GDP figures (Part 2): Cross-country comparisons and a total for the world
In an earlier piece (7 January), I criticised staffers at the Financial Times (FT) for persistently quoting misleading figures for the GDP of a range of developing countries. In all these (as in many earlier) cases, the error arose from taking current dollar exchange rates, rather than purchasing power parity (PPP) converters, as the basis for cross-country comparisons. The flaws involved in exchange-rate-based comparisons of GDP have long been recognised, while the use of PPP-based comparative data has been endorsed by national statistical offices across the world.
As noted in my earlier piece, FT staffers, in a routine formula, mistakenly described India as ‘Asia’s third largest economy’; and I recently came across yet another instance in an article dated 12 February 2015 (by Borzou Daragahi). However, I am able to report that this and other such exchange-rate-based characterisations of countries may now be on their way out. Here are three instances where FT staffers whom I criticised in my earlier posting have now made due and exclusive reference to PPP-based comparative GDP data.
- In the FT for 12 January, in a joint piece, the authors (Victor Mallet, James Crabtree and Amy Kazmin) referred to India, correctly, as ‘the world’s third largest economy, after the US and China, in purchasing power parity terms’.
- On 25 January, in an article on Indonesia, David Pilling noted, likewise correctly, that ‘in purchasing power parity terms, its economy is roughly on a par with Britain’s’.
- On 9 February, in article by Michael Peel and David Pilling, an accompanying diagram quoted figures for GDP per head in Myanmar which were ‘at PPP rates’. Although for reasons that I will come to in a later piece the term ‘rates’ is inappropriate, the numbers given were indeed PPP-based. The figure that was quoted for 2010, of $3,430, is over three times the corresponding exchange-rate-based figure which might well have featured in an older-style FT presentation.
Arriving at a total for world GDP
As it happens, that same issue of the FT (9 February 2014) yielded an exchange-rate based comparison which takes the argument further, beyond individual cases. In a letter to the editor, Simon Walker, Director General of the (London-based) Institute of Directors, described the EU and the US between them as ‘accounting for nearly half of world gross domestic product’.
This is not correct. It is true that if the GDP for 2013 of these countries, and of the world as a whole, is valued in $US at the prevailing exchange rates of that year, the EU/US share for 2013 comes to 46 per cent ($34.3 trillion, out of a world total of $74.7 trillion). However, such valuations do not yield a measure of comparative GDP, since they make no allowance for cross-country differences in the prices of the goods and services that enter into GDP in the various countries covered. Valid cross-country comparisons, and hence a meaningful figure for world GDP, can be derived only when such differences have been allowed for: here lies the purpose and rationale of PPPs.
Taking the PPP-based figures for 2013, the combined GDP of the EU and the US is virtually unchanged, at just over $34.3 trillion at international prices expressed in $US. But for many other countries much higher numbers are involved, so that the total for world GDP now appears as $101.9 trillion, 36.7 per cent higher than the $74.7 trillion quoted above. Of this higher total, the combined share of the EU and the US comes only to just over one-third.
Why should this higher figure for world GDP – $27 trillion higher – be preferred? The underlying reason can be illustrated by the case of China.
Evidence from China
The estimated GDP of China for 2013, valued at current $US, is $9.47 trillion, whereas the corresponding PPP-based figure is about 70 per cent higher, at $16.15 trillion. China therefore accounts for around one-quarter of the difference between the above two rival figures for world GDP. Why is the PPP-based figure for China – and by analogy, for a good many other countries where the two measures of GDP diverge markedly – to be preferred? An answer emerges from direct evidence relating to the comparative output of China and the US.
For 2013, the ratio of Chinese to US GDP appears as 56.5 per cent if the exchange-rate-based valuation is taken, and 96.1 per cent with PPP. In weighing the relative plausibility of these two widely separated numbers, the following evidence for that year is relevant:
- Estimated consumption of primary energy in China was over 25 per cent higher than in the US.
- The total value of exports of goods and services from China exceeded by an estimated 7.7 per cent the corresponding figure for the United States.
- Production of vehicles in China was almost exactly double that in the US.
- Chinese production of steel was almost nine times that of the US.
In the face of such evidence it is hard take seriously the notion, which the exchange-rate-based comparison points to (and Mr Walker’s FT letter implicitly endorsed), that the economy of China is currently over 40 per cent smaller than that of the US.
Evidence of the same kind could be put together for other countries where PPP-based estimates of GDP appear as much higher than those that are exchange-rate-based. For some of these countries, the proportionate divergence is even greater than the 70 per cent quoted above for China: examples are India, Indonesia, Iran, Egypt and Pakistan.
In these two posts, I have focused on comparisons of real GDP as between countries with widely different levels of GDP per head. But it is not only in this context that exchange-rate-based comparative numbers yield misleading results. To quote again from a ten-year old paper of mine:
‘Exchange rates do not enter into the definition or measurement of either output (real GDP) or changes in output over time.’
This generalisation applies, first, to all cross-country comparisons of GDP, including those featuring only rich countries, and second, to individual countries. In Part 3 of this series I will take the argument further.