The Impact of the Digital Economy on Educational Income Inequity: Evidence from Household Survey in China

Abstract

The rapid development of the digital economy has brought profound changes related to income inequity, while its impact on educational income inequity remains unclear, and this paper attempts to fill the gap. The purpose of this paper is to investigate the impact of the digital economy on educational income inequity. Using data from the Chinese General Social Survey, we employ the Theil index to measure educational income inequity, capturing income disparities associated with education. Our findings, based on a two-way fixed effects model, suggest that the digital economy mitigates educational income inequity in China, with a stronger effect at the family level than at the individual level. Mechanism analysis indicates that this reduction is primarily driven by the positive economy effects of digitalization. Additionally, the digital economy not only narrows disparities in labor income but also enhances equity in opportunity and effort. Finally, the effect of the digital economy on educational income inequity is more pronounced in cities with stronger economic foundations, better public services, and more advanced telecommunications infrastructure.

1. Introduction

Income inequity remains a persistent and critical issue in economic discourse. Recent economic development and technological advancements have spurred household income growth. However, this growth has also contributed to changing income inequity. As shown in Figure 1 and Figure 2, although the Gini coefficient and the urban–rural income ratio have gradually stabilized, the disposable income across five distinct residential groups from 2016 to 2022 reveals persistent income disparities, despite overall increases in disposable income. These trends indicate that income inequity is an evolving phenomenon.

Education plays a crucial role in determining income and income inequity. The expansion of education has been supported by policies such as the popularization of compulsory education and the expansion of college entrance examination enrollment in China. However, as the direct manifestation of education’s return, the question of whether expanding education leads to a reduction in income inequity remains unresolved [1,2,3]. The relationship between education and income inequity, which this paper terms educational income inequity (EII), warrants further exploration. As the digital economy (DE) continues to expand, the relationship between education and income inequity is likely to become more complex, presenting new challenges and changes. This paper aims to explore this evolving dynamic.

Inequity is often measured within distinct groups, making the definition of these groups essential. Research on income inequity, including the urban–rural income gap, frequently uses hukou as a grouping variable. However, despite the growing importance of education, there is limited discussion on the relationship with income inequity. This paper defines this relationship as EII, which refers to income disparities resulting from differences in educational attainment, similar to urban–rural income inequity. Specifically, EII captures income disparities among residents caused by educational divergence within cities. This framework allows for a quantifiable measure of how education impacts income inequity. To fully explore this impact, both individual educational attainment and parental education levels are considered. Parental education plays a significant role in children’s educational outcomes, and as such, individual education cannot be fully understood without considering the family background. Therefore, comparing individual education and family education provides additional insights into the impact of the DE on EII.

The DE has become a crucial driver of high-quality economic growth in China [4,5]. In recent years, the DE has gained significant attention, marked by milestones such as the full integration of the industrial internet across major sectors by 2023. Additionally, 5G applications directly contributed CNY 1.86 trillion to total economic output (Data from the reports in http://www.gov.cn/, accessed on 3 November 2024). As these data demonstrate, the DE has significantly influenced household incomes, making the relationship between the DE and income inequity an important topic of inquiry [6,7]. While the DE offers numerous benefits, digital literacy is essential for individuals to fully leverage its potential [8]. However, the digital divide remains a significant barrier, preventing some individuals from benefiting from DE. Digital skills are strongly linked to educational attainment, which creates disparities in how individuals benefit from the DE [9]. Therefore, education, income inequity, and the DE are deeply intertwined. However, limited studies have focused on the impact of the DE on EII.

Given the limited literature directly addressing the impact of the DE on EII, existing studies have only occasionally noted that education or skill levels may mediate the effects of the DE on income inequity. Furthermore, the relationship between the DE and income inequity remains inconclusive, largely due to complex factors such as the Matthew Effect, which is closely linked to digital skills related to education [10,11]. Although the body of research on the DE and income inequity is expanding, much of it focuses on disparities driven by urban–rural divides or gender differences [12,13,14]. Building on these studies, there remains a need for a more systematic investigation into how education influences the heterogeneity of the DE’s effects on income inequity. In light of these discussions, this paper introduces education as a key grouping factor for income inequity and focuses on the DE’s effect on EII. In the context of the DE’s rapid expansion, understanding its impact on education-driven income inequity is vital for informing policy decisions, particularly in promoting broader access to education as a means to achieve the goal of common prosperity.

This paper examines how the DE within macro-regions affects micro-level EII among residents in cities in China, using a two-way fixed effects model with panel data from China. The study contributes to the literature introduces in several ways. First, it presents the relationship between education and income inequity, termed EII, through the Theil index, with education as the key grouping variable. This method builds on the approach used to measure urban–rural income inequity and establishes a foundational framework for further analysis. Second, the study incorporates both individual and household characteristics in measuring EII. This approach captures the dual influences of public education systems and family background, allowing for a more comprehensive assessment. By comparing the impact of the DE on both forms of EII, the analysis indirectly reveals the transmission channels through which the DE operates. Lastly, the study identifies the underlying causes of inequity as significant mechanisms through which the DE affects EII. By decomposing EII into income and inequity components, this study introduces a novel analytical perspective that departs from existing literature. Furthermore, empirical evidence is provided to show that the DE influences EII by generating effects on employment, income, and capital flows.

The structure of this article is organized as follows: Section 2 reviews relevant studies on the DE and EII, proposing corresponding hypotheses. Section 3 focuses on model construction. Section 4 presents the results and interpretations. Section 5 offers analyses and results of mechanisms. Section 6 discusses heterogeneity analysis. Finally, Section 7 summarizes discussion and a brief conclusion.

2. Literature Review and Hypotheses

2.1. The Meaning and Measurement of EII

The calculation of EII is inherently based on income inequity. Existing studies measure income inequity using various methods, including the urban–rural income gap, inequity index, and others [4,15]. Some scholars have aimed for greater precision by adjusting for income distribution and optimizing models [6,16]. The Gini coefficient and Theil index are the most commonly used as tools to measure income inequity [17,18,19].

Methodologically, the Gini coefficient is based on cumulative differences between income and population ratios sorted by income levels [11], while the Theil index calculates entropy changes based on income and population ratios within distinct groups [12,20]. The Gini coefficient represents the proportion of total income attributed to unequal distribution across the population. In contrast, the Theil index captures both within-group and between-group contributions to overall inequity.

