Yearbook

The Most Fundamental Puzzles in China's Labor Yearbook 

The first table in China Labour Statistical Yearbook (between 1999 and 2010) has posed two fundamental puzzles: What is the difference between "Unit Employment"(单位就业人员)  and "Staff and Workers"(职工)?  What is the composition of "Urban Employment"(城镇就业人员)?

Firstly, "staff and workers"(职工) applies to those who work in government or public institutions or state-own-enterprise. People who work in private enterprises, or rural areas are not in this catalogue. Therefore,  "staff and workers" means "urban staff and workers" implicitly. Similarly, the unit employment is the urban unit employment  (there is no "rural unit employment"), as shown explicitly in the yearbook 1999 and earlier, and the yearbook 2014 and later. For unknown reasons, it is not stated clearly between yearbook 2000 and 2013. "unit"(单位) typically means non-private sector. Therefore, "staff and workers" and "unit employment" is the same.

So now let's go back to our initial question. To understand the difference, we turn to the last page of the yearbooks for term explanation. Terms are explained in both Chinese and English. However, only "Staff and Workers" is explained in the English version. In the Chinese version, both terms are explained. But don't feel unfair. The explanation is so blurred that even a Chinese couldn't understand.

In short, "Unit Employment"(单位就业人员)  equals "Number of Staff and Workers"(在岗职工) plus other employment of the state or government. The sum of "Unit Employment"(单位就业人员) and "Employment in Urban Private Enterprises and Individuals"(城镇私营和个体就业人员) SHOULD equal "Urban Employment"(城镇就业人员). These relations are presented in the graph below.

To see these, we need to go back to early yearbooks. Before the yearbook 1995, there is only two components under the urban employment: "urban staff and workers" and  "employment in urban private enterprises and individuals" and they add up as the urban employment.

However, in the yearbook 1994, these two figures begin not to add up. To solve this problem, the NBS (National Bureau of Statistics of China) begin to introduce "others" in the "staff and workers" catalogue. This method makes figures add up for two years. In the yearbook 1997, the figures don't add up again and the gap is widening. Therefore, in the yearbook 1999, the NBS gave up. They simply combine the "others" with "staff and workers" to form "unit employment". 

I guess NBS can measure "staff and workers" with good accuracy. However, as China is reforming, a lot of previous SOEs are becoming hard to define whether they are public or not. Therefore NBS's measure of "others" is increasingly off (the measure of the private sector could be also hard). As a result, the sum of "unit employment" and "private sector" is also increasing off from the total urban employment. In 2003, this sum only equals 60% of the total urban employment. Finally, in the yearbook 2010, NBS stops to report  "staff and workers" and only reports "unit employment".

There is also good news. Since 2003, the NBS has improved its measurement. Recently, the sum between "unit employment" and "private sector" can account for 95% of the urban employment.

Employment in Manufactural Sectors

In China's statistical system, all sectors in the tertiary sector are regarded as major sectors. Therefore, employment data on the education, science, and health sector are relatively abundant. On the other hand, all manufactural sectors in the secondary sector are regarded as minor sectors and often reported as an aggregated "Manufactural Sector". Therefore, employment data in the textile, wood, and machinery sector are much harder to find.

There are two resources of employment data in yearbooks. The labor yearbook reports the "staff and works" of these sub-sectors. During the economic reform, this type of labor is declining over the years. The industry yearbook reports the "annual average employment" which includes on-contract employers and describes the real situation better. However, the employment is for “SOE and other enterprises above designated size”.  This category is not bounded by urban or state-related——Employers in this category could be either urban or rural. Employers in large private firms are counted. On the other hand, not all employers in state-related firms are counted. Only those in SOE are counted.

Besides, in some years, this data is not available and need to be calculated from the labor productivity (I some years, both the employment and the labor productivity are reported and they are consistent). 

Finally, various attempts have been made to adjust the employment or estimate the total employment. One of them is by 陈诗一(Chen). In previous research, I have extracted his data into an Excel form. The figure below shows the employment for the textile industry to illustrate the difference between these data.

The Curious Case of GDP of Beijing 

Due to the change in definitions of statistics, NBS will adjust historical GDP over time. However, such adjustments were only available at the national or provincial level. GDP at the prefecture or county level was not updated in most cases. Therefore, it is common to spot kinks when dealing with county-level data.

Take Beijing as an example. The GDP of Dongcheng district has two obvious kinks: 2000/2001 and 2004/2005. The first kink does not reflect in the prefecture-level GDP (total GDP of Beijing). However, the sum of all districts and counties does not match the aggregated GDP by a lot.

Why is that? One speculation is that some GDPs are not assigned to districts before 2001. For example, the GDP of a state-owned enterprise in the Dongcheng district may not be counted as the local government has no control over the company. However, it is GDP is counted at the prefecture level.

After 2001, these missing GDPs has been counted for each district, creating a huge jump between 2000 and 2001 but also making the sum of districts and counties closer to the aggregated value.

The second kink between 2004 and 2005 is due to the first national economic census in 2004. It resulted in a lot of kinks at county-level GDP and most local governments did not release revised historical data.