In this article, we present a series of stylized facts on education spending in China. Our summary is based on research conducted at Peking University's China Institute for Educational Finance Research (CIEFR).
As part of our education series, we also have an overview and introduction to China's education industry (here), and a summary and discussion of China's new K-12 education reforms that have affected the likes of high profile giants New Oriental Education, TAL Education Group, and Gaotu Techedu (here).
Click here to subscribe and stay tuned for future updates!
Prelude
We mention here some important background information before proceeding on to the analysis.
School Expenses
In China, tuition for primary school and middle school are free, with students being assigned to a school based on their housing district. Pre-school and high school are paid but subsidized by the state. High school tuitions in China range from RMB600 per year in rural areas to RMB3,000 per year in top city districts, with admissions being score-based through a written exam.
One Child Policy
The students surveyed by the CIEFR are almost certainly all only children, so the spending amounts shown reflect the amount of money each family invests in their one and only child.
City Classifications
Chinese cities are classified as falling into a range of five tiers, with higher tier cities being more prosperous than lower tier cities. First tier cities exclusively include Beijing, Guangzhou, Shanghai, and Shenzhen, while other economically or politically important cities are classified as second tier cities. Going down the hierarchy, third tier cities include poorer cities with large populations that have somewhat developed infrastructure but no economical or political significance, while fourth tier cities are cities that are still in the developing and urganizing stage. Lastly, fifth tier cities are the poorest cities in China. Further information is provided in our China City Classifications and Income Factsheet.
A Xiancheng (Chinese name: 县城) refers to an administrative centre of a rural county, and is essentially similar to a small rural town. Xianchengs are assigned to a city as part of statistical reporting and are hence not given their own individual city category.
Urban vs. Rural Classification
A city can have both urban and rural areas, although naturally the percentage of a city's population that lives in rural areas is lower in higher tier cities.
Average Calculation
The CIEFR calculates the population arithmetic mean (i.e. average) in their work using a population-weighting method rather than equal-weighting for each category. More details are provided in Appendix 1.
How Much Do Chinese Families Spend on Education?
In 2017, the CIEFR surveyed 14,000 students nationwide to determine how much Chinese families spend on education, including tuition, school supplies, and supplementary classes. We summarize the findings below.
NATIONWIDE
FACT #1: The average Chinese family spends RMB8,143 on education each year, with high school education spending (RMB16,900) far outweighing the spending on other levels of education.
*average across all education levels denoted by dashed red line
Why Is High School Education Spending So High?
One possibility is that high school is paid, although we don't think this reason alone accounts for the huge jump in spending. Rather, there is a combination of additional factors at play, including (1) school supplies may be more expensive at the high school level (e.g. students need to buy more textbooks with mock exams), (2) supplementary classes become more crucial and these classes are also more expensive at higher levels of education, and (3) boarding programs become much more common. With regards to the third point, many rural areas do not have high schools, so students from these areas would need to attend a boarding school at the closest Xiancheng. At the same time, urban students might also join boarding programs if the school they want to attend is far from home, or if there is no convenient high school nearby.
Note: boarding schools in China can also refer to school-week boarding, where parents can pick up their children on Friday afternoons or Saturdays and drop them back at school on Sunday evenings.
URBAN VS. RURAL
FACT #2: The average urban family spends 2.5X more than the average rural family on education each year.
*average across all education levels denoted by dashed red line
How Much Does Education Account For in a Chinese Family's Household Budget?
As a continuation of the study cited above, the CIEFR examines how much of a Chinese family's budget is allocated to education spending, with the following findings.
NATIONWIDE
FACT #3: The average Chinese family allocates 13.2% of their household spending to education each year, with high school spending (26.7%) far outweighing the spending on other levels of education.
*average across all education levels denoted by dashed red line
Note: we think that the slight drop going from pre-school to primary school is probably due to the free nature of primary school in comparison to the paid nature of pre-school.
URBAN VS. RURAL
FACT #4: Urban families have a smaller financial burden at lower levels of education, and greater financial flexibility at higher levels of schooling.
On average, urban households allocate 10.6% of their family spending to education, compared with 14.3% for rural households. It's not surprising that education would constitute a larger proportion of household spending in rural areas given the lower income levels. However, what's interesting is that education spending as a percentage of total household spending is lower for urban families at all education levels except high school, where the average urban household allocates 5.3% more budget to education than do rural households. We think this captures how when it's crucial to spend, higher income households in urban areas have more spending power and greater financial flexibility.
