China Wikileaks: GDP numbers are false; alternative measures available

To a Sinophile or anyone who has spoken with a university economist in China, this is not news. In fact, this is less than news. To any brave traveller who has made it past the east coast cities and poked around the Chinese interior, this is also not particularly interesting.

What is interesting is to hear this from the man who will be China’s next Premier in 2012. (See the WSJ 12/6 article entitled, “Chinese Leader Called Data 'Man-Made'”).

Here are the facts from the cables as reported by Ambassador Clark Randt:

1. Li Keqing refers to China’s GDP numbers as being ‘manmade’, and “for reference only”
2. He further suggests looking at electricity consumption, rail freight volume, and lastly loan disbursement/ interest rates charged
3. The reason for selecting these is based upon supposed accuracy over-and-above other measures.

A number of large firms in the US that track Chinese data use similar measures to estimate actual economic growth in China. A recent working paper published by the School of Public Policy at George Mason University suggested focusing on satellite imagery and urban lights as a way to measure growth. The WSJ also wrote about this in their "Real Time China Report" (See "Bright Lights, Big Cities: A China GDP Alternative?")

The nagging question: how can one accurately measure GDP growth?

There is clearly a need for stand-in indicators. Ideally, these should take into account the present/ immediate past, the short-term future, and medium-term outlook. Electricity consumption might be a fairly robust present or immediate past indicator. Rail freight seems like it could be a good stand-in for the short-term future (it could represent commodities to be sold in the near-term like foodstuffs, agricultural products, or intermediate goods). Finally, the loan disbursement represents the medium-term (or longer term) investments that should compose GDP growth.

How about other potential measures? Water consumption? Water is the bottleneck of industrial production. It follows that the government and industry would want extremely accurate figures on water. Water consumption is a good way to measure certain types of industrial production.

A friend of mine recently suggested looking at loan defaults. This would provide a good barometer of economic trouble.

How about using satellites to count the number of large industrial ships that enter and leave China in a given period? This could then be compared to historical trends.

There are all sorts of interesting possibilities.