Does Education Affect Economic Growth? A Re-Examination of Empirical Data from China

Abstract

From the perspective of the spatial differences of China’s educational development, exploring the impact of education on economic development is of great importance for alleviating the main contradiction between China’s inter provincial education and economic development. By collecting the spatial panel data from 31 provinces in China between 2011 and 2020, the study adopted spatial autocorrelation analysis and spatial Dubin model and conducted an empirical exploring the impact between China’s education development and economic growth. The findings of the paper include the following: (1) From the perspective of spatial correlation of the education development level, the improvement of education in China’s provinces affects and promotes each other nationwide; the spatial correlation of the improvement of education level presents a situation of strong combination. In provinces with a higher education level, education resources are relatively concentrated. (2) From the perspective of the role of education in promoting the economy, the improvement of education has a significant impact in promoting China’s economic growth, but it is not significant in the western region of China. Based on the research results, efforts should be made to improve the level of spatial correlation among provinces and cities. This study further suggests that, in the future, vigorously developing industries and vocational education in the western region should be considered as an important measure for China in promoting the coordinated development of the inter provincial economy in the future.

1. Introduction

In September 2000, 189 countries signed the United Nations Millennium Declaration at the United Nations Summit, which established eight goals to eradicate poverty and illiteracy, giving high priority to the two following major issues: the roles of education and economic development. In September 2015, the United Nations Sustainable Development Summit adopted the 2030 Sustainable Development Agenda. It specified 17 Sustainable Development Goals (SDGs), pointing out the need to “provide inclusive and equitable quality education and lifelong learning opportunities for all,” “promote sustained, inclusive and sustainable economic growth,” and “reduce inequalities within and among countries”. The SDGs emphasize the equally important positions of education and the regional economy. In 2021, the United Nations Educational, Scientific, and Cultural Organization released the report “Reimagining our futures together: A new social contract for education.” The report noted that “inequalities within countries have risen at different rates” and that there is a need to “see education as our path to a sustainable common future and to realize its full transformative potential” [1]. China is actively implementing the United Nations declarations, agendas, and compacts, contributing to the Chinese strength and solutions for achieve global goals. The data show that China’s adult literacy rate increased from 84.2% to 95% between 2000 and 2021, and the number of years of education per capita increased from 7.2 to 10.9 years. Additionally, the proportion of the population with a university or higher education increased from 4.66% to 15.5%. China’s gross domestic product (GDP) jumped from US $100,300 to US $1,143,700 [2]. It can be noted that China’s economy and education have both developed rapidly in the past 20 years. However, owing to China’s vast size and the great differences among regions, uneven and incohesive development is the main problem facing China’s education and economy in the new era. Against this background, it is particularly important to re-examine the spatial differences in China’s educational development and explore the impact of education on the economy.

2. Literature Review

Education is a technical means for economic development, and the economy is a necessary guarantee of education. An overview of the literature at home and abroad shows that research has focused primarily on the relationship between the two.

2.1. The Contribution of Education Improvement to Rapid Economic Growth

Some scholars have demonstrated that educational development serves the needs of economic growth [3] and that education improvement raises human capital, which improves labor quality and productivity. This directly or indirectly contributes to economic growth [4,5,6,7,8]. Different levels of education contribute differently to the economy, and higher education is considered as an important factor in enhancing economic competitiveness [9]. Other studies have concluded that secondary education contributes far more to the economy than primary and university education [10]. In China’s system, compulsory education and cognitive skills do not significantly directly or indirectly contribute to economic growth. In contrast, high school education indirectly contributes to economic development through improved innovation capacity [11]. Some scholars have also compared the impact of high school, vocational, and university education on China’s economy, noting that high school education as a primary selective mechanism may contribute relatively little to human capital. In contrast, vocational and university education returns are larger, similar to trends seen in the United States [12]. Moreover, improvement in the education level and changes in education quality have different economic effects [13], with the quality of education being a significant factor affecting economic development [14]. Similarly, education’s contribution to the economy varies across different states of economic development. Average years of schooling does not contribute significantly to economic growth in developing countries, being effective only in developed countries [15]. In the same vein, the mere increase in education quality does not contribute strongly to economic growth in developing countries [16]. Moreover, there remain considerable differences in the returns from education across regions and populations and differences in economic impact [17]. Some even argue that education’s contribution to economic development is influenced by multiple factors such as social consciousness [18], gender perceptions [19], and return on investment [20].

2.2. The Interaction between Education and Economic Development

The relationship between education and the economy is not a one-way influence but a two-way interaction. This conclusion is unanimously accepted by domestic and foreign scholars [21]. In the economic growth process, the government promotes education by increasing education expenditure. In contrast, improving education can feed the economy through knowledge spillover, such as enhancing innovation capacity [22]. However, some scholars have argued that there is a non-reciprocal relationship in the interaction between educational and economic development: the effect of the economy’s promotion effect on education is significantly greater than the effect of education’s promoting effect on the economy, while the improvement of the education level has a positive effect on economic development only when a certain threshold is breached [23]. Some studies have argued that in specific economic development periods, the rapid development of education’s rapid development has an inhibiting effect on the economy [24]. The interaction between different education levels and the economy exists, with significant mutual promotion between primary and secondary education and economic growth. However, only reverse causality exists between higher education and economic growth [23,25]. Different economic development stages require different education systems to promote growth [26,27]. Moreover, improving higher education’s quality significantly contributes to economic growth when the economic development level is high. In contrast, the development of in-service education has a stronger impact on the economy with a low level of development [28]. Therefore, only the symbiotic and coupled development of education and the economy can be mutually beneficial [29,30].

To sum up, the academic research on the relationship between education and economic development has been relatively rich, mainly focusing on the impact of education development level, different levels of education, different types of education, education quality, different economic development states, social consciousness, and other factors in the same space or region on economic development, as well as the interaction mechanism between economy and education. This provides an abundant database and theoretical support for this study. In fact, factors such as geographical location, resource endowment, the cultural tradition and system, and resource supply in different regions directly affect the development of education and economy among regions. Space or region should be an important factor in discussing the relationship between education and economic development. In particular, with the development of social conditions, regional differences in education and economic development have become increasingly evident. However, there are few research results on the spatial heterogeneity of education and economy. Based on this, the paper introduces space into the study of the impact of education development on economic growth, builds a spatial econometric model, and comprehensively explores the impact path of education on economic growth.

