The evolution of China's state sector

Since the 1990s, the Chinese public sector has shrunk considerably. In 1992, during his southern tour, Deng Xiaoping 邓小平 justified the development of a private sector in China, paving the way for the emergence of private enterprises. Private enterprises developed mainly in two ways: when individuals from the public sector started their own businesses (known as xiahai 下海, or “jumping into the sea”), or when entire state or collectively owned units were privatized. In 1997, the 15th National People's Congress officially proposed the privatization of SOEs, hastening the shift towards private enterprise. From 1998 to 2016, the number of state-owned and state-controlled industrial enterprises declined at a rate of 6.6 per cent per year, while the number of non-state industrial enterprises increased by 7.3 per cent per year.Footnote ¹⁸ By 2016, state-owned and state-controlled industrial enterprises accounted for only 17.1 per cent of industrial profits, a significant decrease from 47.3 per cent in 1997.

The decline of the public sector is also evident in the realm of employment, as illustrated in Figure 1. The proportion of urban employment in state-owned units drastically decreased from 76.8 per cent in 1980 to 33.3 per cent in 2018.Footnote ¹⁹ Figure 2 displays the breakdown of total employed workers by employer type.Footnote ²⁰ Out of the 776 million employed workers in 2018, 56 per cent were based in urban areas. Among urban employees, only 13.2 per cent were employed in the public sector.

Figure 1. The Share of Employment in the State Sector

Source: Authors’ calculations using China Statistical Abstract 2018; the Urban Household Survey 1988–2009; China Health and Nutrition Survey 2006; 2009; 2011; 2015; China Family Panel Studies 2010; 2012; 2014; 2016; and Chinese General Social Survey 2008; 2010; 2013.

Figure 2. China's Employed Persons by Sectors in 2018

Source: Authors’ calculations using China Population & Employment Statistics Yearbook 2019. There are 775.9 million workers in total.

As the public sector workforce shrank, the Chinese government and SOEs became more selective in their recruitment of entry-level staff. In the 1980s, the Chinese government acknowledged the necessity of cadres with the right human capital to grow the economy. To address this, Deng Xiaoping introduced reforms that mandated the recruitment of younger, educated and professional individuals into the system, leading to an increased demand for highly educated employees.Footnote ²¹ Reforms in China's SOEs in the 1980s and 1990s brought in modern corporate governance and incentive systems, contributing to rising profitability.Footnote ²² The reforms of the 1990s resulted in the shutdown or privatization of most unprofitable small- and medium-sized SOEs and the restructuring of the remaining large SOEs, raising the relative demand for highly educated workers and increasing the appeal of SOE jobs.Footnote ²³

College-educated workers

In China, high school students are required to take the National College Entrance Examination, also known as the gaokao 高考, in order to be admitted to college. The scores needed to gain entry to better or elite colleges are higher.Footnote ²⁴ This paper employs a commonly accepted definition of elite colleges in China, specifically those that are part of the 211 Project, which comprises around 100 leading universities in the country. Additionally, nearly all college students are required to take the College English Test (CET) prior to graduation.

Since the post-reform era, college graduates have been free to choose their employers. However, data indicate that a significant portion of college-educated workers have continued to be employed in the state sector (Figure 1). We use various data sources to estimate the percentages of college-educated workers in the state sector between 1988 and 2009 (see the figure notes for details). As depicted in Figure 1, college-educated workers were heavily concentrated in the state sector until the late 1990s, and although the proportion has decreased in recent years, as of 2016, 53.7 per cent of college-educated workers were still working in the state sector.

While the state sector has significantly decreased as a percentage of the overall labour force, it still employs a substantial number of college-educated workers. As a result, we can anticipate that a considerable state sector job market for college graduates will persist in the future, and it is important to examine who desires and secures these positions.

The Chinese College Student Survey

This study utilizes data from the Chinese College Student Survey (CCSS), which was conducted annually by the China Data Center at Tsinghua University between 2010 and 2015. One of the authors directed the survey. The CCSS sample was constructed by randomly selecting 100 colleges from a list of 2,300 colleges in China. The selection was stratified according to five geographical regions (North-East China, East China, Central China, West China, and Beijing/Shanghai/Tianjin) and seven tiers of colleges. Each college was weighted by its student population, making the sample representative of the population of all college students in China.Footnote ²⁵

To carry out the surveys, we selected two to three survey administrators for each college. These administrators were typically responsible for registering students, teaching or managing student affairs. To control the quality of the survey, we held several days of training sessions in Beijing for all survey administrators. Returning to their own colleges, the survey administrators selected a random sample of students. They then gathered all sample students in one location and asked them to individually and anonymously fill in their questionnaires. The completed questionnaires were then coded and mailed to Beijing for data entry and cleaning to guarantee consistency and quality.

