A new bias? The “lucky store effect”
The Monkey Cage cites a paper by Jonathan Guryan of the Graduate School of Business, University of Chicago, and Melissa Kearney of the University of Maryland that tries to explain the following phenomenon:
In the week after a large-prize winning ticket has been purchased at a given store, that store experiences a 12 to 28% relative sales increase in lottery ticket sales. This increase fades over time, but the store’s lottery ticket sales remain elevated for up to 40 weeks. This effect increases with the size of the jackpot and with the economically disadvantaged proportion of the population.
What’s interesting about this finding is that previous research has shown that people decrease the amount of money bet on certain lottery numbers after those numbers come up winners. So they go to the store that sold the last lottery ticket, but don’t pick any of the numbers from the last jackpot? How can these apparently contradictory findings be resolved? Guryan and Kearney come up with an idea they call the “lucky store effect” in which “consumers erroneously increase their estimate of the probability a ticket bought from the winning store will itself be a winner.”
The lucky store effect, the authors say, may derive from the beliefs people draw from events that have only occurred a (relative) few times. These beliefs can differ. On the one hand, when people see a slot machine spit out a jackpot, they may try to avoid it for the rest of the day (a behavior known as the “gambler’s fallacy”). On the other hand, when people see a basketball player make six shots in a row, they may say he has a “hot hand” and expect him to make the next shot. While the empirical results make the hot hand fallacy tempting to ascribe to lottery stores, the situation is not comparable because these stores have not exhibited a streak of winnings. The increase in lottery ticket sales is generated entirely from a single instance of winning.
What is happening at lottery stores may be related to how people think the winning numbers were chosen. Psychological research has shown that people tend to think inanimate objects like machines generate outcomes differently than humans. In other words, a roulette wheel that lands on red 5 is a random event, but that sharp guy in the nice suit that chose red 5 – aha! – he knew something! People may think that the bettor who went to the winning lotto store knew something special, and they ought to follow her lead.
The authors caution that this explanation is mostly speculation right now. But it’s interesting speculation. The American Economic Review is about to publish the paper, but you can read an old working paper version here if you want.