The rest of this study is structured as follows: Section 3 explores the spatial differences in education development among China’s regions from different dimensions, such as the population’s education level, education equity, and years of education per capita. Section 4 elaborates on the construction of a spatial econometric model to explore the relationship between the improvement of education and the promotion of regional economic growth, and compares and analyzes the effects between different regions from the national macro-perspective and the subregional micro-perspective. Finally, Section 5 summarizes and discusses the study and proposes the areas for future research.

3. Spatial Differences in the Development of Education in China

Over the past 40 years of reform and opening up, China has made significant advances in education. The government and academic community have continuously and actively explored educational development models and adjusted policies, injecting new vitality into China’s education development and promoting its continued progress. The balanced development of education between provinces and regions is a common goal of the Communist Party of China and the Chinese people and an inherent requirement for achieving social equity and promoting coordinated social development. However, while education is developing rapidly, disparities between regions still exist. This study focuses on the spatial differences in the development of China’s educational undertakings from different dimensions, such as the population’s education level, the Gini coefficient of education, and the average years of education.

3.1. Spatial Differences in Education

3.1.1. Significant Increase in Education

The education status of the population is usually regarded as an important indicator for measuring regional or national economic development [31]. The illiteracy rate reflects the educational status of regional or national populations, the degree of cultural and educational popularization and development, and the degree of economic development of a country. In this study, the illiteracy rate of the population aged 15 and above was chosen to reflect the education level of different regions. The illiteracy rate of the population aged 15 and above from 2011 to 2020 was used to draw a line graph (Figure 1), showing that the education level of China’s overall population gradually increased, and the illiteracy rate decreased. The share of the illiterate population decreased from 5.21 in 2011 to 3.26 in 2020, a decrease of almost two percentage points, and the population’s education level significantly improved.

3.1.2. Analysis of Differences in Regional Education

To simplify the analysis and account for the time interval, this study chose two years as the temporal cross-section and plotted the illiteracy rate stratification of the population aged 15 and above in 2012, 2014, 2016, 2018, and 2020 for each region of China (except Hong Kong, Macao, and Taiwan), with a higher regional stratum indicating a lower education level of the relevant population. Figure 2 shows the following. (1) Throughout the study period, the only province located in the fourth stratum since the beginning is the Xizang, indicating that the Xizang has the lowest level of education in the country. This situation has not changed in recent years. (2) From 2012–2018, most of China’s western and southwestern regions were in the third stratum, and the population was significantly less educated than in other regions. China’s western and southwestern regions consist mostly of mountains, plateaus, and basins, and the topographical characteristics affect many factors such as transportation and economy; furthermore, the western and southwestern regions have concentrations of ethnic minorities, and these cultural differences also affect the acceptance of education. (3) The highly educated population mostly resides in China’s northeast and southeast. These regional economies are significantly more advanced than other regions, as is their degree of openness. (4) By 2020, China’s illiteracy rate had been significantly reduced, with most of the country’s regions located in the first stratum, indicating that the literacy initiative and the nine-year compulsory education policy have achieved positive results. (5) Both the Xinjiang and the Guangxi are autonomous minority regions, but the people residing in these regions are significantly more educated than those in other such regions.

3.2. Spatial Differences of Regional Education Equity

3.2.1. Measures of Educational Equity

How to measure the degree of inequality in educational development among regions is a focus of academic research. After an in-depth literature review, the education Gini coefficient was chosen as an indicator to measure the inequality of regional educational development [32,33]. The Gini education coefficient is derived based on the principle of Gini coefficient measurement. According to international organizations, a Gini coefficient below 0.2 indicates absolute equity, between 0.2 and 0.3 fairness, between 0.3 and 0.4 reasonable equity, between 0.4 and 0.5 a gap, and 0.5 or above a large gap. This criterion also applies to the Gini education coefficient; the higher the education Gini educational coefficient, the lower the degree of educational equity, and vice versa. The calculation formula is as follows.

$$G_e=1-{{\displaystyle \sum }}_{i=1}^n\left(X_i-X_{i-1}\right)\left(Y_i+Y_{i-1}\right)$$

(1)

where G_e denotes the Gini coefficient of education; n denotes the number of education levels; i denotes one of the education levels; X_i denotes the ratio of the educated population accumulated in education level i to the total population; and Y_i denotes the percentage of years of education accumulated in education level i to the total number of years of education in the region. Meanwhile, according to the current school year and educational systems in China, there are five education levels: the first level, no schooling, with zero years of education; the second level, primary school, with 6 years of education; the third level, junior high school, with 9 years of education; the fourth level, high school or junior college, with 12 years of education; the fifth level, college and above, with 16 years of education.

3.2.2. Measurement Results

Again, to simplify the analysis, the distribution of the Gini coefficient of education for 2012, 2014, 2016, 2018, and 2020 was plotted for each province (except Hong Kong, Macao, and Taiwan) according to the mapping method in 1.1.2 above. The different Gini coefficient scores were used to indicate different degrees of distributional equity according to the relevant research (

Table 1. Gini coefficient of education equity.

Education Equity	Gini Coefficient
Absolute fairness	Ge ≤ 0.2
Relatively fairness	0.2 < Ge ≤ 0.3
Relatively reasonable	0.3 < Ge ≤ 0.4
Relatively large	0.4 < Ge ≤ 0.5
Large gap	Ge > 0.5

), and the final figure is shown in Figure 3.

Figure 3 shows the following. (1) All regions in China have achieved a reasonable distribution of educational resources, with no regions with large or varying disparities. Additionally, most regions are located in areas with a more equitable educational resource distribution due to China’s recent investment in education and the importance attached to education by the government. China has always given priority to educational development and educational power. Meanwhile, to ensure the fair distribution of educational resources, China has successively introduced numerous policies, such as promoting educational equity through information dispersal, guaranteeing educational equity using beneficiary policies, and maintaining educational equity through standardized management.

All these policies have guaranteed fair educational resource distribution and laid the human foundation for China’s social development. (2) Beijing, Tianjin, Heilongjiang, Liaoning, and Jilin provinces are in regions where the distribution of educational resources is reasonable, indicating less variability in the regional development of education levels. Beijing and Tianjin have historically been regions with highly-educated populations, especially Beijing, which has the highest level of educational development in the country and the primary concentration of qualified talent in China. (3) The regions with a reasonable educational resource distribution are mostly China’s minority autonomous regions, such as the Xinjiang and Xizang. The levels of educational development and resource allocation for ethnic minorities are easily influenced by ethnic culture and regional transportation. (4) However, it is worth noting that equity in educational resource distribution in Henan and Hebei provinces has been declining in recent years. Henan and Hebei have large populations and are also the major provinces regarding college entrance examinations. The degree of equity in education is still gradually decreasing, especially in Henan Province, where the pressure to advance to higher education has led to a greater concentration of educational resources and teachers in urban and more economically developed areas, exacerbating educational inequity.

