The Effect of Intelligent Development on Green Economy Efficiency: An Analysis Based on China’s Province-Level Data

Abstract

As the main driving force of the new technological revolution, intelligent development is the key to promoting high-quality economic development. This paper empirically examines the nonlinear influence of intelligent development on green economy efficiency and its action paths using provincial panel data of China from 2009 to 2021. The result provides significant evidence of a U-shaped relationship between intelligent development and green economy efficiency, indicating that intelligent development initially leads to green economy efficiency decreases before ultimately increasing. Additional analysis confirms that environmental regulation, green finance, and industrial agglomeration positively moderate the impact of intelligent development on green economy efficiency. Furthermore, heterogeneous tests reveal that in the eastern region and after the release of “Made in China 2025” in 2015, the nonlinear effect of intelligent development on green economy efficiency is more pronounced. The findings of this paper provide a beneficial reference for how to leverage intelligent technology to release new kinetic energy for green economic growth under the new development concept.

1. Introduction

Since the introduction of reform and opening-up, the Chinese economy has experienced sustained high-speed growth, positioning itself as the world’s second-largest economy. However, while the extensive growth mode facilitated rapid economic expansion, it concurrently engendered a series of ecological challenges, such as increasingly stringent resource constraints and uneven regional development [1]. Developing a green economy is an inevitable requirement for breaking China’s resource constraints and accelerating the transformation of its economic development mode. Furthermore, developing a green economy is regarded as an effective way to achieve green economic growth. Balancing the equity and efficiency inherent in green economy development aligns seamlessly with China’s aspirations for high-quality economic advancement. This alignment is crucial for fostering a harmonious coexistence between energy conservation and economic growth. As a result, researching ways to realize the development of a green economy has become one of the most important academic topics at present.

As a strategic technology at the forefront of the new technological revolution and industrial transformation wave, intelligent development catalyzes global technological advancement and economic growth. Simultaneously, the advancement of intelligent technologies such as big data, cloud computing, and the Internet of Things can achieve the transformation and replacement of traditional technologies, fostering the emergence of more environmentally friendly and cleaner advanced technologies [2,3]. These innovations will elevate the level of energy conservation and consumption reduction in production processes and establish a crucial foundation for enhancing environmental governance, ultimately facilitating the green transformation and achieving sustainable development. Consequently, investigating whether intelligent development can generate intrinsic motivation to promote the development of a green economy holds significant theoretical and practical importance, as it strives to balance ecological sustainability with economic progress while fostering high-quality economic development, which is beneficial to the unification of economic benefits, social benefits, and environmental benefits.

Former literature has explored various factors influencing green economy efficiency, such as public policy [4], openness [5], agglomeration effects [6], and technological innovation [7], among others. Furthermore, research on how intelligent development affects energy conservation and emission reduction is increasing. Existing studies have shown that adopting an intelligent development strategy can reduce the total volume of carbon emissions [8], improve carbon emissions efficiency by enhancing energy use efficiency and encouraging technical advancement [9,10], and improve urban carbon emission performance [11]. Additionally, green industry development in neighboring areas is promoted by the positive spatial spillover effect of industrial intelligence [12]. Studies also delve into the nonlinear effect of intelligent development on environmental issues. For instance, Dong et al. (2024) demonstrated an inverted U-shaped relationship between artificial intelligence and carbon emissions [13]. Meanwhile, many scholars have found that the use of intelligent development can promote corporate green innovation [14], enhance green knowledge management capabilities [15], and facilitate the green transition of businesses [16], thereby injecting new impetus into the development of a green economy. As the study of the topic has expanded, many academics have developed an interest in the impacts of intelligent development on green economic issues. According to Zhang and Wu (2021), intelligence effectively promotes the manufacturing industry’s green total factor productivity [17], and the promotional effect is more significant in eastern regions [18]. Simultaneously, intelligent manufacturing brings technical advantages to China’s high-quality and energy-efficient economic growth [19]. Nevertheless, some other scholars hold the opposite view that industrial intelligence does not significantly impact green total factor productivity [20].

