Image Source: George Pak

When policies reshape industries, not all majors can adapt. Field-of-study frictions can shape unequal graduate outcomes in China.

By: Emily Liu

Edited By: Hiya Kumar

Since 2000, China has seen a substantial rise in the number of new graduates from college. According to official reports by the Chinese government, “China rolls out sweeping employment push for record number of college graduates,” the number of college graduates in the country was about 950,000 in 2000, reached 5,754,000 in 2000, and by 2022, the number exceeded 10 million for the first time. By 2025, this number is expected to reach 12.22 million. Moreover, from 2021 to August 2025, China created 59.21 million urban jobs in total.

Looking at the macroeconomic data, China’s job market does not appear to be showing weakness, nor is there a huge difference between the supply and demand sides of the labor market. However, new graduates are experiencing tremendous employment pressure, while the labor market shows variations in employment conditions for different majors. In other words, when too much focus is on the macro-aggregate labor market equilibrium, the demand for labor in individual industries and the supply of specialized graduates or workers that correspond to it are often overlooked.

Although public data about the number of new jobs created in a specific industry is difficult to obtain, we can use the employment ratio to estimate the labor market conditions of the industries corresponding to different majors. This data comes from the 2023 Blue Book of China’s Undergraduate Employment Report published by MyCos. In this report, the employment ratio is calculated as the number of undergraduate graduates employed in a particular type of industry divided by the total number of undergraduates employed in the same period nationwide. The report takes into account the employment rates of fifty majors ranging from the class of 2018 to the class of 2022, and this article will use the representative majors in social sciences, STEM (Science, Tech, Engineering, and Math) and business, respectively, to organize and analyze the trend graphs. The results of major industries of employment for undergraduate graduates compiled from 2018 to 2022 are shown in Table I, and the trend graph in Figure I.

TABLE I: Undergraduate graduates compiled from 2020 to 2024

Category	Representative industry	Class of 2024	Class of 2023	Class of 2022	Class of 2021	Class of 2020	The change in five years	Representatives of corresponding majors
Social Science	Education	13.9	13.6	13.0	14.0	17.0	-3.1	Education, Psychology, Language, Sociology
	Government and Public Administration	6.3	6.7	6.8	6.4	6.2	0.1	Political Science, Public Administration, Social Work, Law
STEM (Science, Tech, Engineering, Math)	Electronic &Electrical Equipment Manufacturing	6.6	7.0	7.7	8.6	9.0	-2.3	Electrical engineering, Automation
	Machinery & Equipment Manufacturing	3.6	3.1	2.9	2.6	2.5	1.1	Mechanical Engineering, Materials Engineering, Manufacturing
Business & Economics	Finance	7.6	7.4	7.3	7.2	7.5	0.1	Economics, Finance, Accounting
	Professional Consulting & Design Services	4.6	4.9	5.0	5.3	5.1	-0.5	Business Consulting, Marketing

As shown in Figure I, we can see that the percentage of undergraduates employed in the education sector has experienced a very significant decline, and although it shows a slow rise after 2022, it will be difficult to return to the high level of 2020. The year 2021 saw the implementation of the Double Reduction Policy, which led to a crackdown on out-of-school tutoring classes focused on primary and secondary school students, in an attempt to reduce the pressure on primary and secondary school students to take classes outside of school. An implementation of an exogenous policy like the Double Reduction Policy at this time in the education industry has not just short-term, but long-term impact.

Non-engaged teachers facing fewer employment avenues as a result of the Double Reduction Policy have turned their attention to the government and public administration fields, which are also open for employment in the social science sector.

From 2020 to 2022, the Education industry employment share continues to decline, while the Government & Public Administration’s employment numbers show an upward trend during the same period. After 2022, employment in the education sector shows a slight increase, while the Government & Public Administration’s employment share shows a slight decrease during the same period. This isn’t a simple coincidence, rather a result of the complementary nature of industries within the same broad category. In other words, when an industry within a broad category shrinks or expands due to a policy or other external shock, it leads to a flow of talent. The same trend of talent flow occurs in the financial and economics job market. We can see that even though there is no significant change in the employment numbers in the Finance sector and the Professional Consulting & Design Services, they are still complementary. The same situation is true for the two representative industries for STEM.

We are assuming here that talent in the same broad category is more likely to choose other industries in the same broad category instead of other industries in a different broad category. For example, a person who majors in Chinese language in college is more likely to choose to work in the government and public organizations after the implementation of the Double Reduction Policy, rather than entering the financial industry to do data analysis. It is worth noting that, however, even if an industry possesses the ability to absorb talent flowing out from another industry within the same broad category, this ability does not guarantee that every new graduate who abandons their professionally relevant industry due to exogenous policy shocks will be able to successfully enter other industries in the same broad category or in a different category.

To test this assumption, the article sums the five-year rates of change in employment numbers for all the social science majors listed in the Blue Book to see if the results were equal to or close to zero.

The result is that the combined five-year rate of change in employment numbers for all social science-related majors is -2.5%, well below the ideal level of 0%. The five-year rate of change in employment share for all STEM-related majors combined is -1.3%, which is slightly better compared to the social sciences, but still far from ideal. The broad category closest to the ideal rate is the Business and Economics sector, where the combined five-year change in employment share for all Business and Economics-related majors is only 0.1%.

We can find that the interconversion rate of talents among industries in the broad category of social sciences is lower than that of industries in the other two broad categories, perhaps we can use a phenomenon that is prevalent in China to understand why social science majors are more likely to experience negative conversion rates, which is correlated with a serious frictional problem hidden in Chinese society and the job market, that is, many Chinese graduates are overly reliant on entering the government and public organizations as an alternative path after failing to find an ideal job or becoming unemployed. However, as shown in Table 1 and Figure I, the government and the public organizations have a very limited capacity to absorb skilled talent from other industries.