3.3. Differences in Average Schooling Years

The average number of years of education refers to the average number of years of academic education (including adult education, excluding training) received by a population group in a certain time period and region. The main goal of China’s economic and social development in the 14th Five-Year Plan period is to increase the whole population’s education level and the average number of years of education for the working-age population to 11.3 years. Improving the average years of education of the working-age population is fundamental to improving the overall quality of a country’s population and its development. Simultaneously, raising the average years of education for the working-age population provides a solid human foundation for building a knowledgeable, skilled, and innovative workforce, which is important for China to become a modern country and an educational powerhouse.

3.3.1. Measuring Average Schooling Years

Based on previous studies [34,35], the formula for calculating the average years of education is set as follows.

$$\mathrm{Average} \mathrm{years} \mathrm{of} \mathrm{schooling}=\frac{\mathrm{H}1+6\mathrm{H}2+9\mathrm{H}3+12\mathrm{H}4+16\mathrm{H}5}{\mathrm{H}1+\mathrm{H}2+\mathrm{H}3+\mathrm{H}4+\mathrm{H}5+\mathrm{H}6}$$

(2)

Referring to the classification of educational attainment in the China Statistical Yearbook, this model defines five education levels: illiterate or semi-literate, elementary school, junior high school, senior high school, and college and above. It assumes that adult education qualifications at the same level are the same as those obtained from general education and self-learning exams. The five education levels are the following: H1: illiterate or semi-literate population. This is the total of the illiterate, literacy class, and elementary school dropout populations. In this model, the average years of education are set as 1 year. The second education level is H2: primary school literate population. The average years of education are set to 6 years. H3 and H4 are the middle school and high school literate populations, with an average period of education of 9 years and 12 years, respectively. H5: the college and above population. This includes college, undergraduate, master’s, and doctoral students. Since the available statistics do not subdivide this education level, this formula ignores differences in the years of education for college, undergraduate, master’s, and doctoral students. It sets the average years of education of the population at this level as 16 years.

3.3.2. Measurement Results

To simplify the analysis, the average years of education of the labor force population in 2012, 2014, 2016, 2018, and 2020 for each province in China (except Hong Kong, Macau, and Taiwan) were collated and stratified using 6 years and below, 6–8 years, 9–12 years, and 12+ years in

Table 2. Average years of education of the population in all provinces and municipalities in China.

Year	<6 Year	6–8 Year	9–12 Year	>12 Year
2012	Xizang	Hunan, Guangxi, Chongqing, Sichuan, Guizhou, Yunnan, Gansu, Qinghai, Hebei, Anhui, Fujian, Jiangxi, Shandong, Henan, Ningxia, Xinjiang, Zhejiang	Beijing, Tianjin, Shanxi, Nei Mongol, Liaoning, Jilin, Heilongjiang, Shanghai, Jiangsu, Guangdong, Hubei, Shaanxi, Hainan
2014	Xizang	Guangxi, Sichuan, Guizhou, Yunnan, Gansu, Qinghai, Hebei, Anhui, Fujian, Jiangxi, Ningxia	Beijing, Tianjin, Shanxi, Nei Mongol, Liaoning, Jilin, Heilongjiang, Shanghai, Jiangsu, Zhejiang, Guangdong, Hubei, Shaanxi, Shandong, Hunan, Henan, Hainan, Chongqing, Xinjiang
2016	Xizang	Guangxi, Sichuan, Guizhou, Yunnan, Gansu, Qinghai, Anhui, Fujian, Jiangxi, Henan	Hebei, Tianjin, Shanxi, Nei Mongol, Liaoning, Jilin, Heilongjiang, Shanghai, Jiangsu, Zhejiang, Guangdong, Hubei, Shaanxi, Shandong, Hunan, Hainan, Chongqing, Ningxia, Xinjiang	Beijing
2018		Guangxi, Sichuan, Guizhou, Yunnan, Gansu, Qinghai, Anhui, Jiangxi	Hebei, Tianjin, Shanxi, Nei Mongol, Liaoning, Jilin, Heilongjiang, Shanghai, Jiangsu, Zhejiang, Guangdong, Hubei, Shaanxi, Shandong, Hunan, Fujian, Hainan, Chongqing, Ningxia, Xinjiang, Henan	Beijing
2020		Guizhou, Yunnan, Gansu, Qinghai	Hebei, Tianjin, Shanxi, Nei Mongol, Liaoning, Jilin, Heilongjiang, Shanghai, Jiangsu, Zhejiang, Guangdong, Hubei, Shaanxi, Shandong, Hunan, Fujian, Hainan, Sichuan, Anhui, Jiangxi, Chongqing, Ningxia, Guangxi, Xinjiang, Henan	Beijing

.

Table 2 shows that (1) the years of education in China’s labor force have been increasing. In 2016, the years of education exceeded 12 years and those with less than 6 years were absent; the labor forces in most provinces have a junior high school education or above, indicating that education has progressed. The human capital structure has also improved, injecting additional high-quality talent into economic development. (2) The years of education of the Xizang’s labor force were the lowest in the country, with fewer years of primary education from 2012 to 2016, while the average years of education in Beijing were the highest. This indicates that educational inequity in China remains widespread, with huge disparities between provinces and regions. The focus is on the development of subsequent education careers.

4. The Impact of Education on the Economy

4.1. Research Methodology

4.1.1. Spatial Autocorrelation Analysis

The “first law of geography” states that any element is spatially correlated. Therefore, global and local spatial autocorrelation can be used to analyze spatial correlation [36]. Global spatial autocorrelation is used to analyze the spatial correlation of geographical observations across a region and is measured using Global Moran’s I, which is calculated as follows:

$$Mora{n}^{’}s I=\frac{{{\displaystyle \sum }}_{i=1}^n{{\displaystyle \sum }}_{j=1}^n{w}_{ij}\left(x_i-\overline{x}\right)\left(x_j-\overline{x}\right)}{s^2{{\displaystyle \sum }}_{i=1}^n{{\displaystyle \sum }}_{j=1}^n{w}_{ij}}$$

(3)

In Equation (3), I is the global spatial autocorrelation coefficient. The larger the value of I, the greater the spatial correlation between economic growth and industrial structure development; x_i and x_j are the observed values of the geographical attributes of changes in economic growth and industrial structure; and n is the sample size. $\overline{x}$ and s² are the mean and variance of s. W_ij is the spatial weight matrix constructed based on the proximity criterion.