The literature above has discussed green economy efficiency and intelligent development from multiple perspectives. However, there is a notable gap in research on the relationship between intelligent development and green economy efficiency. In-depth explorations of the nonlinear relationship and the underlying mechanisms influencing this relationship are rather scarce. Compared to the existing literature, the potential marginal contributions of this paper primarily consist of the following: First, it elucidates the internal mechanisms through which intelligent development affects green economy efficiency within the context of China’s severe environmental challenges and resource constraints. This paper delves into the nonlinear impact of intelligent development on green economy efficiency by quantifying the level of intelligent development, thereby extending the research framework concerning the relationship between intelligent development and green economy efficiency. Second, it expands the research field concerning the relationship between intelligent development and green economy efficiency by moderating variables such as environmental regulations, green finance, and industrial agglomeration into the analytical framework to examine how various internal and external environments moderate this relationship, providing valuable insights and methodological guidance for advancing green economic growth through intelligent development.

The remainder of this paper is organized as follows: Section 2 puts forward the hypotheses of the study. Section 3 presents the models’ construction and variables selection. Section 4 analyzes all the empirical results. Section 5 discusses the conclusions, and provides policy implications.

2. Research Hypothesis

2.1. Nonlinear Impact of Intelligent Development on Green Economy Efficiency

Intelligent development represents an advanced form of productivity that emerges on the substratum of informatization, digitization, and networking. Consequently, the influence of intelligent development on green economy efficiency aligns with the comprehensive logical framework of the influence of technological advances on sustainable development. Additionally, the impact of technological advancement on green economy development is intricate and multifaceted [21], suggesting that intelligent development may exert dual and contrasting effects on green economy efficiency.

On the one hand, intelligent development enhances green economy efficiency. From the production perspective, the advancement of intelligent development can bolster production efficiency and economic output by simulating production processes in advance, optimizing the coordination of production factors across departments [22] and accurately calculating resource requirements to facilitate resource reuse and minimize energy consumption [23]. From the perspective of daily life, extensive computer utilization fosters e-commerce modes such as online shopping, telemedicine, and remote work solutions [24], which reduces information search costs and transportation expenses while enhancing supply chain efficiencies. From the regulatory perspective, intelligent development improves the efficacy and precision of information transmission, promptly conveying environmental data to government, enterprises, and the public for swift decision-making, which reduces resource wastage and ecological degradation stemming from information asymmetry [25]. Furthermore, the advancement of intelligent technology enables precise monitoring of waste emissions during production processes, lowering labor costs and pollutant emissions, thus augmenting green economy efficiency.

On the other hand, intelligent development may lead to diminished green economy efficiency. The progression toward greater intelligent development necessitates substantial energy support; fossil fuels like coal remain indispensable due to their abundant reserves and stability. The energy rebound effect and energy consumption coupled with an inefficient factor structure may also exacerbate reliance on non-clean energy sources [26]. Moreover, to facilitate the advancement of intelligent development, manufacturers are likely to escalate investments, which potentially constrain productive factors, leading to reduced economic returns, thereby hindering enhancements in green economy efficiency.

Drawing upon the above analysis and the theoretical framework of the Environmental Kuznets Curve, it is plausible to deduce that the influence of intelligent development on green economy efficiency may exhibit different effects at different stages. In the early stage of intelligent development, endeavors primarily concentrate on infrastructure development, and significant fossil fuel consumption persists, resulting in heightened pollutant emissions. However, owing to constraints such as limited technical expertise, an absence of core technologies, and funding shortages, the green economic benefits brought by intelligent development are challenging to align as offsets against negative externalities, and this may negatively impact green economy efficiency. As intelligent development attains a particular threshold, elevated resource utilization efficiencies and clean energy adoption rates emerge along with better coordination among various factor endowments. The advancement of intelligent development keeps economic output constant while consuming fewer resources and decreasing pollutant emissions, thereby augmenting the green economy efficiency. Based on this analysis, this paper proposes the following hypothesis:

Hypothesis 1.

Intelligent development exhibits a U-shaped effect on green economy efficiency.

2.2. The Moderation Effect of Environmental Regulation, Green Finance, and Industrial Agglomeration

Environmental regulation is a policy instrument that the government employs to achieve a relative equilibrium between economic growth and ecological preservation. In the immediate term, the ’compliance cost’ theory posits that to adhere to environmental regulations, enterprises incur heightened costs for pollution control, which consequently crowd out investments in production and research [27]. This somewhat delays the progression and implementation of intelligent technologies, diminishing the anticipated benefits of intelligent development to enhancing green economy efficiency. Conversely, according to the Porter Hypothesis [28], in the long run, well-structured environmental regulations yield an ’innovation compensation’ effect. The government consistently motivates enterprises to pursue green technological innovation through financial subsidies, tax benefits, and other incentives. Such resource-saving and environment-friendly technological advancement facilitates the mitigation of environmental pollution, and the positive impact of ’innovation compensation’ gradually offsets the negative impact of ’compliance costs’ [29]. Simultaneously, through innovations in production technologies and methodologies, intelligent development is promoted to enhance resource utilization efficiency to counteract the augmented production costs associated with adhering to environmental regulations, ultimately amplifying the positive effects of intelligent development on green economic growth. Based on this analysis, this paper proposes the following hypothesis:

Hypothesis 2a.

