A Dynamic Impact Evaluation of the High-Quality Development of China’s Construction Industry Using the Panel Vector Autoregressive Model

Abstract

Based on the theory of new structural economics, this research aims to explore the dynamic correlation among high-quality development, environmental regulation structures, and property rights structures in China’s construction industry. A panel vector autoregressive model (PVAR) is employed to conduct an empirical study of 30 provinces from 2008 to 2022. To further explore regional heterogeneity, K-means clustering is utilized to categorize the country into three types of regions. The results indicate that strict environmental regulation structures have a beneficial effect on the high-quality growth of the construction sector, which is most pronounced in Region III with a lower degree of construction development. Conversely, state-owned ownership structures are an impediment, and their influence is the greatest in Region I with a higher degree of construction development. Additionally, environmental regulation structures and property rights structures demonstrate a certain interactional effect. The dynamic correlation between these indicators varies in distinct regions. Various regions in China should combine their development characteristics and advantages to reasonably adjust environmental regulation structures and property rights structures. This research provides a direction for structural adjustments for the high-quality development of the construction industry.

1. Introduction

China’s economy has transitioned from a period of high-speed growth to a phase of high-quality development. Currently, it is in a critical phase of modifying its development model, upgrading its economic structure and altering its growth drivers. The economic advancement of China will inevitably depend on the attainment of high-quality development. The 20th National Congress of the Communist Party of China advocated for a commitment to advancing high-quality development, establishing a novel development model and prioritizing the enhancement of the quality and efficiency of developments [1]. In recent years, China’s economic progress has entered a new stage, characterized by structural issues as the primary contradiction in economic growth. Promoting supply-side structural transformation is a significant and innovative approach to adapt to the current economic conditions and effectively respond to the impacts of the global financial crisis. To facilitate supply-side reform, it is essential to focus on the key aspect of “structural reform” of supply-side reform [2]. Yifu Lin, an expert in new structural economics, similarly asserts that China’s new economic normal may last for an extended period and that structural reforms are necessary for the country’s economy to return to a stable state [3]. Numerous production domains, both domestically and internationally, have accomplished significant advancements in quality development by structural reform, such as the agricultural [4,5] and energy [6,7] industries.

The construction industry is a fundamental sector of China’s national economy [8]. Its high-quality development is crucial to the high-quality growth of the national economy. Furthermore, it serves as an essential condition and a safeguard for the high-quality growth of other sectors and industries within the country. In recent years, the added value of the building sector has been maintained at more than 6–8% of the GDP [9], highlighting its significantly growing importance in supporting China’s national economy. The Ministry of Housing and Urban–Rural Development of China has released “The 14th Five-Year Plan for the Development of the Construction Industry”, which aims to achieve a substantial improvement in the quality and efficiency of the construction industry by 2035. Nevertheless, the building sector currently faces numerous developmental challenges, including an excessive dependence on resource consumption, high costs, a lack of innovation drive, and the inefficient management of carbon emissions [10]. In addition, the high-quality growth of the building sector exhibits uneven distribution characteristics across various regions of China [11]. Therefore, how do we correctly guide the building sector from extensive development to high-quality growth? This has become one of the major challenges in China’s new stage of development.

New structural economics is a neoclassical framework that specifically emphasizes the transformation in economic development’s structure. Its primary objective is to foster sustainable economic expansion and mitigate income inequalities among developed and developing nations [12]. It is essential to consider the structural characteristics of various stages in economic development [13]. The theory of new structural economics provides a robust theoretical foundation for developing nations to achieve sustainable economic growth, eliminate poverty, and narrow the income gap with wealthier nations. Given that there are many defects in the construction industry, it is crucial to comprehensively understand the current state of high-quality growth in both the national and regional building sectors. By analyzing and addressing issues from a structural perspective, appropriate economic policies can be developed for the construction sector, and the government’s initiatives can be adjusted accordingly. Only in this way can we promote the green, digital, and intelligent transformation and development of the construction industry [14,15], thereby improving the high-quality level of the building sector in various regions.

In the context of the “carbon peak and carbon neutrality strategy [16,17]” and the significant adjustments within a market economy [18], it is of great practical significance to select environmental regulatory structures and property rights structures. Among the various structural factors influencing high-quality development, these factors are representative and may exert a more pronounced impact. Based on the above discussion, this study uses a panel dataset covering 30 provinces in China’s construction industry from 2008 to 2022. Unlike previous studies, this research uses a panel vector autoregression (PVAR) model to explore the dynamic correlation appraisal of high-quality development in conjunction with environmental regulation structures and property rights structures. This provides empirical evidence from China based on the theory of new structural economics. In terms of regional differences, a K-means cluster analysis was employed to categorize the provinces into three distinct regions, unlike traditional geographic divisions. Subsequently, based on the theory of new structural economics, the heterogeneous effects of different structural factors on high-quality development in China’s national region and the three sub-sample regions are further investigated. This research provides a direction for structural adjustments for the balanced development of the construction industry in the next decade.

This paper is structured as follows: Section 2 explains the theory of new structural economics and reviews the relevant literature. The modeling methodology and data are presented in Section 3. Section 4 classifies the regions by K-means clustering. The empirical results of the PVAR model can be found in Section 5. The Section 6 and Section 7 are a discussion and conclusions. Figure 1 illustrates the structural framework of this study.

2. Theoretical Analysis and Literature Review

2.1. New Structural Economics

New structural economics applies a neoclassical analysis to investigate the factors that influence a country’s economic structure and the underlying causes of any changes in it. Its goal is to explain why the development process involves continuous adjustments and modifications of the economic framework [19]. The essence of economic development is continuous structural change. Developing countries are catching up with developed countries, ostensibly because their income levels are converging with those of developed economies. However, at a deeper level, this convergence also involves an alignment of industrial structures. Ultimately, at a deeper level, the factor endowments and their structures must also catch up with those of developed economies [20].

The theory of new structural economics posits that numerous developing nations have long been ensnared in the cycle of low-income stagnation or have persistently found themselves trapped in the middle-income country pitfall, similar to the case of Latin American countries. The common characteristic among these nations is the lack of effective structural adjustments and the upgrading of their economic structure [21]. This theory summarizes the experiences of developing countries such as China. It focuses on factor endowments and their structures as entry points, highlighting the inherent disparities and evolving nature of economic structures at various stages of development. It is believed that each stage of economic development is best aligned with a corresponding economic structure. Compared with existing theories, it can summarize and generalize the development laws of the economy and society from a higher dimension [19].