Local spatial autocorrelation is mainly used to measure the geographically observed variables and is analyzed using Local Moran’s I, which is calculated as follows:

$$I_i={\mathrm{z}}_{\mathrm{i}}{{\displaystyle \sum }}_{j=1}^n{w}_{ij}{z}_j$$

(4)

In Equation (4), z_i and z_j are the normalized values of the observations of regions i and j. I_i is the local spatial correlation between the economic growth and industrial structure development of region i and region j, which can be classified into four aggregation types: high–high, low–high, high–low, and low–low.

4.1.2. Spatial Durbin Model

The spatial Durbin model (SDM) incorporates the spatial lag model (SLM) and the spatial error model (SEM). Therefore, this study first considers introducing the SDM, whose general paradigm is as follows:

Y = ρWY + Xβ + θWX + αl_n + ε

(5)

In Equation (5), X denotes the explanatory variable; Y denotes the explained variable; ρ denotes the spatial autocorrelation coefficient; W denotes the constructed spatial weight matrix; WX and WY denote the spatial lag term; α denotes the constant term; l_n denotes the n × 1 unit weight matrix; θ and β denote the regression coefficients; and ε denotes the spatial error term.

When there is a spatial lag term in the model, the regression coefficients will no longer reflect the explanatory variables’ effect on the explanatory variables. Therefore, the total effect must be decomposed into direct and indirect effects. For this purpose, the above equation is decomposed into a matrix form using partial differential form decomposition as follows.

Y = [I − ρW]⁻¹cl_N + [I − ρW]⁻¹[X’β + WX’β] + [I − ρW]⁻¹ε*

(6)

In this equation, ρ is the spatial autocorrelation coefficient; W is the spatial weight matrix; Y is an N × 1-dimensional vector of explanatory variables; c is a constant quantity; l_N is an N × 1-dimensional vector whose elements are all N × 1-dimensional; and X’ is an N × K-dimensional matrix of all explanatory variables. ε is the error term. We derive the partial differential matrix of the explanatory variable Y with respect to the kth explanatory variable at time t as:

$${{\displaystyle \left[\begin{array}{ccc}\frac{\partial {{\displaystyle Y}}_i}{\partial {{\displaystyle X}}_{1K}}& \cdots & \frac{\partial {{\displaystyle Y}}_i}{\partial {{\displaystyle X}}_{NK}}\end{array}\right]}}_t={{\displaystyle \left[\begin{array}{ccc}\frac{\partial {{\displaystyle Y}}_i}{\partial {{\displaystyle X}}_{1K}}& \cdots & \frac{\partial {{\displaystyle Y}}_1}{\partial {{\displaystyle X}}_{NK}}\\ ⋮& \ddots & ⋮\\ \frac{\partial {{\displaystyle Y}}_N}{\partial {{\displaystyle X}}_{1K}}& \cdots & \frac{\partial {{\displaystyle Y}}_N}{\partial {{\displaystyle X}}_{NK}}\end{array}\right]}}_t={{\displaystyle (I-\rho W)}}^{-1}{{\displaystyle \left[\begin{array}{cccc}{{\displaystyle \beta }}_K& {{\displaystyle {{\displaystyle \alpha }}_K\omega }}_{12}& \cdots & {{\displaystyle {{\displaystyle \omega }}_{1N}\alpha }}_K\\ {{\displaystyle {{\displaystyle \omega }}_{21}\alpha }}_K& {{\displaystyle \beta }}_K& \cdots & {{\displaystyle {{\displaystyle \omega }}_{2N}\alpha }}_K\\ ⋮& ⋮& \ddots & ⋮\\ {{\displaystyle {{\displaystyle \omega }}_{N1}\alpha }}_K& {{\displaystyle {{\displaystyle \omega }}_{N2}\alpha }}_K& \cdots & {{\displaystyle \beta }}_K\end{array}\right]}}_t$$

(7)

where the direct effect is represented by the mean of the elements on the main diagonal of the rightmost matrix β_k, which is expressed as the output elasticity of the region’s independent variable to the region’s dependent variable; the indirect effect is expressed as the mean of the elements of the matrix at the right end except for element β_k on the main diagonal, which reflects the effect of the neighboring region’s independent variable on the region’s dependent variable. The total effect is the sum of the direct and indirect effects.

4.2. Research Hypothesis, Indicators, and Data

4.2.1. Research Hypothesis

The logical starting point for this study is the general form of the Cobb–Douglas production function.

Y = AK^αL^1−α

(8)

The sources of economic growth depend mainly on capital and labor inputs and the increase in total factor productivity. Additionally, numerous other factors affect economic growth, such as the level of education development, industrial structure, foreign investment, government intervention, and innovation capability. Based on this, this study proposes the following theoretical hypotheses:

H1. 

An increase in education level has a significant contribution and positive spillover effect on economic growth.

H2. 

The development of industrial structure has significant positive and spillover effects on economic growth because the upgrading of industrial structure promotes economic growth through industrial agglomeration and structural dividend [37,38].

H3. 

Foreign investment has a positive effect on regional economic growth with spatially significant spillover effects. Regional economic growth is promoted in different ways, such as accelerating domestic technological progress, introducing advanced management experience, and technology spillover through foreign investment [39,40].

H4 

. Government intervention has a positive effect on regional economic growth, with a significant positive spillover effect. The government influences economic growth by changing the path of fiscal spending [41,42].

H5 

. Innovation capacity contributes significantly to regional economic growth, and the positive spillover effect is spatially significant. The “horizontal“ and “structural” effects of innovation capabilities generate inter-regional spillover effects and thus increase inter-regional economic growth [43,44].

4.2.2. Indicators and Data Sources

Based on the above assumptions, the following indicator system was established (

Table 3. Selection of variable indicators.

Variable	Meaning	Evaluating Indicator
Interpreted variable	GDP	Regional GDP
Core explanatory variables	Edu	Per capita years of education
Control variable	K	Capital stock
	L	Employed population
	FDI	Amount of foreign capital actually used
	Gov	Government expenditure
	PPAT	Number of patent applications
	STR	Industrial structure upgrading coefficient

).

(1)

Explained variable: regional economic growth (GDP). This study uses regional economic output GDP to express the development of the regional economy.

(2)

Core explanatory variable: educational development level (Edu). This study draws on the literature and uses the regional average years of schooling to express this.