EII measures income inequity due to education. Education, both at the individual and family levels, is a key factor in determining wages. Family influence cannot be overlooked. On one hand, intergenerational education shows a strong link between father and child. On the other hand, a father’s education is related to genetic factors or abilities that strongly correlate with children’s educational attainment. In China’s education system, public and family services of education work together. Public education ensures basic fairness in educational opportunities, while family education preferences, resources, and capabilities influence further educational development. Thus, EII is calculated considering both individual and family perspectives. EII from an individual perspective directly reflects the impact of one’s own educational attainment on income, while EII in relation to family captures income disparities among offspring that are driven by the educational attainment of their parents. The latter highlights the influence of educational factors beyond individual effort, emphasizing the role of inherited advantages in shaping income outcomes.

The Theil index is used to calculate EII by considering urban–rural income inequity, with education as the grouping factor. This method is sound for measuring regional EII. According to Roemer’s framework on equality of opportunity, inequity arises from the following two sources: inequalities of environment (opportunity) and effort, consistent with the “two-fold distinction” concept [21]. The relationship between education and income is not simply linear. Regions, urban–rural differences, and occupation can lead to income inequity even among individuals with the same education level. Higher education does not necessarily result in higher income.

The decomposition of EII helps reveal the nature of inequity. Inequity of effort refers to inequity arising from individual effort within the same educational group. Inequity of opportunity pertains to inequity due to environmental differences across groups, beyond individual effort [22]. This method offers two main advantages: it reasonably shows income differences caused by education, and it avoids errors from group regression comparisons, thus improving the reliability of the results.

2.2. The Impact of the DE on EII

To comprehensively review existing literature, DE research covers a wide range of variables, including the DE itself, Digital Inclusive Finance [23], the internet [9,15], and inclusive finance [24]. Given the limited direct research on the relationship between the DE and EII, and since EII in this study is based on income inequity, this literature review primarily focuses on the impact of the DE on income inequity, particularly in studies involving education or skills.

EII, based on income inequity with education as the grouping variable, measures inequity attributable to education. Before examining the impact of the DE on EII, it is useful to consider how the DE affects income inequity. Current research on this topic can be categorized into two main approaches.

The first category uses an indirect approach. It starts with changes in actual income and examines how various factors affect income across different groups, thereby influencing income inequity [6,7]. However, these factors’ impact varies across groups, meaning the DE affects some people but has little effect on others [4,25]. This discrepancy contributes to further income inequity, but the effect remains ambiguous, with both positive and negative outcomes. Education and skills are often considered key variables for analyzing the heterogeneous impact of DE. Generally, the effect of the DE on income inequity varies based on educational attainment [15,20,26].

The second category uses a direct approach, focusing on income disparities and examining specific factors. For instance, a U-shaped relationship is identified between the DE and the urban–rural income gap [27]. The financial inclusion system, combined with spatial factors, further exacerbates the urban–rural income divide [11]. Internet access contributes to widening income disparities [8], while broadband internet reduces the inter-city income gap [9]. Digital Inclusive Finance narrows the urban–rural income gap by reducing disparities in wage, property, and transfer income, though it has a limited impact on the net operating income gap between urban and rural areas [28]. To date, few studies have explicitly adopted the concept of EII to capture income disparities resulting from education. However, a substantial body of literature has demonstrated that the heterogeneity impact of the DE on income inequity is closely related to education or skill levels. These findings offer indirect yet valuable support for the foundation of this study.

Moreover, both individual effort and family background—particularly educational attainment within the family—have received considerable attention in discussions of educational outcomes. Notably, individual and household educational attainment are not linearly correlated [29]. In the context of China’s educational development, there has been a marked improvement in education levels from one generation to the next. However, income levels are not solely determined by education. Broader structural factors, such as social and economic progress, have contributed to overall income growth. Whether such progress has substantially altered the pattern of income inequity, however, remains an open question. Against this backdrop, comparing EII based on individual and family educational attainment is meaningful. Therefore, to comprehensively examine the impact of the DE on EII, this paper incorporates EII with both individual and family into the analytical framework.

H1. 

The DE may significantly reduce EII, but its impact may vary between family and individual.

2.3. The Economic Effect Brought by DE

DE has generated a wide range of positive economic effects. However, these effects exhibit significant heterogeneity across groups differentiated by education or skill levels, particularly regarding income inequity. For instance, the DE has been shown to promote employment among low-skilled workers, enhance income for high-skilled individuals, and strengthen entrepreneurial intentions among non-agricultural workers. These differentiated impacts, in turn, contribute to evolving patterns of income inequity. This heterogeneity forms the theoretical foundation that changes in income serve as a meaningful mechanism [12,13,14,30]. To investigate the relationship between the DE and EII, this paper focuses on three primary economic effects (displayed in Figure 3).

First, the DE promotes economic growth by opening new avenues for development, which boosts overall income levels [31,32]. As a new economic paradigm, the DE has become a key driver of China’s economic expansion, reflected in the rising GDP and resident incomes. Although educational attainment may not experience significant short-term improvements, the income growth spurred by the DE can directly affect EII.

Second, the DE has fostered the emergence of new industries and forms of employment, generating more job opportunities and reducing unemployment [23,33]. The development of the DE has facilitated industrial upgrading—particularly in sectors such as e-commerce—and introduced flexible work arrangements that transcend geographic and temporal constraints [34]. These developments have been especially beneficial for providing employment opportunities among low-skilled groups. Given the centrality of employment to economic development [30], the heterogeneous employment effects across demographic groups indirectly influence EII.

Finally, the DE has enhanced financial inclusion by expanding the accessibility and reach of online financial services [33]. Through the proliferation of internet-based financial products, the DE has lowered barriers to financial access and reduced information asymmetries in investment [35]. This expansion provides stronger financial support for entrepreneurial activities, thereby increasing income and potentially reshaping EII.

H2. 

The DE helps reduce EII by generating positive economic effects, including promoting employment, increasing income, and enhancing financial accessibility.

2.4. The Cause of EII

EII measures the inequity of income resulting from education. This implies the following three key relationships: education, income, and inequity. Education is deeply discussed in the context of EII. To explore how the DE affects EII, understanding the causes of EII in relation to income and inequity is important (displayed in Figure 3).

Income can broadly be classified into three categories as follows: labor income, property income, and transfer income. Labor income is linked to workers’ skill levels and employment status. Property income includes returns from assets such as interest, rents, dividends, and capital gains, typically associated with accumulated wealth. Transfer income refers to cash benefits or subsidies from government or social programs. All forms of income contribute to economic disparities and should be considered when discussing inequity [10,11]. However, some scholars [2,9] focus only on labor income or wages, excluding other income sources.