*average across all education levels denoted by dashed red line
FACT #5: High school education spending constitutes almost 1/3 of the average urban family's budget and 1/4 of the average rural family's budget.
In the sample surveyed, high school education constitutes 30.9% of household spending in urban families and 25.6% in rural families. These are extremely high numbers, especially given that they reflect the level of spending for one child, and very much exemplify the financial burden of education that has become a growing social problem in China.
How Many Students Participate In Supplementary Classes?
In 2019, the CIEFR surveyed over 12,000 students across China to gauge how many students attended additional classes outside of school and how much each student's family spent on supplementary education. We summarize the findings below.
NATIONWIDE
FACT #6: On average, 24.4% of Chinese students attend additional classes outside of school, with participation rates being highest at the middle school level and lowest at the high school level.
*average across all education levels denoted by dashed red line
Why Are Supplementary Education Participation Rates So Low in High School?
A possible explanation for the drop in participation rate at the high school level is that education is no longer free (compared with primary school and middle school), so families have to cut down on the spending on extra classes. However, we don't believe this is entirely the case, as high school tuition does not constitute a substantial proportion of household income. The average 2020 disposable income in China is RMB17,131 in rural areas and RMB43,834 in urban areas, compared with high school tuitions ranging between RMB600 per year in rural places to RMB3,000 per year in the expensive districts of large cities. Assuming both parents work (which is typically the case in Chinese families), the high school tuition burden would only be about 2% of each family's income.
Instead, we think that the reduced participation rate is due to the substantial presence of boarding schools and self-study sessions at the high school level. Many high schools offer self-study sessions, sometimes under the guided supervision of teachers, for students to attend in the afternoon and evening. As mentioned previously, numerous high schools throughout the country also offer a boarding program, where, in many instances, boarding students score significantly higher on the college entrance exam than day students at the same school do due to the increased discipline and teacher supervision that come with boarding school.
In this regard, we think that the Chinese government's objective of increasing students' reliance on regular school (see the new education reforms here) by extending school hours and offering additional study sessions could be an effective solution to alleviating the financial and time burden of supplementary classes for parents.
URBAN VS. RURAL
FACT #7: Supplementary education participation rates decrease at the high school level for both urban and rural families, but the drop in participation rate going from middle school to high school is higher for rural students (43.8%) than for urban students (36.4%).
We think this reflects how both urban and rural students have a tendency of substituting regular school options (e.g. study sessions or boarding school - albeit with additional boarding expenses) for supplementary classes, with rural students in particular being more likely to do so due to their higher budget sensitivity. In contrast, some urban students from more well-off families may still choose to pay for additional classes (e.g. one-on-one or small group tutoring) if they think this option is more useful.
*average across all education levels denoted by dashed red line
FACT #8: Urban students are 2.17 times, 2.28 times, and 2.58 times more likely to participate in supplementary classes than rural students at the primary school, middle school, and high school levels respectively.
The increasing divergence in participation rate reflects the differences in purchasing power and budget sensitivity between urban and rural families. Supplementary classes tend to be more expensive at the high school level, and poorer families are more likely to utilize regular school substitutes for external classes when available.
CITY TYPE
FACT #9: Almost 1/2 of the students in first tier cities attend extra classes, compared with 1/3 of the students in second tier cities and 1/5 of the students in Xianchengs.
*average across all education levels denoted by dashed red line
How Much Do Chinese Households Spend on Supplementary Education?
An extension of the CIEFR study above examines the subsample of students who do participate in supplementary classes to determine how much Chinese families spend on education outside of regular school. Both the arithmetic mean (denoted as "average") and median are provided, which gives rise to interesting analysis.
NATIONWIDE
FACT #10: The average (i.e. mean) level of household spending on supplementary education is RMB8,438 per year, compared with a median of RMB3,600.
*average across all education levels denoted by dashed red line
*median across all education levels denoted by dashed green line
FACT#11: The average (i.e. mean) amount of household spending on supplementary education is at least twice the median amount across all education levels, where average spending increases with education level but the median value barely changes.
The large and persistent difference between the average spending amount and the mean spending amount on supplementary education exemplifies the income disparity between the wealthy and less wealthy in China. Wealthier families are able to dedicate significantly larger sums to their children's education, especially as their children's education level goes up, whereas the average Chinese family does not share this financial flexibility. In this regard, the evidence here supports the common criticism that supplementary education exacerbates social inequality (see here).
URBAN VS. RURAL
FACT #12: Urban households spend 3X more on additional classes than rural households, with the mean-median gap persisting in both urban and rural areas.
