During a public lecture at the School of Public Policy on November 16, CEU Professor Botond Kőszegi offered a behavioral economist’s perspective on big data and privacy. He began by reviewing some of the insights about human behavior that we can learn from literature – specifically from Sherlock Holmes’ “The Adventures of the Blue Carbuncle” in which Sherlock Holmes looks at a man’s hat and deduces a number of things about a man he has never met: that he was fairly wealthy but is no longer, probably an alcoholic, that his wife no longer loved him, that he is middle-aged and sedentary, etc. “You leave a lot of traces about yourself as you go about your daily life,” observed Kőszegi.
He went on to note that Big Data makes it possible for many people to know a lot about you (often without your knowledge). Kőszegi said that this affects the relationship between firms and consumers. For example, firms use Big Data to make extraordinarily accurate predictions about consumer behavior which they then use to target particular consumers with offers that they will find especially appealing. This situation raises serious ethical, privacy, and related issues that are being extensively researched – but only very rarely from the perspective of behavioral economics.
Kőszegi explained that the methodology of classical economics, which has been used to analyze the behavior of markets and organizations, “starts from extremely simple assumptions about human behavior.” When analyzing privacy issues in particular, classical economics assumes that consumers are rational, and that the information that firms acquire is related to consumers’ preferences. This in turn implies that “the best possible world is one in which they [firms] know everything.” Kőszegi then gave the example of a monopolist airline to show that “even if no privacy is good for society, it’s bad for consumers – especially for rich consumers.” A simple solution to this would be to require that firms secure consumer’s consent before using their data. “You can think of Facebook in this way,” said Kőszegi. “You assign a value to it and so are willing to give up rights to data about yourself.”
Kőszegi explained that, in general, behavioral economics extends and modifies classical economics in at least two ways: what motivates people; and where and when consumers “make mistakes.” Privacy is one of those things that people say they value. There are numerous examples, however, to show that “people’s preference for privacy is malleable.” Kőszegi went on to point out that information sharing services (such as Google, Facebook, and LinkedIn) are much more successful than information protection services.
Another issue that Kőszegi explored in some length is what he called “consumer naivete” – the fact that many consumers don’t seem to fully understand how much they are paying for particular services such as gym memberships. Kőszegi gave an example of a study analyzing gym-goers’ choices between two membership options: customers can either pay per visit or pay a larger monthly amount for unlimited visits. Most customers choose to pay the larger amount even though it would be cheaper for them to pay per visit. Why is this? It’s because, according to Kőszegi, consumers routinely assume they will go the gym more often than they do.
Most people are also over confident and/or ignorant when it comes to predicting how much credit card debt they will incur; how much it will cost them to own and use a printer; how and when they will use their cellphones; etc. “Profit-seeking companies of course take advantage of this type of consumer naivete,” explained Kőszegi.
In their research, Kőszegi and Paul Heidhues, ESMT distinguished affiliate professor and a professor of behavioral and competition economics at Dusseldorf Institute for Competition Economics (DICE), are exploring naivete-based discrimination. Kőszegi pointed out that unlike perfect preference-based discrimination, which maximizes welfare, in many situations perfect naivete-based discrimination minimizes welfare. Another way in which naivete-based discrimination differs from preference-based discrimination is that “consent does not help because naïve consumers don’t know they are naïve.”
Commenting on the implications of his work on naivete-based discrimination for public policy, Kőszegi said that he did not know what the right public policy was but he was certain that the current focus on transparency and control is “grossly insufficient.”