(3)

Control variables: these primarily include capital stock, labor force, level of openness to the outside world, industrial structure, level of government intervention, and innovation capacity. Among them: (i) capital stock (K) is calculated by the perpetual inventory method using 2011 as the base period, and a depreciation rate of 6% [45]; (ii) labor force (L) is expressed by the number of employees at the end of the year in the region; (iii) the level of external openness (Open) is expressed by the amount of actual foreign investment utilized by the region in the current year; (iv) industrial structure (STR) is expressed by the industrial structure upgrading coefficient, calculated as STR = R₁ + 2R₂ + 3R₃, where R₁, R_2, and R₃ denote the proportion of output value of primary, secondary, and tertiary industries in GDP respectively [46]; (v) the level of government intervention (Gov) is expressed by the amount of government fiscal expenditure; (vi) the level of innovation (PPAT) is expressed by the annual number of patents granted in a region.

The above data are taken from the China Statistical Yearbook (2012–2021).

4.3. Empirical Results and Analysis

4.3.1. Analysis of Spatial Autocorrelation Results

(1)

Global autocorrelation analysis

The prerequisite for using the spatial model is to ensure that the explanatory variables are spatially autocorrelated. Therefore, this study used data on the level of education development and GDP of 31 Chinese provinces (excluding Hong Kong, Macao, and Taiwan) from 2011 to 2020 and calculated their Moran’s I indices using the geographical adjacency matrix (

Table 4. Moran’s I index of education level and GDP.

	2012	2013	2014	2015	2016	2017	2018	2019	2020
Edu	0.298	0.251	0.234	0.286	0.287	0.259	0.246	0.260	0.275
GDP	0.243	0.241	0.239	0.241	0.253	0.243	0.236	0.259	0.262

). The results from the data show that the development of the education level in each region has spatial autocorrelation within the study years. All pass the 5% significance test, indicating that their development is significantly correlated and will be influenced by neighboring regions; that is, the improvement of education level and economic growth in one region will lead to the improvement of education level and economic growth in the surrounding regions.

Since Moran’s I index has substantial limitations in testing spatial autocorrelation, there may be a failure to scientifically reveal local spatial correlation and heterogeneity in each direction. Therefore, when further considering whether there is local spatial clustering of observations, the local indicators of spatial association are needed to portray the local spatial interdependence more intuitively and the heterogeneity characteristics of each region’s industrial structure. Therefore, a two-year time cross-section is used to simplify the analysis to show the national education development level’s local autocorrelation level. The results show spatial variation in the level of education in China over the years studied; the provinces and cities located in the high and high agglomeration quadrant are mostly the more economically developed regions. However, the results also show an inseparable relationship between education and economic development, presenting a pattern of a strong economy leading to a strong education or a strong education leading to a strong economy (

Table 5. Local spatial autocorrelation state of education level.

Year	High-High Aggregation	High-Low Aggregation	Low-High Aggregation	Low-Low Aggregation
2012	Beijing, Tianjin, Shanxi, Nei Mongol, Liaoning, Jilin, Heilongjiang, Shanghai, Jiangsu, Zhejiang, Hubei, Hainan, Shaanxi	Hebei, Anhui, Fujian, Jiangxi, Shandong, Henan, Ningxia	Hunan, Guangxi, Chongqing, Sichuan, Guizhou, Yunnan, Xizang, Gansu, Qinghai	Guangdong, Xinjiang
2014	Beijing, Tianjin, Shanxi, Nei Mongol, Liaoning, Jilin, Heilongjiang, Shanghai, Jiangsu, Zhejiang, Shandong, Henan, Hubei, Hunan, Guangdong, Hainan, Shaanxi	Hebei, Anhui, Fujian, Jiangxi, Ningxia	Guangxi, Chongqing, Sichuan, Guizhou, Yunnan, Xizang, Gansu, Qinghai	Xinjiang
2016	Beijing, Tianjin, Shanxi, Nei Mongol, Liaoning, Jilin, Heilongjiang, Shanghai, Jiangsu, Zhejiang, Hubei, Hainan, Shaanxi, Ningxia	Hebei, Anhui, Fujian, Jiangxi, Shandong, Henan	Guangxi, Chongqing, Sichuan, Guizhou, Yunnan, Xizang, Gansu, Qinghai, Xinjiang	Hunan, Guangdong
2018	Beijing, Tianjin, Shanxi, Nei Mongol, Liaoning, Jilin, Heilongjiang, Shanghai, Jiangsu, Hubei, Hainan, Shaanxi	Hebei, Zhejiang, Anhui, Fujian, Jiangxi, Henan, Ningxia	Shandong, Guangxi, Chongqing, Sichuan, Guizhou, Yunnan, Xizang, Gansu, Qinghai	Hunan, Guangdong, Xinjiang
2020	Beijing, Tianjin, Shanxi, Nei Mongol, Liaoning, Jilin, Heilongjiang, Shanghai, Jiangsu, Zhejiang, Hubei, Hainan, Shaanxi	Hebei, Anhui, Fujian, Jiangxi, Shandong, Henan, Ningxia	Hunan, Guangxi, Chongqing, Sichuan, Guizhou, Yunnan, Xizang, Gansu, Qinghai, Xinjiang	Guangdong

).

4.3.2. Spatial Econometric Estimation of the Impact That Education on Economy

(1)

Spatial Econometric Model

The spatial autocorrelation test reveals that the economic development of the Beijing–Tianjin–Hebei urban agglomeration has a significant spatial correlation. Therefore, its spatiality cannot be ignored when studying its relationship to economic growth, and the introduction of spatial econometric models must be considered. The commonly used spatial econometric models include the SLM, SEM, and SDM. Their differences mainly originate from the spatial matrix and the interaction terms of variables. Among the three models, the SDM is the standard one. It provides sufficient feedback on the spatial correlation problems caused by the explanatory and explained variables and their interaction terms and also captures different variables’ spatial spillover effects. Therefore, the SDM is used for estimation here. First, the following SDM is constructed.

GDP_it = ρWgdp_it + α₁STR_it + α₂K_it + α₃L_it + α₄FDI_it + α₅Gov_it + α₆ppat_it + α₇Edu_it + β₁WlnSTR_it +
β₂WK_it + β₃WL_it + β₄WFDI_it + β₅WGov_it + β₆Wppat_it + β₇WEdu_it + ε_it

(9)

In Equation (9), ρ is the spatial autocorrelation coefficient; W is the spatial weight matrix; α and β are the autoregressive coefficients; ε is the spatial error term.

