The GDP mystery
Gross domestic product, GDP, is the total value of all goods and services produced in a country per year. Along with unemployment, the interest rate, the budget deficit, and public debt it’s one of the big five economic indicators that regularly breaks out of economics and into public life.
I think GDP is brilliant. Yes, it has its problems – covered very nicely in Diane Coyle’s “GDP: A Brief but Affectionate History” – but it turns out to be a brilliant indirect measure of so many of the things that really matter in society. Recently though I’ve stumbled upon a problem with GDP that I can’t get my head around.
It started with this paragraph by Giles Wilkes in the FT.
I wanted to check this. It felt wrong to me. So I got the World Bank’s data and looked at which economy grew quicker from 2000 to 2010. The answer is clearly Germany; its economy overtook the UK by growing by 49% to the UK’s 31%.
So I asked Giles what was going on. We danced around a bit.
What about absolute GDP growth per capita instead of percentage GDP growth per capita? Germany still outgrew the UK.
What about GDP growth of the whole country, instead of per capita? Germany still outgrew the UK.
What about at market exchange rates, or a constant exchange rate, instead of current exchange rates? Germany grew quicker.
So Giles dug out the figures he’d used and pointed me towards them. They'd been covered in The Economist so he was very confident in them.
Giles was absolutely right to be confident. Ask the IMF instead of the World Bank and the UK economy grew by 39% to Germany’s 36% in the decade to 2010.
What is going on?
The two measures that we’re using here are as close as possible to being identical. GDP per person, adjusted for prices in each country (PPP), then expressed in a common currency (international $) to allow comparison.
The only big difference is that one dataset is produced by the IMF and one by the World Bank. Yet they give very different answers to the same question.
But wait there's more!
Enter the OECD (this was actually where Giles’ figures came from). They publish tables of historical annual Real GDP growth and when you make a growth chart out of them it’s really clear that the UK outgrew Germany in the decade to 2010.
So to finish off I decided to look at the OECD's data on GDP/capita (PPP) for Germany and the UK since 2000. This shouldn't surprise you by now; it shows something completely different.
What’s the solution?
Since I’ve started looking into this lots of smart people have suggested new measures that would give me the truth. I should look at constant price time series (done that, same problem), get a Bloomberg terminal (I don’t have the money), or only use figures from each country’s national statistics agency (are the OECD, IMF, and World Bank really that unreliable? And anyway, what method of comparing different currencies should I use?).
I’ve spent spare time for nearly a year reading explainers on how to compare GDPs. Real, nominal, current prices, constant prices, local currencies, constant international dollars, current international dollars, PPP vs PPS, I’ve read it. I'm not an economist, but should I have to be to understand something as simple as this?
At the end of it all I’m no closer to understanding what’s really going on. After a year of trying, I’m not sure whether to carry on or give up.
Data’s great power to me is to challenge my own opinions. It forces me to think again about how the world works. If I can reasonably pick a dataset that confirms any existing opinion I have then the data is worthless to me.
In case you’re interested, my vote for the best explainer of all that I’ve read is this on Constant US dollar definition, current vs. constant explainer from the World Bank.