China’s Statistical Failures
For China’s onlookers, it is no surprise that official data can sometimes befuddle even the most capable of minds. Yet, China’s National Bureau of Statistics (NBS) is not blind to its reputation for less than spotless data collection and release. A brief exploration into China’s economic indicators can reveal much insight into the dilemmas the bureau continues to face even as reforms for statistical aggregation and analysis are implemented.
The gross domestic product (GDP) is an important indicator of the level of economic growth and health in any country as well as a powerful snapshot of the overall performance in an economy. By zooming in on this particular indicator, the hurdles and peculiarities of China’s statistics could be exposed. To see how China measured, I went through the nominal quarterly GDP numbers from Q4 1999 to Q2 2011 for data collected directly by the NBS and those reported by the provinces. The aggregate numbers should come out to be the same, if not with only minor rounding differences. In a simple calculation, we compare the national GDP growth rate released by the NBS to the summed provincial weighted growth rates reported by local authorities. Unsurprisingly, the calculations show that data reported by the provinces are higher than those collected by the NBS. This difference can be explained by the political priorities of the different actors. The NBS has the responsibility to keep track of economic data and to ensure market participants, as well as the leadership, a clear image of China’s economic performance. However, local authorities are politically motivated to report higher numbers as most get promotions based on economic performance—incentivized since higher growth rates mean better job and income creation for the Chinese population, which ensures the stability of the country and the Chinese Communist Party’s (CCP) grip on power. The NBS numbers are regarded with more accuracy because of the bureau’s access to data on tax revenues, electricity consumption and goods turnover. Provincial governments would also double count productivity numbers from industries with facilities in other provinces; thus, inflating the summed provincial GDP numbers.
Figure 1: National vs. Provincial Growth Rates (%)
Source: NBS and provincial Chinese governments
China’s GDP is divided into three major categories of primary industries (agriculture), secondary industries (mining, heavy industries, manufacturing and utilities) and tertiary industries (services). To isolate the origin of this anomaly, I peeked into the raw nominal quarterly GDP numbers collected by the NBS and those reported by the provinces, which showed even greater discrepancies in China’s data. The raw numbers revealed some major differences between what the NBS reports in one year versus what provinces report for the varying industries. Differentials to the order of tens of billions of renminbi (RMB) are already significant, but in some quarters it reaches trillions of RMB.
Figure 2: Data Difference: National Minus Provincial GDP Numbers (RMB bn)
Source: RMB, provincial governments, RGE
The brief foray into China’s sector GDP numbers reveals an astonishing conundrum whereby services data collected by the NBS are generally higher than those reported by provincial governments, at least before the end of 2009. One explanation for this difference is the lack of attention paid by provincial governments towards the services sector until after 2009, when the central government began emphasizing the growth of the services industry. Another explanation relates to the difficulties for China to collect and report data on services. China’s classification of services industries might include companies and institutions affiliated with or heavily contracted by the central government and/or state-owned enterprises; thus, it is easier for the NBS to include companies working on national contracts compared to provincial authorities. Looking at the secondary industries, the NBS reports a significantly lower number after 2004 compared to provincial governments, and that gap has only grown in recent quarters. A strong explanation for this is the provincial authorities desire to maintain a higher target growth rate even as the global economy slows due to the 2008 financial crisis and a growing European sovereign debt crisis. The only measure of certainty is the primary industry GDP, where data discrepancies between the NBS and provincial governments are small to none.
The large spikes in differences between the NBS data and provincial data reached their peak in Q3 2004, when about two-thirds of provincial governments did not report any GDP and sector breakdown data. In other quarters prior to 2008, some provincial authorities have failed to report their quarterly data. The differentials between 2000 and 2008 can also be explained by continued reforms in statistical reporting and classification by the NBS, leading to major revisions in the GDP and the respective industries. Another explanation for major quarterly differences could be access to other proxies of economic growth such as energy consumption data. Energy consumption data could be higher for provincial authorities than the NBS because of off-grid electricity generators, which could influence provincial level statistical bureaus to estimate different economic output numbers because of production-based data collection.
Even as China’s statistical reporting continues to change for the better, other proxies of economic growth will need to be observed to better understand China’s economy. Li Keqiang, who is presumed to replace Premier Wen Jiabao, said that the GDP numbers are unreliable and that the observance of electric consumption, transportation logistics and bank lending would provide a better picture of China’s growth story. It might not be surprising for emerging markets to have statistical failures but advanced economies are not too far ahead when it comes to statistical accuracies. Even the U.S. has suffered the plague of statistical failures, as pointed out by Felix Salmon from Reuters.
Comments are closed.
