China's Puzzling Numbers - WSJ.com
No one doubts that China is an important driver of global growth, with a major impact on the financial markets. But many observers doubt that China's economic data can be trusted.
In his new book, "Understanding China's Economic Indicators," The Wall Street Journal's Beijing-based columnist Tom Orlik guides readers through data on everything from gross domestic product to bond market yield curves.
The following excerpt explains the political and technical challenges facing statisticians measuring the size and growth rate of the world's second largest economy after the U.S.:
In 1998, the Asian Financial Crisis brought the region's economy grinding to a halt. A closed capital account protected China from the speculative attacks that crippled Thailand, Indonesia, and South Korea. But with major trade partners sliding into recession, China was not immune to the effects. Falling growth in energy consumption, airline passenger numbers, and imports all pointed to a sharp slowdown in growth. But if the economy was indeed sliding into recession, it was not evident to the National Bureau of Statistics. Official data for the year shows GDP growth of 7.8%, down only slightly from 8.8% in 1997 and within spitting distance of the magic 8% that is believed to be the minimum required to maintain social stability in China.
The 1998 GDP data has generated a storm of controversy. Academic economists have expended much energy in either defending the NBS calculation or, more common, attacking it and offering their own alternative estimates. Professor Harry Wu of Hitotsubashi University in Tokyo and the late Professor Angus Maddison were among the most stern, concluding on the basis of their own index of industrial production that China's GDP grew just 0.3% in 1998 (minus-0.1% in Professor Wu's recent updated results). Professor Carsten Holz of the Hong Kong University of Science and Technology has weighed in forcefully on the other side, concluding that although there are inherent difficulties with calculating the GDP of a large and rapidly developing economy, it is difficult to identify systematic biases in the NBS data or to arrive at compelling alternative estimates.
The government has never admitted any problem with the data. Indeed, in a revision to the historical GDP data as a result of the 2004 Economic Census, the 1998 figure was the only one that was left untouched. But it has come close. The story that has trickled out in speeches and articles in the official press points the finger of blame at an excess of enthusiasm from local officials. Caught between the reality of an economy in crisis and the dream of career progression that depends on delivering growth hitting the 8% mark, officials engaged in rampant falsification of production data. Premier Zhu Rongji spoke of a 'wind of embellishment and falsification' that swept through the statistical system. An article in the NBS in-house magazine, Economics Education, even developed a game theory model to explain how exaggeration by one official could trigger a wave of exaggeration by other officials, a kind of Chinese bureaucratic version of keeping up with the Joneses.
In the years that have followed, controls on local statistics offices have been improved, more data is reported directly to the NBS headquarters in Beijing (bypassing the embellishing hand of local leaders), and more surveys and other checks on the accuracy of data collected at a local level are now conducted. These checks and balances have freed the national-level data from the impact of local exaggeration—or 'adding water,' as it is called in China.
But local-level GDP data, especially for more backward provinces, remains deeply unreliable. In the Wikileaks scandal that broke at the end of 2010, it was revealed that even China's premier-in-waiting, Li Keqiang, had little faith in the provincial GDP data. Speaking to the U.S. ambassador in 2007, when he was still party secretary of Liaoning province, Li said that the data was 'man-made' and, therefore, unreliable. To keep a handle on the growth rate of the Liaoning economy, he relied on tracking electricity consumption, rail cargo volume, and bank lending.
But lying local officials aren't always the problem with China's official growth data, and the NBS doesn't always err on the upside. As a second episode in the troubled history of China's GDP calculation shows, the sheer size and complexity of the Chinese economy can also defeat the statisticians, and the official numbers can understate the true size of the economy.
In 2004, China conducted an Economic Census, including a thoroughgoing attempt to get to grips with one of the most slippery parts of the economy: the services sector. With legions of statisticians on the ground in street-side cafes, foreign-language schools, and IT support centers, the NBS was able to get a more accurate gauge of the total output of this important but overlooked sector of the economy.
The result was an upward revision of GDP for 2004 of a whopping 2.3 trillion yuan ($360 billion), adding 16.8% to the size of the Chinese economy. How could the regular annual accounting exercise have overlooked such a large chunk of China's output? Part of the reason is that the services sector is made up of many small enterprises and informal operations. Adding up the output of a million local hairdressers, half of them with no accounting ledger and half with a ledger that conceals as much as it reveals, is a bigger challenge than taking stock of the output of the big modern firms that dominate the industrial sector.
But that's not the entire story.
The NBS now uses the United Nation's approved System of National Accounts as the basis for calculating GDP. But in the early reform era, the Material Product System, borrowed from the old U.S.S.R., was employed. As might be expected from a scheme devised in Soviet Russia, the Material Product System is rather good at measuring physical outputs (the tons of steel and cement valued by central planners) and less good at measuring the intangibles produced by the services sector. Some of the biases in the old system linger in the new, and that's another reason the NBS missed such a large volume of services output.
After the embarrassment of the 2004 census, the NBS made a serious attempt to overcome the deficiencies of its coverage of the services sector. New service sector industries, accounting for most of the output that was missed in the years before the census, were brought into the fold of the annual survey. But counting the output of a rapidly evolving services sector remains a challenge for China's statisticians. In 2008, the next round of the Economic Census discovered another 1.3 trillion yuan in GDP that had previously been overlooked, adding 4.4% to the estimated size of the Chinese economy. Once again, the lion's share of the addition came from the services sector. A 4.4% addition to GDP in 2008 is considerably smaller than the 16.8% addition in 2004; the NBS is getting better at counting service sector output. But undercounting the services sector remains the most serious methodological problem for China's national accounts.
Suspicions about the reliability of China's data continue to focus on lying officials 'adding water' to bias the GDP numbers upward. But the more real risk is that that a large chunk of a rapidly changing economy has again been overlooked by the statisticians, and the official data understates the true size of the Chinese economy."
China's Puzzling Numbers - WSJ.com