Years of steep and sustained falls in the cost of storing data either online or on local drives will transform human interactions. As The Data Economy neatly defines it, data is information that has been digitized, meaning that a computer can then read it. And data that is of most economic interest concerns human activity. When data is cheaper to store, computers can do more with it. We are perhaps most familiar with structured data, numbers and associated fields that are easily organized in a table, such as a spreadsheet or long format with many fields. But unstructured data, such as images and sounds, which are also more at an individual level, can now also be digitized at scale and so also exploited by computers. At this scale, it is naive to think that data is stored as an end in itself. Data is not exogenous technological progress: it is a manufactured input and is shaped by incentives. The objective of data is that one party can react more quickly and precisely to another party. Information processing companies have accessed these vast stores of data to create responses that rival our own intelligence. More human activity that is so measured can become more integrated with the data economy.

While the book is well organised and well written, a general reader will not find this the best introduction to how data is transforming the world (though I cannot suggest a single alternative). The book is subtitled Tools and Applications for a reason: its purpose is to explain how academic economists have modelled the data economy. The ideal reader of this book would be a student who wants to organize their thoughts around a mathematical modelling framework. The approach is to formalise the complex data economy into optimising problems and then to describe and interpret the solutions. Data is shown to be most commonly modelled as a factor input into a micro-founded macroeconomic model of the economy. Chapter 2 on statistical tools is a useful collection of methods. Other chapters in the book review how we might model investment, market power, advertising, feedback and so on, adapting those methods to account for the role of data. There are many normative policy questions to do with data.

Most are mentioned at the end of the book, though some get far more attention than others. There is a useful chapter at the end on valuing data. What the book reveals is that we are playing catch up. Ten years ago, most of us would have said that more shared data will make things better. After all, acquiring information results in the reduction of uncertainty, just as how a published price data sends out the signal to usefully clear markets. Yes, there are incentives to manipulate information, but surely the best would emerge from what is clearly a vigorous marketplace of ideas. I would guess that nearly all the models that economists built about data had predicted outcomes that were welfare improving. The attention was on the positive questions to do with more data, such as deriving the implications of more data for the behaviour of prices. Now, we are not so sure that the data economy is all that benevolent. What is happening needs to be picked apart, even if it is too late to change course. For example, the book introduced me to the concept of a statistical inference externality, which is when data about one person provides a signal on others with the same characteristics. We may then consider that too much data is being revealed and even that statistical discrimination is worsened.

I think there are two other ways in which data affects the economy that are not discussed in the book. One is how public data affects political choices over economic and corporate policies. Another is in how the data economy influences the allocation of our time by making some things but not everything easier. Baley and Veldkamp’s book invites us to think that the data economy should feature more in economics. Certainly, the standard production function of labour and capital making widgets looks outdated after reading this.