Before proceeding with model construction, it is important to determine whether the variables are independent. There are many methods to detect multicollinearity, and the (variance inflation factor, VIF) VIF value in regression analysis is often used; the larger the VIF value, the more serious the multicollinearity. It is generally believed that when the VIF is greater than 10, the model has a serious collinearity problem. Here, we use the VIF value to carry out the collinearity test. The results in

Table 6. Collinearity test among variables.

	Edu	K	L	FDI	Gov	PPAT	STR
VIF	1.31	4.20	3.76	3.01	6.04	3.98	1.02

show that the VIF values among the variables are less than 10, indicating that the variables are independent of each other.

Additionally, the model must be tested for reasonableness based on the model discriminatory approach proposed by Anselin [47]. Using the LR test and the Wald test, the data in

Table 7. LR and Wald test results of SDM model.

Test	Result	p
LR test -spatial- lag(SEM)	38.04	0.00
Wald-spatial -erro(SEM)	14.54	0.02
LR test -spatial- lag(SLM)	38.04	0.00
Wald-spatial -error(SLM)	16.24	0.02

show that both pass the 5% significance level test, justifying the use of the SDM.

In the selected SDM, it is necessary to discern which estimation method is more appropriate, random or fixed effects. In this study, the hausman test was used, and the results showed that the estimate of the hausman test was 28.11, and it passed the 1% significance level test. The fixed effect estimate was higher than the random effect; therefore, the fixed effect results were used for the analysis here (

Table 8. Model estimation results.

Variable	Common Panel Model		Random Effect Model		Fixed Effect Model
	Coef	p	Coef	p	Coef	p
Edu	−1248.77	0.04	−7.22	0.99	754.04	0.00
K	0.09	0.00	0.08	0.00	0.07	0.00
L	4.47	0.00	1.82	0.00	1.21	0.00
Str	6.02	0.21	5.00	0.27	−0.49	0.94
FDI	0.09	0.23	0.07	0.32	0.47	0.00
Gov	1.97	0.00	2.05	0.00	1.82	0.00
PPAT	0.03	0.00	0.04	0.00	0.08	0.00
C	355.20	0.95	1833.93	0.75
Wx-edu			−1941.10	0.00	−209.51	0.00
Wx-l			3.88	0.00	2.24	0.00
Wx-ppat			−0.01	0.14	0.03	0.01
R2	0.93		0.91		0.98
ρ			0.05	0.23	−0.23	0.01
Hausman test	28.11(0.00)

).

(2)

Analysis of Overall Effects

The results of the fixed effects of the SDM show that the increase in the level of regional education contributes significantly to economic growth. At the same time, its lags also indicate that its development has a significant spillover effect on the development of neighboring regions. The value of ρ from the spatial model also shows that the regional economy has a significant spatial spillover effect during education development. Accompanying educational development, capital investment, foreign investment in labor, and the increase in innovation level all significantly contribute to China’s economic growth.

(3)

Analysis of the Results of the Decomposition of Spatial Effects

The spatial Durbin model contains both spatial lagged terms of the dependent variable and the independent variable, leading to some possible bias in the estimated coefficients of SDM in Table 7. Here, the total effect of the independent variable’s effect on the dependent variable is decomposed into direct and indirect effects, using the effect decomposition method (

Table 9. Decomposition results of spatial effects.

Variable	Direct Effect		Indirect Effect		Total Effect
	Coef	p	Coef	p	Coef	p
Edu	784.57	0.01	−336.12	0.06	448.45	0.01
K	0.07	0.00	−0.01	0.01	0.06	0.00
L	1.09	0.00	1.68	0.00	2.77	0.00
Str	−0.82	0.91	0.12	0.93	−0.69	0.91
FDI	0.48	0.00	−0.09	0.02	0.39	0.00
Gov	1.84	0.00	−0.35	0.00	1.49	0.00
PPAT	0.08	0.00	0.01	0.02	0.09	0.00

).

The results show that the role of education development in promoting economic growth is significant. The improvement in education level changes the composition structure of regional human capital at a certain level, and the improvement of human capital causes a better and more effective combination of labor and capital [36,46]. Human capital improvement also promotes economic growth through technological innovation and other means. The research results further support the view of some scholars that the improvement of education level promotes economic growth [36], while the improvement of human capital also promotes economic growth through technological innovation and other means. The findings further support some scholars’ view that the improvement of education level promotes economic growth [48,49]. Regarding indirect effects, the increase in education level has a negative spatial effect, but its extent is less than its contribution to the region. Scholars’ findings on the spillover effect of education level on economic growth vary, with some studies finding that the increase in education level promotes inter-regional population mobility and thus changes the labor market in surrounding areas, indirectly promoting neighboring regions’ economic growth [50]. Some scholars also argue that human capital is affected by the increase in education level in an imperfectly competitive market. High-quality talent squeezes the survival space of low-quality talent. Education is not sufficiently sensitive to changes in the market economy. The resulting educational development level has a significant time lag. Meanwhile, the labor force transfer between urban and rural areas, regions, and economically developed and less developed areas caused by urbanization also results in negative spillover to neighboring regions regarding the increase in education level.

Despite China’s increasing innovation drive in recent years, the miracle of China’s economic growth remains the result of factor inputs [51,52]. The study finds that capital stock has a significant boosting effect, but the results suggest a negative spatial spillover effect. This is because significant competition exists among China’s regions, and the increase in investment promotes regional development. In addition, the well developed regions tend to have a strong “siphon effect”, which attracts more talent and foreign investment into the region, thus demonstrating a negative spillover effect on neighboring regions’ economic growth.

Research has shown that labor contributes significantly to regional economic growth and has a significant positive spillover effect. Labor is the fundamental driver of economic growth and structural transformation, and a certain level of labor input leads to rapid economic growth. Simultaneously, inter-regional labor mobility also causes a positive spillover effect between them. However, it is worth noting that the “leveling effect” in labor supply has diminishing returns, which makes it impossible for regional development to rely on the “demographic dividend” in the long run. Nevertheless, this is one of the ways to achieve economic transformation and upgrading [53,54].

The study also found that the industrial structure’s promotion effect on China’s economic growth and the spillover effect on neighboring regions are not significant, which may be related to the existence of many irrational factors in the development of China’s industrial structure. Therefore, the transformation and upgrading of the industrial structure remain the focus of future development [46].

