Why internet/GDP ratios make no sense

By Felix Salmon
April 19, 2012
the Economist reprinted a chart from a BCG report, which purported to show the contribution of "the internet" to the total GDP of various different countries.

" data-share-img="" data-share="twitter,facebook,linkedin,reddit,google" data-share-count="true">

On Monday, the Economist reprinted a chart from a BCG report, which purported to show the contribution of “the internet” to the total GDP of various different countries. Britain comes out on top, with an internet-to-GDP ratio of 8.3%; it’s followed by Korea, China, Japan, USA, India, and Australia. After the UK, the highest-ranking European country is Germany, on just 3.3%, while Canada lags far behind the US.

All of this was rather puzzling to me, so I spoke to BCG’s Paul Zwillenberg, one of the authors of the report. And the main thing I wanted to know, of course, was how on earth you could turn “the internet” into an annual dollar amount divisible by national GDP.

“It’s like electricity. It’s part and parcel of the fabric of daily life,” Zwillenberg said, almost before I could ask my question. “It’s touching every part of the economy.” I’m inclined to agree — but you’d never dream of measuring different countries’ electricity-to-GDP ratios. So what’s he doing here?

Zwillenberg did say that in ten years or so, “you won’t need to measure the internet economy because it will be totally pervasive”. But for the time being, he’s determined to measure the internet. And the way he’s doing it is very web 1.0.

Remember the dot-com boom of the 1990s, when everybody got excited about the internet because it was a new way to buy stuff? That’s basically what BCG is measuring here. They’re taking total consumer expenditure in each country, and working out how much of that expenditure is online. As in, buying a hardback from Amazon, or a Beanie Baby from eBay. Then they add in the amount you pay your ISP to get online each month. And then they add a certain amount for investment by private enterprise in internet infrastructure, and a bit more for what they call “net exports” — the Czech Republic, for instance, apparently has a big internet security software sector.

The exports bit helps to explain why Canada’s number is low: a lot of Canadians, for obvious reasons, like to buy things on Amazon and other U.S.-based e-commerce sites. And every time they do, under BCG’s methodology, the Canadian internet economy is decreased by that amount. (It’s an export of the U.S., and an import of Canada, and the calculations add up net, rather than gross, exports.) As a result, it’s theoretically possible, in BCG World, for the internet to account for a negative proportion of GDP, in some countries.

But more generally, it seems to me that BCG’s not really measuring the internet here — it’s not measuring the hours spent watching YouTube, or interacting with friends on Facebook and Tumblr, or spreading news on Twitter, or even checking your stock portfolio or updating your billing information somewhere. It’s measuring e-commerce, primarily, which is an interesting subset of the internet, and one of the oldest, but ultimately just a fraction of what it can be.

And when you’re measuring e-commerce, you’re measuring lots of things which aren’t really internet-related at all. For instance, e-commerce is naturally going to be more common in countries with a reliable postal service; in countries where gasoline prices are very high (because the cost of gasoline is a hidden tax on shopping in real life); and in countries where the cost of retail property is very high. The UK, of course, scores very highly on all those counts, while the U.S. doesn’t. But what those things say about the internet’s contribution to the economy as a whole is, I think, limited.

On May 17, if Alexia Tsotsis is to be believed, Facebook is going to go public, and might well find itself with a $100 billion valuation. Almost none of that valuation can be seen in the kind of things that BCG is measuring. So let’s not get too carried away with the idea that it’s easy to quantify the degree to which countries are diverging in terms of digital uptake. It isn’t. And the BCG methodology really isn’t a step in that direction.

4 comments

We welcome comments that advance the story through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can flag it to our editors by using the report abuse links. Views expressed in the comments do not represent those of Reuters. For more information on our comment policy, see http://blogs.reuters.com/fulldisclosure/2010/09/27/toward-a-more-thoughtful-conversation-on-stories/

Hold on a minute Felix, I’m not disagreeing with you per se, but just pointing out a couple of points you made that I don’t see any logic to.

First of all, the Facebook valuation generates GDP in the markets, that isn’t ecommerce. The advertising potential is what the valuation is based on, and that can be measured and put into the internet related GDP subset.

Secondly, the main reasons for the rapid adoption of internet commerce are a unified banking system, while the need for convenience and time saving is a motivator more than the cost of housing or fuel prices. You have worked hard to have the facts fit the hypothesis, but I don’t buy it.

Posted by FifthDecade | Report as abusive

‘it’ being the reasoning.

Posted by FifthDecade | Report as abusive

Sounds like all sorts of methodology problems here, including another one not mentioned – the numerator and denominator are apples and oranges. GDP is a concept of value-added, while a company’s revenue is not. Using a concrete example, Amazon’s revenue is far greater than its contribution to GDP, as much of the GDP contribution comes from the production and transportation of goods prior to Amazon entering the picture. Roughly speaking, a company’s contribution to GDP is the total of the wages paid to its employees plus its operating profit.

Posted by realist50 | Report as abusive

Felix writes: “you’d never dream of measuring different countries’ electricity-to-GDP ratios.”

Really? And certainly that is wrong with respect to changes in an individual country’s watt/GDP ratio over time.

Elctricity-to-GDP seems to be a widely looked-at metric, particularly with respect to cross-checking Chinese GDP numbers: http://www.google.com/search?hl=en&q=chi na+electrical+consumption+GDP

Posted by SteveHamlin | Report as abusive