Why we need data unions — the economics of data

Published in

Unbanx

5 min readDec 21, 2021

The market for personal data is worth billions. While personal data is in theory perfectly distributed among individuals, big corporations reap the benefits of monetising our personal data. This is to some extent due to underlying economic characteristics of data, like non-rivalry and non-excludability. To earn from data while avoiding negative privacy externalities, individuals need to participate directly and collectively in the bargain for their data. Data unions like Unbanks are a way to make this happen.

Data is an economic good

The Economist hailed data “the new oil” on its cover in 2017, manifesting the new business imperative to collect, hoard and monetise data of all sorts [1]. While this analogy transported a powerful message, it actually makes little sense in economic terms. It’s well understood that data behaves nothing like oil. Let’s look at some of the distinct characteristics that make data unique . This helps us understand why the our data economy is so highly unequal and inefficient, and how we can reverse this in the future.

Rivalry and excludability

Economic goods can be broadly distinguished by their excludability (other persons can be excluded from consumption) and their rivalry (one person’s consumption of a good reduces the amount available for others). Most tangible goods like food, clothes, or cars are rival and excludable. They are called private goods because they can only be used by one party at a time and others can be prevented from using them.

Rival and non-excludable goods are called “common goods”. A popular example is the fish stock in our oceans, which is clearly rival (depletable), but also non-excludable to a large extent. This leads to the “tragedy of the commons” we hear about in climate change so often, where common goods are over-consumed to the detriment of nature and society.

Data is anti-rival

Data does not fit into either of these boxes, because it is not a rival good. If one party uses a specific dataset, that does not reduce its usefulness to another party, and data can be shared at close to zero marginal cost. Some even argue data is anti-rival, meaning the use by one party increases its value to other parties [3]. Anti-rival in this context basically means network effects — few data points are worth litte, but when aggregated they start becoming non-proportionately more valuable. This is quite intuitive and makes sense from a statistical viewpoint as well. A data set on something like shopping behaviour is only representative (i.e. valuable) if it covers a large and diverse population of individuals. The result is a strong tendency towards scale economies in data markets and explains why it makes no sense for folks to sell their data individually.

Data is (partially) non-excludable

But is data excludable? In a perfect world where individuals are in full control of their data, it should be an excludable good. But in reality, the answer is not as straightforward and probably depends on the kind of data at hand and who controls it. It’s nearly impossible to take part in everyday life without giving away loads of personal data (think web browsing, social media, location services, search, etc.). To a large extent, our personal data might be non-excludable because there are so many channels through which it is extracted. This is such a common perception that most people lost track and interest in protecting their data at all (turning them into privacy cynics).

But non-excludability is also apparent in the sheer inability of organisations to safeguard personal data (331 data breaches were reported to EU regulators in 2020 on average, per day [4]). So we can assume that large parts of our personal data are simply “out there”, being traded without our knowledge or consent (you can check if your personal data has been breached in the past here). The problem that arises for individuals when other parties transact with their data is a well known in economics and called “negative externalities”.

Personal data trade can create negative privacy externalities

A negative externality in simple terms is a “cost or benefit for a third party who did not agree to it” [5]. Again, we find most examples for negative externalities in environmental science — companies that pollute their surroundings but don’t bear the environmental cost they cause. In the data sphere, sharing and trading non-anonymised personal data causes similar externalities in the form of privacy loss [6]. The problem is that third parties transact with our personal data but do not bear these privacy costs. The solutions — let people take direct part in the bargain for their data. Let them choose the degree of anonymity and the compensation they require for providing their data.

The status quo is neither efficient nor fair

To sum up, personal data is anti-rival, partially non-excludable and can produce negative privacy externalities. The result of non-rivalry are strong scale effects which explain why the data economy is dominated by huge third-party data brokers, financial institutions, and tech monopolies. However, these data aggregators do not bear the privacy externalities of data trade. Regulators have also long recognised this and increasingly press data buyers and sellers to prove the legality and provenance (a.k.a. consent) of their data sets. As a result, we need new ways for data buyers and individuals to establish a direct and fair data economy.

Data unions to the rescue

Data unions can be a superior alternative that address both non-rivalry and non-excludability of data. They are associations of individuals that pool their data to collectively create value from it. The pooling of data enables value creation, while the direct participation and reward enables informed consent, internalisation of privacy costs, and by the way — a way to make some extra money. Welcome to the age of universal data income (UDI).

One of the pioneers in this space is Unbanks, who enable direct, secure, and anonymous monetisation of your valuable banking data. Check them out and sign up for the waitlist here.