We conducted the survey in May or June of each year, just before students graduated from college, when many had already begun their job search or received job offers. The questionnaires covered a wide range of variables, including socioeconomic characteristics (individual and family background) and educational background (for example, student ability, school performance and college quality), as well as variables related to the students’ future employment, such as job search behaviour, the characteristics of their desired job and the characteristics of their best job offer.Footnote ²⁶

In our sample of 40,911 students,Footnote ²⁷ 67.6 per cent planned to start working after graduationFootnote ²⁸ and 65.1 per cent reported actively job searching. Among the job seekers, 95.2 per cent provided information on their desired jobs (25,346 students, 62 per cent of the total sample), while 72 per cent provided information on their best job offers (19,176 students, 46.9 per cent of the total sample). The students who had received job offers were also asked to provide details such as the job type, location, employer size and compensation package.

In our survey, we directly asked students to report their preferred ownership type and actual ownership type of their potential employer. We asked the following questions: “What type of work do you aspire to?” and “What type of work does your best job offer belong to?” Students were provided with the following answer choices:

1. 1) Party and government offices;

2. 2) Public organizations or social groups such as the Union, the Communist Youth League, the Women's Federation;

3. 3) Public institutions (shiye danwei 事业单位);

4. 4) State-owned enterprises;

5. 5) Collective enterprises;

6. 6) Foreign enterprises;

7. 7) Private enterprises;

8. 8) Self-employment;

9. 9) Others (please specify).Footnote ²⁹

Variables

Table 1 presents a summary of our sample. In Panel A, we observe that 55.6 per cent of the students were male, 46.5 per cent had urban hukou prior to college, and 30.8 per cent were Party members.

Table 1. Descriptive Statistics

Source: Authors’ calculations using CCSS.

Notes: State sector includes the Party and government offices, public institutions and SOEs. “EFML” refers to economics, finance, management and law; “science major” includes science, agriculture and medical science; “social sciences/arts major” includes philosophy, education, literature, history and the arts; “engineering major” includes engineering and military science.

Panel B in Table 1 displays information on the educational background of our sample students. Of these students, 41.8 per cent attended elite universities. The most popular major was engineering, with 41.4 per cent of students, followed by EFML (economics, finance, management and law), science (science, agriculture and medical science), and social sciences/arts (philosophy, education, literature, history and the arts).Footnote ³⁰ Our university sampling method resulted in a proportional distribution of students across regions.

We collected several measures of student academic quality. The first measure is the students’ standardized gaokao scores, which range from -9.9 to 11.2, with an average score of 0.4.Footnote ³¹ The second measure is the Grade Point Average (GPA) earned in college coursework, a standardized score ranging from 0 to 4. The average GPA in our sample is 3.1. The third measure is a student's rank in class, which we operationalize using a dummy variable called “Top 20 per cent.” This variable is equal to 1 if students reported that their academic achievement was in the top 20 per cent of the class and 0 otherwise; 46 per cent of students reported that they were ranked within the top 20 per cent of their class. The fourth measure is the student's English language ability, as measured by the CET. It is important to note that all four academic measures are self-reported and may be subject to non-classical measurement error or non-random recall bias, potentially leading to noisy and/or biased estimates.

In Panel C, we present a summary of the family background characteristics of the sample students. We found that 18.6 per cent of the students had at least one parent who held a job with administrative rank such as a civil servant or bureaucrat; 25.4 per cent of the students had at least one parent who completed college; and the average annual household income of the students was 81,609 yuan (USD 13,000 in 2015 prices). Moreover, 65.9 per cent of the students reported having at least one relative or close friend who worked in the state sector.

Strong, consistent preference for state sector jobs

Despite China's state sector reforms over the past three decades, Chinese college students still exhibit a strong preference for state sector jobs. As presented in Table 2, the percentage of college students who listed their ideal job as being in the state sector was consistently above 60 per cent between 2010 and 2015. Further analysis reveals that of those students who preferred state sector jobs, 18 per cent preferred jobs in Party and government offices, 26 per cent preferred jobs at SOEs and 20 per cent desired jobs at public organizations over the six-year period.

