The Chinese statistical authorities recently revised the Chinese GDP level and real growth rate for the period 1993–2004 following China's first national economic census for 2004. However, the methodology used in their revision is opaque. Using a trend- deviation interpolation approach, this study has managed to replicate the basic procedures of the revision and reproduced the official estimates. Through this exercise, we have found that the estimates that could be obtained by the straightforward interpolation procedures were significantly modified. Based on a political economy argument, we attempt to explain why the revision had to leave the growth rate of 1998 intact and why it had to bypass the price issue and directly work on the real growth rate revision. Based on previous studies and other observations, we also question the census results on non-service industries.