Environmental regulation positively moderates the U-shaped relationship between intelligent development and green economy efficiency.

Green finance refers to the financial activities that use innovative financial instruments to achieve sustainable development of the ecological environment [30]. In the initial phases of green finance development, phenomena such as ’greenwashing’ constantly emerge due to the absence of relevant regulatory frameworks, which impede the promotion of green and low-carbon initiatives and exacerbate challenges for small and medium-sized enterprises in securing green investments. Consequently, this brings about adverse effects on green economy efficiency. As the development of green finance progresses, optimizing financial resource allocation enables the green industry to secure essential funding while simultaneously tightening financing constraints on polluting enterprises. This process directs social capital away from high-energy consumption and high-pollution industry toward greener and low-carbon industry [31], effectively elevating the quality of environmental information disclosure and evaluation through integration with intelligent technologies. These advancements contribute to the enhancement of green economy efficiency. Additionally, the evolution of green finance provides financial support for promoting intelligent development within the green industry [32]. By augmenting product-added value through intelligent development, the green industry gains a fresh competitive edge that attracts increased green investment. It accelerates the flow and aggregation of green capital elements while further amplifying the positive impact that intelligent development has on enhancing green economy efficiency. Based on this analysis, this paper proposes the following hypothesis:

Hypothesis 2b.

Green finance positively moderates the U-shaped relationship between intelligent development and green economy efficiency.

During the early stages of industrial agglomeration development, the expenditure on infrastructure construction temporarily imposes production burdens. Coupled with inadequate intelligent development, industrial agglomeration tends to result in expanded pollution emissions, and the concentration of homogeneous enterprises can also simplify industrial structure [33], thereby negatively affecting the enhancement of green economy efficiency. However, as industrial agglomeration progresses, increasing numbers of manufacturers within the market can share the costs associated with intelligent development, promoting resource coupling and collaborative research and development among departments [34] while reducing expenses and facilitating knowledge spillover through reliance on industrial agglomeration. Moreover, it promotes the intelligent transformation of various supply chain segments, including production, logistics, and transactions. Ultimately, this augments the core competitiveness of intelligent development in green production and achieves economic benefits while fostering environmental sustainability. Based on this analysis, this paper proposes the following hypothesis:

Hypothesis 2c.

Industrial agglomeration positively moderates the U-shaped relationship between intelligent development and green economy efficiency.

3. Model Construction and Variable Selection

3.1. Model Construction

Based on the characteristics of green economy efficiency, using ordinary OLS will result in biased or inconsistent parameter estimates. This paper uses the Tobit regression model to verify the nonlinear impact of intelligent development on green economy efficiency. The baseline regression model can be expressed by Equation (1):

$${GEE}_{it}={\alpha }_0+{\alpha }_1{IT}_{it}+{\alpha }_2({IT}_{it}{)}^2+{\alpha }_{3}Contro{l}_{it}++{\epsilon }_{it}$$

(1)

Among them, subscripts i and t represent provinces and years, respectively. GEE_it stands for green economy efficiency. IT_it stands for intelligent development. Control_it stands for control variables, and ε_it represents the random perturbation term.

In order to examine whether environmental regulations, green finance, and industrial agglomeration moderate the process of intelligent development impact on green economy efficiency, interaction terms between environmental regulations, green finance, and industrial agglomeration and intelligent primary and secondary terms are introduced based on model (1). The specific model settings are shown in Equation (2):

$${GEE}_{it}={\beta }_0+{\beta }_1{IT}_{it}+{\beta }_2({IT}_{it}{)}^2+{\beta }_3{D}_{it}+{\beta }_4{IT}_{it}\times D_{it}+{\beta }_5({IT}_{it}{)}^2\times D_{it}+{\beta }_{6}contro{l}_{it}++{\epsilon }_{it}$$

(2)

Among them, Dit stands for the moderating variables. The meanings of other variables are the same as Equation (1).

