For many of us in developing countries, vehicular pollution is an everyday nuisance. The high level of pollution in cities surely bothers the drivers of those vehicles as well. But while deciding whether to install low-pollution engines, they weigh the additional cost with the benefits of lower pollution. And while the benefit of lower pollution is spread across millions of people in the city, the cost is concentrated with each driver. Put simply, she is not incentivised to invest in engines or fuels that pollute less. Therefore, vehicle owners continue polluting and the rest of us continue suffering. As a result, the amount of pollution is way beyond the socially optimal level.
You might be wondering why am I am talking about vehicular pollution in a post about big tech. Well, because data shows similar characteristics. When I create ‘data exhaust’ — browsing history, transaction logs, etc. — I weigh the cost and benefits of leaving this data behind. When the costs are obvious, I make better decisions. For example, I restrict where I give my phone number because of the risk of pesky calls and annoying messages. But the negative consequences of other data I generate is not so obvious. Does my social media data really matter? Yes — but how do I internalise the cost of more targeted advertising and echo chambers that it creates? More importantly, my data tells the business something about the characteristics of others — my family, my friends, my neighbours, my colleagues etc. If I earn a lot, my neighbour is likely to earn a lot too. If I work in IT, my friends will have a higher probability of being in IT as well. If I’m brown and commit a crime, the AI engine will likely increase the probability it assigns to brown-skinned people committing crimes. None of this, of course, enters my cost-benefit calculus. So I blithely carry on being an endless creator of these digital trails.
The ‘societal value’ of an individual’s data is different from the ‘personal value’. For corporations, the value of our combined data is usually far higher than the sum of the value of our individual data.
However, this ‘data exhaust’ is valuable for collectors of behavioural surplus like Google and Facebook. Such platforms collect from us data (both sensitive and innocuous), and combine billions of such data points to create powerful platforms that affect the very core of our society. What enables them to do so easily is the economic phenomenon of externality — put simply, our actions (or our data) affect those around us. The ‘societal value’ of an individual’s data is different from the ‘personal value’. For corporations, the value of our combined data is usually far higher than the sum of the value of our individual data. This also results in the creation of what are being called ‘network monopolies.’ These organisations display monopolistic behaviour because, once millions are on such platforms, there is no way for competitors to emerge. This provides platforms like Google and Facebook tremendous power over our markets, our societies and our lives.
Data is the raw material that feeds this engine. Therefore, platforms seek to incentivise me to generate more and more data. And I, with my limited worldview of the value of my data, tamely submit. Unless we close this gap in economic incentives between us and businesses, this problem is unlikely to go away. Data protection regulations might alleviate the situation, but are unlikely to solve them unless they strike at the heart of this world.
A popular approach to address externalities is quantity-based regulation, wherein the product can only be created in a pre-determined quantity. Such an approach is the fulcrum of the battle against climate change, wherein each nation has committed to caps on its greenhouse gas emissions as part of the Paris Agreement. However, such an approach is likely to be ineffective for data because of the obvious difficulty in enforcing ‘data caps’ across billions of individuals. We need more innovative approaches. Here are some early ideas that I’ve been thinking about, and I would love to hear your reactions -
- Well-enforced individual data rights: Assigning and enforcing clear property rights is a popular approach to solving the problem of externalities. For example, what prevents you from making a copy of a book in your possession, and selling it for a profit? The fact that you would be violating the author’s copyright, and are likely to be penalised. Similarly, data protection regulations like GDPR and CCPA seek to provide individuals rights over their data. These include the rights to access, rectification, and erasure. If enforced effectively, such rights will allow individuals to control their data, breaking the asymmetric power dynamic between us and the platforms. However, such a copyrights-based approach will fail to solve two fundamental problems. First, that we still won’t internalise the impact of our data on others. Second, that companies do, and therefore value our data more.
- Individual-centric data intermediation: Data intermediation is a promising approach that strikes at the very root of this asymmetry between individuals and corporations. The idea is simple — create data intermediaries who collect data from individuals and ‘sell’ that to corporations on their behalf. This way, consumers trade their data as a collective, rather than as individuals. This increases the bargaining power of individuals and helps them get better terms from businesses. Economists have shown than such data brokerage can reduce corporations’ tendency to collect extra data. When such intermediaries are set up, designers need to think about the economic incentives and whether those align with the interests of the individual.
- Differential taxation: A popular approach to correcting an externality problem is to impose a tax or subsidy on the externality-causing problem. In the case of data, this could take the form of higher tax rates on profits that firms reap from data harvesting. Economists have shown that this will reduce the problem of excessive collection of data. Businesses will respond to this by shifting towards revenue streams that do not rely on unfettered exploitation of data. A similar approach is to impose higher regulatory costs, risks and fines on larger data collectors. These measures will increase costs for businesses as they grow larger and realise the benefits of network monopolies. This is an approach that most data protection regulations are taking, wherein costs are higher for larger data processors.
- Platform Interoperability: The fourth and most disruptive idea is to create interoperability between the platforms, striking at the very root of their monopoly power. There are two kinds of interoperability. One is data interoperability, where an individual is able to take her data from one platform to another. Most data protection legislations, including GDPR,CCPA and India’s draft bill, have strong provisions for data portability. However, this is not likely to solve the issue because tech giants draw their market power from their scale, in addition to the amount of data they hold. For example, even if I take my data from Facebook, is there any other social media platform where I can take that data to avail the same service? Therefore, what we need is platform interoperability. These platforms should speak to each other. This requires a fundamental redesign of the internet. For example, a gmail user can freely send and receive emails from hotmail, yahoo and other email providers. However, a MySpace user cannot send a message to a Facebook user. Or a Lyft user cannot book a cab on Uber. The setup of the email market prevents gmail from exploiting its market dominance in the way that a facebook or twitter can. Such a market structure arises because emails are built on top of a standardised protocol called SMTP. If we can create such protocols for other tech platforms, we may be able to check their unfettered power over our lives. For example, an individual could raise specify the kind of cab ride she needs, and platforms like Uber and Ola can then compete to fulfill that demand. This will bring in the kind of interoperability and competition that will empower the individual vis-à-vis platforms.
Externality is a market failure that strikes at the very root of how our markets are structured. It makes us produce more data than is societally desirable, and enables some organisations to emerge as the arbiter of our collective destinies. Solving the externality problem will therefore lead to a re-balancing of market power between individuals and corporations. This will help us create more competitive markets and healthier societies.