With respect to relevant scholarly works, Wang and Wu [22] studied the regional differences and driving mechanism of the high-quality growth of the building sector in China under the constraint of carbon emissions. Their findings indicate that there are regional variances in the level of the high-quality development of China’s building sector, and the regional heterogeneity will exist for a long time. Therefore, considering the distinctive regional features of China and the development status of the building sector, this paper further categorizes this nation into distinct sub-samples. From the perspective of “new structural economics”, this research studies the different regions as different development stages and economic structures. It plans to select environmental regulation structures and property rights structures as key structural influencing factors, analyzing the heterogeneous impacts on the high-quality growth of China’s building sector at both national and regional levels.

2.2. The Relationship between Environmental Regulation and High-Quality Development

Neoclassical economics posits that environmental regulations impose additional financial burdens on enterprises through the internalization of environmental externalities. This compels firms to alter their initial optimal production choices, which in turn diminishes their capacity for innovation and competitiveness [23,24]. Porter [25,26] argues, based on a dynamic perspective, that reasonable and strict environmental regulations can serve as an impulse for technological innovation. This impulse can lead to innovation compensation, which helps to offset the expenses associated with compliance with environmental regulations. Consequently, it promotes enhanced productivity and competitiveness. However, Porter’s hypothesis is based on developed countries, such as the United States, which are at the forefront of the global economy and technology. The mechanisms of their development, transformation, and operation differ greatly from those of developing countries, not only in terms of quantity but also in terms of quality and structure. It is possible that the laws of developing nations will not be properly reflected if Porter’s hypothesis is tested in places like China.

Therefore, the Chinese scholar Zheng et al. [27] innovatively proposed “the New Structural Porter hypothesis” based on the theory of new structural economics. Through structural analyses, he found that the optimal structure of environmental regulations varies depending on the stage of economic development. Furthermore, they observed that the influence of environmental regulations varies at different stages of economic development. Ultimately, he concluded that environmental regulations primarily impact high-quality development mainly through structural change. The main mechanism that influences the innovation and high-quality development of an economy in developing countries is the process of structural change in environmental regulations. “Reasonable and strict environmental regulations” depend not only on the strength of environmental regulations but also on their structure. Due to the different functions and roles of different environmental regulatory tools, this involves the issue of optimal environmental regulation structures. Established studies have divided the structure of environmental regulations in different ways according to different research needs [28]. Based on existing research, it can be concluded that most of them revolve around two categories: command-and-control and market-economy environmental regulatory tools. The environmental regulation structure in this paper refers to the different types of combinations of environmental regulatory instruments. Based on the above analysis, particularly with regard to the construction industry, a hypothesis, H1, is put forward.

Hypothesis 1: 

The dynamic impact of high-quality development in conjunction with environmental regulation structures varies in regions with distinct development levels of the construction industry.

2.3. The Relationship between Property Rights Structures and High-Quality Development

Zhou [29] argued that the property rights structure provides appropriate incentives for individuals’ economic behavior, thus ensuring the efficiency of resource allocation and use. Yang and Zeng [30] put the property rights structure into their analysis of the development of the construction industry and empirically tested the impact of the property rights structure on market performance in the Yangtze River Economic Belt. Their findings reveal a strong inverse relationship between the proportion of the output value from state-owned firms and market performance. Drawing on the theory of new structural economics, self-sufficiency can only be achieved by a corporation when the industry it works in aligns with the comparative advantage determined by the factor endowment structure of the economy [31].

The effect of property rights structures on the high-quality development of the building sector has two aspects. Firstly, enterprises with a higher proportion of the state-owned economy often have a larger operating scale, strong capital, and more available resources, which support the progress of the building sector with more capital protection and a strong risk-resistance ability. The second issue is that the incentive constraints of the distribution system for enterprises with a higher proportion of the state-owned economy are not effective enough, and the driving force for accelerating the digital transformation and innovative high-quality growth of the building sector is insufficient [32]. The rational arrangement of property rights is advantageous for the market performance of the building sector. But the excessive percentage of state-owned construction companies’ total output value can lead to an unreasonable property rights structure, which will result in an inefficient resource distribution and the loss of enterprise benefits. Hypothesis 2 is proposed based on the analysis provided above.

Hypothesis 2: 

The dynamic impact of high-quality development in conjunction with property rights structures varies in regions with distinct development levels of the construction industry.

Combined with the analysis provided above, the theoretical framework of this study is shown in Figure 2.

3. Main Variables and Model Specification

3.1. Main Variables

3.1.1. Level of High-Quality Development (HQ)

In recent years, China’s domestic discussions on high-quality development in the building sector have been more extensive, driven by the concept of high-quality development itself. Several researchers have provided their own interpretations on the notion of this novel development mode in the building industry. Li and Wang [33] redefined the concept of high-quality development within and beyond the city, offering a fresh outlook on its meaning. Sun et al. [34] constructed a theoretical framework for high-quality development by combining construction economics and the theory of high-quality development. They found that high-quality growth requires not only continuous optimization and upgrading but also the achievement of coordinated development with the external environment. Wang et al. [35] innovatively established a comprehensive evaluation system for high-quality development based on quality requirements, stability guarantees, and the latest “Five Development Concepts”. They proposed fresh ideas for the high-quality growth path of the Chinese building sector by defining the connotation of high-quality development in this sector. Based on this, this paper proposes a five-dimensional evaluation index system according to new development concepts (innovation, coordination, green, openness, and sharing). This study presents a comprehensive evaluation index system that comprises five primary indicators and twenty-one secondary indicators, which include the industry scale, innovation drive, green development, open development, and efficiency level (

Table 1. Comprehensive evaluation index system of HQ.