External factors influencing income inequity can be categorized into three types. First, differential factors arise from the living environment, such as region, urban–rural divide, and ethnicity [13]. Second, family background factors include family assets, parental education, and genetics [3]. Third, individual factors such as education, gender, and skills are also influential [14,36,37].

Based on Roemer’s framework, EII is divided into “opportunity” and “effort” components, aligning with the nature of the Theil index. Exploring the mechanisms through which the DE affects EII is meaningful. In studies of income inequity using other methods, opportunity inequity is influenced by external factors, such as urban–rural distinctions [11] and household characteristics [38]. Effort inequity is influenced by internal factors, such as income cohorts [10,38], education, and skills.

H3. 

The DE reduces EII by addressing its underlying causes, including disparities in income sources, and inequity of opportunity and effort.

2.5. The Heterogeneity Effect of EII Brought on by the DE

Education is a key factor in understanding how the DE affects income inequity. In some studies, education is identified as a critical variable explaining the heterogeneous effects of the DE on income inequity [15,20,26]. The effect of the DE is generally found to vary with digital skills. The digital divide exacerbates inequities among groups without digital skills. Moreover, the environment often mediates the impact of DE. For instance, urban groups tend to have better access to funding and employment opportunities compared to rural counterparts, even with the same educational background. However, there is a lack of research directly examining the impact of the DE on EII.

In cities or provinces, the effect of the DE on income inequity may vary based on characteristics such as economic development [20], urbanization [12], regions [4,11,15], and internet usage [8,9]. Similarly, the effect of the DE on EII may vary based on economic infrastructure, public services, and telecommunications services. In cities with well-developed economic foundations and public facilities, education levels tend to be higher. In cities with strong telecom industries, more people have internet access, and the impact of the DE increases accordingly. It is important to examine whether differences among cities lead to variations in the impact of the DE on EII, given the close relationship between EII, income, and education.

H4. 

The effect of the DE on EII varies with the level of economic infrastructure, public services, and telecommunications services in cities.

3. Materials and Methods

3.1. Model Design

The objective of this study is to analyze the impact and mechanisms through which the DE in cities influences EII among residents. Industrial structure, the number of enterprises, and the employment situation are closely interconnected [23], while financial pressure directly affects the level of infrastructure development. Population density is strongly associated with the demand for the DE, educational attainment, and competition in the labor market. Naturally, a wide range of factors are relevant to both the DE and EII. However, based on the foundation established by existing studies [13,23] and in line with the objectives of this paper, particular attention is given to these specific variables. Panel data at the city level are utilized to construct a two-way fixed effects model (Equation (1)).

$${Theil}_{it}={\beta }_0+{\beta }_1{DE}_{it-1}+{{\sum }_{j=1}^m{\beta }_{j}X}_{it}+{\theta }_i+{\phi }_t+{\epsilon }_{it}$$

(1)

In this model, ${Theil}_{it}$ denotes EII in cities, which includes education factors both family and individual; ${DE}_{it-1}$ denotes the DE in cities, measured as a comprehensive index; i and t denote city and year, respectively; $X_{it}$ represents the set of control variables, which include population density, the proportion of employees in the secondary industry, the number of industrial enterprises above the designated size in the city, and financial pressure. The terms ${\theta }_i \mathrm{and} {\phi }_t$ denote the city and year fixed effects; ${\epsilon }_{it}$ denotes random error terms, clustered at the city level; and the coefficients ${\beta }_0,{\beta }_1,\mathrm{a}\mathrm{n}\mathrm{d} {\beta }_j$ are to be estimated. It is important to note that, in order to mitigate potential endogeneity issues arising from reverse causality, the DE is introduced into the model with a one-period lag.

3.2. Calculation of EII and the DE

To calculate the income inequity attributed to education, this paper uses the Theil index. Groups are based on the education levels of both family and individual. This approach allows for the decomposition of EII into “opportunity” and “effort” components. In addition, educational attainment in family is typically measured using the educational level of the household head (huzhu) [3]. In the Chinese context, this role is most commonly held by the father. The father’s education level serves as a proxy for family education, while the individual’s own education level measures individual education.

Family education is categorized into four groups as follows: illiterate, below or including compulsory education, high school education, and higher education. Individual education is classified into three groups as follows: below or including compulsory education, high school education, and higher education. There is no illiterate category for individual education.

$${Theil}_{\mu }={\sum }_{j=1}^m{p}_i\left({\displaystyle \frac{n_i}{\overline{X}}}\right)ln{\displaystyle \frac{n_i}{\overline{X}}}=theil(\overline{\mu })+{\sum }_{n=1}^N{p}_{m}theil\left({\mu }^m\right)$$

(2)

In this model, $\mu$ denotes different groups, classified according to the education levels of the family and individual;$p \mathrm{and} n$ denote the proportion and number of personnel per group; $i \mathrm{and} m$ denote members and groups of the set $\mu$; $\overline{X}$ is the mean income within each group; $theil\left(\overline{\mu }\right)$ reflects inter-group inequity, representing the inequity of opportunity; and $\sum _{m=1}^M{p}_{m}theil\left({\mu }^m\right)$ captures the inequity within each group, signifying the inequity of effort.

Given the diverse sources of income, total income is used in EII for both family and individual. Labor EII and non-labor EII are used to analyze the causes of income in the mechanism. Both family and individual education are calculated to explore the influence of education on income inequity.

The DE is a multifaceted variable. It includes infrastructure components such as internet access and evaluates its role in fostering economic growth through financial services and products. Accurately assessing regional DE levels is essential for understanding its impact on EII. This study constructs an indicator system (

Table 1. The indicator system of the DE.

Variable	Context
Internet penetration	Internet users per 100 people
Number of Internet practitioners	The proportion of computer service and software professionals
Interne-related output	Per capita total telecommunications services
Number of mobile Internet users	Number of mobile phone users per 100 people
DIF	The index of digital inclusion finance developed by Peking University

) based on existing literature. It draws on the framework developed by Zhao et al. [39], which constructs an indicator system using PAC for DE, focusing on internet infrastructure, information development, and the level of inclusive financial services. Principal component analysis is then used to evaluate DE levels across cities.

