China Struggles to Publish Accurate Economic Data

The government moves to shed its reliance on local officials
A worker controls the production line of textile at Babei Group in Shengzhou, Zhejiang province, China Photograph by Feng Li/Getty Images

The figures that go into China’s gross domestic product are “man-made” and “for reference only,” warned Chinese politician Li Keqiang back in 2007. The comments by the then-regional party head, who is expected to become premier next spring, were revealed in a diplomatic cable published by WikiLeaks in late 2010. Li’s remarks are especially relevant now as Chinese officials are expected to announce on July 13 that GDP grew by less than 8 percent in the second quarter, a three-year low.

But how low exactly? Investors, bankers, and economists face a host of difficulties in interpreting the numbers from China’s statistics bureau. Combine all officially reported provincial GDP numbers for last year and you come up with a total exceeding national GDP by about 10 percent, Ma Jiantang, head of the National Bureau of Statistics, said in February. Ma said that is due partly to double counting of activities like factory production, and that his bureau was trying to correct the problem.

The fact that China’s registered urban unemployment has barely budged, moving between just under 4 and 4.3 percent for the last decade, is harder for officials to explain. Also perplexing: why growth in electricity consumption has slowed much faster than growth in official GDP (dropping to about 4 percent growth in June, according to Chinese media) when usually they move more in tandem. That has some analysts wondering whether GDP figures are being skewed upward in the runup to a leadership transition this fall. “Out of the black box comes a number, and that number doesn’t always line up with the other numbers,” says Andrew Batson, Beijing-based research director at macroeconomics consulting firm GK Dragonomics. “I wouldn’t be surprised if the GDP numbers this year are smoothed.”

One legacy of the planned economy is that bureaucrats are given targets by the central government for everything from steel production to harvests and local GDP. These same officials traditionally have been promoted on their success in making their numbers. “We have a saying in China: The cadres produce the data, and data produces the cadres,” says Jin Yongjin, a professor of statistics at Renmin University.

There are some pieces of the economy that the government is pretty good at measuring, says Louis Kuijs, an economist at the Fung Global Institute, a think tank in Hong Kong. Industrial production and profits are considered relatively accurate, he says. Those numbers benefit from a nationwide system of corporate reporting instituted decades ago to help central planners steer the economy. “In the old system it was crucial to have that reporting system in place,” says Kuijs. “China’s industrial survey is giving us quite good numbers.”

The country’s statistical system is far less capable of measuring the service economy or getting an accurate reading of consumption by the middle class. “If you look at the number of cars produced in China, you will probably get a more or less accurate number,” says Stephen Green, regional head of research for greater China at Standard Chartered. “But the system isn’t good at measuring how many karaoke nights people have had or restaurant meals they have had.” Measuring retail sales is more oriented to counting purchases by government offices and big state enterprises, says Green. That may explain why retail sales haven’t fallen during downturns: State entities are ordered to keep buying even when individual consumption has likely fallen.

Investment banks have searched for the indicator that will predict an economic turning point. Standard Chartered looked at sales of earth-moving equipment before deciding it was a lagging rather than a leading indicator. Bank loans, as well as electricity consumption and rail cargo volume, all cited by Li Keqiang as more reliable than GDP, are still a good proxy for economic activity, says Green. UBS Securities has informally surveyed local developers to get a handle on real estate trends. Perhaps the most ambitious effort is the recently launched China Beige Book, a quarterly survey of some 2,000 bankers and company executives, modeled on the U.S. Federal Reserve’s Beige Book. It measures growth in eight key industries across China’s major regions, says Leland Miller, president of CBB International, which publishes the book.

Chinese policymakers are trying to address the government’s statistical shortcomings. More data are now directly reported to Beijing, as opposed to being first filtered through local party offices. The statistics bureau has moved to standardize data collection by China’s many ministries and industrial associations. And it has worked with the United Nations, the International Monetary Fund, and the Organization for Economic Cooperation & Development to improve its tracking of the economy.

China still tends to treat its data gathering as a national secret, says Anne Stevenson-Yang, co-founder of Beijing-based equities analysis firm J Capital Research. She cites the government’s refusal to release the weighting of goods tracked to compile its consumer price inflation index. “Why would you ever lift the hood and show people how you do it? That only reduces your ability to change numbers if you need to,” she says.

Before it's here, it's on the Bloomberg Terminal.