Table 2. Students’ Ideal Jobs by Sector (%)

Source: Authors’ calculations using CCSS.

Despite the Chinese government's recent anti-corruption campaign, which began in 2012, Chinese college students have maintained a consistent preference for state sector jobs, which is surprising. The campaign resulted in significant restrictions on working and living benefits (for example, housing and transportation) for state sector employees. This could have potentially reduced the appeal of pursuing state sector careers among college students. However, our data reveal that, despite some job preference reshuffling in the initial aftermath, interest in employment within the state sector remained consistently high after peaking in 2012. Between 2012 and 2013, interest in Party and government office jobs and SOE jobs decreased by 3 and 4 per cent respectively, while interest in public institutions increased by over 5 per cent. However, in the years following 2013, interest in these three categories gradually returned to levels similar to those before the campaign.

Variations across student background

Table 3 presents an analysis of how students’ preferences for state sector jobs vary according to their observable characteristics. Despite differences in these characteristics, we find that the majority of students consistently preferred state sector employment, with over 60 per cent across all categories. To test for differences in preferences between groups, we use t-tests and set the null hypothesis as the absence of a difference in means between the groups.

Table 3. Percentages of Ideal Jobs in the State Sector (by Sample Student Characteristics)

Source: Authors’ calculations using CCSS.

Notes: Gaokao scores are standardized by province, year and educational track (arts or sciences). CET score measures the English ability of undergraduate students. A t-test is used to determine whether the means of two groups are equal to each other. The null hypothesis is that the difference in group means is zero. * significant at 10% level; ** significant at 5% level; *** significant at 1% level.

Students’ personal characteristics played a role in their job sector preferences. Female students were more likely than males to prefer a career in the state sector, with a difference of 7 percentage points. This may be because a larger proportion of female students desired to work in public institutions compared to male students. Surprisingly, students’ place of upbringing (urban versus rural) did not significantly affect their job preferences. Both students who grew up in rural areas (i.e. those with rural hukou before going to college) and students raised in urban environments (i.e. holding urban hukou) had a similar desire for a job in the state sector (64 per cent for both groups). Our t-test analysis does not find a statistically significant difference in job preferences between students raised in rural versus urban areas.

Party members, on average, preferred state sector jobs more than non-members. Joining the CCP is a costly and time-consuming process, and Party membership may indicate a desire to participate in the Chinese political system or the state sector more broadly. Furthermore, Party membership may be useful in obtaining a state sector job, and Party members may be more exposed to information and experiences that might encourage them to pursue a state sector career. However, over 61 per cent of non-members also desired a career in the state sector, indicating that this trend was not limited to Party members.

The state sector was appealing to students of different academic abilities – we do not observe significant differences in state sector job preferences based on students’ academic quality. The proportion of students at elite and non-elite colleges who desired state sector employment was almost the same. Similarly, there is no significant difference in preferences for state sector jobs between students with above- and below-median gaokao scores (63 versus 64 per cent) or between students with above- and below-median GPAs (64 versus 63 per cent). Students in the top 20 per cent of their class were more likely to prefer a career in the state sector by four percentage points (66 versus 62 per cent). However, students with above-median CET scores showed a lower preference for state sector jobs, although this is likely owing to their desire to work for foreign enterprises (21 to 13 per cent). Although there are variations in preferences, the state sector was still attractive to students with varying levels of academic ability.

Interestingly, however, student preferences vary with parental political capital and socioeconomic status. Children of Party cadres were 4 per cent more likely to prefer a state sector job and nearly 10 percentage points more likely to want to work in Party and government offices than other students (25.0 versus 15.7 per cent). Relatedly, students with social connections in the state sector were 5 percentage points more likely to desire employment in the state sector. Finally, parents’ socioeconomic status mattered, while their college attendance did not.

Household income makes a difference in students’ preference for state sector jobs. Students from below-median income households were 4 per cent more likely to prefer a state sector job than their wealthier counterparts. This figure is mainly a result of the difference in the proportion of students who aimed to work at SOEs, which was 28.4 per cent for lower-income students compared to 24.6 per cent for higher-income students. SOE jobs are perceived as very stable, low-risk and high-paying, which may be particularly attractive to lower-income students.Footnote ³² In contrast, graduates with more financial flexibility may be more inclined to pursue riskier ventures or switch between private sector jobs early in their careers.