3.3. Data Source and Descriptive Statistics

The income data used in this study are sourced from the Chinese General Social Survey (CGSS) for the years 2012, 2013, and 2015. The CGSS is the earliest and most comprehensive continuous academic survey project in China. It collects data across multiple societal levels, including community, family, and individual. Using CGSS data allows for a more accurate measurement of EII in this study. However, due to limitations in city code availability, only data from 2012, 2013, and 2015 are used. Additional data are obtained from the City Statistical Yearbook and the Center for Digital Finance at Peking University.

Descriptive statistics for the main variables are presented in

Table 2. Descriptive statistics.

Variable		Context	N	Mean	SD	Min	Max
1	EII with family	The Theil index Calculated by fathers’ education groups	244	0.35	0.34	0.01	1.98
2	Labor EII with family		244	0.35	0.34	0.00	2.05
3	Non-labor EII with family		244	3.04	1.67	0.14	8.07
4	EII with individual	The Theil index Calculated by individual education groups	244	0.34	0.36	0.01	2.40
5	Labor EII with individual		244	0.34	0.34	0.01	2.23
6	Non-labor EII with individual		244	2.46	1.27	0.12	5.87
7	DE	Digital economy calculated by PAC	244	13,731.87	26,958.39	545.61	230,291.00
8	peomidu	Density of population	244	491.63	404.70	21.25	2648.11
9	erchanry	The proportion of employees in the secondary industry	244	44.24	12.41	4.46	76.23
10	qynum	Number of industrial enterprises above the designated size in the city	244	1539.70	1778.35	55.00	9962.00
11	yali1	Financial pressure	244	0.37	0.24	−0.37	0.90
12	tax	Per capita tax revenue (CNY)	243	5368.54	7763.08	−2045.35	87,990.42
13	loan	Per capita loan balance (CNY)	244	86,996.92	122,809.29	1004.58	978,258.56
14	empolyment	Employment rate	244	0.14	0.14	0.01	1.38
15	empolyment1	Proportion of employment in the information sector	244	0.00	0.01	0.00	0.05
16	wages	Average wage per employee (CNY)	244	43,737.00	12,369.89	19,176.07	103,400.41
17	gdpper	Per capita GDP (CNY)	244	44,028.12	25,369.12	8877.00	149,495.00
18	bookper	Books per capita in libraries	244	27.01	83.30	0.00	920.03
19	rodeper	Road area per capita (square meters)	244.	11.11	5.56	3.34	38.27
20	broadbandper	Households with broadband access per capita	244	0.17	0.17	0.00	1.33
21	telecommunicationper	Per capita telecommunication expenditure (CNY)	244	970.76	1227.00	15.89	13,976.82

Note: The data of EII here is the true value. yali1 (Financial pressure) = (Local general public budget expenditure-Local general public budget revenue)/Local general public budget expenditure. Role 1 and 4 are explained variables, role 7 is an explanatory variable, role 8–11 are control variables. Variables Role 2–3, 5–6, 11–17 are variables used to discuss the influence mechanism. Role 18–21 are variables involved in the heterogeneity analysis.

. The difference in EII between family and individual is relatively small. However, significant differences are observed between labor and non-labor income, both between family and individual. And the non-labor EII in relation to family is bigger than that in relation to just individuals. Additionally, there is notable heterogeneity in the levels of DE. The other variables in the model serve as control variables.

4. Results and Interpretations

4.1. Baseline Regression on the Impact from the DE on EII

The results from the two-way fixed effects model confirm that the DE can reduce EII. It has a direct effect on reducing overall income inequity (the result is shown in Appendix A,

Table A1. DE reduces income inequity.

	(1)
VARIABLES	Income Inequity
DE	−9.155 *
	(−1.82)
control	YES
city FE	YES
Year FE	YES
N	244
R-squared	0.518
Adjust R-squared	0.239

Note: *** p-value < 0.01, ** p-value < 0.05, * p-value < 0.1. Cluster standard errors at the city-level are presented in parentheses. The results of OLS is presented in the table. The calculation of income inequity follows Equation (2), with quartile income as groups. To demonstrate the coefficients, the Theil index is enlarged by 1,000,000 times in all regression.

). Given the small magnitude of the Theil index, the coefficients are scaled up for practical estimation, as shown in

Table 3. Basic regression.

	(1)	(2)	(3)	(4)	(5)
VARIABLES	EII with Family	EII with Individual	DE	EII with Family	EII with Individual
DE	−5.419 *	−7.327 **	-	−9.308 ***	−8.336 ***
	(−1.88)	(−2.56)	-	(−3.32)	(−3.51)
Iv1			−0.007 **
			(−2.23)
Iv2			0.003 ***
			(5.85)
peomidu	4691.382 ***	4832.006 ***	116.530 ***	5207.928 ***	4878.062 ***
	(5.25)	(5.29)	(7.77)	(5.75)	(5.42)
erchanry	−100.572	−4950.655	−170.634 *	−1421.323	−5067.700
	(−0.02)	(−1.07)	(−1.96)	(−0.31)	(−1.14)
qynum	−259.706	−250.860	0.456	−253.995	−206.096
	(−1.58)	(−1.42)	(0.22)	(−1.47)	(−1.20)
yali1	−437,822.616	−682,829.567	−2035.517	−478,093.004	−724,566.631
	(−1.32)	(−1.29)	(−0.37)	(−1.46)	(−1.38)
city FE	YES	YES	YES	YES	YES
Year FE	YES	YES	YES	YES	YES
N	244	244	242	242	242
R-squared	0.487	0.486	-	0.087	0.100
Adjust R-squared	0.190	0.189	-	0.0597	0.0735
First-stage F	-	-	-	18.317	18.317
Hansen J	-	-	-	0.989	0.159
Hansen J p-value	-	-	-	0.320	0.690

Note: *** p-value < 0.01, ** p-value < 0.05, * p-value < 0.1. Cluster standard errors at the city-level are presented in parentheses. To demonstrate the coefficients, the Theil index is enlarged by 1,000,000 times in all regression. Due to the absence of data for Zhangzhou, two observations are missing in columns (3)–(5). Columns (1)–(2) use OLS regression, and columns (3)–(5) use 2SLS method.

. According to the results in columns (1) and (2) of Table 3, the coefficient of the DE indicates that a one-unit increase in the DE is associated with a 5.419-unit reduction in family EII and a 7.327-unit reduction in individual EII. H1 has been confirmed. In practical terms, the mean DE is 13,731.87, with a standard deviation of 26,958.39. A change in the DE by one standard deviation from the mean would reduce family EII by 0.146 and individual EII by 0.198. The means of family and individual EII are 0.35 and 0.34, respectively. These findings are consistent with previous research [19,28].