Primary Indicators	Secondary Indicators	Unit	Indicator Attribute
Industry scale	Gross output value of the construction industry	CNY ten thousand	+
	Number of construction enterprise units	individual	+
	Number of contracts signed by construction enterprises	Percentage of CNY million	+
	Completion rate of building area		+
Innovation driven	Technical equipment rate	RMB/person	+
	Power equipment rate	KW/person	+
	Labor productivity	RMB/person	+
	Total year-end power of self-owned construction machinery and equipment	Million Kilowatts	+
Green development	Sewage treatment rate	Percent	+
	Greening coverage of the completed area	Percent	+
	Steel consumption with a total output value of CNY 100 million	Tons/CNY 100 million	-
	Wood consumption with a total output value of CNY 100 million	m3/CNY 100 million	-
	Cement consumption with a total output value of CNY 100 million	Tons/CNY 100 million	-
Open development	The proportion of private enterprises	Percent	+
	The proportion of foreign-invested enterprises	Percent	+
	The proportion of the output value of private enterprises	Percent	+
	The proportion of the output value of foreign-invested enterprises	Percent	+
Efficiency level	Value-added of the construction industry	CNY ten thousand	+
	Profit margin on production value	Percent	+
	Employees at the end of the period	Person	+
	Asset liability ratio	Percent	-

Note: The above indicators are all from the construction industry. Positive indicators are represented by +. Negative indicators are represented by -.

). But the attributes of these indicators are inconsistent and can be categorized into two types: positive indicators and negative indicators. Positive indicators are characterized by a positive correlation between the indicator value and the result value, whereas negative indicators signify that larger values indicate lower levels. In order to homotrend the force of all indicators in the results, this paper adopts the reciprocal form for all negative indicators [36].

In this paper, the principal component analysis (PCA) is applied to assess the extent of the high-quality development. The PCA is a statistical method that reduces the dimensionality of multiple indicators by transforming them into a smaller set of indicators. It is possible to substitute the original multi-dimensional indicator variables with a small number of primary components while assuring a minimal loss of the original data information. Consequently, this not only decreases the number of system variables but also preserves most of the valuable information from the original dataset, thereby simplifying the data structure significantly [37]. This method has been extensively employed in both the natural and social sciences. To eliminate the negative effects obtained by the PCA and to facilitate the future dynamic econometric analysis, the formula $Y_i^t=H+Y_i$ is used by coordinate translation [38].

3.1.2. Environmental Regulation Structures (ERSs)

In this paper, the environmental regulation structure is selected as a structural heterogeneity variable and is calculated by the entropy method in order to measure the environmental regulation structures in various regions of China. Drawing on the research findings of Zhao [39] and Peng [40], an environmental regulation can be categorized into two types: an explicit environmental regulation and an implicit environmental regulation. An explicit environmental regulation can be further categorized into three main types: a command-and-control environmental regulation, a market incentive environmental regulation, and a voluntary environmental regulation. Since voluntary environmental regulations are still in the early stages of development in China and play a small role [41], this article does not take into account the effects of voluntary environmental regulations. As for implicit environmental regulations, Pargal and Wheeler [42] believed that these can be measured by three indicators: the per capita income level, the education level, and population density. Li, RYM et al. [43] discovered that there is a certain degree of “overlap” between citizens’ sustainability awareness and legal regulations. The complete separation of law and morality proves ineffective in the environmental regulation of modern society. If nobody complies with a legislation, this will lead to the failure of enforcement [44]. Based on their research conclusions, we can infer that the level of sustainability awareness is of great significance for policy development. Sustainability consciousness can be viewed as an implicit form of environmental regulation, and there is a significant association between it and the advancement of environmental legislation. Consequently, this article considers the involvement of local inhabitants in environmental matters, which reflect the local sense of sustainability. Considering the availability of data, we utilize the quantity of petitions received regarding environmental concerns as a metric. Based on the information provided, this paper constructs the following evaluation index system of environmental regulation structures (

Table 2. Evaluation index system of ERSs.

Primary Indicators	Secondary Indicators	Meaning of the Indicator
Explicit environmental regulations	Command-and-control environmental regulations	Investment in environmental pollution control/CNY 100 million
	Market incentive environmental regulations	Income from sewage charges/CNY ten thousand
Implicit environmental regulations	Public participatory environmental regulations	The total number of environmental petitions/ individuals

).

3.1.3. Property Rights Structures (PRSs)

Drawing on previous research studies [32,45,46], in this study, the index of property rights structures is weighted by three indicators: investment, employment, and the income output value. The specific formula is as follows: $PRS=0.4PR1+0.3PR2+0.3PR3$. PR1 is defined as the ratio of the investment in fixed assets made by state-owned construction units to the overall investment in fixed assets of the construction industry. PR2 represents the ratio of state-owned construction units to the total number of construction businesses. PR3 is the ratio of the revenue output of state-owned construction units to the total construction revenue. Based on the characteristics of the above indicators, the PRSs mentioned later are all state-owned property rights. That is, a greater value of PRSs corresponds to a higher proportion of state-owned property rights.

To summarize, the chosen research variables and their corresponding definitions are displayed in

Table 3. Variable definitions.

Variable	Variable Symbol	Meaning
Level of high-quality development of the construction industry	HQ	Comprehensive evaluation of the five indicators by principal component analysis
Environmental regulation structure	ERS	Entropy weight method for calculating environmental regulation structure
Property rights structure	PRS	The proportion of investment, employment, and revenue output of state-owned units in the national total

.

3.2. Data Sources and Descriptive Statistics

This study primarily focuses on data collected from 30 provinces (including municipalities and autonomous regions directly under the central government) in China, spanning from 2008 to 2022. The dataset does not include Tibet, Hong Kong, Macao, and Taiwan. The data are mainly from the China Statistical Yearbook [47], the China Environmental Yearbook [48], and the China Construction Industry Statistical Yearbook [49]. Missing data are interpolated using interpolations. The results of a descriptive statistical analysis of specific variables are listed in

Table 4. Descriptive statistical tables.

Variable	Obs	Mean	Std. dev	Min	Max	Skewness	Kurtosis	Jarque–Bera
HQ	450	2.0819	1.1053	0.0160	6.3750	0.7470	3.6928	50.8501
ERS	450	0.1792	0.1200	0.0506	0.9067	2.5362	11.7488	1917.5764
PRS	450	0.1656	0.0954	0.1299	0.5443	1.3309	4.9930	207.3230

.

