When the total employment can't continue to grow, the reallocation of employment among industries will become an important driver of economic growth. In China, any single non-specialized employment statistic can only satisfy two of the three key characteristics—industry segmentation, period continuity and scientific method—resulting in an imperfect data foundation for studying employment allocation. Consequently, scholars have devised compromise methods that cross-complements data from multiple employment statistics to construct subdivided industries employment data fulfilling all three characteristics. Through an analysis of employment statistics, this paper reveals significant discrepancies in survey designs and inconsistent statistical outcomes across employment statistics, and forced cross-complementary can lead to serious accounting biases. Although the employment data of the demographic survey has the disadvantage of being collected every five years, its industry classification is more detailed and the data are more accurate. We integrate the demographic survey data with the input-output tables to reconstruct the employment and economic growth database of 69 subdivided industries in China from 2000 to 2020. Then, using the criterion of whether employment allocation across subdivided industries promotes total factor productivity growth, this paper evaluates the appropriateness of employment allocation. The analysis demonstrates that employment allocation across subdivided industries is generally appropriate in China. The reallocation of labor from agriculture to non-agricultural industries contributes most significantly to enhancing allocation appropriateness, and its contribution is still rising. Recently, the reverse allocation of employment in some non-agricultural industries—where employment "increases when it should decrease" and "decreases when it should increase"—has intensified. Skills mismatch and disparities in different industries' labor income shares are identified as key barriers to improving the appropriateness of employment allocation.

BACKGROUND

When the total employment can't continue to grow, the reallocation of employment among industries will become an important driver of economic growth. In China, any single non-specialized employment statistic can only satisfy two of the three key characteristics—industry segmentation, period continuity and scientific method—resulting in an imperfect data foundation for studying employment allocation. Consequently, scholars have devised compromise methods that cross-complements data from multiple employment statistics to construct subdivided industries employment data fulfilling all three characteristics, but they did not systematically demonstrate the feasibility of those compromise methods. In addition, even if we have high-quality employment data for specific industries, there is a lack of an analytical framework that helps us determine whether the employment structure among industries is reasonable as well as helps us point out the optimization direction of employment allocation among industries.

OBJECTIVE

First, this paper wants to clarify whether different types of employment statistics have the possibility of cross-complementarity. Second, we clarify the potential accounting errors in the existing compromise methods for employment accounting in subdivided industries. Third, this study reconstructs the employment and economic growth database in subdivided industries. Fourth, we propose a framework for analyzing the appropriateness of employment allocation among subdivided industries.

METHODS

Firstly, this paper systematically sorts out the differences among five commonly used employment statistics in terms of statistical objects, frequencies, methods, employment definitions, industry classification, and statistical results, to analyze the possible biases that may arise from the two compromise methods of employment accounting in China's subdivided industries. Secondly, we select the employment statistics with the most comprehensive and accurate employment data in subdivided industries and merge them with the "China Input-Output Table" to reconstruct the employment and economic growth data of China's subdivided industries. Finally, we use the criterion of whether employment allocation across subdivided industries promotes Total Factor Productivity growth to evaluate the appropriateness of employment allocation.

RESULTS

On the one hand, this paper reveals significant differences among five employment statistics in survey designs and inconsistent statistical outcomes, and forced cross-complementary can lead to serious employment accounting biases. If we split the employment of the three major industries from the labor force survey into subdivided industries based on the industrial structure of urban non-private units, we will overestimate the employment scale of industries dominated by formal employment. If we split the employment of the three major industries from the labor force survey into subdivided industries based on the industrial structure of the population census or economic census, the industry structure differences and the contradiction of total employment scale between the labor force survey and unconventional employment statistics will be ignored. On the other hand, the analysis demonstrates that employment allocation across subdivided industries is generally appropriate in China. The reallocation of labor from agriculture to non-agricultural industries contributes most significantly to enhancing the allocation appropriateness, and its contribution is still rising. Recently, the reverse allocation of employment in some non-agricultural industries—where employment "increases when it should decrease" and "decreases when it should increase"—has intensified. Skills mismatch and disparities in different industries' labor income shares are identified as key barriers to improving the appropriateness of employment allocation.

CONCLUSIONS

The research on employment accounting of subdivided industries should be more cautious in using different types of employment statistics, scholars should try to use single-source data for analysis as much as possible to avoid accounting errors caused by cross-complementarity of multiple employment statistics. In addition, there are three ways to improve the appropriateness of employment allocation in subdivided industries: First, it’s important to keep the main role of the employment reallocation from agriculture to non-agricultural industries and implement differentiated employment allocation optimization plans in agricultural subdivided industries Second, eliminating the skills mismatch problems faced by employees migrate from industries whose employment should be reduced to industries whose employment should be increased. Third, increasing the share of labor income in industries whose employment should increase to help alleviate the problem of the declining appropriateness of employment allocation in non-agricultural industries.

CONTRIBUTION

In terms of data comparison, this paper systematically compared the characteristics and differences of five employment statistics for the first time, and we clearly pointed out that the forced cross-complementary of employment statistics to construct employment data of subdivided industries would cause serious biases. In terms of data reconstruction, although the employment data of the demographic survey has the disadvantage of being collected every five years, its industry classification is more detailed, and the data is more accurate. We integrate the demographic survey data with the input-output tables to reconstruct the employment and economic growth database of 69 subdivided industries in China from 2000 to 2020. In terms of data application, this paper empirically evaluates the appropriateness of employment allocation in China’s 69 subdivided industries, and we analyze the reasons for the inappropriateness of employment allocation among subdivided industries and propose policy recommendations for improvement.