Are Chinese statistics manipulated?

In his comment on one of my previous blog posts, CrisisMaven asked me whether or not Chinese statistics are manipulated to meet political objectives. Indeed, it is not unusual for a foreign observer to question Chinese data like that. Given the importance of this question—if the underlying data is manipulated, all analysis based on it will be distorted—I would like to use this blog post to answer it.

Is the data manipulated by the authorities to meet political objectives? My answer to it, at least regarding the data in recent years, is a firm NO. How do I know that? The reason is pretty simple: in such a big economy as China, it is extremely hard, if not impossible at all, to massage the data without leaving any trace—any manipulation in the data will be quickly spotted by observers—and I personally haven’t found any signs of data manipulation in recent years.

In China, there are a whole lot of government departments, semi-government organizations, and private sector institutions providing various statistics. For example, in addition to the National Statistical Bureau (NBS) which is the main official data provider in China, there are dozens of other government branches including the People’s Bank of China (central bank of China), Ministry of Finance, Ministry of Commerce, and National Development and Reform Commission compiling and releasing their own data. Meanwhile, most sector associations like China Iron and Steel Association and China Association of Automobile Manufacturers have their data available to the public. Moreover, some companies like Soufun and CLSA also report data compiled by them. This is not to mention the high quality price data generated by various markets (Shanghai Futures Exchange, Dalian Commodity Exchange, etc.), everyday as China has become a market-based economy. Just like how different parts of the economy are connected to each other within an organic whole, data released by different institutions is interrelated and need to be consistent with each other. Hence, any attempt to manipulate data—say, flattering the GDP number a little bit—requires concerted efforts of various government departments as well as private sector institutions. If there are some participants in the circle not doing this kind of massage to their own data, inconsistency arises, leaving footprints of manipulators in statistics.

To be concrete, I’ll use the V-shaped recovery of China’s GDP growth in the last several quarters as an example. As shown in figure 1 with the dark blue line, China’s GDP growth started to bottom out in the second quarter of 2009. During that time, however, a lot of foreign observers believed the growth number was overstated and the recovery was unreal. In fact, to tell whether this improvement in the GDP growth is real or not, one only need compare the GDP series with other related indicators. As shown in figure 1, the uptick in the GDP number is corroborated by value added to industry, electricity consumption, and value added tax revenue, which all posted V-shaped recovery almost simultaneously. If one is not satisfied with the series I listed above, there are more left at his/her disposal, ranging from domestic transportation and product price numbers to international trade data released by other countries. In fact, all these data all show similar V-shaped patterns, suggesting the improvement in the GDP number is driven by underlying growth momentum. This example shows how one can detect manipulations by comparing related indicators.

As a consumer of a wide range of Chinese statistics, I carry out exercises like what is shown above fairly frequently. I personally haven’t found any significant inconsistency in Chinese data over the last decade. It makes me believe that data manipulation is not a big problem in China in recent years. Indeed, this is the consensus in the academic field as most researchers on this topic agree that there is little sign that China is manipulating its economic data.