3.2. Variable Selection

3.2.1. Dependent Variable

Green economy efficiency (GEE). Green economy efficiency refers to economic efficiency that considers resource utilization and environmental pollution, and it has become an essential indicator for the development of green economy. This paper constructed the SBM-DEA (Slack-Based Measure Data Envelopment Analysis) model proposed by Tone (2004) [35], which includes resource consumption, financial benefits, and unexpected output to measure it. The model is expressed as:

$$min\rho ={\displaystyle \frac{\frac{1}{m}{\sum }_{i=1}^m\frac{\overline{x}}{x_{ik}}}{\frac{1}{r+s}\left({\sum }_{p=1}^r\frac{\overline{y}}{y_{pk}}+{\sum }_{q=1}^s\frac{\overline{\mathrm{z}}}{z_{qk}}\right)}}$$

(3)

$$\begin{array}{c}s.t.\left\{\begin{array}{c}\overline{x}⩾{\displaystyle \sum _{j=1,\ne k}^n}{x}_{ij}{\theta }_j;\overline{y}⩽{\displaystyle \sum _{j=1,\ne k}^n}{y}_{ij}{\theta }_j;\overline{z}⩾{\displaystyle \sum _{j=1,\ne k}^n}{z}_{ij}{\theta }_j\\ \overline{x}⩾x_k;\overline{y}⩾y_k;\overline{z}⩾z_k;{\theta }_j⩾0\\ i=1,\cdots ,m;j=1,\cdots ,n;r=1,\cdots ,p;s=1,\cdots ,q\end{array}\right.\end{array}$$

In Formula (3), n represents the number of decision-making units (DMU), and it refers to 30 provinces in this paper. Each DMU is characterized by m types of inputs, r types of expected outputs, and s types of non-expected outputs. The elements x, y, and z denote input, expected output, and unexpected output. The term ρ denotes the regional green economy efficiency. Four input variables were selected: labor, capital, energy, and water. Labor input is represented by the number of employees in each region. Concerning Scott and Edward (1990), this paper adopted the perpetual inventory method to calculate the capital input [36]. The base year is 2008. Energy input is indicated by the electricity consumption in each province. Water input is represented by water consumption in each province. The output index consists of expected output and non-expected output. The expected output was measured by the GDP in each province. The total amount of industrial sulfur dioxide emissions, industrial wastewater emissions, and industrial smoke (dust) emissions in each province were selected as non-expected output.

3.2.2. Explanatory Variable

Intelligent development (IT). At present, there is no unified indicator for measuring intelligent development. Previous studies have commonly used indicators such as industrial robot application density, inventory, input, and import volume as proxy variables. Still, these indicators are relatively singular and lack relevant data at the provincial level. This paper drew on the research of Sun and Hou (2019) [37] and Liu et al. (2020) [38], and the indicator evaluation ideas and design framework of the “National Innovation Index Report 2022–2023”, mainly measuring intelligent development from three perspectives: fundamental investment, technology application, and market efficiency. Explicitly speaking, fundamental investments such as funding, talent, and facilities are the essential guarantees for promoting intelligent development, knowledge innovation and material innovation are the keys to measuring the level of intelligent development, and market efficiency mainly reflects the benefits and efficiency of intelligent development. The specific indicators are shown in

Table 1. China’s Provincial Intelligent Development Evaluation Index system.

Primary Indicator	Secondary Indicator	Specific Indicators
Fundamental investment	Fund investment	R&D expenditure of high-tech industries/R&D expenditure of all enterprises
	Talent investment	Number of high-tech industry employees/number of employees
		Number of scientific and technical employees/number of employees
	Facility investment	Number of R&D institutions in high-tech industries
		Length of long-distance optical cable lines/regional area
		Number of Internet Broadband Access Ports
Technology application	Knowledge innovation	Number of invention patent applications
		Number of scientific papers published
	Material innovation	Number of R&D projects in high-tech industries
		Number of new product development projects
Market efficiency	Enterprise income	Total profits of high-tech industry/number of high-tech enterprises
		Primary business income of high-tech industry/number of high-tech industry employees
		New product sales revenue of high-tech industry/number of high-tech industry employees
	Capital operation	R&D investment intensity

. This paper utilized the entropy approach to comprehensively evaluate the intelligent development level at the provincial level in China from 2009 to 2021 and obtain the intelligent development index. The higher the index value, the greater the level of intelligent development.