The results show that foreign investment and government intervention have a significant role in promoting China’s economic growth and a negative spillover effect on neighboring regions’ economic growth. Foreign investment facilitates economic and trade exchanges between countries worldwide and introduces advanced management experience while promoting inter-regional competition and cooperation, which significantly contributing to economic efficiency [39,40]. The negative spillover effects of foreign investment and government intervention further indicate that China’s degree of regional competition is relatively high and the cooperation is insufficient. In a subsequent development, more attention must be given to balanced and coordinated development among regions.

The results here similarly show that innovation capacity has a significant contribution and positive spillover effect on China’s economic growth. An increase in the national innovation level promotes economic efficiency and the development of high-tech industries [55]. Technological innovation is a core driver supporting the Chinese economy’s healthy and sustained growth. The outward spillover effect of innovation capability is recognized by many scholars [44], and the innovation level increase accelerates knowledge and capital spillovers between regions. Additionally, it promotes neighboring regions’ innovation level improvement.

(4)

Decomposition of Subregional Spatial Effects

China is vast, and there are huge development differences between regions. To make this study more realistic, we divide 31 provinces and cities into four regions (

Table 10. Division of four regions in China.

The east region	Beijing, Tianjin, Hebei, Shanghai, Jiangsu, Zhejiang, Fujian, Shandong, Guangdong, Hainan
The middle region	Shanxi, Anhui, Jiangxi, Henan, Hubei, Hunan
The west region	Nei Mongol, Guangxi, Chongqing, Sichuan, Guizhou, Yunnan, Xizang, Shaanxi, Gansu, Qinghai, Ningxia, Xinjiang
The northeast region	Liaoning, Jilin, Heilongjiang

) and use them to construct a spatial model and spatially decompose their effects. The results are shown in

Table 11. Four region results of regional spatial effects.

	The East Region		The Middle Region		The West Region		The Northeast Region
	Direct Effect	Indirect Effect	Direct Effect	Indirect Effect	Direct Effect	Indirect Effect	Direct Effect	Indirect Effect
Edu	5147.61 ***	−1580.27 ***	4431.50 ***	1954.39	188.27	−38.31	5718.68 *	608.74
K	0.07 ***	−0.08 ***	0.08 **	−0.01	0.06 ***	−0.01 **	0.05 *	0.01
L	4.19 ***	2.17 **	1.80 *	−3.75 *	1.22 ***	−0.22 **	44.39 ***	19.15 **
STR	−10.01	6.40	1684.11	2,9899.97	−342.49	14.46	283.44 ***	260.87 **
FDI	0.09	0.13	1.34	1.04	1.35 ***	−0.25 **	−0.63	0.07
PPAT	0.05 ***	0.01	0.03	0.16 *	0.13 ***	−0.02 **	−0.30 ***	0.03
Gov	2.90 ***	2.42 **	6.50 ***	5.34 **	0.83 ***	0.97 ***	−0.35	0.04

Note: *** represents p ≤ 0.01; ** represents 0.01 < p ≤ 0.05; * represents 0.05 < p ≤ 0.1.

.

The results in Table 11 show that education development significantly promotes economic growth in the eastern, central, and northeastern regions; the effect of improving the education level on the western region’s economic development has not reached the expected effect, mainly because it is mostly a remote mountainous area and a gathering area for ethnic minorities. Education development is affected by factors such as transportation, residents’ awareness, and financial investment. The education development level has a negative spillover effect only in the eastern region, which has a highly developed economy and strong internal competition. Improving the education level, especially the concentration of high-quality talent, is key to improving human capital, which is also an important factor influencing China’s economic growth.

Both capital and labor force show significant regional promotion effects, indicating that they are the main drivers of regional economic growth. However, capital has a significant negative spillover between eastern and western regions. This suggests that more attention should be paid to inter-regional coordination and cooperative development rather than increasing inter-regional competitiveness. A win–win situation is more suitable for inter-regional development, especially in western regions, where the labor force also shows significant negative spillover.

The industrial structure has shown a significant boost and positive spillover effect only in the northeast. As China’s historical industrial base, this region’s economic development primarily relies on the development of secondary industries. However, with the recent policy of industrial revitalization, the industrial structure of the northeast region has gradually changed, dominated by tertiary industries, and the industrial structure has been upgraded and developed. As a result, its role in economic promotion has been increasing.

Foreign investment has also significantly contributed to economic growth only in the western region. This is because the western region is one of China’s less developed economic regions, where capital investment may be lacking. The amount of foreign investment can compensate for the lack of capital investment to a certain extent. With the western region’s unique resources and development status, foreign investment’s advanced management capabilities and technological innovation promote regional economic development [56]. However, it is also important to note the threshold effect of foreign investment. In the initial stage of economic opening, foreign trade promotes economic growth, and after exceeding a specific threshold, further increases in trade openness may reduce the economic growth rate [57]. Therefore, more attention should be paid to quality in the subsequent selection of foreign investment.

The innovation level has also not had the expected effect in the central and northeastern regions. Generally, locations with stronger economic development have strong innovation dynamics. New industries and economies grow rapidly, providing a new and strong impetus for economic development, a smooth transformation of old and new dynamics, and more robust and “resilient economic growth” [58], such as the developed coastal provinces and cities in the Yangtze River and Pearl River Delta regions, which have taken the lead in adaptation. For example, the developed coastal provinces and cities in these regions have led adaptation to leadership in the new national economic normal. They have nurtured new growth momentum in the development transition. The characteristics of the new economic normal, with “speed change, structural optimization, and power transformation” as the core, are becoming increasingly obvious; central region economic growth at this stage remains dependent on the pull of capital and labor, and capital investment has been gradual. Recently, however, capital investment has been increasing. Against the background of the new normal, the traditional path of economic growth driven by strong factor inputs and high energy consumption is difficult to adapt to the new development situation and requirements. At this stage, the key to economic development is to transform the mode of economic development and increase innovation capacity to enhance the proportion of the new economy, new industry, and new kinetic energy in the traditional industry. The historical industrial areas in the northeast have not entered the new normal of development, with the old kinetic energy accelerating decline and the new kinetic energy remaining unformed [58]. Pre-innovation capital investment is time-lagged and has a certain inhibitory effect on economic growth [59], but technological innovation remains a key factor.

Government intervention causes spatial differences in economic growth, but the data show less significance in the northeast. Local governments are likely to have different economic effects in driving economic development [60]. In the early stage of “revitalizing the old industrial bases in Northeast China,” the “three strong and three weak” states of strong government and weak enterprises, strong government and weak market, and strong government and weak society often appeared in Northeast China [61]. Still, as the government continued to delegate powers to lower levels, the “three strong and three weak” state was improved. Although fiscal spending in the northeast has not had a significant effect thus far, government actions such as infrastructure construction and industrial restructuring will provide a strong impetus for subsequent economic development [62,63].