4.2. Re-Examining the Effect of the DE on EII Based on Endogeneity Tests

Although a two-way fixed effects model is used and the DE is lagged by one period to mitigate reverse causality, endogeneity remains a concern. This issue is caused by omitted variables and measurement errors. To address this, two instrumental variables (IVs) are constructed. IV1 is the interaction term between the distance from a city to the provincial capital [40] and the average DE of other cities. IV2 is the interaction term between the historical number of banks in cities [41] and the average DE of other cities within the same province.

The number of banks in 1984 serves as a historical variable. It reflects past infrastructure and financial development but is not directly correlated with the current DE level. However, it laid the groundwork for the current state of DE. Additionally, urban development is generally correlated with proximity to provincial capitals. Cities closer to provincial capitals tend to have higher levels of development. To enhance the time variation in the IVs [42] and avoid excessive correlation, the average DE of other cities or cities within the same province is included in the construction of the IVs.

The 2SLS estimation method is used to test the baseline regression. The first-stage F-statistic in the 2SLS results is 18.371. The Hansen J test p-values are 0.320 and 0.690, indicating that both IVs are valid. The coefficients for the DE show minimal variation compared to those in columns (1) and (2) of Table 3. This suggests that, after addressing potential endogeneity, the DE significantly reduces EII. The effect is stronger for family EII than for individual EII. Therefore, H1 is still confirmed with the results shown in Table 3. Based on the results presented in Table 3, the DE significantly affects EII. Moreover, the magnitude of this effect differs between family and individual levels. Specifically, the impact of the DE is more pronounced on EII with family. These findings provide strong empirical support for Hypothesis 1. Consequently, a linkage can be established among DE, education, and income inequity.

4.3. Robustness Test on the Impact of the DE on EII

The robustness tests in this paper address potential biases in the calculation of EII and continue to use IVs to tackle endogeneity concerns. The results are in

Table 4. Robustness tests.

	(1)	(2)	(3)	(4)	(5)	(6)
	Shrink the Tail by 5%		Standardized		[0, 1] Range
VARIABLES	EII with Family	EII with Individual	EII with Family	EII with Individual	EII with Family	EII with Individual
DE	−8.218 ***	−6.741 ***	−29.254 ***	−21.733 ***	−6.991 ***	−5.529 ***
	(−3.67)	(−4.12)	(−3.55)	(−3.20)	(−3.40)	(−4.18)
control	YES	YES	YES	YES	YES	YES
city FE	YES	YES	YES	YES	YES	YES
Year FE	YES	YES	YES	YES	YES	YES
N	242	242	242	242	231	227
R-squared	0.080	0.121	0.090	0.099	0.023	0.053
Adjust R-squared	0.053	0.095	0.063	0.072	−0.007	0.023
First-stage F	18.317	18.317	18.317	18.317	18.887	19.147
Hansen J	0.484	0.287	0.655	0.032	0.198	0.238
Hansen J p-value	0.486	0.592	0.418	0.858	0.656	0.625

Note: *** p-value < 0.01, ** p-value < 0.05, * p-value < 0.1. Cluster standard errors at the city-level are presented in parentheses. The results of 2SLS methods with IVs are presented in the table. To demonstrate the coefficients, the Theil index is enlarged by 1,000,000 times in all regression. The standardized EII with family and individual in columns (3)–(4) has expanded by 10^6 times, to intuitively demonstrate the effect of DE on EII.

as follows. First, the income data from CGSS use education as a grouping variable, which may introduce errors in the EII calculation. Second, the population distribution is not uniform, potentially causing bias in the EII estimate [16]. Third, although the exact range for the Theil index is not defined [43], it is generally agreed that the inequity measure lies within the range [0, 1] [44].

To ensure robustness, several methods are applied. First, to mitigate the influence of extreme values, the top and bottom 5% of EII are truncated [26]. The empirical results show that the DE still significantly reduces EII, with a stronger effect on family EII than individual EII. Next, EII is standardized to reduce the impact of uneven population distribution. The results again demonstrate that the DE significantly lowers EII. Finally, samples with EII values outside the [0, 1] range are excluded to remove extreme outliers. The results confirm that the impact of the DE on EII remains significant.

5. Mechanisms for the Impact of the DE on EII

Based on the above analysis and assumptions, the impact mechanisms of the DE on EII are examined from the following three perspectives.

5.1. The Economic Effect of the DE

The development of the DE has generated various economic effects, such as promoting income, employment, and capital flows. Empirical evidence indicates that the DE has led to significant economic changes, reducing EII in

Table 5. The economic effect of DE.

	(1)	(2)	(3)	(4)	(5)	(6)
	Employment		Wages and Economy			Loan
VARIABLES	employment	employment1	wages	gdpper	tax
DE	0.019 **	0.002 ***	0.113 **	0.183 **	0.103 ***	0.899 ***
	(2.28)	(6.80)	(2.32)	(2.34)	(2.92)	(3.21)
control	YES	YES	YES	YES	YES	YES
city FE	YES	YES	YES	YES	YES	YES
Year FE	YES	YES	YES	YES	YES	YES
N	242	242	242	242	241	242
R-squared	0.414	0.682	0.133	0.250	0.070	0.518
Adjust R-squared	0.397	0.673	0.107	0.227	0.042	0.503
First-stage F	18.317	18.317	18.317	18.317	18.282	18.317
Hansen J	1.122	3.619	5.263	17.397	0.227	4.502
Hansen J p-value	0.290	0.057	0.022	0.000	0.634	0.034

Note: *** p-value < 0.01, ** p-value < 0.05, * p-value < 0.1. Cluster standard errors at the city-level are presented in parentheses. The results of 2SLS methods with IVs are presented in the table. To demonstrate the coefficients, the Theil index is enlarged by 1,000,000 times in all regression. The values of employment and employment1 in columns (1)–(2) has expanded by 10,000 times, to intuitively demonstrate the economy effect of DE.

. First, the DE can increase the employment rate, particularly in the information sector. This, in turn, raises household income and reduces EII [7,45]. Second, the DE boosts wages, increases tax revenues, and stimulates economic growth [32]. These effects, including direct wage increases and tax adjustments, help mitigate EII. Finally, the DE significantly increases loans, facilitating capital flows and liquidity, which further reduces EII [5,46]. H2 can be confirmed.

