The Wall Street Journal
January 24, 2006

Chinese statistics still suffer from two shortcomings: technical difficulties and political manipulation, writes economics professor Thomas G. Rawski.

We all now know what mischief lurks in the depths of accounting. In countries, as in companies, close examination may uncover pervasive inconsistencies. Critics are now swarming over allegations that Chinese official statistics have significantly exaggerated recent economic performance. In fact its growth rate since 1997 is probably less than half the official figure of 7.6%. The tip-off that there may be something fishy about China’s official statistics is that they don’t match the experience of other Asian economies. In Japan in the late 1950s, Taiwan in the late 1960s, Korea in the late 1970s and China itself in the late 1980s, cumulative growth in real output exceeded 30%. Each expansion sparked higher energy consumption, increased employment and rising prices. But during the 1997-2001 period, China’s energy use, employment and prices all fell. So how how could real output have grown by one-third, as Chinese officials claim? Unless China has embarked upon an unprecedented economic trajectory that combines high-speed growth with rising unemployment, sluggish demand, massive excess capacity, glutted commodity markets, mild deflation and low expectations, Beijing’s math is fuzzy. In Beijing’s defense, it should be said that its economic reporting has vastly improved over the past 20 years. The National Bureau of Statistics has replaced Soviet methods of accounting with a system modeled on international practice. But Chinese statistics still suffer from two shortcomings: technical difficulties and political manipulation. The technical problems are unavoidable in any large, low-income economy undergoing structural change. And in China’s case, cuts in statistical personnel and a dynamic but hard-to-measure private business sector – whose omission injects a downward bias into all GDP figures – has also contributed to making its economic statistics less reliable.