3.3. Panel Vector Autoregressive Model

In this research, the panel data vector autoregressive (PVAR) model proposed by Love and Zicchino [50] is employed for the analysis. PVAR integrates the advantages of conventional VAR models with a panel data analysis, effectively addressing the issue of individual heterogeneity by utilizing panel data without the need for satisfying the long-term series requirement of VAR models [51]. The PVAR model exhibits distinct characteristics compared to other panel data models. Instead of distinguishing between endogenous and exogenous study variables, this model assumes all variables as endogenous and incorporates the influence of lagged factors on other variables [52]. The PVAR model comprehensively considers individual fixed effects and time effects, analyzing the correlation between different variables through model impulse response and variance decomposition. In addition, the PVAR model is an appropriate instrument for analyzing macroeconomic dynamics, making it highly relevant for examining economic issues in various regions or countries [53]. The derivation and construction method of the PVAR model are as follows:

First, we consider a set of panel data $Y_{it}$:

$$Y_{it}=\left[\begin{array}{c}{y}_{1it}\\ y_{2it}\\ \begin{array}{c}{y}_{3it}\\ ⋮\\ y_{nit}\end{array}\end{array}\right], i=\mathrm{1,2},3,\cdots ,n; t=\mathrm{1,2},3,\cdots ,T$$

(1)

We can obtain a matrix form of the PVAR model by introducing the panel data into the VAR model.

$$Y_{it}={\sum }_{j=1}^k{\phi }_j{Y}_{i,t-j}+m_i+n_t+{\epsilon }_{it}$$

(2)

where

$${\phi }_j=\left[\begin{array}{ccc}\begin{array}{c}\begin{array}{cc}{\phi }_{11}\left(j\right)& {\phi }_{12}\left(j\right)\end{array}\\ \begin{array}{cc}{\phi }_{21}\left(j\right)& {\phi }_{22}\left(j\right)\end{array}\end{array}& \cdots & \begin{array}{c}{\phi }_{1n}\left(j\right)\\ {\phi }_{2n}\left(j\right)\end{array}\\ ⋮& \ddots & ⋮\\ \begin{array}{cc}{\phi }_{n1}\left(j\right)& {\phi }_{n2}\left(j\right)\end{array}& \cdots & {\phi }_{nn}\left(j\right)\end{array}\right], j=1,2,3,\cdots ,k,{\epsilon }_{it}=\left[\begin{array}{c}\begin{array}{c}{\epsilon }_{1it}\\ \begin{array}{c}{\epsilon }_{2it}\\ {\epsilon }_{3it}\end{array}\end{array}\\ ⋮\\ {\epsilon }_{nit}\end{array}\right].$$

(3)

i denotes the region; t denotes the year; and $Y_{it}$ is a column vector consisting of the high-quality development level (HQ), environmental regulation structure (ERS), and property rights structure (PRS). $Y_{i,t-j}$ is a column vector composed of the high-quality development level, environmental regulation structure, and property rights structure in region i with a lag of the j period. k represents the lag order. ${\phi }_j$ represents the slope coefficient of the lag of the j period. $m_i$ is the individual effect vector, $n_t$ is the time effect vector, and ${\epsilon }_{it}$ is the random disturbance term.

According to the research variables in this paper, the PVAR model in Equation (2) can be rewritten as follows:

$${HQ}_{it}=m_{1i}+n_{1t}+{\epsilon }_{1it}+\sum _{j=1}^k{a}_{1j}{HQ}_{i,t-j}+\sum _{j=1}^k{b}_{1j}{ERS}_{i,t-j}+\sum _{j=1}^k{c}_{1j}{PRS}_{i,t-j}$$

(4)

$${ERS}_{it}=m_{1i}+n_{1t}+{\epsilon }_{1it}+\sum _{j=1}^k{a}_{2j}{HQ}_{i,t-j}+\sum _{j=1}^k{b}_{2j}{ERS}_{i,t-j}+\sum _{j=1}^k{c}_{2j}{PRS}_{i,t-j}$$

(5)

$${PRS}_{it}=m_{1i}+n_{1t}+{\epsilon }_{1it}+\sum _{j=1}^k{a}_{3j}{HQ}_{i,t-j}+\sum _{j=1}^k{b}_{3j}{ERS}_{i,t-j}+\sum _{j=1}^k{c}_{3j}{PRS}_{i,t-j}$$

(6)

4. Regional Analysis by K-Means Clustering

4.1. K-Means Clustering

A cluster analysis can be employed under unsupervised conditions to categorize samples with high similarity into the same class of clusters by learning the intrinsic distribution structure of the data. Conversely, it can also be utilized to group distinct samples into separate class clusters. It is one of the main techniques of machine learning and data mining. K-means clustering, an iterative method for cluster analyses [54], is a fuzzy clustering technique first proposed by MacQueen [55]. It is computationally efficient and easy to implement. But there are also defects, such as a sensitivity to noise and outliers, the requirement to pre-determine the number of clusters, and so on. It is necessary to combine the characteristics of specific issues with the selection of suitable similarity calculation methods in specific applications.

Based on existing research and the literature [56,57], the fundamental procedures of the K-means algorithm are as follows: Initially, the dataset is partitioned into K groups. The next step is to assign each sample in turn to the category to which it is most closely related, based on calculating the similarity between the samples and the centers of the categories. The third step then entails recalculating the cluster center values for each category for which samples were obtained and using these updated cluster centers as the baseline for the next iteration of clustering. This process is repeated in a loop until the value of the clustering center remains constant after a number of consecutive rounds of iterations, resulting in the final clustering result.

4.2. Cluster Analysis Results

This study employs Matlab R2021b to perform a K-means clustering analysis on a dataset consisting of 30 provinces and municipalities. The analysis utilizes the average values of three variable indicators for the years 2018–2022 as the input data.