According to Hongbin Li and his colleagues, “Job Preferences and Outcomes for China’s College Graduates,” while approximately 63.8% of Chinese college students wished to work in the state sector, only 41.8% actually received offers from the government and public organizations. Also, Yanzhou Huang in “Secure Jobs, Uncertain Future: China’s Growing Appetite for Government Employment” reports that the proportion of new graduates preferring jobs in government or public organizations rose from 51% in 2019 to 73% in 2024. It is difficult to ascertain from the available literature and data whether most of these individuals are new graduates who have been discouraged by other social science positions. But we can still look at the aggregated upward trend to prove our theory that the gradual weakening of the government’s ability to create government and public sector jobs will lead to more employment pressure on new graduates who aspire to go into branches of the government sector. Taking the civil service exam and entering the government sector is not a more relaxed or low-competition alternative when another industry shrinks.

When students in the social sciences can’t make the leap to STEM or business and economics sectors, because of skill barriers, such as not having experience in programming or statistical analysis, other social sciences industries become their saving grace. However, when other social sciences industries (e.g., the public sector) don’t have the desired ability to accept them, they may become part of that -2.5%.

So, what makes interconversion higher in STEM and the business and economics category? In other words, why haven’t STEM and the business and economics accumulated as much frictional unemployment as the social sciences?

The answer is that China’s national policies are more skewed toward technology, economics and business.

According to the World Bank’s report, “Innovating China: New Drivers of Growth,” China’s innovation and technology development strategy focuses on producing a large pool of STEM graduates. In 2013, STEM majors accounted for more than 40% of China’s total tertiary student population. At the same time, while producing a large number of STEM graduates, China has been stimulating job creation in related fields. It has been using high salaries to attract them into the workforce.

The monthly income of the new graduates class of 2022-2024 in various disciplines six months after graduation provided in the 2025 China Undergraduate Career Report, published by MyCos allows us to compare the post-graduation salaries of STEM, business and economics majors, with social science majors. The data is organized in the TABLE II below, where the unit is RMB.

Table II: Average Monthly Income of Undergraduate Graduates by Discipline (Classes of 2022-2024)

Discipline	Category	Class of 2024	Classes of 2023	Classes of 2022
Engineering	STEM	6841	6709	6610
Economics	Business & Economics	6280	6088	6003
Science	STEM	6115	5964	5759
Management	Business & Economics	6075	5962	5843
Literature	Social Science	5789	5641	5509
History	Social Science	5445	5279	5134
Education	Social Science	5085	5002	4867
National Average (All Undergraduates		6199	6050	5990

TABLE III: Monthly Income and Growth Rate of Undergraduate Graduates by Discipline (Class of 2019)

(unit: RMB yuan, %)

Discipline	Category	Monthly Income Five Years After Graduation	Monthly Income Half a Year After Graduation	Growth Rate (%)
Engineering	STEM	12265	5809	111
Economics	Business & Economics	11102	5519	101
Management	Business & Economics	10475	5350	96
Science	STEM	9998	5392	85
Literature	Social Science	9325	5234	78
Education	Social Science	8131	4778	70
National Average (All Undergraduates)		10619	5440	95

Within the same cohorts, it can be seen that STEM and Business and Economics related majors have higher monthly salaries over a six month period and also have a greater increase relative to the social sciences, and the income level of social sciences is more likely to be below the national average level.

In order to demonstrate the relationship and mechanism between majors and labor market salary distribution, the article introduces the model built by Sharon Traiberman in “Occupations and Import Competition: Evidence from Denmark.”

Traiberman suggests that the students’ fields of study in Denmark determine the range of occupations they can enter in the labor market, which in turn determines the level of their income after graduation. Moreover, Denmark has a strict education stratification mechanism, where each stage of education leads directly to a different profession and thus to a different level of wages. This is why the paper raises a similar problem in Denmark as in China, namely, low occupational mobility. The difference, however, is that matching is more efficient in Denmark, while it is much less so in China, especially in the social sciences.

When we compare these two countries, we will find that although they both have clear education systems, career-oriented education has not been fully popularized within Chinese colleges and universities; it has only been kept clear career-oriented education in vocational schools. In addition, China’s generalized bias against vocational education makes it even more difficult for undergraduate graduates to adapt to the friction between their undergraduate education and the skills required in the real world once they enter the job market. Especially when it is difficult for their field of study to benefit from the national policies.

In summary, the real challenge facing Chinese university graduates does not stem from a macro-level “shortage of positions,” but rather from the persistent mismatch between the educational structure and labor demand. Policy shocks from the government can instantly alter an industry’s absorption capacity, leading to frictional unemployment accumulation in some sectors while causing excessive contraction in others. However, the deeper issue lies in the skill barriers between different disciplines, which hinder talent from freely moving between specializations. STEM and business majors exhibit higher conversion rates and faster salary growth because these fields align closely with China’s development goals, and industrial expansion can continuously absorb new labor. In contrast, social science graduates face simultaneous challenges: industry contraction, difficulty in cross-disciplinary skill transfer, and limited absorption capacity within government departments. Consequently, they are more prone to accumulating structural unemployment. To alleviate graduate employment pressures, a nation must not only focus on short-term industry policies but also address the mismatch between educational outcomes and labor market demands. This requires shifting the emphasis from “what students learn” to “what they can do.”