Figure 1 illustrates the changes in intelligent development levels across China in recent years. As depicted in Figure 1, there are significant disparities in intelligent development levels among various provinces, regions, and cities. The result is broadly consistent with the findings of the China Regional Science and Technology Innovation Evaluation Report 2023. This alignment supports the feasibility and rationality of the intelligent development evaluation index system established in this study.

In order to delve deeper into spatial heterogeneity in intelligent development across different regions of China, the research categorized its study samples into eastern, central, and western regions based on geographical classifications. Figure 2 reveals a distinct spatial distribution pattern, with the eastern region exhibiting the highest level of intellectualization, followed by the central and western regions, respectively. Over the period spanning 2009 to 2021, all three areas demonstrated an upward trend in their respective intelligent development indices. Notably, the eastern region experienced a substantial surge, while the central region gradually outpaced it in growth magnitude over time, despite initially mirroring the western region.

3.2.3. Moderating Variables

Environmental regulation (ENV), green finance (GFIN), and industrial agglomeration (IAGG) were selected as moderating variables. Following Li and Cao (2024), this study measures environmental regulation as the proportion of pollution investment within each province’s GDP [39]. Green finance has played an essential role in environmental protection and promoting economic development. Referring to the approach of Xu et al. (2024), the entropy method is used to construct indicators from four dimensions: green credit, green securities, green investment, and green insurance [40]. Industrial agglomeration is measured by the employment density in each province, referring to the study of Wang et al. (2024) [41].

3.2.4. Control Variables

These control variables are as follows: (1) Financial development (FIN), represented by the ratio between the proportion of financial industry-added value to regional GDP and the proportion of national financial industry-added value to GDP. (2) Industrial structure (IND), represented by the ratio of the tertiary industry’s added value to the secondary industry’s added value. (3) Government intervention (GOV), represented by the proportion of regional fiscal expenditure to regional GDP. (4) Informatization (INFOR), represented by the proportion of total postal services to the regional gross domestic product. (5) Environmental infrastructure (GREEN), represented by the proportion of green space area to the built-up area. (6) Foreign Direct Investment (FIND), represented by the proportion of foreign direct investment to the regional gross domestic product.

3.3. Sample Selection and Data Source

This paper took 30 provinces in China (excluding Xizang, Hong Kong, Macao, and Taiwan) as the research object and selected panel data from 2009–2021 to investigate the relationship between intelligent development and green economy efficiency. The data for each variable is extracted from the CSMAR database, the official website of the National Bureau of Statistics, “China Statistical Yearbook” (2009~2021), “China Statistical Yearbook on Environment” (2009~2021), and “China Statistical Yearbook on Science and Technology” (2009~2021). Missing individual values have been addressed through interpolation methods. Descriptive statistics for each variable are presented in

Table 2. Descriptive statistics of variables.

Types of Variables	Variables	Obs	Mean	St. Dev	Min	Max
Dependent variable	Gee	390	0.6549	0.1406	0.5005	1
Explanatory variable	IT	390	0.1187	0.0929	0.0154	0.7136
Moderating variables	ENV	390	0.1213	0.1260	0.0009	1.1034
	GFIN	390	0.3496	0.1419	0.0745	0.7086
	IAGG	390	0.2575	0.3750	0.0039	2.1707
Control variables	FIN	390	0.9824	0.4605	0.2584	3.0265
	IND	390	1.2757	0.7202	0.5271	5.2440
	GOV	390	0.2541	0.1119	0.1050	0.7583
	INFOR	390	0.0632	0.0517	0.0143	0.2896
	GREEN	390	0.0121	0.0323	0.0001	0.2213
	FIND	390	0.0196	0.0156	0.0001	0.0819

.

4. Results

4.1. Baseline Regression Results

Table 3. Baseline regression results.

Variable	(1)	(2)	(3)	(4)
IT	−0.925 ***	−0.853 ***	−0.682 ***	−0.704 ***
	(−4.18)	(−3.81)	(−3.12)	(−3.26)
IT2	1.289 ***	1.155 ***	1.208 ***	1.265 ***
	(4.92)	(4.67)	(4.74)	(5.00)
FIN		0.034 ***	0.037 ***	0.029 ***
		(3.28)	(3.60)	(2.66)
IND		0.003	0.017	0.028 *
		(0.19)	(1.09)	(1.76)
GOV			−0.210 ***	−0.203 ***
			(−3.77)	(−3.66)
INFOR			0.414 ***	0.358 ***
			(3.06)	(2.76)
GREEN				0.309 *
				(1.67)
FIND				0.744 ***
				(2.76)
Constant	0.693 ***	0.585 ***	0.489 ***	0.422 ***
	(17.24)	(6.54)	(5.23)	(4.46)
Observations	390	390	390	390