Before performing K-means clustering, it is essential to select the optimal number of clusters, denoted as K. In this research, the elbow method is employed to determine the optimal K value. The elbow method [58] is a way to determine the optimal K value using a graph of SSE and K values. The idea of the algorithm is that as K continues to increase, the data are segmented in more detail, and the SSE gradually decreases. When the K value is smaller than the actual number of clusters, the SSE value changes greatly. When k is equal to the actual number of clusters, the SSE value changes more gradually as the K value increases. Consequently, the graph between the SSE and K values is an “elbow-shaped” line chart, with the “elbow” being the optimal K value. In Figure 3, the blue line represents the relationship between SSE values and the number of clusters. And the number of clusters is determined to be K = 3. The clustering findings of the 30 provinces and municipalities are displayed in Figure 4, where three distinct colors represent the three different clustering results. To make the results clearer, the provinces clustered in different regions are marked with different colors in Figure 5. The specific provinces and their characteristics are detailed in

Table 5. Provinces and characteristics of different regions.

Region	Provinces and Municipalities	Characteristics
I	Beijing, Shanghai, Jiangsu, Zhejiang, Shandong, and Guangdong	High-quality level of development, high degree of environmental regulations, and high proportion of state-owned property rights.
II	Tianjin, Hebei, Liaoning, Shanxi, Fujian, Anhui, Henan, Hubei, Hunan, Sichuan, Chongqing, and Shaanxi	Medium level of quality development, medium level of environmental regulations, and medium share of state ownership.
III	Inner Mongolia, Jilin, Heilongjiang, Jiangxi, Guangxi, Hainan, Guizhou, Yunnan, Gansu, Qinghai, Ningxia, and Xinjiang	Low quality development, low environmental regulations, and low share of state ownership.

.

5. Empirical Results and Analysis

5.1. Unit Root Tests

To prevent the appearance of “pseudo-regression” in the model, it is essential to conduct a unit root test on the variables to ascertain their stationarity. Since the data used in this research are short panel data with a high number of observations (n) and a low number of time periods (T), the IPS test was chosen, so as to evaluate the stationarity of the three variables in the four samples (China, Region I, Region II, and Region III).

As is apparent from

Table 6. IPS unit root test for raw data.

Region	HQ	ERS	PRS
China	−4.2426 ***	1.8874	−4.4705
I	−1.2332	1.5030	0.3013 *
II	−1.6397 *	−2.3341 ***	−1.0812
III	−1.4258 *	−2.8866 ***	−1.0113

***, and * demonstrate significance levels at 1%, and 10%.

, some of the data do not pass the IPS unit root test, indicating that the raw data are not stationary. Next, the IPS unit root test was performed on the first-order differential data to verify whether the differential data are stable.

As is apparent from

Table 7. IPS unit root test for first-order differential data.

Region	∆HQ	∆ERS	∆PRS
China	−9.9700 ***	−9.2871 ***	−10.4187 ***
I	−3.5356 ***	−5.2992 ***	−4.8679 ***
II	−4.5091 ***	−11.8601 ***	−5.4196 ***
III	−5.5755 ***	−7.2085 ***	−6.0903 ***

*** demonstrates significance levels at 1%.

, the null hypothesis can be declined at least at the 1% level of significance for each variable after first-order differences, indicating that the data, after first-order differences, are stable.

5.2. Cointegration Test

It is evident that all three variables are first-order single integers, and there is a same-order single-integer relationship between variables. Further cointegration tests can be conducted on variables to examine whether there exists a sustained equilibrium relationship between the variables. The Pedroni cointegration test was conducted on three variables, and the findings of the cointegration test are presented in

Table 8. The Pedroni cointegration test.

Region	Modified Phillips–Perron Test	Phillips–Perron Test	Augmented Dickey–Fuller Test
China	3.2957 ***	−6.1020 ***	−6.0076 ***
I	2.0704 **	−1.4420 *	−1.7893 **
II	1.4164 *	−5.7065 ***	−5.8123 ***
III	2.4678 ***	−6.5818 ***	−2.4213 ***

***, **, and * demonstrate significance levels at 1%, 5%, and 10%.

.

From the above table, it can be seen that the null hypothesis can be declined with a significance level of at least 10%. This implies that there is a cointegration association between the variables, and a PVAR model may be constructed.

5.3. PVAR Estimation

5.3.1. Lag Selection

Before using the PVAR model, it is essential to ascertain the most effective lag order for each model. An excessively large lag period will result in a significant reduction in the degrees of freedom and will compromise the accuracy of the model’s estimates. Conversely, a lag period that is too small will lead to a substantial loss of sample data. Therefore, using Stata17, the lag order of the model was calculated based on the Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC), and Hannan–Quinn Information Criterion (HQIC), and the results are displayed in

Table 9. Selection of optimal lag orders.

Region	Lag	AIC	BIC	HQIC
China	1	−8.7747 *	−7.7680 *	−8.3756 *
	2	−8.1743	−7.0084	−7.7107
	3	−7.7812	−6.4342	−7.2439
I	1	−7.9531 *	−7.1373 *	−7.6265 *
	2	−7.7236	−6.5853	−7.2705
	3	−7.3495	−5.8566	−6.7596
II	1	−8.7328 *	−7.8530 *	−8.3754 *
	2	−8.1453	−7.0316	−7.6927
	3	−6.3332	−4.9573	−5.7741
III	1	−10.3161 *	−9.1680 *	−9.7570 *
	2	−10.0477	−8.8197	−9.6904
	3	−9.9334	−8.9402	−9.4809

* indicates the optimal lag order selected by the criterion.

. Grounded on the ideas of minimizing AIC, BIC, and HQIC [59], the optimal lag order of the PVAR model in China is 1 order, and Region I, Region II, and Region III are 1 order.

5.3.2. PVAR Model Regression

Once the lag order was determined, the PVAR model was examined using the System Generalized Moment Estimation (System GMM) method [60]. The regression results for the PVAR model are displayed in

Table 10. The results of PVAR model regression.

Region	Variable	HQt−1	ERSt−1	PRSt−1
China	HQ	0.0865	8.1120 ***	−5.3871 ***
	ERS	0.0001	0.7311 ***	−1.1451 ***
	PRS	−0.0024	0.1689 ***	0.5566 ***
I	HQ	0.0883	1.4157	−16.2496 **
	ERS	0.0003	0.9152 ***	−0.0772
	PRS	0.0019	−0.0785 **	0.2432
II	HQ	0.0377	7.9106 ***	−10.4267 ***
	ERS	−0.0016	0.5320 ***	−0.4703 ***
	PRS	−0.0067	0.2384 ***	0.5261 ***
III	HQ	−0.0913	34.7050 ***	−0.4553
	ERS	0.0033	0.3668 ***	−0.0635 **
	PRS	−0.0268	2.7607 ***	0.7782 ***

***, and ** demonstrate significance levels at 1%, and 5%. HQt-1, ERSt-1, and PRSt-1 indicate the time lag of GTFP, DE, and IS, respectively.

