Back to
English Articles
Publié le
Jeudi 09 Juin 2016
On 31 May 2016, France Stratégie and CEPII organized a debate on whether the digital economy is accounted for accurately in GDP. Sir Charles Bean, professor of economics at the London School of Economics and former deputy governor for monetary policy at the Bank of England, was invited to discuss the challenges facing the measurement of production in the wake of his report on the digital economy and statistics for the UK government, published in March 2016. The following article is based on remarks delivered by Fabrice Lenglart at the event.
Measuring Production in the Digital Age

Without a doubt the rise of the Internet over the past two decades has profoundly reshaped the way we live our lives. Information and communications technology (ICT) have facilitated everything from the way we travel to the way we work.

In short, our economies have become digitalized – and seemingly much more productive.

Yet despite this common perception, our figures for gross domestic production (GDP) do not reflect this. In fact, they show a negligible impact, if any, of the digital economy on aggregate economic activity as reflected in national accounts.

This apparent lack has led many to question whether new statistics need to be developed. In this context, London School of Economics (LSE) professor Charles Bean recently carried out an independent review of the UK’s economic statistics for the British government.

The final report concludes the digital economy has made it harder to measure economic output due in large part to the transition from capital-intensive to knowledge-intensive production, with intangible assets proliferating.

It posits that if this new output were accurately measured, between one-third and two-thirds of a percent could be added to the growth rate of the UK economy.

But does GDP really underestimate the value of the digital economy in production?

In the case of e-commerce, output is measured by margins, as with brick-and-mortar retailers. The only change is the location of the company selling the goods (e.g. purchasing products on the American website Amazon in Europe). This raises the question of the quality of the source of information used for measuring imports and exports. Assuming it’s reliable, output will not be undervalued.

When it comes to the so-called sharing economy, again, current statistics in principal capture the economic activity generated. Two runaway success stories, Uber and Airbnb, illustrate this.

With Uber, French GDP is increased by the taxi services its drivers provide to France’s residents. Whereas the service Uber provides of putting French customers in contact with drivers is recorded as an import as it is a non-resident company.

Contrary to transport services, housing services are part of personal consumption expenditures (PCE) and are consequently already accounted for in GDP. Not only is the value of the rent tenants pay estimated in national accounting, but a value is imputed for homeowners equivalent to the rents they would be charged as tenants for housing similar to their own.

The implications for Airbnb are that a tenant or owner renting out their home for a few weeks should not increase GDP. The problem arises in the fact that people generally rent their homes out for short periods of time at values much higher than what they pay as tenants (and by extension the rental value imputed to owners).

And herein lies the rub when it comes to the digital economy and economic output: it presents a measurement problem to national accounting and not necessarily a conceptual one.

In fact, services have long posed a problem with respect to GDP for the simple reason they are much more difficult to measure than goods. In particular, person-to-person services such as babysitting, gardening and home renovations have long been difficult to measure accurately due to a lack of information.

Websites financed by advertising further illustrate the problematic nature of measuring production. It can be argued the service produced by, say, The New York Times’ website to the benefit of its readers is not captured in statistics. But television broadcasting shows this problem is anything but new.

The production of a TV show is not registered in national accounts as such. That is not to say it isn’t accounted for: the advertising fees sponsors pay to a broadcaster are reflected in the price of their products or services. So ultimately, the value created is captured in GDP. The same can be said of any number of websites, including The New York Times’.

Of course, the question of accuracy of measurement can be raised. Bean suggests a ‘service’ of attention is produced by households and could be purchased by companies as an intermediate consumption: the latter would no longer pay the broadcaster but rather the household watching the ad, who in turn would pay the broadcaster for watching the show. Its income would be increased and consequently GDP.

But if this were so, what about the attention ‘produced’ when an individual is looking at an advertisement when riding the metro or simply walking down the street?

Lastly, the value of the bundling of personal data collected from individuals – which in aggregate forms part of big data – and sold on for marketing purposes to finance web companies has been called into question. While such data certainly has market value, does it mean an individual actually ‘produces’ it?

In some instances, such as an Uber driver profile, personal data can be considered a produced intangible asset. But the opposite is true for data generated by simply chatting or interacting on social media websites. In this case, they would be intangible non-produced assets.

The above notwithstanding, the digital revolution has certainly increased the number of intangible assets in our economies and radically changed the way services are provided. But has it ultimately led to a marked increase in the value of economic output? This is far less certain.

That said, as Bean points out, the digital revolution does raise price measurement issues that may lead GDP to be underestimated in volume if not correctly addressed by statisticians.

But issues such as the location of production, quality price adjustments and new sources of information it has brought to the fore are not new.

In the end, the digitalization of our economies has most probably increased our well-being much more than it has what we actually produce. And though this is far from negligible, it shouldn’t fundamentally change the way we measure production.

Edited by Richard Venturi

Click here to subscribe to our newsletter, In Brief


Fabrice Lenglart, commissaire général adjoint de France Stratégie
Type d'image: 
Anciens auteurs de France Stratégie
Tous nos travaux sur  :