5.2. Labor and Non-Labor Income of EII

Income consists of both labor income from work and non-labor income such as rent and interest. Both types can contribute to EII, as shown in

Table 6. Labor and non-labor income inequity in EII.

	(1)	(2)	(3)	(4)
	EII with Family		EII with Individual
VARIABLES	Labor Income	Non-Labor Income	Labor Income	Non-Labor Income
DE	−9.054 ***	16.913	−6.002 ***	−15.200
	(−3.09)	(0.91)	(−2.81)	(−0.98)
control	YES	YES	YES	YES
city FE	YES	YES	YES	YES
Year FE	YES	YES	YES	YES
N	242	242	242	242
R-squared	0.080	0.002	0.077	0.069
Adjust R-squared	0.052	−0.028	0.049	0.041
First-stage F	18.317	18.317	18.317	18.317
Hansen J	2.288	0.947	1.055	0.102
Hansen J p-value	0.130	0.330	0.304	0.749

Note: *** p-value < 0.01, ** p-value < 0.05, * p-value < 0.1. Cluster standard errors at the city-level are presented in parentheses. The results of 2SLS methods with IVs are presented in the table. To demonstrate the coefficients, the Theil index is enlarged by 1,000,000 times in all regression.

. Analyzing the effects of the DE on labor and non-labor income helps clarify how the DE influences overall EII. The results in Table 5 show that the DE significantly reduces labor income-related EII, while its effect on non-labor income is not statistically significant. The reduction in inequity is more pronounced for family EII than individual EII. This suggests that, in terms of income composition, the DE primarily reduces EII by increasing labor income—specifically, by narrowing wage disparities within groups of the same education attainment. In contrast, non-labor income exhibits a weaker dependence on education. The DE has broken down spatial barriers, fostering new industries and job opportunities. This creates greater employment and income prospects for low-education groups [9], thereby increasing labor income and reducing EII. However, access to DE benefits still requires some educational attainment [27], making the impact on family EII stronger than on individual EII. Part of H3 can be confirmed.

5.3. Inequity of Effort and Opportunity

Empirical results show that the DE reduces both opportunity and effort inequalities in

Table 7. Inequity of opportunity and effort in EII.

	(1)	(2)	(3)	(4)
	EII with Family		EII with Individual
VARIABLES	Inequity of Opportunity	Inequity of Effort	Inequity of Opportunity	Inequity of Effort
DE	−4.677 ***	−4.631 ***	−2.469 *	−5.867 ***
	(−2.64)	(−2.80)	(−1.95)	(−2.80)
control	YES	YES	YES	YES
city FE	YES	YES	YES	YES
Year FE	YES	YES	YES	YES
N	242	242	242	242
R-squared	0.101	0.057	0.105	0.104
Adjust R-squared	0.074	0.029	0.078	0.078
First-stage F	18.317	18.317	18.317	18.317
Hansen J	1.507	0.209	0.651	0.004
Hansen J p-value	0.220	0.647	0.420	0.948

Note: *** p-value < 0.01, ** p-value < 0.05, * p-value < 0.1. Cluster standard errors at the city-level are presented in parentheses. The results of 2SLS methods with IVs are presented in the table. To demonstrate the coefficients, the Theil index is enlarged by 1,000,000 times in all regression.

. Specifically, for individuals, the DE has a greater impact on reducing inequity of effort than inequity of opportunity. For families, however, the effects on both opportunity and effort inequities are nearly identical. Inequity of effort is linked to inequalities individuals may overcome through personal effort. Inequity of opportunity stems from factors beyond individual control, such as family background and regional disparities [29].

From the individual perspective, educational attainment often reflects digital skills, which are crucial for benefiting from DE. Thus, the DE has a stronger effect on inequity of effort than on inequity of opportunity for individuals. From the family perspective, EII reflects the living and growth conditions of offspring [47]. The development of the DE provides greater opportunities for social mobility, particularly for groups with limited family backgrounds. Hence, the DE significantly impacts both opportunity and effort inequities for families. The whole of H3 can be confirmed.

6. Heterogeneity Analysis

DE and EII are closely tied to city development. As a new economic development system, China’s DE has quickly emerged, supported by a solid economic foundation. Given the uneven development of cities in China [24], it is crucial to examine whether the impact of the DE on EII is influenced by this disparity as H4. This approach better reflects China’s unique conditions. Therefore, this paper conducts a heterogeneous analysis of the effects of the DE on EII. It examines the following three perspectives: economic infrastructure, public services, and telecommunications services. The goal is to provide theoretical insights based on the trends in city development across China.

6.1. Economic Infrastructure

There is significant unevenness in the economic foundations of cities. Since both the DE and EII are closely linked to economic infrastructure [17], this paper explores heterogeneity from this perspective. Generally, cities with stronger economic foundations experience faster DE development. These cities often have more developed labor markets, higher industrial demand, and better educational attainment. All these factors can influence how the DE shapes EII. This paper uses geographical location [48] and urbanization level [12] as proxies for economic infrastructure. The empirical analysis in

Table 8. Comparison of Geographical Location.

	(1)	(2)	(3)	(4)	(5)	(6)
	EII with Family			EII with Individual
Groups	East	Middle	West	East	Middle	West
DE	−12.576 ***	9.999	−29.828	−12.127 ***	−108.568	−44.040
	(−3.13)	(0.07)	(−0.74)	(−3.15)	(−0.76)	(−1.13)
control	YES	YES	YES	YES	YES	YES
city FE	YES	YES	YES	YES	YES	YES
Year FE	YES	YES	YES	YES	YES	YES
N	74	96	72	74	96	72
R-squared	0.221	0.046	−0.128	0.155	−0.221	−0.033
Adjust R-squared	0.140	−0.029	−0.249	0.067	−0.317	−0.144
First-stage F	15.141	0.629	1.896	15.141	0.629	1.896
Hansen J	0.392	0.193	0.316	0.773	3.211	0.217
Hansen J p-value	0.531	0.660	0.574	0.379	0.073	0.641

Note: *** p-value < 0.01, ** p-value < 0.05, * p-value < 0.1. Cluster standard errors at the city-level are presented in parentheses. The results of 2SLS methods with IVs are presented in the table. To demonstrate the coefficients, the Theil index is enlarged by 1,000,000 times in all regression. The eastern provinces include Beijing, Tianjin, Hebei, Shanghai, Jiangsu, Zhejiang, Fujian, Shandong, Guangdong, and Hainan; the central provinces encompass Shanxi, Anhui, Jiangxi, Henan, Hubei, and Hunan; and the western provinces consist of Inner Mongolia, Guangxi, Chongqing, Sichuan, Guizhou, Yunnan, Shaanxi, Gansu, Qinghai, and Ningxia. The regions of Xinjiang, Tibet, Hong Kong, and Macau are excluded from the scope of this discussion.

finds that the DE significantly reduces EII in eastern regions, but no significant impact is identified in central and western regions [27,49]. A change in the DE by one standard deviation from the mean would reduce family EII by 0.340 and individual EII by 0.327 in eastern regions. Furthermore, the impact of the DE on EII is more pronounced in cities with higher urbanization rates in

Table 9. Comparison of Urbanization Development Levels.