5. Discussion and Conclusions

5.1. Discussion

From the perspective of space, using spatial autocorrelation analysis and spatial Dubin model, this study empirically analyzes the impact of China’s education development and economic growth by combining macro and micro research paradigms, and explores the mechanism of China’s education level improvement on economic growth, which has certain theoretical and practical significance. Compared with previous scholars’ research, it has certain innovation. In terms of theoretical significance, this study introduces the element of spatial impact as a factor to discuss the influence of education on economy, providing a new research perspective, and verifying that improvement of education has significantly different impacts on economic growth under different spatial conditions. This is conducive to enriching the theoretical achievements of the research on the relationship between education and economy, expanding the scope of research in this field, and making people more dialectically and multi-dimensionally view the relationship between the improvement of education level and economic growth. In terms of practical significance, this study uses the spatial panel data of 31 provinces in China from 2011 to 2020 as the data support to measure the education level, education equity and average years of education of the population in different regions of China, analyzes the different impact of the improvement of education level in different regions on economic growth, promotes the optimal allocation of regional resources and education reform of relevant government departments, and enhances the quality of education in China. Balanced and coordinated development provides important data support and suggestions. In fact, the unbalanced and uncoordinated spatial contradiction between education and economic development is a common issue in developing countries. This empirical research and analysis of China, which is a typical case, can provide inspiration for economic and educational reform in other developing countries.

5.1.1. There Is Spatial Correlation in the Improvement of China’s Education Level

The improvement of education level in China has a significant spatial correlation nationwide. When improving the education level of each province, we cannot ignore the influence among regions. How to focus on improving the level of spatial correlation among provinces and cities, creating more spatial spillover paths, and promoting the education level of northwest regions and southwest ethnic minority regions with the help of the east and middle regions that have high education levels is the focus of our attention. On the one hand, it is necessary to establish and improve the investment guarantee mechanism of educational resources, improve the transfer payment of educational resources allocation, and promote the balanced development of regional education. There is a large imbalance in the level of regional education development in China [64]. The regional imbalance of education development may further aggravate the imbalance of regional talent supply, thus hindering regional economic coordination and social equity development [65]. Therefore, the central government of China should further increase the support for low level education, improve the transfer payment and other means of education resource allocation, and adjust the current policy of China’s education resource allocation and regional positioning to promote the positive interaction between regional education development and local economy. On the other hand, the government should strengthen the policy guidance and funding for the coordinated development of education among regions. The government can issue relevant policy documents to encourage schools, enterprises, and organizations to actively participate in the development of education in regions with low education levels, increasing social and economic support for them, so as to promote the coordinated development of education among regions.

5.1.2. Education Improvement Has Not Played a Significant role in Promoting Economic Development in the Western Region

Through the empirical research, we can see that the impact of education improvement on the economy is similar to the research of many scholars, but the spatial effect on economic growth in different regions is different, especially in Northwest China; the impact of education improvement on economic growth is not significant. Since the beginning of the 21st century, China has paid more attention to education than ever before. The state has successively introduced the Plan for Revitalizing Higher Education in the Central and Western Regions (2012–2020), The Guiding Opinions of the General Office of the State Council on Accelerating the Development of Education in the Central and Western Regions, the Several Opinions on Revitalizing Higher Education in the Central and Western Regions in the New Era, and other policies, which has vigorously promoted the improvement of education in the western regions. With the implementation of these measures and policies, the average length of schooling of the population aged six and above in the western region has basically exceeded 9 years, and the education quality has been significantly improved. However, the loss of high-quality and high-level talents in the western region is relatively serious, which directly affects the role of education development in promoting economic growth. High-level talents are characterized by extensive loss, more elites, youth, and more key positions [66]. Therefore, to change the status quo that the education development in the western region does not significantly promote economic growth, we need to start from two aspects. On the one hand, we should further optimize the industrial structure of the western region. The optimization and upgrading of industrial structure is the key to the gathering of highly-educated scientific and technological talents [67]. However, compared with the east, the industrial structure in the west has been solidified for a long time, and it is unable to meet the market demand in a timely manner, leading to the loss of a large number of graduates trained by western universities [68]. Therefore, the western region should actively grasp the national support strategy, strive to optimize the industrial structure, and provide sufficient space and high-quality opportunities for the return of human capital. On the other hand, we should vigorously develop vocational education. For the western region, the development of vocational education is the internal driving force to promote regional economic development, the inevitable choice to eliminate the structural contradiction of labor force, the important cornerstone of building a modern education system, and an effective way to maintain social security and stability [69]. The relatively low income and relatively high incidence of poverty in the western region also determine the necessity of developing vocational education [54]. The brain drain in the western region makes most of the talents employed in the region are junior and high school graduates. The development of vocational education can provide them with “skills” and can also serve the economic development of the region better, thus improving the role played by education in promoting the economy.

5.1.3. Study limitations

This study also has limitations. There are many factors influencing economic growth. However, some factors based on data limitations were not considered, such as the agglomeration level of the economy, local government policies, and natural resources. Additionally, the measurement indicators for education level improvement were not sufficiently comprehensive, and the influence of highly-qualified personnel with master’s degrees and above on economic development was not considered. This study also did not consider the degree of difference between education levels on economic promotion. Therefore, in subsequent studies, researchers are advised to find more perfect indicators for improving education levels and to divest the promotion effect of different education levels on the economy.

5.2. Conclusions

This study used spatial panel data from 31 Chinese provinces in 2011–2020 to empirically analyze the relationship between education development and economic growth in China. The primary conclusions are the following. First, regarding the spatial correlation of education development levels, improving education levels in Chinese provinces mutually influences and promotes nationwide economic development. The spatial correlation of improving education levels also appears to be strong, with higher-level provinces generally showing greater clustering. Second, regarding education’s promotion effect on the economy, the improvement of the education level has a significant effect on China’s economic growth and a significant negative spillover effect. Apart from the education level’s less significant promotion effect on the western economy, it significantly promotes economic growth in the other three major regions and has a significant negative effect on the eastern region. Third, capital and labor remain the main drivers of economic growth. The advanced industrial structure demonstrates a significant promotion effect on the northeastern region with a significant positive spillover. Fourth, foreign investment only shows a positive promotion effect on the western region but with a negative spillover effect. The increase in innovation level has different significant effects on different regions, and the promotion effect is more significant in the eastern and western regions. Finally, government intervention only fails to achieve the expected promotion effect in the northeastern region.