Note: * and *** indicate significance at the 10% and 1% statistical levels, respectively. T-values in parentheses.

presents the baseline regression results. Column (1) demonstrates the impact of intelligent development on green economy efficiency in the absence of control variables. The estimated coefficient of the first-order term of intelligent development is negative, and that of the second-order term is positive at the 1% significance level. As control variables are introduced in Columns (2) through (4) stepwise, the regression outcomes remain substantially unchanged. These results indicate that the impact of intelligent development on green economy efficiency initially decreases and then increases, showing a U-shaped characteristic that verifies Hypothesis 1. The possible explanation is that in the early stage of intelligent development, the positive externality of intelligent development is unable to effectively offset the shortage of raw materials and supply equipment caused by the expansion of its scale, which will exceed the carrying capacity of local natural resources and the economy. When the level of intelligent development exceeds a certain threshold, it encourages interactive information and resource sharing, gives rise to the sharing economy model, and improves the efficiency of optimal resource allocation, all of which contribute to enhancing green economy efficiency. Based on calculation, the “U-shaped” inflection point in the relationship between intelligent development and the green economy efficiency is 0.278, falling within the range of the intelligent development values [0.0154, 0.7136]. In light of the intensity of intelligent development in various provinces from 2009 to 2021, most samples did not exceed the threshold value. This demonstrates that China’s total degree of intelligent development falls short of what is required to enhance green economy efficiency.

4.2. Robustness Test

With the aim of verifying the stability of the baseline regression results, this paper conducted the following robustness tests in

Table 4. Robustness tests.

Variables	(1)	(2)	(3)	(4)	(5)
	One-Year Lagged Explanatory Variables	Change Model	Recalculate Explanatory Variables	Winsorization of Samples	Substitute Control Variables
IT		−0.654 ***	−0.139 *	−0.599 **	−0.863 ***
		(−3.08)	(−1.90)	(−2.25)	(−3.95)
IT2		1.164 ***	0.415 **	1.108 **	1.430 ***
		(4.82)	(2.15)	(2.52)	(5.64)
L.IT	−0.826 ***
	(−3.18)
L.IT2	1.663 ***
	(4.89)
Constant	0.452 ***	0.647 ***	0.244 ***	0.339 ***	0.706 ***
	(4.39)	(16.67)	(2.83)	(3.47)	(20.70)
Controls	yes	yes	yes	yes	yes
Observations	360	390	390	390	390
R-squared		0.4521

Note: *, ** and *** indicate significance at the 10%, 5% and 1% statistical levels, respectively. T-values in parentheses.

: (1) Lagging explanatory variables by one period. This study examined one-year-lagged explanatory variables to investigate if intelligent development and its squared term impacted green economy efficiency in the previous period. (2) Utilizing the OLS model to investigate the effect. (3) Recalculating the explained variables with principal component analysis. (4) Applying a 2% winsorization to all samples. (5) Substituting control variables. This study substituted the financial development level with social purchasing power (measured by per capita consumption expenditure) and the informatization level with the level of technological market development (measured by the proportion of technological market transaction volume in regional GDP) as new control variables for regression analysis.

The results shown in Table 4 are consistent with the conclusion of baseline regression, which further assures us that Hypothesis 1 is valid.

4.3. Moderation Tests

The research undertaken by Haans et al. (2016) elucidates that the assessment of moderation in a U-shaped relationship necessitates concurrent attention to variations in curve slope gradients and displacements in the inflection point [42]. In the context of slope gradient alterations, if the interaction term coefficient between the quadratic term of intelligent development and the moderation variable is significantly positive, it indicates that the moderation variable will steepen the U-shaped curve and vice versa. The inflection point is calculated by the formula −β₁/2β₂ (Symbol derived from Equation (2)).

Table 5. Moderation tests.

