By Diane Coyle & Wendy Li
There can be no doubt about the growing importance of data in the economy. The importance of data will only continue to expand. The era of 5G and the Internet of Things (IoT), the combined forces of artificial intelligence (AI) and big data, are reshaping how and where goods and services are produced and distributed.
This makes data policy a hot topic everywhere, and across different areas of policy, from competition (where the accumulation of troves of data has been identified as a source of Big Tech’s market power) to trade agreements (with growing concern about the location of data as well as cyber-security). There is also growing evidence that the effective use of data is driving a wedge between the most productive, data-intensive companies and the so-called ‘long tail’ of less productive ones. Online platforms based on AI and data have disrupted each industry sector they have entered, with their know-how in using data. When an online platform enters an industry sector, traditional incumbents with a higher degree of digital capability will adjust better than their less data-savvy counterparts. The COVID-19 pandemic has accelerated the pace of digital disruption in almost every aspect of daily life and business, pushing more traditional firms to accelerate their digital transformation.
However, most data transfers are unobserved. Firms can trade data through data brokers but this is relatively rare. There are also cross-border flows that are unmeasured, as many transfers involve a data centre overseas. So policymakers have some unanswered questions. How big is the market for data? How fast is it growing in scale and scope? Does the growth vary by industry and by country? Can small developing countries rely on accessing global data markets via the cloud, given the expense of data centres?
Understanding the market size for data is therefore critical, but the measurement of market size is challenging because there is no data on most data transfers. There are few market prices, these are not transparent where they exist, and moreover most data are collected for firms’ own use. Many firms are unwilling to compile or release data on data for reasons of commercial confidentiality or competitive advantage.
In our new paper, we develop a novel methodology for measuring the size and growth of markets for data by sector. It involves comparing the organizational capital of data-intensive and less data-intensive firms. Entry of an online platform using data leads to rapid depreciation of the organizational capital of incumbents. This gives us a measure of how much firms should be willing to pay to maintain the value of their know-how derived from data. We can thus use the loss of the value of organizational capital to measure firms’ willingness to pay for the data. We use the hospitality industry as an example to apply this measurement approach. Our preliminary, conservative estimate of the market size for data in the global hospitality industry is US $43.2 billion, with a growth rate currently doubling market size every three years.
We also develop a typology of trade in data and digital goods and services at the country level. We identify six different country categories. As the size and growth of markets for data will differ across industries and countries, both industry-level and country-level market size for data and the associated growth can affect the international division of labor. Population size, availability of high-tech talent, financing availability, digital infrastructure, foreign direct investment, and institutional contexts in different countries will therefore affect trade in data, and in digital goods and services.
Two categories, the two net data importers, have only one member: the US in the developed world and China in the developing world. All others are net data exporters. This reflects the dominance of large US and Chinese tech companies. However, although the value created by using data to produce digital goods and services depends also on skills and financing, which is concentrated in a few countries, all countries, even low-income countries, will be essential data providers for businesses serving their markets. Developing countries are unlikely to be able to invest in costly digital infrastructure such as data centers and rely heavily on international investors to incubate and develop e-commerce and other platforms, as well as to build costly logistics and warehouse networks to support e-commerce.
The policy trend in many countries is for increasingly tight restrictions on data transfer. However, these policies should be informed by the implications of digitalization and data trade for international comparative advantage. Understanding how big the market for data is and how fast it might grow is a first step. Data imports and exports are not directly related to the distribution of the value of data or the creation of the value of data while, unlike a physical supply chain, local clusters of related suppliers are not necessary for a data value chain. Both upstream inputs and downstream products are tradable, at least where there is no restriction on data transfer across borders or between firms. So digital goods and services can be produced anywhere as long as there is access to the data. As the value of data is created through its utilization, not its ownership per se, an open data-sharing ecosystem with an appropriate framework for access rights could unlock innovation, eventually resulting in the increase in productivity and economic wellbeing.
As all firms will need to access data, regulations on how data can be shared within borders and be transferred across borders, will need careful thought. It is hard to compete with the tech giants, but equally hard to do without them.
Read the full ESCoE Discussion Paper here.
Diane Coyle is Bennett Professor of Public Policy at the University of Cambridge.
Wendy Li is Executive Director at the Moon Economics Institute.
ESCoE blogs are published to further debate. Any views expressed are solely those of the author(s) and so cannot be taken to represent those of the ESCoE, its partner institutions or the Office for National Statistics.
