In their fascinating paper on “The Data Economy: Market Size and Global Trade” for the Economic Statistics Centre of Excellence
(part of the UK’s National Institute of Economic and Social Research), Diane Coyle and Wendy Li talk about the growing “data gap” between global Big Tech and potential competitors, disruptors and innovators. They argue (convincingly) that this data gap is a a barrier to entry that affects not only businesses but also aggregate innovation, investment and trade:
Large data holdings, rich in volume and variety, thus give large online platforms a significant competitive advantage, powered by network effects and the virtuous cycle between data and the AI algorithms improving the services and increasing revenues.
This advantage means that the platforms obtain insights about adjacent sectors and can then enter them more easily. Potential new competitors without access to that data and the advantage that is confers therefore must struggle to enter the markets. This has been obvious for some time. Indeed, the EU’s “A Europe fit for the Digital Age” initiative launched in 2020 made a central observation that the market power provided by the data advantage allows a handful of large players to unfairly leverage into new markets.
That all sounds qualitatively correct, but is there quantitative evidence?
That’s a difficult question to answer. Most data flows around out of the view of statisticians and there is no way to measure the value the of data. Statistics Canada tried to estimate the value of the country’s data in 2019 and came up with somewhere between c$160-220 billion. For comparison, that would make the value of all the data in America something in the region of $1.4-2 trillion (which would be nearly 5% of America’s stock of private physical capital).
xxx
A good place to start analysing any economy is by measuring it. A robust methodology has yet to be developed, but the data economy is already large. Statistics Canada, a government agency, last year tried to estimate the value of the country’s data (its stock plus related software and intellectual property in the field). The result was between c$157bn and c$218bn ($118bn and $164bn). If that number is close—a big “if”—the value of all the data in America, whose gdp is 12 times that of Canada, could amount to $1.4trn-2trn, which would be nearly 5% of America’s stock of private physical capital.
From A deluge of data is giving rise to a new economy | The Economist:
xxx
xxx
Relatively crude measures exist for estimating the size of the data economy, yet for the USA, estimations are that it was between $1.4trn-2trn in 2019 (including software and intellectual property) (“A Deluge of Data Is Giving Rise to a New Economy,” 2020), while projection for the EU27 data economy by the European commision are at €829bln in 2025 (up from €301bln in 2018) (European Commission, 2020b).
xxx
That’s a pretty significant sum. But, as The Economist notes, that while the data economy is clearly large, a robust measurement has yet to be developed. This is why Diane and Wendy’s paper was so interesting to me. They propose a new and consistent “impact-based” approach to estimate the size of the overall data market in a sector by comparing the values of data with and without the entry of an online platform. They use the impact of AirBnB on the hospitality sector an example, and calculate the market size for data in the global hospitality sector as USD $43 billion in 2018, growing on average at the rate of one-third per annum. But as their calculations show, the benefits of this growing market are not distributed evenly.
That rate of growth and its distribution would seem to confirm that the online platforms’ disruption of incumbents’ firm-specific knowledge is fast and significant. This may be why a number of online platforms are adopting the “super-app” approach pioneered in the China because the ability to combine data from multiple sources is even more advantageous than we might thing, meaning that the platforms impact may be “accelerated or multiplicative”.
Super-apps are yet another confirmation that data isn’t the new oil. It doesn’t get used up as it is refined. Instead the use of data produces even more data and the more this reservoir data grows the more it feeds innovations and generates positive externalities.
Bits and Borders
This methodological analysis gets even more interesting when applied at the aggregate level, as it shows that there is one large developed country that is a net importer of data and a net exporter of digital goods: yes, as you would guess, its the United States. America’s online platforms collect data from around the world to feed centralised digital production in the US. China is also a net data importer and digital goods exporter although more of its digital consumers are domestic. The UK, by comparison, has no dominant local platforms. We are net data exporters with a sound digital infrastructure but a disadvantage in data and given that data is what we need to produce added-value digital goods, it’s a pretty serious disadvantage.
Meanwhile, many countries are revising their data policies and localisation rules in a thrust for “digital sovereignty”. This is a useful catch-all descriptions of the many ways that governments try to assert more control over their digital infrastructure. It has long been a concern in supply chains, affecting the kinds of hardware and software available in a given market. Now, of course, this form of resurgent nationalism is fragmenting the cloud. Governments around the world are passing measures that require companies to host infrastructure and store certain kinds of data in local servers. Some also require companies that operate within their borders to provide the government with access to data and code stored in the cloud.
These policies are not currently based on economic calculation, which is why the measurement data markets should help policymakers understand the dynamics and contribute to localisation policies. But also, it must be said, because the data imports and exports are not directly related to the creation and distribution of the value of data. Data flows do not reflect the value derived from data.
(In the paper Diane and Wendy illustrate this point using the example of Taiwan. Google has two data centers in Taiwan to support its operation in Asia. This means that there are large cross-border data flows between Taiwan and other countries, but Taiwan is unlikely to
receive most of the benefits.)
Blocking data flows is sort of metaverse mercantilism that benefits no-one. The European Council on Foreign Relations (ECFR, a prominent think tank) published a call for action on “Defending Europe’s Economic Sovereignty” last year in which that called for the EU (and the UK) to agree data free-flow with the US, prohibiting forced sensitive data transfers and introducing a dispute settlement mechanism. Data sharing will enable a greater degree of entry and competition, driving more innovations.
Diane and Wendy conclude that an open data-sharing ecosystem will increase productivity and therefore economic wellbeing. From my inexpert perspective, I could not agree more. We have entered an era of data hoarding where companies are now storing any and all of the data that they can get their hands on just in case it will be worth something in the future. But as it sits in these hoards, that data is not working to the greater good.
This, to my mind, is yet more support for the idea of taking on the data misers and forcing them to share data to the benefit of competition and the economy as whole. We already know what do it. Open banking was a good place to begin the attack on data hoarding by incumbents and we are learning a lot as the paradigm spreads. Open finance is an obvious next step. But policymakers should be set the firm goal of open data in all sectors (except possibly national security) in mind and start working towards it. The fintech sector’s demands should be maximal: open everything.
Snippings 