Variables	(1) Environmental Regulation	(2) Green Finance	(3) Industrial Agglomeration
IT	−0.597 ***	−0.442 **	−0.451 *
	(−2.70)	(−2.45)	(−1.84)
IT2	1.316 ***	1.096 **	0.870 **
	(5.06)	(2.09)	(2.30)
D	0.608	0.041	0.053
	(0.12)	(0.37)	(0.62)
IT × D	−1.918 *	−2.644 **	−2.273 **
	(−1.90)	(−2.30)	(−2.49)
IT2 × D	9.74 **	3.675 *	3.987 *
	(2.41)	(1.82)	(1.81)
Constant	0.392 ***	0.323 ***	0.219 ***
	(4.17)	(2.69)	(2.91)
Controls	yes	yes	yes
Observations	390	390	390

Note: *, ** and *** indicate significance at the 10%, 5% and 1% statistical levels, respectively. T-values in parentheses.

displays the empirical findings regarding the moderating effects of environmental regulation, green finance, and industrial agglomeration on the relationship between intelligent development and green economy efficiency. The results reveal that the interaction terms between the quadratic term of intelligent development and the moderation variables are significantly positive, indicating that environmental regulation, green finance, and industrial agglomeration amplify the U-shaped relationship between them, thereby steepening the U-shaped curve. Furthermore, the U-test results suggest that after incorporating environmental regulation, green finance, and industrial agglomeration and their interaction terms, the inflection point advances from 0.278 to 0.227, 0.202, and 0.259, respectively. This result demonstrates that including these moderating factors can expedite the manifestation of the positive effects of intelligent development on green economy efficiency. Consequently, environmental regulation, green finance, and industrial agglomeration exert positive moderating effects in the U-shaped relationship between intelligent development and green economy efficiency. Hypothesis 2a, Hypothesis 2b, and Hypothesis 2c are confirmed.

The main reason is that environmental regulations can boost research and development vitality, improve the efficiency and innovation capacity of green technologies, reduce the cost of pollution reduction and carbon reduction, and achieve high-quality economic development. Green finance expands the financing options for environmental protection firms and enhances the financing efficiency of corporate green projects, thus improving the green economy efficiency. Industrial agglomeration promotes the growth of economies of scale, while accelerating the generation, dissemination, and accumulation of knowledge, thus generating technological emission reduction effects, which can stimulate the impetus of green technology innovation of enterprises, subsequently boosting the green economy efficiency.

4.4. Heterogeneity Tests

4.4.1. Regional Heterogeneity Tests

The development and application of intelligent development necessitates substantial investments in technology and resources. Given the variations in institutional frameworks, resource endowments, and social cultures across diverse regions within China, notable regional variations in intelligent development are evident. Consequently, the effect of intelligent development on the green economy efficiency may also manifest regional heterogeneity in the three major regions of China. Columns 1–3 of

Table 6. Heterogeneity tests.

Variables	(1)	(2)	(3)	(4)	(5)
	East	Central	West	2009–2015	2016–2021
IT	−0.705 ***	−4.148	2.011 ***	0.509	−0.934 **
	(−4.27)	(−1.35)	(3.04)	(1.58)	(−2.16)
IT2	1.327 ***	13.119	−6.083 ***	−1.289 *	1.601 ***
	(7.57)	(1.02)	(−2.66)	(−1.70)	(3.78)
Constant	0.604 ***	0.536 ***	0.487 ***	0.411 ***	0.539 **
	(8.39)	(5.61)	(10.72)	(3.53)	(2.35)
Controls	yes	yes	yes	yes	yes
Observations	143	104	143	210	180

Note: *, ** and *** indicate significance at the 10%, 5% and 1% statistical levels, respectively. T-values in parentheses.

indicate the results of intelligent development on green economy efficiency in the eastern, central, and western regions, respectively. The influence of intelligent development on green economy efficiency in the eastern region displays a U-shaped relationship with an inflection point at 0.266. Regression results for the central region did not meet the significance level. In the western region, the coefficients of intelligent development are significantly opposite to the baseline regression results.

These empirical findings align with the current situation in China. The eastern region exhibits a distinctly superior economic development level to the central and western regions, endowed as it is with abundant resources and well-established institutional frameworks that provide ample support for intelligent development in technology, talent, capital investment, and infrastructure. Consequently, intelligent development commenced earlier and has attained a relatively advanced stage in the eastern region, which has brought about a more pronounced impact of intelligent development on its green economy efficiency. Conversely, owing to the constraints of their physical placements and resource endowments, the central and western areas have comparatively lagged in their intelligent development, resulting in a less evident influence on their green economy efficiencies.