*average across all education levels denoted by dashed red line
*median across all education levels denoted by dashed green line
CITY TYPE
FACT #13: First tier families spend 2.5X the national average, second tier families spend 1.2X the national average, and Xiancheng families spend 0.5X the national average on supplementary education.
*average across all education levels denoted by dashed red line
*median across all education levels denoted by dashed green line
How Do Spending Patterns Vary Across City Types?
First tier city families are overall the most wealthy in China and are able to invest a substantial amount in their children's supplementary education since primary school. Second tier city families tend to be slightly above average in terms of income, but are still less financially flexible than first tier city families, resulting in moderate spending levels that start picking up pace with middle school education. In contrast, Xiancheng families try to invest more in education at each academic level, but are clearly still facing an income constraint. The largest mean-median divergence occurs at the Xiancheng high school level, where some families are able to invest more than others, but the average family still lacks the means to do so.
How Much of Their Household Income Do Chinese Families Allocate to Supplementary Education?
We extend the analysis by calculating the level of supplementary education spending as a proportion of household income. To do so, we use the supplementary education spending data from the 2019 CIEFR survey, and approximate each household's income by multiplying two to the average disposable income of each family's city type. In most Chinese families, both parents work, so we think this is a reasonable approximation.
Note: we don't include a national average for the proportion of average household income that is allocated to education spending because the resulting statistic is likely to be biased upwards. We explain this further in Appendix 2. Other calculation details are provided in Appendix 3.
FACT #14: On average, urban and rural households do not differ significantly in terms of their income allocation to supplementary education.
Despite the vastly different raw monetary amounts spent on supplementary education shown in Figure 9, as a percentage of household disposable income, the urban and rural area spending levels on supplementary education are actually quite similar.
*average across all education levels denoted by dashed red line
FACT #15: First tier families allocate over 20% of their household income to supplementary education, while second tier families are still constrained at 15%.
First tier and second tier city families have long been the main clientele of supplementary education organizations, due to their higher income levels. Interestingly, we find that apart from the four first tier cities (i.e. Beijing, Guangzhou, Shanghai, and Shenzhen), other economically developed Chinese cities that constitute the second tier still exhibit a cap in terms of the amount they are willing and able to pay for supplementary education.
*average across all education levels denoted by dashed red line
Note: we don't have separate values for Xianchengs because Xianchengs are classified as part of various lower tier cities in statistical reporting.
Click here to subscribe and stay tuned for future updates!
Education Series
Regulation
Glossary
Appendix
1. Average Calculation
The CIEFR calculates the population arithmetic mean (i.e. average) in their work using a population-weighting method rather than equal-weighting for each category. So, the average supplementary education participation rate would *not* be calculated as (Average for Primary School + Average for Middle School + Average for High School)/3, which is why the average values denoted by the dashed red lines do not equal the averages of the three categories. Instead, population weighting is used, for example, if 1/4 of the population survered were primary school students, 1/4 were middle school students, and 1/2 were high school students, then the overall average would be calculated as 1/4*(Average for Primary School) + 1/4*(Average for Middle School) + 1/2*(Average for High School).
2. National Average Calculation for (Supplementary Education Spending/Average Household Income)
We don't include a national average for the proportion of average household income that is allocated to supplementary education spending because the resulting statistic is likely to be biased upwards. Specifically, the average education spending amount in the survey is likely to be positively biased since the sample includes families from first and second tier cities, who have higher tuition fees to pay and are also the main clientele of supplementary education organizations, with only some representation from families in poorer cities. In contrast, the reported average disposable income figure takes into account the whole of China, where poorer cities represent the majority and weight the average figure downwards. Hence, the national average value for education spending as a percentage of household income would be biased upwards, with a higher level of spending that is reflective of first and second tier cities divided by a denominator that is weighted more heavily towards the poorer cities.
3. Calculation of Average Disposable Income in First/Second Tier Cities
There are no official statistics for disposable income according to city tier.
We calculate the average disposable income in first tier cities by calculating an equal-weighted average of the 2019 disposable incomes in the four first tier cities: Beijing, Guangzhou, Shanghai, and Shenzhen.
We proxy for the average disposable income in second tier cities by using our 2020 income table for China (here) which includes a sample of 17 representative cities from the total of 30 cities that have been classified as second tier cities in 2021. Since the income numbers are for 2020, we calculate a 2019 equivalent by discounting the numbers using the growth rate in average national income going from 2019 to 2020.
Commentaires