. Firstly, for HQ, row 1 of the table indicates that there are two coefficients that are statistically significant, and both of them are significant at the 1% level. That is, HQ is very much influenced by ERSs and PRSs, with one-period-lagged ERSs having a significant positive effect on HQ, while one-period-lagged PRSs have a significant negative effect on HQ. This implies that as the environmental regulatory structure increases, HQ construction will increase, while as the state ownership structure increases, this will inhibit the growth of HQ construction. Secondly, regarding ERSs and PRSs, one-period-lagged versions of both variables have a significant effect on their own, and both are also significant at the 1% level with respect to each other.

Due to the dynamic nature of the PVAR model, the regression results do not adequately capture the relationship between ERS, PRS, and HQ. Therefore, we conducted an impulse response analysis and a variance decomposition analysis to more accurately depict the interactions among the three variables.

5.4. Granger Causality Test

The granger causality test is used to test whether there is a granger causality between HQ, ERS, and PRS. The prerequisite for the granger causality test to be used in panel data is that the variable sequence is stationary [61]. The panel data used in this paper passed the IPS heterogeneous unit root test and cointegration test, and the appropriate lag order was selected, so there are conditions for the application of the granger causality test on the panel data. The results are displayed in

Table 11. Granger causality test results.

Region	Variable	HQ	ERS	PRS
China	HQ		64.542 ***	24.511 ***
	ERS	0.002		13.355 ***
	PRS	0.357	12.412 ***
I	HQ		0.670	4.754 **
	ERS	0.005		0.342
	PRS	0.663	4.458 **
II	HQ		70.387 ***	32.809 ***
	ERS	0.080		38.492 ***
	PRS	1.730	20.111 ***
III	HQ		33.488 ***	0.336
	ERS	2.188		6.410 **
	PRS	2.257	25.936 ***

***, and ** demonstrate significance levels at 1%, and 5%.

. In the national region and Region II, there is bidirectional causality between ERS and PRS. Unidirectional causation is from ERS to HQ and from PRS to HQ. For Region I, the unidirectional causality can be seen from ERS to PRS and from PRS to HQ. In Region III, the same bidirectional causality exists between ERS and PRS, and the same unidirectional causality exists between ERS and HQ.

5.5. Model Stability Test

The purpose of the model stability test is to ensure that the coefficient estimates of the PVAR model maintain statistical significance under sampling variations and to avoid a singularity of the solution of the model [62]. Only when the model itself is stable will the subsequent dynamic effect analysis of the impact response analysis and variance decomposition be reliable and meaningful. Therefore, using the PVAR model to create an AR root plot, if the modal values of all characteristic roots (unit roots) are strictly less than 1, that is, located in the unit circle, then this indicates that the PVAR model is stable [63]; otherwise, the model is unstable. Figure 6 shows the test results of the nation, Region I, Region II, and Region III, and all feature roots fall within the unit circle. So, this confirms that the constructed PVAR model is stable at both national and regional levels, which lays the foundation for the subsequent dynamic impact analysis.

5.6. Impulse Response

The regression results of the PVAR model analyzed the relationship between ERSs, PRSs, and HQ in the building sector but do not reflect the dynamic adjustment process shown by the dependent variable when the independent variable is impacted [64]. In order to reveal the dynamic conduction mechanism between variables, the impulse response function (IRF) analysis method was used [65]. The impulse response analysis can simulate the dynamic change path of a single explanatory variable during both the present and future phases once the variable exhibits a standard deviation, so as to intuitively show the dynamic interactional effect between the variables and their duration [66]. Specifically, through running the Monte Carlo simulation 500 times, this paper obtains the 0–10 impact response process of key factors to the HQ development of the construction industry. The prediction spans a duration of 10 years, commencing from the present year of 2022 and extending from 2023 to 2032. It should be noted that if the response curve is above the horizontal axis, the positive impact of the explanatory variable will cause a positive dynamic adjustment of the explained variable; otherwise, if the response curve is below the horizontal axis, this means that the positive impact of the explanatory variable will lead to the negative dynamic adjustment of the explained variable [67].

5.6.1. Impulse Response in China

Figure 7 reports the impulse effects for China. It is evident that the negative effect of HQ on PRS peaks in the first period and the positive effect on ERS peaks in the third period and then tends to plateau. That is, the growth of HQ will contribute to the improvement of environmental regulations and will encourage the development of non-state business ownership. This is because with the development of high-quality growth in the building sector, it will provide more favorable compensation conditions for technological innovation, and enterprises will prefer to choose “environmental cost” over “survival of the fittest” [68,69].

For ERS, as can be seen in the figure, the positive impact on PRS peaks in the second period, followed by a slow rise to a stable level. The response of a one-standard-deviation shock to HQ growth is also positive in years 1–10, peaks in the first year, and then tapers off. This suggests that the environmental regulation structure is one of the influential factors in promoting HQ development, which is also consistent with the findings of some previous empirical studies [70]. For this reason, sound environmental regulations can provide “innovation compensation” to industry entities [71], thereby improving firm performance and industry efficiency [72,73].

The negative effect of PRS on HQ peaks in the first period and then gradually decreases to a plateau. Consistent with the theory of ownership property rights [74,75], the structural variables of state-owned property rights have obvious harmful effects on industrial performance. This is due to the fact that the presence of a significant quantity of state-owned firms in some regions and industries has led to a significant issue of inefficient allocations of resources. Several zombie enterprises occupy production resources with the help of external forces, hinder the growth and expansion of high-efficiency enterprises, and even eliminate advanced production capacities during periods of economic adjustments, causing serious resource misallocations and damaging the growth potential of the economy [76].

