October 24, 2005

You are on the invidual archive page of Making China's numbers add up. Click Simon World weblog for the main page.
Making China's numbers add up

Last week China reported another stunning GDP growth number of 9.4%. But as we've found numerous times before, the numbers underlying the GDP calculation don't add up. Either China's consumers went on strike or fixed asset investment has been over-estimated. Jake van der Kamp is on the case and reaches an unsurprisingly but important conclusion (below the jump).

Other reading

Brad Setser calls it China's crazy numbers.
Big Picture also doesn't trust the numbers, saying I simply don't trust the commies to release the real data


When the National Bureau of Statistics in Beijing occasionally admits that the figures it publishes are not entirely reliable, I know exactly what it means. We had another case of it last week with the release of third-quarter provisional statistics for gross domestic product. Strong investment and robust private consumption pushed economic growth to 9.4 per cent in the quarter, said the bureau.

Oh yeah? Let us take a closer look at this.

The red line on top of the first chart shows a four-quarter rolling total of reported GDP in dollar-of-the-day terms. For the year to September 30, the mainland's economy registered an output of slightly less than 15 trillion yuan. Now we turn to the components of GDP. The biggest by far is fixed asset investment and firm figures for this were also published last week. Fixed asset investment for the year to September was 8.2 trillion yuan, 24.5 per cent greater than the previous year and the equivalent of 54.8 per cent of total GDP.

These are huge numbers, incidentally. It may be possible to find another country with fixed asset investment of more than half of GDP but I cannot think of any. For contrast try Britain at only 16 per cent or the United States at 19 per cent. In Hong Kong the figure is 21 per cent.

Let us take the fixed asset investment figure out of the mainland's GDP, however, to see what is left. We now have the red line on the chart. The rest of the economy turned out 6.8 trillion yuan in the year to September. The next step is to take out net trade in goods and services. The mainland enjoys a soaring trade surplus at the moment. It is not huge by the benchmark of the overall economy but deduct this component of GDP and you get the green line on the chart. We are now down to 6.1 trillion yuan. Finally the line in the ugly colour. Government consumption spending is another component of GDP and detailed figures are also published. Take it out as well and we are down to 3.4 trillion yuan. Notice here that this final line for what is left now curves distinctly down.

What is left is two items. One, change in inventories, is now an insignificant component of GDP and we shall ignore it. The other is private consumption spending, normally the biggest single component in most economies but certainly not in the mainland, it seems.

It seems, I say. Take that line in the ugly colour, plot it as a year-over-year growth rate and you get the red line in the second chart. It seems from this that in the year to September the man on the street spent 17 per cent less on daily necessities and toys than he did the previous year. But this is not what other official statistics say. They say that retail spending for the year to September was 13.6 per cent greater than it was the previous year (the blue line) and that this retail spending alone was almost twice as great as the remainder number we calculated for all personal consumption spending.

How is it possible?

It is not. The latest GDP figures from the mainland simply do not add up. I hesitate to use the word "rubbish" to describe them but I am starved of a better one.

I think the enormous discrepancy most likely results from an overstatement of fixed asset investment. Capital spending probably is much less than the National Bureau of Statistics says it is. This would imply something else again, however. It would suggest that a vast amount of money earmarked for capital projects was embezzled by corrupt officials and used instead for personal spending on luxury services and toys.

I shall not suggest that this surprises you.

Every second anecdote from the mainland tells you it happens every day. All I have done is put some possible numbers to the scale of it, a very big scale indeed. But I do suggest to the National Bureau of Statistics that it adopt a brand new approach for checking statistics, a new one to the bureau that is. The next time it publishes data it might want to check that the sum of the parts adds up to a given total.

If it does not, and no further work can make it so, then the bureau should not bother to mislead us with grossly and obviously false information. The round bin under the desk is where these sorts of statistics belong.

Determining economic statistics is notoriously difficult at the best of times. But such wildly inaccurate numbers make scary reading. Why? If you're steering the world's fourth largest ship and your navigational data is "rubbish", you're going to end up doing plenty of damage not just to yourself but everyone around you.

posted by Simon on 10.24.05 at 09:51 AM in the China economy category.


TrackBack URL for this entry:

Send a manual trackback ping to this post.

Is China's "Economic Miracle" a Lie?
Excerpt: A growing number of macroeconomists seem to think so...
Weblog: A Stitch in Haste
Tracked: October 26, 2005 12:48 AM


Very interesting. I wonder whether inventories might also be a lot bigger than expected?

posted by: HK Dave on 10.24.05 at 10:53 AM [permalink]

I think inventories are a relatively minor part of GDP, so even if they were widely different the impact on the larger number would be small.

posted by: Simon on 10.24.05 at 10:56 AM [permalink]

Don't forget that what China tells the world, tells its people and tells its officials are three entirely different things. I have no doubt that the Chinese officials know exactly how much they are really spending/growing/stealing/etc... I don't think navigation or bad finacial data is as real a danger as these figures express. The rampant corruption on the other hand, is another story entirely.

posted by: DudeMiester on 10.24.05 at 03:47 PM [permalink]

It is very surprising to me that none have carefully checked the data. The data of the fixed investment used includes private and government investment, and the amount of the government consumption is the government spending, which also includes government investment. The fact is that the government investment has been subtracted twice. How can the remainder be equal to the private consumption, when you subtract government investment twice?!

posted by: Quan Song on 10.27.05 at 04:19 AM [permalink]

Your calculation is totally wrong and misleading because you doesn't understand the terminologies used by the Chinese statistical authorities. Investment expenditure you use is so-called the completed investment in fixed assets, which is not gross capital formation in the national accounts. Investment in fixed assets involves double accountings. The difference between the completed investment in fixed assets and capital formation is much bigger than that between retail trade and final consumption. The Chinese statistical authority has never released GDP components quarterly, which are only available in yearly figures several months after the year. The problem with China GDP statistics is not "adding up", which is easy to do, rather is whether they reflect the performance of the Chinese economy. Professor Thomas Rawski from University of Pittsburgh has discussed several issues about China GDP statistics (www.pitt.edu/~tgrawski/papers2001/gdp912f.pdf).

posted by: Leilei on 10.27.05 at 12:01 PM [permalink]

The problem with China GDP statistics is not "adding up", which is easy to do, rather is whether they reflect the performance of the Chinese economy.Either way, there's a big problem.

posted by: Simon on 10.27.05 at 12:06 PM [permalink]

At the base, China does not have an accounting system that clearly distinguishes between value-added and intermediate consumption at the micro level. So, the problem goes even deeper than the analysis here. Bottom up compilation is not coherent.

posted by: parisdude on 11.16.05 at 10:25 PM [permalink]

Post a Comment:


Email Address:



Remember your info?