Wednesday, October 26, 2005

Is China's data reliable?

This is not a new topic, but surprisely heated up currently, and, therefore, drew my attention. West academics and commenters, such as Barry Ritholtz, Brad Setser and Simon World, made their new claims based on the latest number, even the respected professor James Hamilton joined the boat. All of them used similar official data and charts created by Simon World. They argued that China's data did not add up based on:

GDP = I + G + C + Net export

It is very surprising to me that none have checked the data. The data of the fixed investment used includes business 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 added twice.

There is official data I got from National Statistics Bureau, Ministry of Finance, and China Customs, which is expected as same as Simon World's.

2005 H1 GDP 6742.2 billion yuan (National Statistics Bureau)
2005 H1 Fixed Investment 2796.7 billion yuan (National Statistics Bureau)
2005 H1 Sales 2960.97 billion yuan (National Statistics Bureau)
2005 H1 Net export 327.3 billion yuan (China Customs and the exchange rate $1 = 8.27 yuan used)
2005 H1 Government Spending 1242.1 billion yuan (Ministry of Finance)


Annualized data, billion yuan:

2005 GDP 13484.4
2005 Fixed investment 5593.4
2005 Sales 5921.94
2005 Net export 654.65
2005 Government Spending 2484.20

In order to compare with the claim made by Simon World and other, let’s subtract GDP in the same wrong way, in which I subtract government investment twice.

GDP Less fixed investment = 13484.4 – 5593.4 = 7891
Less net trade = 7891 – 654.65 = 7236.35
Less government spending = 7236.35 - 2484.20 = 4752.15

Obviously, the remainder is far less than private consumption of 5921.94. How can it be equal to 5921.94, when you subtract government investment twice?!

Now let’s subtract GDP in the right way.

GDP Less government spending = 13484.4 – 2484.2 = 11000.2
Less net trade = 11000.2 – 654.65 = 10345.56
Less private consumption = 10345.56 – 5921.94 = 4423.61

It means the business investment including the fixed investment and the inventory investment should be 4423.61 billion yuan. Annualized fixed investment is 5593.4 billion yuan, and needs to subtract the government investment. However, the data of total government investment is not available. I get the following components of the fixed investment from National Statistics Bureau:

Component 1: 2005 H1 fixed investment on water conservancy project, environment project, and public facilities: 214.447 billion yuan (428.89 annualized)
Component 2: 2005 H1 fixed investment on education: 74.372 billion yuan (148.74 annualized)
Component 3: 2005 H1 fixed investment on health care, social security and welfare: 19.245 billion yuan (38.49 annualized)
Component 4: 2005 H1 fixed investment on public administration: 85.596 billion yuan (171.19 annualized)

Fixed investment less component 1 = 5593.4 – 428.89 = 5164.51
Less component 2 = 5164.51 – 148.74 = 5015.77
Less component 3 = 5105.77 – 38.49 = 4977.28
Less component 4 = 4977.28 – 171.19 = 4806.09

The difference between the remainder of the fixed investment and the business investment derived from GDP is only 382.48 billion yuan, which is likely below the government investment on transportation infrastructure. In the other words, when you subtract the remainder by the government investment on transportation infrastructure and then add the business inventory investment, the result may not be very different from zero.

China's data may not be very accurate, but it seems add up. When there is no creditable case against it, it is not fair to say that the data is unreliable and unbelievable, unless you hold prejudice.