5.6.2. Impulse Response in Region I

Figure 8 reports the impulse effect for Region I. After facing a one-standard-deviation shock from HQ, the impulse effects of ERS and PRS reach a peak of beneficial influence in the first period and subsequently diminish over time, until they reach a stable level. In contrast to the above, in Region I, PRS shows a weak positive response to shocks facing HQ.

The implementation of ERS has a favorable influence on HQ, reaching its highest point in the third phase and subsequently declining progressively while still maintaining a consistently high degree of beneficial impact in the short term, showing a typical “inverted U-shaped” relationship [77,78].

The impulse effect of PRS on HQ fluctuates and generally shows an inhibitory effect, with the largest negative peak when compared to other regions. That is, the short-term suppression of HQ by PRS is most pronounced in Region I. This is due to the higher level of economic development in Region I and the greater competition among enterprises; so, there is a greater need to allow for the non-state capital to participate in state-owned enterprises, including the foreign capital, the private capital, and the entry of other legal persons. The diverse ownership structure will provide greater vitality for the development of construction firms and will truly break the pattern of the government’s “one share of the big one”. Liu et al. [79] concluded that non-state economic components have a positive effect on enterprise innovation investments and that the diversification of property rights can improve the enterprise innovation environment and enhance the core competitiveness of enterprises.

5.6.3. Impulse Response in Region II

Figure 9 reports the impulse response for Region II. Both ERS and PRS on HQ peaked positively versus negatively in the first period, and HQ on PRS peaked negatively in the first period, after which they all leveled off, and the results are consistent with the trend of the national response map.

The difference is that Region II, when faced with a one-standard-deviation shock to HQ, exhibits some dampening effects on ERS in the early stages. It peaks negatively in the first period and then rises gradually until it has a boosting effect on ERS in the later stages.

5.6.4. Impulse Response in Region III

Figure 10 reports the impulse response for Region III. Consistent with the national trend graphs, HQ shows promoting and inhibiting effects on ERS and PRS, respectively, and ERS has a positive effect on HQ, which reaches a maximum in the first period. In particular, the peak positive response of ERS to HQ is largest in Region III, and this indicates that the impact of ERS on HQ is strongest in the provinces with less stringent environmental regulations in Region III. This aligns with the observed outcomes of [80]. The reason for this is that increased environmental regulation in low-regulation provinces requires companies to implement measures aimed at mitigating or managing the environmental damage. On the one hand, this will compel enterprises to augment their present environmental management expenses in order to adapt to alterations in environmental regulations [81]. This requires enterprises to actively promote technological innovation and transformational development. On the other hand, this will compel enterprises to enhance their environmental consciousness during the process of selecting and producing goods [82]. As a result, in Region III, HQ is more affected by environmental regulations. These results correspond with the conclusions of Lin et al. [83] and Zheng et al. [27] and verify the theory of new structural environmental regulations and the new structural Porter hypothesis.

And when faced with a one-standard-deviation shock to PRS and after reaching a positive peak in the first period, HQ first shows a small positive response. And then there is a stronger negative response in the later stages. Analyzing reasons for this from the theory of new structural economics, it can be noted that, in Region III, the economic development is relatively low, the technology is relatively backward, the talent is relatively scarce, and the market vitality is insufficient; so, it does not have a structural comparative advantage. If the proportion of state-owned firms increases, it may increase the financial resources and viability of businesses in the near future. However, in the long term, the incentives and constraints of the enterprise distribution system will not be effective enough, and the innovation and development of enterprises will be inhibited to a greater extent [84].

5.7. Variance Decomposition

Unlike the impulse response function, the variance decomposition of the prediction error (FEVD) allows for another measure of the strength of the effect of each variable from the other side of the equation. It measures the significance of the effect of individual variables on the remaining endogenous variables. This approach corroborates and complements the results of previous regression and impulse response analyses, allowing for a more comprehensive understanding of the impact mechanisms. Based on this, in order to more accurately assess the extent of the impact of ERSs and PRSs on the HQ of the construction industry, this paper adopts the variance decomposition method, with a forecasting period of 10 years for the four samples. This paper reports the weight of the contribution of ERS and PRS to the changes in the HQ at periods 2, 4, 6, 8, and 10. It provides insights into the dynamic process of the variation of the impact effects. The pertinent findings are displayed in

Table 12. Variance decomposition of the PVAR model (%).

Region	Period	HQ	ERS	PRS
China	2	81.52	11.80	6.68
	4	70.88	16.88	12.24
	6	69.06	17.25	13.69
	8	68.75	17.24	14.01
	10	68.71	17.23	14.06
I	2	66.15	0.41	33.81
	4	62.29	2.08	35.62
	6	60.88	4.28	34.83
	8	59.90	5.82	34.28
	10	59.20	6.91	33.89
II	2	69.61	21.43	8.96
	4	61.80	20.68	17.52
	6	60.74	20.84	18.42
	8	60.62	20.98	18.39
	10	60.61	20.99	18.39
III	2	74.59	25.35	0.06
	4	70.82	27.04	2.14
	6	69.88	26.84	3.28
	8	69.57	27.01	3.41
	10	69.53	27.05	3.41

.

By comparing the variance of the results, the following can be clearly seen: At the national and regional levels, the interpretation of the change of HQ construction is always the largest, but the self interpretation will gradually decrease over time. In terms of the impact of the ERS, it explains more of the change in HQ construction than the PRS in China, Region II, and Region III, and this influence is continuing to grow. This finding affirms that the swift advancement of environmental regulations within a dual-carbon framework is significantly and deeply contributing to the improvement of high-quality growth in the building sector. It can also be seen that PRS is another important factor affecting HQ. Particularly in Region I, the extent of its influence exceeds that of ERS and explains a large and rising share of HQ.

6. Discussion

There are numerous studies on the high-quality development of China’s construction industry, but few scholars have examined this issue from a structural perspective. This paper innovatively applies the theory of “new structural economics” to the field of construction in China, a developing country. This approach is well-suited for the economic position of China and aligns with the focus areas of China’s supply-side structural reform program. Due to China’s enormous expanse of land and varying levels of development in the construction industry, researchers have commonly categorized the country into three areas, eastern, central, and western, by geographical location. In this paper, the K-means clustering algorithm was employed to scientifically divide China into three samples for analysis, utilizing three data indicators: HQ, ERS, and PRS. This method enhances the relevance of the research and represents one of the innovations of this paper.

