Blurring the line: are digital technologies disrupting the measurement of GDP?


Blurring the line: are digital technologies disrupting the measurement of GDP?

By Diane Coyle

One of the programmes of work in the ESCoE will be looking at the challenges of measuring the modern economy, challenges arising from digital transformations and highlighted in last year’s Bean Review. New digitally-enabled technologies, from smartphones to automation to high-tech tunnelling techniques, are everywhere – but not apparent in the economic statistics.

In a new discussion paper set out a framework for analysing one aspect of the digital changes in our lives, looking at what economists call the ‘production boundary’. This is the conceptual line between what we do count in GDP and what we don’t. In principle it is clear: anything that can be bought and sold in the market is inside, and non-market transactions are outside. So going out to a restaurant is in GDP but cooking a meal at home is not. In practice, the distinction is not so clear. Government spending on services is included (considered as collective consumption). So is ‘imputed rent’, the amount home owners would have to pay themselves if they were renting their home in the market.

There are other activities where people could – just as with the buy/rent decision on housing – switch between purchasing and self-providing, such as childcare, gardening or cleaning. These are defined to be outside the production boundary, however, and are recorded in the ‘household satellite account’.

What does childcare have to do with the digital economy? The link is that digital technology is blurring the production boundary. There are three reasons for this.

One is that people are increasingly providing some previously purchased intermediation services for themselves. This do-it-yourself digital production includes things like banking and travel. Although there are still intermediaries, these are online, and consumers are saving more time in total but using their own time and their own computers or smartphones and Wi-Fi to undertake transactions. They are also likely to have more choice and find a better match for their preferences.

A second is the use of so-called ‘sharing economy’ platforms. This is a small sector but has been growing rapidly. In addition to the growth of employment through these platforms, these are likely to provide again a wider choice and also lower prices. They also involve the use of household capital equipment such as computers, houses or cars. If people are making an income from their participation, there is no issue in principle in including it in the standard measures such as GDP but it might pose a practical problem in terms of data collection. However, to the extent that people are, say, swapping homes or using Airbnb rather than paying for a standard holiday, the lower prices (and greater choice) they face are not being captured.

The third reason is the increase in the provision of free digital products created by people in their unpaid time. One example is Wikipedia, but there are plenty of others, ranging from software packages such as R and blogs for entertainment to advice forums. In many ways, these activities are like conventional volunteering in a school or charity shop, which lie firmly outside the production boundary. However, this kind of digital volunteering seems likely to be substituting to some extent for paid-for products that are including in conventionally measured GDP – encyclopaedias, proprietary business or educational software, magazines and so on. If consumers save money on these purchases, they will be able to spend it on other goods and services, so nominal GDP will be unaffected, but the substitution to free goods will not be captured by the price deflators used to calculate real GDP.

Each of these in themselves might be small in their effects, but combined with other types of substitution inside the production boundary (for example, from buying a camera to using a free app on a smartphone – covered in a separate paper) might potentially add up to a significant measurement impact on the GDP deflator and real GDP.  Figure 1 below shows the divergence around 2007 between the paths of nominal and real GDP per hour worked for the UK (and the same pattern is evident for other countries). This is when in many cases the ‘productivity puzzle’ of a halt in the growth of GDP per hour worked is first observed.

There will without doubt be several factors contributing to the puzzle, but it would be ironic perhaps if one piece of the jigsaw lies in the way digital technologies are making activity at home part of the productive economy in a way they have not been since the Industrial Revolution.

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.

About the authors

Research Projects

Related publications


News and Blogs