Student preferences vary significantly between different majors. History (78.6 per cent), law (78.5 per cent) and medical science (77.3 per cent) students exhibit the strongest preferences for state sector jobs.Footnote ³³ On the opposite end of the spectrum, literature (62.0 per cent), management (61.2 per cent) and arts (57.8 per cent) students expressed the least interest in state sector employment. Still, the primary takeaway is that technical and non-technical majors, preprofessional and non-preprofessional majors – that is, the majority of students in all majors sampled – sought a state sector job.

Reasons for preferring state jobs

Our results suggest that most Chinese students preferred to work in the state sector; the question is why? To answer this, we compare the wages and benefits (such as pensions, health insurance and housing subsidies) of state and private jobs.

Based on our survey data, it appears that entry-level positions for college graduates in the state sector offered higher pay than their counterparts in the private sector. Table 4 displays the mean wages by sector from 2010 to 2015, and for each year, a t-test result demonstrates that the average monthly wages offered by the state sector are greater than those provided by the private sector. Specifically, in 2015, the average monthly wage in the state sector was over 12 per cent higher than that in the private sector. These findings align with previous research indicating that China's state sector tends to offer higher wages.

Table 4. Monthly Wages and Benefits in State and Non-state Sectors

Source: Authors’ calculations using CCSS.

Notes: Standard deviations are in parentheses. A t-test is used to determine whether the means of two groups are equal to each other. The null hypothesis is that the difference in group means is zero. * significant at 10% level; ** significant at 5% level; *** significant at 1% level.

In line with prior research, our analysis utilizing a series of t-tests indicates that jobs in the state sector generally offer more extensive benefits compared to those in the private sector, as presented in Table 4.Footnote ³⁴ Specifically, our findings reveal that state sector jobs were more inclined to provide a wider offering of benefits, including local urban hukou, pensions, various types of insurance and housing benefits, in comparison to private sector jobs.

In addition to considering compensation and benefits, students may also take into account the implicit benefits associated with various jobs. This hypothesis is supported by the existing literature, which suggests that state jobs are often associated with greater job security and higher levels of prestige (as discussed above).

The role of the state in employing college students

The market for entry-level state jobs is highly competitive. Our data show that 64 per cent of college graduates aimed to secure a job in the state sector, but only 42 per cent were successful in doing so (as illustrated in Table 5, Panel A). Furthermore, even among students who expressed a preference for state sector jobs, only half were ultimately able to secure employment in this sector (as presented in Table 5, Panel B).

Table 5. Students’ Best Job Offers by Sector (%)

Source: Authors’ calculations using CCSS.

Steady preferences for state sector jobs, combined with a declining labour demand in the state sector, have intensified competition for state employment. Our data show a decreasing trend in labour demand for state sector jobs, as illustrated in Panel A of Table 5. From 2010 to 2012, the proportion of college students who received a job offer from the state sector remained stable, at approximately 45 per cent. Of these, around one-third secured jobs in SOEs, while 12 per cent received offers from government and public institutions. However, the proportion began to decline in 2013, and in 2014, the proportion of college graduates who obtained state sector jobs decreased by 11 percentage points compared to the previous year. This significant drop is primarily attributed to the reduction in the employment share of public institutions (from 12.4 per cent to 6.3 per cent) and SOEs (from 26.6 per cent to 21.9 per cent). The proportion in Party and government offices decreased as well. Eventually, the share of total students who received a job offer from the state sector remained less than one-third in 2015.

In contrast, the proportion of college students who were self-employed increased by two-thirds, rising from 6.8 per cent in 2013 to 11.0 per cent in 2014. This trend continued in 2015, with the proportion remaining at 9.0 per cent. This increase was likely owing to the government policy implemented in 2014, which aimed to encourage college graduates to start their own businesses. While the share of employment in foreign enterprises declined over time, there was a significant increase in the proportion of students who received job offers from domestic private enterprises. In fact, private enterprises provided more than half of all job offers to college graduates in 2015 (as depicted in Table 5, Panel A).

The data also indicate a decreasing trend in the number of state sector jobs received by graduates who preferred to work in that sector. Panel B of Table 5 reveals that only 51 per cent of students who wanted to find a job in the state sector were able to obtain an offer. Moreover, since 2014, the share of employment in the state sector has fallen below 40 per cent. The next section will examine the factors that influence which of the students who prefer to work in the state sector can ultimately obtain an offer.

