The “urban-and-TVE” series in chart 2 and table 4 suggests that China had 109 million manufacturing employees by yearend 2002, whereas the officially compiled series reported 83 million, a difference of 26 million. Which series is correct or, at least, more nearly correct?
On the one hand, if the yearend-2000 totals from the two series in the chart and the table are compared with 2000 census-based estimates of manufacturing employment, it becomes evident that the census found about 8 million more manufacturing workers in China than the NBS-Labor Ministry compilation that year (see table 3), but 20 million fewer than the TVE and urban manufacturing employment total for that date. The census number is closer to the official series, which would argue in favor of that series.
On the other hand, the urban-and-TVE series agrees with the results of the 1995 industrial census, which counted 147 million workers in industry (gongye) nationwide. China’s official administrative data series for the whole country (not only for TVE’s) reported that, in 1995, 89 percent of all of China’s “industry” employees were in manufacturing. Applying this percentage to the industry employment data from the 1995 industrial census would translate into 131 million manufacturing workers nationwide that year.45 The total of urban and TVE manufacturing employment for yearend 1995 was 128 million. Therefore, evidence from the 1995 industrial census of China argues for using the urban and TVE series to estimate total manufacturing employment in China. 46
More information is needed to determine which series is more nearly correct for total manufacturing employment in China during 1990-2002. Tentatively, it would appear that the urban and TVE series on manufacturing employment would be more useful than the official series. One reason is that the NBS and the Labor Ministry have little usable information associated with their figure of 45 million “rural” manufacturing workers. Nothing is published except that one number, and no information is available on how this number was derived or estimated. By contrast, a branch of the Agriculture Ministry has gathered and published some information on the TVE manufacturing employees and on the earnings paid to them. The only regular reporting system that collects detailed data for rural enterprises, including manufacturing enterprises, is the data collection system run by the Township Enterprise Bureau of the Ministry of Agriculture:
The Township Enterprise Bureau of the Agriculture Ministry collects detailed data on township-run and village-run collective enterprises, and basic statistics on co-operative enterprises, private enterprises and individual-owned enterprises. The NBS assists the Township Enterprise Bureau in the design of the statistical reporting forms. The Township Enterprise Bureau collects the data and makes some of them regularly available to the NBS. The NBS relies on the Township Enterprise Bureau’s data and has no independent regular reporting system for rural enterprises. 47
If the urban and TVE series is the more nearly correct one, then the 2000 census long form may have underestimated the true number of manufacturing workers in China on November 1, 2000, by about 20 million (108,000,000 – 88,000,000). (See tables 3 and 4.) One study has pointed out that the census questionnaire and the instructions for enumerators may tend to bias the responses of rural households in favor of reporting all members as engaged in agricultural employment, even if a family member is working in