	(1)	(2)	(3)	(4)
	EII with Family		EII with Individual
Groups	High	Low	High	Low
DE	−12.679 ***	−5.656	−8.828 **	−7.830
	(−3.28)	(−0.94)	(−2.51)	(−1.15)
control	YES	YES	YES	YES
city FE	YES	YES	YES	YES
Year FE	YES	YES	YES	YES
N	122	120	122	120
R-squared	0.175	0.142	0.124	0.208
Adjust R-squared	0.125	0.089	0.070	0.159
First-stage F	238.437	30.485	238.437	30.485
Hansen J	1.629	0.793	2.291	0.003
Hansen J p-value	0.202	0.373	0.130	0.956

Note: *** p-value < 0.01, ** p-value < 0.05, * p-value < 0.1. Cluster standard errors at the city-level are presented in parentheses. The results of 2SLS methods with IVs are presented in the table. To demonstrate the coefficients, the Theil index is enlarged by 1,000,000 times in all regression. A city is classified as having a high level of urbanization if its urbanization rate exceeds the median urbanization rate of all cities in that year; otherwise, it is considered to have a low level of urbanization.

. A change in the DE by one standard deviation from the mean would reduce family EII by 0.342 and individual EII by 0.238 in cities with higher urbanization rates. The cities with stronger economic foundations, higher DE levels and better digital technology adoption lead to a more significant reduction in EII. This conclusion is consistent with H4.

6.2. Public Services

Significant disparities exist in the level of public services across cities due to factors such as fiscal capacity and resident population. Public services include both soft and hard infrastructure components [50,51]. Soft infrastructure refers to services like libraries, hospitals, and science museums or stadiums [52], while hard infrastructure includes road networks, park green spaces, and other physical assets. Cities with more resources typically offer higher public service levels. This, in turn, creates more employment opportunities, investment returns, and overall well-being for residents, influencing EII. In this study, per capita library collections represent soft public services, while per capita road area is used as a proxy for hard public services. The empirical results show that in cities with higher public service levels, the DE significantly reduces EII. A change in the DE by one standard deviation from the mean would reduce family EII by 0.257 and individual EII by 0.357 in cities with higher soft public service levels in

Table 10. Comparison of Soft Public Service Levels.

	(1)	(2)	(3)	(4)
	EII with Family		EII with Individual
Groups	High	Low	High	Low
DE	−9.516 *	207.576	−13.242 **	111.387
	(−1.93)	(1.20)	(−2.62)	(0.86)
control	YES	YES	YES	YES
city FE	YES	YES	YES	YES
Year FE	YES	YES	YES	YES
N	121	121	121	121
R-squared	0.137	−0.703	0.149	−0.353
Adjust R-squared	0.084	−0.808	0.097	−0.436
First-stage F	14.458	2.194	14.458	2.194
Hansen J	0.003	2.986	0.001	0.046
Hansen J p-value	0.959	0.084	0.972	0.830

Note: *** p-value < 0.01, ** p-value < 0.05, * p-value < 0.1. Cluster standard errors at the city-level are presented in parentheses. The results of 2SLS methods with IVs are presented in the table. To demonstrate the coefficients, the Theil index is enlarged by 1,000,000 times in all regression. Per capita library collections are used as one proxy for the soft public service level of a city. A city is classified as having a high level of soft public services if its per capita library holdings exceed the median per capita library holdings of all cities in that year; otherwise, it is considered to have a low level of soft public services.

. A change in the DE by one standard deviation from the mean would reduce family EII by 0.248 and individual EII by 0.249 in cities with higher hard public service levels in

Table 11. Comparison of Hard Public Service Levels.

	(1)	(2)	(3)	(4)
	EII with Family		EII with Individual
Groups	High	Low	High	Low
DE	−9.208 ***	−14.129	−9.253 ***	−13.352
	(−3.07)	(−0.87)	(−3.21)	(−0.88)
control	YES	YES	YES	YES
city FE	YES	YES	YES	YES
Year FE	YES	YES	YES	YES
N	121	121	121	121
R-squared	0.227	0.108	0.096	0.159
Adjust R-squared	0.179	0.053	0.040	0.107
First-stage F	38.871	7.150	38.871	7.150
Hansen J	0.002	0.157	1.252	0.007
Hansen J p-value	0.968	0.692	0.263	0.934

Note: *** p-value < 0.01, ** p-value < 0.05, * p-value < 0.1. Cluster standard errors at the city-level are presented in parentheses. The results of 2SLS methods with IVs are presented in the table. To demonstrate the coefficients, the Theil index is enlarged by 1,000,000 times in all regression. Per capita road area serves as one proxy for the hard public service level of a city. A city is classified as having a high level of hard public services if its per capita road area exceeds the median per capita road area of all cities in that year; otherwise, it is considered to have a low level of hard public services.

. This suggests that the DE’s impact may be subject to a threshold effect based on the city’s public service level. This conclusion is consistent with H4.

6.3. Telecommunications Services

Telecommunications services are a key industry supporting DE development [17]. Broadband access, which reflects household internet usage, is a crucial indicator. Variations in telecommunications services can, therefore, influence the impact of DE. Grouping cities based on differences in telecommunications infrastructure to analyze the DE’s effect on EII is of practical relevance. For variable selection, per capita telecommunications usage is used as a proxy for telecommunications services. Additionally, the number of broadband subscriptions per capita serves as a supplementary proxy. The empirical results show that when telecommunications services are higher, the DE’s impact on EII is more pronounced. The effect is stronger for individual EII. A change in the DE by one standard deviation from the mean would reduce family EII by 0.316 and individual EII by 0.340 in cites with higher telecommunications service levels in

Table 12. Comparison of Telecommunications Service Levels.

