Facebook CEO Mark Zuckerberg will answer questions on Capitol Hill this week stemming from the ongoing Facebook-Cambridge Analytica (CA) controversy. This story has it all – big tech, privacy, the Trump campaign, and the never-ending attempt to relitigate the 2016 election. It has even succeeded in pushing Stormy Daniels off the front page.
Some commenters, like Franklin Foer in The Atlantic, believe the behavior of Facebook reinforces the argument for regulating the internet, just as we regulated industries like banking, aviation and agriculture. But before jumping to that conclusion, they should read the extensive literature on those experiences. Numerous studies have shown that consumers did not benefit from traditional industry regulation. Regulating the internet would likely prove to be even more costly.
The outlines of the Facebook-CA story seem to be clear: A Cambridge University researcher developed an app for a quiz taken by 270,000 people. The app gathered data not only on the 270,000 individuals who voluntarily took the quiz, but also on as many as 87 million of their Facebook friends. These data were provided to CA, a political consulting firm that worked for the Trump campaign. Facebook claims the data gathering practices used by CA violated its data sharing policies. The company requested that CA delete the data, which it apparently didn’t.
The markets reacted swiftly and severely. In the span of a week, Facebook shareholders saw their equity value decline by 14 percent. Facebook’s Chief Operating Officer Sheryl Sandberg said in an interview that the company doesn’t look at privacy issues in terms of damage to the stock price, but it’s hard to believe they don’t care about a $75 billion hit.
It’s a good thing if the stock market punishes behavior detrimental to users’ interests, because that will incentivize companies to change those practices. It’s not such a good thing if the market is largely reflecting broader fears of regulation of the tech sector, which probably explains the decline in all the FAANG (Facebook. Apple, Amazon, Netflix and Google) stocks, though Facebook fell the most.
The principal complaint against Facebook seems to be that most of the 87 million people were unaware of the purpose for which their data were being used. But the data were used in ways that are commonplace among marketers and even the politicians now on the warpath.
Online and offline marketers use data from large numbers of people all the time, for commercial products as well as political campaigns. Marketers routinely gather or purchase data on cohorts of people with particular characteristics or who may be interested in purchasing a particular product. Sophisticated data analysis has also become a fact of political life, used by candidates across the spectrum. Consumers and advertisers benefit from better targeting of both advertisements and political messages, which is made possible by the availability of large sets of personal data.
Facebook is paying a price for a lack of transparency, but it’s unclear that most people care if they are part of a particular database. Moreover, it would be virtually impossible to make that information available in a meaningful way. If internet companies, data brokers or marketers were required to do so, the data would be of much lower quality or wouldn’t be available at all. This would raise costs to marketers (commercial and political) of reaching their target audiences. These costs would be passed on to consumers. Consumers would also incur the costs of receiving more mistargeted messages that don’t give them useful information.
What would regulation do? Proposed policy responses involve limitations on the ways data can be shared or used rather than mitigating harm to users. In fact, nobody has identified harms due to the Facebook-CA incident. Without some identified harms to remedy, regulation will not have any benefits.
Some, like Apple CEO Tim Cook, have called for regulations to prevent users’ data from being used in new ways without their knowledge. Cook’s proposal has been a central plank of the privacy agenda in the U.S. and is part of European regulation, which mandates that data should only be collected for a pre-specified purpose. In the world of big data, artificial intelligence and machine learning, however, data are always being used in ways that were not anticipated when the data were collected. It is simply not feasible to inform individuals that this is happening, nor would people appreciate getting the thousands of notices that would be required.
Facebook itself has announced it was cancelling a program that combined Facebook data with data from third-party data brokers to better target Facebook users. But, using data from multiple sources to better target advertising is good for consumers.
These restrictions on the use of data would generate real costs. They affect not only marketing, but also data security, credit markets, health research and other areas, without any offsetting consumer benefits. They would hamper U.S. efforts to compete in developing innovative new technologies that are highly data dependent.
Those who suggest that it is now time to regulate the internet should look at the effects of regulation of other industries. It is difficult to find success stories.