China Research

In the last China Research issue, Hubert Fromlet shared his thoughts on the tricky issue of China’s GDP data reliability. Since then, GDP data for 2014 has been officially announced: 7.4 percent. International media have commented this negatively, with the argument that this growth rate was lowest in 24 years. One might argue that the real tragedy about this number is the only tiny deviation of 0.1 from the growth target propagated in early 2014 and in our lack of trust in this result. Why all the talk about upcoming lower growth, why introducing the hardly attractive concept of the “new normal” state of China, if the target was just failed by 0.1 percent? Have the sinister news of the last months just been some spinning efforts to make 7.4 bad enough, but still credible? Has growth actually been much lower, as Hubert Fromlet’s analysis suggested? Are we back to a situation like in 1998?

In that year, the Chinese government had propagated a growth target of 8.0 percent, a target that was soon threatened by the repercussions of the Asian financial crisis and the huge floods hitting several provinces around the Yangzi during that summer. The government reacted by stressing the dedication to still achieve the 8.0 growth target. The official growth rate of that year was later published as 7.8 percent, arguably a very Chinese effort of “spinning”: Less than the target, thereby increasing the credibility of the number, but still close enough to satisfy internal and external expectations. Economists, both in and outside China, questioned the reliability of these data. The American China economist Thomas Rawski later become famous for his reality check on the GDP data by using alternative measures for growth based on energy, traffic, real estate etc. statistics. The results were devastating, indicating growth rates of around 2.0 percent for China’s economy for 1998 and following years. When his conclusions started to be circulated in popular international media, Chinese media started a counter propaganda strike, questioning the methods of Rawski’s analysis.

Only few years later though, when – in the course of the more recent leadership change – Chinese media tried to stress the economic competence of Li Keqiang, they repeated a story of him questioning reported provincial GRP (gross regional product) data. He was said to use alternative indicators instead to get a realistic picture of provincial economic development, indicators that surprisingly resembled those of Thomas Rawski.

So, are Chinese GDP data just deliberate fabrications? Probably not: There exist a number of technical reasons why Chinese statistical data are troubling, some of which are difficult to avoid: First, China’s gradual economic transition necessitates a gradual adaptation to the internationally practised System of National Accounts (SNA). Progress in this transition, for example by including the service sector into the reporting system, explains a lot of changes in Chinese statistics over the last 35 years. Second, China’s fast growth regularly necessitates adaptations of the units of analysis. For example, if it were appropriate to count gross capital investment above 500,000 Yuan in the 1990s or early 2000s, with investment volumes continuously on the rise such investment projects today are only counted if they exceed 5 million Yuan, leading to a one-time decrease in growth rates of FAI in the year when the adaptation was initiated (2010).

But still, there has been an element of fabrication in GDP data in the past that until today feeds our suspicions: Chinese GDP data is regularly published in the China Statistical Yearbook. Each yearbook publishes the latest nominal GDP data as well as nominal and real GDP growth. Past data for nominal GDP is corrected in the current yearbook if necessary. In most cases since 1992[¹] (YB 2005, 2007, 2009, 2011, 2012, 2013, 2014) these ex post corrections only pertain to the previous year GDP, but in several years (YB 1996, 2003, 2004, 2006, 2008 and 2010) backward correction of nominal GDP data was undertaken for a number of years. The most substantial corrections of this kind happened with the Yearbook of 2006, when – based on a new census – nominal GDP data was corrected back for the years 1992 to 2005. At the same time, real GDP growth rates have seldom been corrected backwards. Only in the YB 2005, 2007 and 2011 could such backward revisions be observed with the most substantial revision for a long time series undertaken in the YB 2011. In economic terms, revising nominal data without adjusting real GDP growth accordingly implies that the implicit GDP deflator (China does not publish a GDP deflator, but also does not use the published consumer or producer price index as deflator) would have to be corrected accordingly. This again assumes that the flaws in original GDP data and the GDP deflator were of the same scope. Such an assumption is hardly convincing from an economic perspective.

To give an example: GDP nominal growth in 2009 according to the data published in 2010 was 13.2 percent, according to later revised nominal data it was just 8.6 percent. Real growth rate for that year is assumed to have been 9.1 percent according to the Yearbook 2010, which was revised to 9.2 percent in the Yearbook 2011. Thus, while the correction of the nominal growth rate was 4.6 percentage points, the correction of the real GDP growth rate was only 0.1 percent. Thus there must have been a major correction of 4.5 percentage point to the implicit GDP deflator!

In the years since 2010, backward revisions of nominal GDP have been comparatively small. This could indicate that GDP data has become more reliable. The problem is that we don’t know for sure. Following the logic of Chinese politics major corrections of GDP may have been deemed inappropriate in times of the global financial crisis and Chinese leadership change. If we assume this logic to be relevant, simple alternative calculation will not work anymore: If the Chinese Premier used alternative calculations in the past himself and if the Chinese leadership today really wants to cook the books in order to persuade Chinese and foreign observers that GDP data may go down, but only slowly so, the leadership would also know how to do the trick more convincingly than in the years around 1998.

It’s a real dilemma: How can the Chinese leadership reinstall trust in the numbers after they have themselves shown their distrust in past reporting? And how can we regain trust in today’s data after there have been so obvious flaws in the past? To me, the Chinese government has not to put efforts into explaining that lower growth rates may be the “new normal” situation for China. It would be much more helpful if the “new normal” stood for transparency and reliability of media and data.

[¹] The calculations presented here are not looking into the years before 1992.

 

 

 

 

 

 

 

Doris Fischer
Professor, University of Würzburg, China Business and Economics

 

 

Back to Start Page