Characteristics of college graduates preferred by the state sector

As previously mentioned, there is a significant gap between the percentage of students who desired state sector employment and those who actually obtained it. In order to shed light on this discrepancy, we examine the characteristics of the students who received state sector job offers in Table 6, where t-tests are used to determine the statistical significance of observable differences between those who received offers and those who did not.

Table 6. Percentages of Best Job Offers in the State Sector by Student Characteristics for Those Who Want to Find Jobs in the State Sector

Source: Authors’ calculations using CCSS.

Notes: Gaokao scores are standardized by province, year and educational track (arts or sciences). CET score measures the English ability of undergraduate students. A t-test is used to determine whether the means of two groups are equal to each other. The null hypothesis is that the difference in group means is zero. * significant at 10% level; ** significant at 5% level; *** significant at 1% level.

We observe significant variation in state sector job offers based on individual characteristics. Despite female students’ overall greater preference for state sector employment, male students were significantly more likely to receive a state sector job offer. In particular, 57 per cent of men landed jobs in the state sector, compared to only 45 per cent of women, even though 7 per cent more women listed a state sector preference. Women faced particular difficulty in landing jobs at SOEs compared to men (26 per cent of women versus 44 per cent of men), although they had some advantages in getting jobs in public institutions such as schools and hospitals (16.0 per cent of women verus 9.7 per cent of men). Urban students were more likely to obtain a state sector job than rural students (56 per cent versus 48 per cent). Party members were also more likely to obtain a state sector job (55 per cent versus 50 per cent), with higher employment in SOEs as well as Party and government offices.

Students with stronger academic backgrounds had better chances of receiving state sector job offers compared to their less academically distinguished peers. Specifically, those who attended elite colleges or had higher than median gaokao scores had higher success rates in obtaining a state sector job offer, with 61 per cent and 57 per cent respectively, compared to 46 per cent and 45 per cent for those who did not meet these criteria. Additionally, students with higher CET scores also had higher success rates in obtaining state sector jobs, with 56 per cent compared to 48 per cent for those with lower CET scores. Among academically elite students, there was a significantly higher likelihood of receiving SOE job offers and a significantly lower likelihood of being self-employed or working in private enterprises.

Surprisingly, students with higher GPAs are less likely to land state sector jobs compared to those with lower GPAs. The negative correlation between GPA and job offers from the state sector persists even after controlling for demographics.Footnote ³⁵ This is potentially owing to the GPA being traded off for other career-related activities such as work experience, networking or obtaining a driver's licence. Additionally, students with high gaokao scores fared better in obtaining state jobs than those with high GPAs. This indicates that state sector hiring practices prioritize academic ability, as indicated by gaokao scores, over college performance, as measured by college GPAs. This negative correlation is also observed when using the “Top 20 per cent” to measure academic performance, where students in the top 20 per cent of their class were less likely to receive a job from the state sector.

Parental political capital and socioeconomic status are significant factors in determining who secured state sector jobs. Students with at least one parent who was a Party cadre had a higher chance of receiving a state sector job offer, with 66 per cent of these students receiving offers compared to 50 per cent of those without a cadre parent. Moreover, students with a college-educated parent, high-income families and social connections to the state sector were also more likely to receive state job offers, potentially through family ties.

There are notable disparities among students in various majors regarding their chances of obtaining state sector jobs. Philosophy and law majors were offered the highest percentage of jobs from Party and government offices, while engineering majors received more offers from SOEs, and medical science and history majors received more offers from public institutions. Agriculture majors were offered the smallest number of state sector jobs. Despite being the students least interested in joining the state sector, art majors (58 per cent expressed preference, 34 per cent received offers) had relatively high employment rates in state sector jobs.

To test the statistical significance of the factors that influence state sector job offers, we employed formal regression models instead of simple t-tests. Our OLS regressions employed the indicator for successfully obtaining a state sector job as the dependent variable, with control variables such as college major, job preference, year and location, in addition to the key factors of interest that were identified in the t-tests of Table 6. Our regression results (not shown) confirm the findings of our prior analysis, indicating that males, urban hukou holders, CCP members, strong standardized test-takers, elite university students and students with higher household income or more elite parental backgrounds were more likely to receive job offers in the state sector.