A short history of China’s doubtful GDP

It is no secret that the quality of Chinese statistics, especially the GDP, is far from perfect. The GDP’s lack of volatility often looks too surreal. As many other China analysts, I often feel frustrated by this too. Perhaps I feel even more strongly because it is somehow embarrassing when foreign investors and scholars always ask me how authentic my country’s data is. Recently I reread American-Chinese historian Ray Huang’s book in which he again raised view that China failed to industrialize before Europe because the old empire was functioning on the basis of philosophical principles instead of solid numbers. For instance, 50 years after the Bank of England was set up, the 18^th century Chinese emperor even had no idea about the exact number of population under his rule and some government officials even argued that it was not necessary to collect the statistics and it would only increase the fiscal burden to do such a census.  

China paid huge costs for its inability to modernize and the old empire almost lost its sovereignty. Of course, things changed dramatically in the 20^th century as China started to industrialize at a relatively fast pace. However, although GDP became the main tool measuring an economy after the WWII, it was only until 1985 that mainland China started to calculate its GDP. Prior to that, China was using Material Product System, which was a method adopted by the former Soviet Union and many other socialist countries before 1991 and still used by North Korea and Cuba today, to calculate its economic output. One of its main distinctions is that the MPS does not include any services.  It was only until 1993 that China completely abandoned the MPS and GDP became the only main measurement of economic output.  In other words, we actually have little idea about China’s actual GDP before 1985 because it was not even calculated back then. The pre-1985 GDP numbers we find in the economic databases today were rebuilt after 1985 so it is reasonable to argue that the data quality is not very high. Moreover, the GDP in late 1980s and early 1990s was still not trusted by the World Bank, who published every country’s GDP in US dollar, because official exchange rate of the Renminbi was very far from the market rate. The World Bank only started to use China’s official data after China’s first exchange rate reform in 1993.

If the scepticism on GDP was largely on methodology or exchange rate before 1993, the concern later became whether the number was cooked and it had a lot to do with the “growth target”. Growth target itself is very Soviet-style but in the 1980s it was not as important as later in the 90s. There could be two reasons. First, China only started to calculate GDP in 1985 so it takes time to become a widely acceptable number. Second, the political event in late 80s heightened the importance of economic growth.  The 8^th Five-Year-Plan released in 1990 only set growth target at 6%, but two years later the party decided to lift growth target from 6% to 8% in the 14^th Party Congress in 1992. The congress report says “it is not only a major economic issue, but also a major political issue that whether our economy can grow even faster”. The more well-known fact is that during the Asian crisis former Premier Zhu Rongji stressed that GDP growth had to reach 8%, but actually 8% was mentioned 6 years earlier than that.  Why 8%?  Many believed that only by growing 8% China would not have massive unemployment, but if that’s the case it would be hard to explain why the party set growth target at 6% in 1990. Clearly in 1990 the party did not think growth had to reach 8%, but later they figured that more political assets were needed and higher growth would help.

To some extents it is fantastic to have a GDP growth target, as long as the target is considered to be reasonable, because the government can finally function on the basis of numbers instead of philosophy, moral standards or relationship. It was also a big adjustment from Mao’s chaotic era when economic output targets were largely based on his personal belief. In fact, China was not a Soviet-style planned economy before 1980s because it never had planning on mathematical basis like the former Soviet Union. It was a one-man command economy which was different from both Soviet Union, who used sophisticated mathematical equations to calculate general supply and demand, and the United States who let the market decide most of the issues. Therefore, having a serious target is a big progress in Chinese history and it helped a lot in achieving high growth. However it also has dark sides as local leaders will have incentives to inflate their data. Data manipulation is natural in such a regime and it would become more severe during bad times. We saw ridiculous number cooking when Mao launched the “Great Leap Forward” movement in late 1950s and early 1960s and the cost was lives of millions. The 90s was much better than Mao’s era, but when Zhu Rongji announced 8% growth target during the Asian crisis, local governments started to cook numbers again simply because 8% was mission impossible.  The numbers reported by local governments were so surreal that even the National Bureau of Statistics would not believe it. The national GDP announced by the NBS, which represented 7.8% growth, was much lower than the sum of local GDP. Since then, it has become a common phenomenon that every year the national GDP is lower than the total of local GDP.  Although the NBS data seemed closer to reality than the local governments’ figures, there were still many people challenging its quality. Their argument was that since the work of the NBS had to be based on the reports from local level, it seemed impossible that NBS could make an accurate calculation if all local data was inflated. Thomas Rawski’s 2002 paper largely summarised the debate on China’s data quality at around 2000 and he believed that China’ 7.8% growth in 1998 was overstated greatly.

