned out correct, and thus, he made a large profit. Breaking the information cascade actually worked in his favor.]
It would be interesting to study those who are breaking information cascades (in this case, the people who bet for candidates at a price much lower than 50 cents on predictit.org). Further, I would be curious to see whether the people who turned out to be correct follow any sort of trends. Maybe these people are more active on social media, or maybe they are receiving some sort of outside information that followers of the information cascade might not have access to.