4.4.2. Temporal Heterogeneity Tests

In 2015, the State Council issued “Made in China 2025”, a document to enhance the national manufacturing innovation capacity and comprehensively implement green manufacturing as a strategic imperative [43]. This document established a policy foundation for advancing intelligent development and green economic growth and can be seen as a critical event in the evolution toward an intelligent transformation. The release of this strategic document indicates that China has begun to attach importance to intelligent development, furnishing policy and essential resources for intelligent innovation, and attracting innovative elements such as R&D talents, which will drive the transformation and upgrading of the industrial structure and accelerate the development process of intelligent development in China. This paper divided the research period into two distinct phases, 2009–2015 and 2016–2021, to examine the differential impacts of intellectualization on green economy efficiency.

According to Columns 4–5 of Table 6, the impact of intelligent development on green economy efficiency from 2009 to 2015 was insignificant. This outcome may be attributed to the relatively low level of intelligent development during this stage, resulting in a less significant influence on green economy efficiency. Conversely, from 2016 to 2021, there is a significant U-shaped relationship between intelligent development and green economy efficiency. This finding demonstrates that the proposal of “Made in China 2025” facilitated the policy and resource support for intelligent development. Various management measures are increasingly improved, while new applications, new business formats, and new models continue to emerge, injecting new impetus into the development of a green economy.

5. Conclusions and Policy Implications

Based on theoretical analysis, this paper empirically examined the nonlinear impact of intelligent development on green economy efficiency and its action paths, using provincial panel data of China from 2009 to 2021 as research samples. The results reveal that: (1) Intelligent development initially leads to green economy efficiency decreases before ultimately increasing, exhibiting a U-shaped relationship. (2) Environmental regulation, green finance, and industrial agglomeration can positively moderate the impact of intelligent development on green economy efficiency. (3) In terms of regional heterogeneity, the nonlinear effect of intelligent development on green economy efficiency is more pronounced in eastern regions. In terms of temporal heterogeneity, the U-shaped impact of intelligent development on green economy efficiency becomes more significant after the proposal of “Made in China 2025”.

The following policy implications are suggested in light of the findings above:

First, systematically promote the development of intelligent development. The government needs to expedite the construction and optimization of the intelligent development infrastructure system and promote the deep integration of intelligent development and green economic growth. An intelligent development evaluation system centered on the development concept of “innovation, coordination, green, openness, and sharing” must be established. Furthermore, China should develop intelligent technologies oriented toward green and low-carbon development. For example, carry out green data center evaluation and construction, optimize low-carbon data center energy efficiency programs, etc. Additionally, enterprises and the general public should be more aware of energy conservation, emission reduction, and green economy development, and build an ecologically friendly and technology-friendly society.

Second, efforts should be dedicated to promoting the institutional construction and enhancement of the competitiveness of environmental regulations, green finance, and industrial agglomeration. The government should optimize the top-level design of environmental regulations. It is essential to intensify supervision within an appropriate range, raise emission standards, and distribute emission reduction subsidies. Meanwhile, the government should reasonably delineate the scope and degree of intelligent technology applications, emphasize the importance of intelligent development in environmental governance, and fully utilize intelligent technology to enhance its surveillance dynamics. More than that, the government needs to strengthen the rational guidance of green finance policies and construct a green finance system supported by intelligent development. By introducing targeted policy guidelines, subsidies, and preferential measures, the government can prioritize support for green projects that incorporate intelligent elements and integrate green finance into technological innovation and green economy development. Moreover, the government can make efforts to actively guide industrial restructuring, enhance the competitiveness of high-tech industrial clusters, break down the communication barriers between industries, and combine the conditions of different sectors to carry out the layout of a green transformation.

Third, advance intelligent development by fully integrating regional characteristics. The eastern region, which has established a relatively complete intelligent industry, ought to leverage its leadership in intelligent development to actively undertake the innovation costs in the process of intelligent development and support the construction of the new development paradigm. Intelligent industry model selection and policy formulation should aim for mutual benefit and cooperation, as well as the exploration of mutual trust mechanisms and the connection of the intelligent technology development network. In order to close the development gap between regions, the central and western regions should take note of the eastern region’s development experience, make full use of the knowledge spillover effect that intelligent development brings to more developed areas, reduce the transmission cost of technological elements and the learning cost in the process of intelligent development, and expedite the formation of a pattern of coordinated green economic development across regions.

Nevertheless, this study also has some limitations. First, future research requires samples from other nations and longer time periods to provide more generalizable results and connect findings to international perspectives or broader trends, with the goal of contributing to global sustainability. Second, based on data availability, this paper only uses the subdivision indicators from high-tech and other related industries to build an intelligent development indicator system. Continued research on different industry types could provide some new insights.