To counter the criticisms on quality of statistics, the NBS made another progress in calculating national account in 2003. China Daily said

China will reform gross domestic product (GDP) calculation and statistics release based on international standards, so that the statistics objectively reflect the performance of the economy.

Li Deshui, director of China’s National Bureau of Statistics (NBS), said here Thursday that the reform plan had been approved by the State Council, and will be implemented by NBS. He pointed out China’s GDP calculation method, initiated in 1985, had been improved gradually. However, the calculations still had some problems due to the lack of a regular adjusting and revising mechanism. Some traditional methods did not comply with international norms, arousing some criticism from the international community.

The NBS’ efforts should get some credits, the reform, however, did not reduce scepticism on China’s statistics, but this time it is on the opposite direction. Since late 2003, economists started to argue that official data understated real growth. The People’s Daily said:

Official Chinese economic data has been viewed skeptically for decades, but one prominent Chinese economist is making a charge rarely heard before.

…Wu Jinglian says the country’s economy actually grew faster during the first half of the year than the official 8.2 percent announced by the government.

Wu, who works for the States Council’s Development Research Center, bases his statements on numbers from the country’s statistics bureau.

Wu’s argument of GDP understatement is very similar to some of the arguments we see nowadays. He pointed out that power generation and IP growth was much higher than GDP growth, which was exactly the same (but in opposite direction) as the arguments last year when analysts said GDP growth was overstated because power generation growth was a lot lower.  

Wu said an economy growing at 8.5 percent a year does not produce data such as the 16.5 percent industrial output growth recorded in the first three quarters of this year.

He also said tax revenue growth, at 21.7 percent in the first nine months, and power generation growth of 15.6 percent were also out of kilter with the National Bureau of Statistics’s top forecast.

Wu was soon proved right that GDP was understated, though the logic was quite different. In December 2004, China did its first national economic census and found out that actual national GDP in 2004 was 16.8% higher than the previous calculation. Bloomberg reported:

A yearlong census revealed millions of companies previously unaccounted for, with combined output valued at about $284 billion in 2004, the National Bureau of Statistics said today in Beijing. Most of those companies are in service industries, which have been growing faster than manufacturing.

Since then, every economic census ended up discovering that overall GDP is a lot higher because service sector is a lot higher than imagined. This is plausible, as every country has a so-called “underground economy” which largely consists of service industries. If China does another census tomorrow I would bet that GDP would again be higher than the number announced previously.   

The latest and perhaps most persuasive evidence that China’s GDP is unreliable comes from comments of the current Premier Li Keqiang. The Economist made a Keqiang index in 2010, saying that

IF CHINA’S deputy Prime Minister, Li Keqiang, succeeds his boss, Wen Jiabao, in 2013, as is likely, he will become his country’s top economic policymaker. But he may not pay much heed to the figures provincial officials feed him. In 2007 he told America’s ambassador that GDP figures in Liaoning, where he was then party chief, were “man-made” and unreliable, according to a State Department memo released by WikiLeaks.

In Mr Li’s honour, The Economist has created a “Keqiang index” for China’s economy, combining his three preferred indicators (see chart). It reveals an economy that is as dynamic as the official figures suggest, but a great deal more volatile. Electricity consumption and cargo traffic both shrank in the final months of 2008 and in early 2009, implying that China’s economy suffered more grievously than the official figures allow. A loan surge in 2009 presaged the rapid recovery that followed.

Indeed, the volatility matters much more than the actual size. It is not even important at all to argue whether GDP is overstated by $1 trillion because property’s weight is too small in the inflation basket, or it is understated by hundreds of billions because many services industries or new technology sectors are not counted in. Even the growth does not matter too much now and what matters is the quality.

20 years after the party set GDP growth target at 8%, which was a big progress by then, do we still need GDP target now? I doubt. GDP target can be very unrealistic and time-lagging that local governments will waste many resources simply to make their boss happy. With soft budget control and distorted pricing in the financial system, China’s local parts can invest endlessly which could make GDP growth extremely high in short term, but in long term the country (especially households) will have to pay for the imbalance caused by inefficient investments in the future. Needless to say the numbers can be cooked. Many argue that China needs high GDP growth to absorb excess labor, but without unemployment data we do not even know whether it is worth polluting the environment to keep GDP growth high. It is said that China started to collect unemployment data since 2005, so why not publish it? We need more data to make the right judgment and it is time to make another progress now. 

Central Banking Seminar

Chen Long