Accounting for China's Growth in 1952-2008: China's growth performance debate revisited with a newly constructed data set

Author Name Harry X. WU  (Institute of Economic Research, Hitotsubashi University / The Conference Board China Center)
Creation Date/NO. January 2011 11-E-003
Research Project Productivity of Industries and Firms and Japanese Economic Growth
Download / Links


Using a "data fundamentalist approach," this study revisits the long debate about China's growth performance by seriously tackling the existing data problems that have been the major obstacles to a proper assessment of China's growth performance. First, this study examines and adjusts the serious break in the official employment statistics in 1990. Second, it provides an adjustment for the numbers employed by a human capital effect. Third, it tests the sensitivity of Maddison's (1998a) "zero labor productivity growth" assumption in gauging the real growth of the so-called "non-material (including non-market) services." Fourth, it further improves the author's earlier physical output-based production index for the industrial sector (Wu, 2002a) by using multiple weights and time-variant value added ratios obtained from the Chinese input-output tables. The likely problem of "product quality" in such a physical measure is examined and rejected. Fifth, it provides a new set of estimates of capital stock for the aggregate economy using alternative deflators and depreciation rates, crosschecked by the author's industry-level capital stock estimates (Wu, 2008b). This completely new data set is used in a Solow-type growth accounting exercise with different factor income share assumptions. The new results—under the full adjustment scenario for the post-reform period using input-output table income weights—show that the estimated annual TFP growth rate is 0.3 percent, which is substantially lower than the estimate of 3.1 percent derived from the official data without any major adjustment. A range of TFP estimates is also provided for each sub-period under different assumptions.