GDP Cannot Explain The Digital Economy
The world’s primary measure of growth is largely irrelevant for the most important part of the economy
The engine of the digital economy is the technology used to deliver finished digital services to businesses and consumers.
Comprising connected devices, broadband infrastructure, cloud software, internet-based services and many other markets, the societal benefits that this technology delivers are measured by proxy, by adding up the total spending by businesses and consumers on finished digital products and services.
The consensus view is that the digital economy currently accounts for around 5% of global GDP.
But if the contribution of the digital economy was calculated differently - based on the absolute value delivered, rather than by using economic measures like spending by consumers and businesses, then the digital economy would be regarded as a far more important part of the overall economy than is currently the case.
The difference between assumed value delivery and actual value delivery is not small: rather than the digital economy accounting for just 5% of global economic value, it might already be delivering 20% of the total value.
The encyclopaedia market provides a stark illustration of just how inaccurate standard economic metrics are when it comes to calculating the value delivered by specific markets within the digital economy:
In 1990 around 100,000 copies of Encyclopaedia Britannica were sold at an average price of USD 1,200, implying annual revenues of USD 120 million. If we assume that the active installed base of the publication was 10 times the annual sales level (an approximate replacement rate of 10 years) then this implies that the value delivered per active user in 1990 would have been about USD 120 million / (10 x 100,000) = USD 120 per active user per year.
If we now assume that value delivered by Wikipedia is half that of Encyclopaedia Britannica (some maintain that Wikipedia is less accurate) and then discount this further by a factor of 3 (to take into account the current widespread availability of other sources of digital information) then the value delivered to a typical Wikipedia user might be estimated in 1990 terms as USD 120 x 0.5 x 0.3 = USD 18 per active user per year.
But as there are 500 million active users of Wikipedia today then this implies that Wikipedia is delivering an equivalent value of 500 million x 18 = USD 9 billion per year in 1990 terms, or about USD 12 billion per year in 2016 terms.
However, a GDP-based valuation approach discounts the annual value delivery of USD 12 billion to zero.
It is absurd for the value delivered by Wikipedia to be discounted from a number that is clearly in the billions to zero, but this is what a GDP-based approach actually does.
We are not alone in recognising the limitations of GDP as a way of measuring economic value and growth: in January 2016, Christine Lagarde, head of the IMF, said at Davos “We have to go back to GDP, the calculation of productivity, the value of things – in order to assess, and probably change, the way we look at the economy.”
The problem is not confined to sites like Wikipedia – but applies across the whole digital economy. Here is another example:
The recorded music industry is today delivering more absolute value to a far larger number of users than ever, and yet conventional economic measures would conclude that value delivery has halved compared with the mid-1980s when the industry was at its economic peak. Again, this cannot be right.
By any standard, the sort of mis-analysis outlined above demands corrective action.
It wouldn’t be so bad if the reasons for such errors were technical or methodological in nature. Unfortunately, the reason is that the use of standard economic measures to define the value delivered by the digital economy is deeply flawed at a conceptual level.
The economist Robert Solow observed in 1987 that "You can see the computer age everywhere but in the productivity statistics.” We believe that the reason why Solow was unable to see any sign of the computer age, was that he was looking at the wrong statistics.
Given that the digital economy is growing at a faster rate than the physical economy and taking into account the disruptive impact of upcoming ‘horizontal’ technologies, such as machine intelligence and how new waves of disruption will impact on new, major ‘vertical’ markets – such as financial services, automotive and healthcare - then in the future the digital economy might not just be an ‘important part‘ of the economy, it might effectively become the economy.
That the digital economy is already delivering far more value and is therefore far more important than most policy makers realise suggests to us that the opportunities and challenges posed by digital technology are far more prescient than commonly believed.
Digital technology allows more to be accomplished with less capital and with fewer human resources. In essence, digital technology represents a ‘productivity accelerator’ where the actual value being delivered is being grossly underestimated by the use of crude measures such as spending by consumers and businesses and, therefore, GDP.
Our analysis is that the pace with which digital technology is developing is such that it is cannibalising assets at a faster rate than they are being replaced and this should be a matter of concern to policy makers worldwide, especially because these effects are only going to accelerate.
While some economists are beginning to wonder whether globalisation is being executed at a faster rate than its effects are being managed one could make a similar argument for the digital economy: perhaps policy makers need to begin by acknowledging the real value being delivered by the digital economy and its rate of development, and then think about where this is leading.
At present, the best that can be said about the trajectory of the digital economy is that it is directionally correct, but a conventional economic analysis means that its impact on society is not just unknown, but unknowable.