	(1)	(2)	(3)	(4)
	EII with Family		EII with Individual
Groups	High	Low	High	Low
DE	−11.707 **	135.928	−12.599 **	100.437
	(−2.51)	(0.57)	(−2.49)	(0.67)
control	YES	YES	YES	YES
city FE	YES	YES	YES	YES
Year FE	YES	YES	YES	YES
N	121	121	121	121
R-squared	0.163	−0.269	0.189	−0.191
Adjust R-squared	0.111	−0.347	0.139	−0.265
First-stage F	17.192	2.345	17.192	2.345
Hansen J	0.188	3.145	0.220	0.161
Hansen J p-value	0.664	0.076	0.639	0.689

Note: *** p-value < 0.01, ** p-value < 0.05, * p-value < 0.1. Cluster standard errors at the city-level are presented in parentheses. The results of 2SLS methods with IVs are presented in the table. To demonstrate the coefficients, the Theil index is enlarged by 1,000,000 times in all regression. A city is classified as having a high level of telecommunications services if its per capita telecommunications usage exceeds the median per capita telecommunications usage of all cities in that year; otherwise, it is considered to have a low level of telecommunications services.

. A change in the DE by one standard deviation from the mean would reduce family EII by 0.252 and individual EII by 0.352 in cites with higher broadband access, as shown in in

Table 13. Comparison of Broadband Access.

	(1)	(2)	(3)	(4)
	EII with Family		EII with Individual
Groups	High	Low	High	Low
DE	−9.355 *	177.016	−13.041 **	56.602
	(−1.89)	(1.16)	(−2.44)	(0.64)
control	YES	YES	YES	YES
city FE	YES	YES	YES	YES
Year FE	YES	YES	YES	YES
N	121	121	121	121
R-squared	0.140	−0.548	0.143	−0.070
Adjust R-squared	0.087	−0.644	0.090	−0.135
First-stage F	14.411	2.259	14.411	2.259
Hansen J	0.005	1.400	0.000	1.006
Hansen J p-value	0.946	0.237	0.992	0.316

Note: *** p-value < 0.01, ** p-value < 0.05, * p-value < 0.1. Cluster standard errors at the city-level are presented in parentheses. The results of 2SLS methods with IVs are presented in the table. To demonstrate the coefficients, the Theil index is enlarged by 1,000,000 times in all regression. A city is classified as having a high level of broadband access if its per capita broadband subscriptions exceed the median per capita broadband subscriptions of all cities in that year; otherwise, it is considered to have a low level of broadband access.

. This suggests that broader the DE coverage leads to a more significant impact. This conclusion is consistent with H4.

7. Discussion and Conclusions

7.1. Discussion

This paper innovatively incorporates education from both household and individual perspectives into the analysis of EII. It offers a more comprehensive understanding of income inequity resulting from education. It directly examines the impact of the DE on EII by analyzing the interplay between DE, education, and income inequity. Existing studies have primarily focused on two approaches. For example, the impact of the DE and (Non-)Cognitive Ability on labor income is analyzed using interaction models [26], with findings indicating that the DE has a significantly positive effect on the labor income of individuals with higher cognitive ability. Additionally, the effect of the DE on income inequity varies across groups with different education levels [9,20]. Although these studies do not directly address the relationship between the DE and EII, they provide empirical evidence that the effect of the DE on income inequity is conditional on education. This indirectly but crucially supports the findings of the present study.

The development of the DE has significant economic effects, especially on EII. However, the direction of this effect—whether positive or negative—remains inconclusive in the literature [10,11,20,53]. This paper examines the DE’s empirical effects in cities, emphasizing its role in promoting employment, increasing income, raising tax revenues, and stimulating loans, all of which influence EII. To further explain the DE’s impact on EII, the paper explores the following two key aspects: the types of income and the sources of inequity. This approach offers a novel perspective on studying the DE’s effect on income inequity, providing a more direct explanation of its impact on EII.

In heterogeneity analysis, the effect of the DE on EII is more pronounced in cities with stronger economic foundations, better public services, and more advanced telecommunications infrastructure. This is primarily because these cities exhibit higher DE, and the educational attainment of their residents is likely to be higher, making the impact more pronounced. This finding is supported by existing research [17,27].

In summary, the hypotheses proposed in this paper have been confirmed. This paper not only integrates education into the study of income inequity but also provides direct insights into the DE’s effect on EII. It analyzes the educational impact from both family and individual perspectives, contributing to existing literature on the influence of household background on income inequity.

However, the DE may also affect EII through channels, such as changing production methods and increasing technological progress rates. Due to data limitations, these issues require further exploration. In the future, we plan to explore longer-term data to further investigate the long-term effects of the DE on EII within the proposed research framework.

The findings of this study also hold significant practical implications. As the DE continues to flourish and the overall educational quality improves, addressing income inequity and advancing the goal of common prosperity have become key economic development objectives. While non-labor EII remains relatively high, the development of the DE does not appear to have a significant effect on this form of inequity. Thus, for individuals, enhancing digital skills and acquiring the necessary vocational competencies to keep pace with techno-logical advancements represent more effective strategies. The stronger impact of the DE on effort-based inequity, as compared to environmental inequity, further supports this view. For developing countries, this study provides theoretical evidence that leveraging the trends of DE growth can help reduce EII and improve overall welfare. For policymakers, in addition to continuing efforts to promote universal education, enhancing urban infrastructure and addressing regional development disparities are crucial for reducing income gap. For individuals, investing in digital skills training and acquiring the vocational skills required to thrive in an evolving job market is a core decision to avoid economic obsolescence.

7.2. Conclusions

Using panel data from cities in China, this paper finds that the DE significantly reduces EII from both individual and family perspectives. The effect remains robust even after addressing endogeneity with IVs. The DE exerts a positive economic impact by improving employment, wages, income, tax revenue, and promoting loans. These factors all contribute to reducing EII. The DE primarily affects labor income, rather than non-labor income, in reducing EII. It also significantly decreases both inequity of opportunity and effort, which are key mechanisms influencing EII. Notably, the DE’s impact on individual inequity of effort and opportunity shows little variation. However, its effect on inequity of effort in families is stronger than on inequity of opportunity. Further analysis reveals that in cities with stronger economic foundations, higher levels of public services, and greater telecommunications infrastructure, the DE’s impact on EII is more pronounced.