Economic Externalities of Autocomplete: Evidence from Financial Markets

Abstract

Research on human biases stemming from the use of decision aids has predominantly dealt with human tendencies to reduce cognitive effort. In this research we examine a bias that results from the human tendency for exploratory behavior. We frame our analysis within a natural experiment setting and analyze the text completion aid (autocomplete) within the financial markets domain. In financial markets, investment options are relatively fixed but may occasionally change as new securities (tickers) are introduced to the markets. Theory on exploratory behavior suggests that when new tickers are introduced, the autocomplete decision aid can facilitate diversion of the users from their initially intended ticker to the newly introduced ticker. We analyze whether such diversions make an economic impact in terms of transaction loss. Consistent with our hypothesis, we provide evidence that upon ticker introduction, turnover of securities that are syntactically similar to the new introduced security is significantly reduced by 3%-5% around the starting day of trade of the new security. Exploratory behavior upon decision making appears to have a significant economic effect on transactions.