The purpose of this paper is to establish a PVAR model that illustrates the dynamic adjustment process of one variable in response to another. This study simulates the dynamic change path and influence degree of different explanatory variables at different regional levels, predicting the dynamic relationship among HQ, ERS, and PRS over the next decade. This analysis aims to provide a more targeted direction for the structural adjustments of China’s construction industry across different regions, ultimately fostering a more balanced and high-quality growth.

K-means clustering divides provinces in China into three regions, each exhibiting varying levels of development in the construction industry. After a complete empirical study, the two research hypotheses, H1 and H1, in this article have been validated. The dynamic impact of high-quality development, in conjunction with environmental regulation structures and property rights structures, differs in regions with distinct development levels of the construction industry. This study expands on the new structural Porter hypothesis and property rights theory. This finding affirms that the swift advancement of environmental regulations within a dual-carbon framework significantly contributes to the improvement of high-quality growth in the building sector. And the degree of influence is the greatest in Region III. Conversely, PRS has the greatest impact in Region I. This is due to the high level of economic development in Region I, where the intensity of environmental regulations is higher, the environmental awareness of citizens is also higher, and the competition between enterprises is greater. Consequently, there is a greater need to allow non-state capitals to participate, such as the foreign capital and private capital, so as to provide greater vitality for the development of construction firms. In contrast, Region III is relatively low, the market vitality is insufficient, and there is no structural comparative advantage. If the proportion of state-owned enterprises is increased, it is possible that in the near future, the financial resources and viability of enterprises will increase. However, in the long run, the “zombie effect” may emerge [85]. Zombie enterprises occupy production resources with the help of external forces, hinder the growth and expansion of high-efficiency enterprises, and even eliminate the advanced production capacity during periods of economic adjustments. This will cause serious resource misallocations, damaging the growth potential of the economy [76]. In order to accelerate economic growth, many provinces in Region III have adopted an extensive development model, which does not adapt to the current new development concept and high-quality development requirements. It is also necessary to change the structure of environmental regulations, forcing enterprises to choose “environmental cost” to achieve a sound development of the construction industry.

This has important implications for policy developments. The promotion of the building industry’s high-quality process should adhere to the principle of adapting measures to local conditions and taking advantage of the situation. To establish and improve the high-quality development system, it is essential to consider not only the overall level of China but to also focus on improvements at the regional level. To improve the development level of various regions, it is imperative to formulate appropriate strategies for high-quality development from the perspective of adjusting various structural factors, taking into account the specific situation of the location. If the local government ignores the coordination between the environment and property rights and blindly increases the intensity of environmental regulations, it could severely hinder growth.

The research presented in this paper also has certain limitations. (1) The mediating effect model may be used in future work to investigate the transmission mechanism of the impact of various variables on the high-quality development of the construction industry. Additional possible follow-up extensions of this study include the inclusion of other impact indicators and other structural impact indicators. (2) This paper takes China, a developing country, as the research object, and the data have certain limitations. A possible extension is to explore the policy of the construction industry at the international level.

7. Conclusions and Contributions

7.1. Conclusions

Based on the theory of new structural economics and the policy of China’s supply-side structural reform, this paper analyzes the problem from a structural perspective. Using panel data collected from 30 provinces in China spanning the period from 2008 to 2022, this study considers the regional heterogeneity among the samples through K-means clustering. PVAR techniques were employed to investigate the dynamic relationship among ERS, PRS, and HQ and to analyze the effects of various structural heterogeneity factors in different regions. The conclusions of this research can be summarized in three major points.

(1)

At the national level, ERS is a positive factor for HQ, whereas PRS is an inhibitor. It can be concluded from the GMM regression results and the granger causality test that there is a unidirectional causality between ERS and HQ and PRS and HQ. According to the regression coefficient, strict environmental regulations will promote the high-quality development of the construction industry, while state-owned property structures will be inhibited at the national level.

(2)

ERS has varying impacts on HQ in different regions. The results from the IRFs indicate that the positive effect of ERS on HQ is most significant in provinces with a lower degree of environmental regulations in Region III. The FEVD results suggest that the future ERS will play a relatively important role in influencing HQ in Regions II and III over the next 10 years.

(3)

PRS has distinct impacts on HQ in different regions. The results from the IRFs show that the short-term dampening effect of the PRS on HQ is most pronounced in Region I, which exhibits a higher degree of construction development. Conversely, PRS will produce a weak positive effect in the near future in Region III. The FEVD analysis demonstrates that the impact of PRS on HQ becomes more significant as time progresses, compared with ERS.

(4)

HQ demonstrates a notably weak differential effect on ERS and PRS across various regions. Additionally, ERS and PRS exhibit distinct interactions, as illustrated by the IRF figures. In different regions, these factors may either promote or inhibit one another. The local government should not ignore the coordination between the environment and property rights. A unilateral increase in the intensity of environmental regulations could significantly impede growth.

7.2. Academic and Practical Contributions

The theory of “new structural economics” specifically addresses the development of the economy and society in developing countries. This research applies this theoretical innovation to the field of a construction industry, and studies different regions of China as distinct stages of development and economic structure. By analyzing the problem from the structural direction, this study not only responds to the policy of “supply-side structural reform” in China but also provides a novel theoretical perspective for the study of the high-quality development of China’s building sector. The PVAR dynamic correlation appraisal enriches Porter’s hypothesis of the new structure and the theory of property rights. Additionally, this research offers novel restructuring concepts for the development of other industries.

Practically, the high-quality growth of the building sector contributes to the improvement of the quality of building products. Furthermore, high-quality development is not only conducive to technological progress, industrial structure upgrading, and the improvement of labor productivity, but it also plays a pivotal and influential role in the development of industries both upstream and downstream. Moreover, with respect to responding to the policy of “supply-side structural reform” in China, it will promote the regional coordinated development of China’s building sector through rational restructuring. The promotion of the building industry’s high-quality development should follow the principle of adapting measures to local conditions and taking advantage of the situation. To improve the development level of various regions, it is imperative to formulate appropriate strategies for high-quality development from the perspective of adjusting various structural factors, taking into account the specific situation of the location.