27 Feb 2017

The negative impacts of big data

Melle writes*

Big data consists of all digital traces that we leave, from social media activity to GPS systems in our phone tracking our every move. Not surprisingly, this information can constitute a valuable resource to companies. If successfully collected and analyzed, businesses, or politicians for that matter, can more effectively assess consumer preferences and behavior. Despite obvious benefits, both the process and outcome of big data use pose a threat to consumer welfare, as they produce negative impacts ("negative externalities").

Before delving into the negative, it would be unfair to neglect any mention of the numerous benefits to big data analytics. A business can enjoy competitive advantage if it can realize the capacity, in terms of physical, human and organizational capital, to extract and subsequently exploit hidden insights on consumer behavior. By utilizing big data consumer analytics, businesses may market to specific groups or areas based on more detailed and individual consumer predictions. Additional benefits relate to tailoring product and price to consumer preferences. Perhaps making the illusive microeconomic principle of first-degree price discrimination possible.

As businesses are increasingly inclined and able to use big data, concerns of privacy and security have grown. Research has shown that a large majority of internet users would ideally opt out of being monitored, as they do not trust companies with their private information. The extraction and storage of private information alone can in certain circumstances be considered an externality. Examples are tracking devices in phones and cars which consumers are often unaware of. Additionally, unstructured data, being amongst other things photos and documents, may contain personally identifiable information and intellectual property. Take for example the case in Italy where three YouTube executives were convicted of violating the privacy of a child who could be seen being abused in a video on the website. Such instances of privacy violations rarely lead to compensation for the affected person.

If you are not convinced yet, consider some of the possible uses for big data that produce negative externalities. A common example is the issue of medical insurance companies using big data to discriminate in premiums against people from certain areas with higher incidence of certain diseases. However, I argue that the potential for consumer fraud extends far beyond such incidents. A powerful tool in big data analysis is psycho-informatics. The practice is based on a branch in psychology called psychometrics, which played a considerable role in Brexit and the US election. Big data can be used to quite accurately determine the personality of any individual based on the Big Five personality traits. These same personality dimensions can be related to addiction potential. Essentially, companies can thus search out the personalities that will be most vulnerable to a gambling, drinking or any other addiction, and target these individuals specifically. The result is that based on information consumers were not willing to supply, individuals are subject to personal advertisements that may carry long-terms costs to their well-being. The unregulated nature of the digital environment allows for such use of valuable information with little cost, and no benefit to the consumer.

Bottom line: The Big Data revolution offers a wealth of opportunities for businesses in terms of consumer behavior research and marketing. Nonetheless, as economists it is also important to analyze the negative externalities that accompany this development. Not only can gathering and storing big data be considered an externality, as it is an invasion of privacy for which consumers are not compensated, the application of the data through targeting vulnerable individuals opens the door to even larger issues of, for example, perpetuation of gambling behavior.

* Please comment on these posts from my environmental economics students, to help them with unclear analysis, alternative perspectives, better data, etc.