Introduction

Carbon emissions from the construction sector are currently 5% higher than before the epidemic and are at an all-time high, and such emissions from buildings and construction have already attracted the attention of the 27th United Nations Climate Conference (UN 2022). For example, infrastructure construction accounts for 80% of all carbon emissions in the United States and 69% in Europe (Dhakal 2009). In particular, the construction industry, a pillar industry in China, is responsible for 41% of the world’s total carbon emissions (Huang et al. 2018). To address this issue, the Chinese government hopes to reduce carbon emissions through green development (General Office of the State Council, PRC 2016). However, China still has serious environmental problems, unbalanced development, uneven allocation of resources, and imperfect policy formulation in terms of the process of reducing carbon emissions. On the one hand, China’s industrial structure is based on heavy industry and a coal-based energy structure, which causes serious pollution (Chen et al. 2017). On the other hand, due to the vastness of China’s territory, it is characterized by unbalanced regional development and inconsistent levels of industrial transformation.

Therefore, the Chinese government has formulated policies related to carbon emission reduction represented by the smart city policy (SCP). The aim of the SCP is to promote the construction of new urbanization and ensure green and low-carbon development (Su et al. 2022). Fortunately, the role of the SCP in alleviating the carbon emission contradictions of construction enterprises has attracted the attention of scholars. The research on the SCP and carbon emissions of construction enterprises is mainly divided into the following aspects. First, there is an aspect that found that SCP and carbon dioxide emissions have a nonlinear relationship, and the effect of the role does not change over time (Yigitcanlar and Kamruzzaman 2018). Second, the implementation of the SCP can significantly improve the low-carbon economy of Chinese cities, and its driving effect increases over time (Fan et al. 2021). Third, heterogeneity analysis finds that the contribution of the SCP to urban green and low-carbon development is more pronounced in large cities and resource-based cities (Cheng et al. 2022). Despite the differences in research methods, research regions, and research results with respect to the above empirical studies, they illustrate the existence of the SCP impact on carbon emissions. Unfortunately, existing SCP-related studies neither target the carbon emissions of construction enterprises in the Yangtze River Economic Belt (YREB) nor fully consider the impact mechanism of SCP on the carbon emissions of construction enterprises in the YREB.

In summary, to investigate the impact of the SCP on the carbon emissions of construction enterprises in the YREB and its mechanism, this paper combines the Political-Economic-Sociocultural-Technological-Environmental-Legal (PESTEL) model and the pollution halo hypothesis, considers the background of the SCP, and constructs a model to evaluate the impact of the SCP on the carbon emissions of construction enterprises. Meanwhile, this paper empirically examines the following four questions using YREB panel data compiled by the Chinese government from 2004 to 2021: Does the SCP affect the carbon emissions of construction enterprises in pilot cities? Are there heterogeneous effects? Do foreign direct investment (FDI) and R&D investment have mediating effects on the mechanism of the impact of SCP on the carbon emissions of construction enterprises? Does new urbanization construction have a moderating effect on the impact mechanism of the SCP on the carbon emissions of construction enterprises?

The main contributions of this paper are as follows: first, the innovative combination of the PESTEL model and the pollution halo hypothesis leads to constructing a model to assess the impact of the SCP on the carbon emissions of construction enterprises, thus expanding the framework of the research related to the policy evaluation with respect to carbon emissions of construction enterprises. Second, current econometric research has not yet conducted an in-depth study on how SCP affects the carbon emissions of construction enterprises, so this paper provides a new way of econometric thinking in terms of assessing the SCP and related research with respect to the carbon emissions of construction enterprises. Third, this paper introduces two variables, R&D and FDI, to explore the transmission mechanism of these variables in the SCP with respect to the carbon emissions of construction enterprises, hence broadening the research perspective of the impact of the SCP on the carbon emissions of construction enterprises. Fourth, this paper analyses the moderating effect of the new urbanization in the SCP on the carbon emissions of construction enterprises, thereby enriching the research related to the policy mix.

Theoretical background and research hypothesis

Theoretical background

PESTEL model

PESTEL is proposed based on the PEST model (Dale 2000), which is used to analyze the macroenvironment of enterprises (Richardson Jr 2006), and it provides a good solution for the sustainable development of enterprises. Previous studies on corporate carbon emissions have mainly explored the factors affecting corporate carbon emissions from the perspectives of the theory of planned behavior (Wang et al. 2022b) and system dynamics (Lai et al. 2017), focusing on organizational variables, policy variables and technological variables. On the one hand, the factors affecting corporate carbon emissions are governmental factors, as government intervention can promote the technological innovation of enterprises to reduce carbon emissions (Xiang et al. 2023); on the other hand, these factors are related to technological variables, as technological efficiency significantly affects the improvement of corporate performance by carbon regulation (Yang et al. 2021), and the pilot carbon emissions trading policy promotes the green technological innovation of enterprises (Ge et al. 2023). However, the literature tends to analyze the factors affecting corporate carbon emissions from a single perspective, such as government intervention (Du and Li 2020), technological innovation (Chen and Wang 2023), and industrial structure, and seldom considers combined multiple perspectives. In contrast to the above theories, Soares (2023) used the PESTEL model to analyze the enabling factors for developing microgrid solutions in Mozambique from six dimensions: political, economic, social, technological, legal, and environmental. Wang and Cao (2022a) constructed an improved multicriteria decision-making framework based on the PESTEL model to explore competitive strategic decisions to improve the green transformation of Chinese coal enterprises. Hence, by considering political, economic, social, technological, legal, and environmental factors, and based on the PESTEL model perspective, this paper identifies important factors affecting the carbon emissions of construction enterprises, i.e., the number of patents, urban economic development, investment in R&D, technological innovation, education level, new urbanization construction, and waste treatment.

In summary, the PESTEL analysis model can analyze the external environmental factors and can identify the important factors affecting the enterprises. Although current relevant studies using the PESTEL model have not yet addressed the issue of carbon emissions in construction enterprises, the model provides a theoretical basis for identifying key factors affecting carbon emissions in construction enterprises from the political, economic, social, technological, legal and environmental dimensions.

Pollution halo hypothesis

The pollution halo hypothesis, which is the antithesis of the pollution haven hypothesis, suggests that investments by enterprises from developed countries help reduce environmental pollution in host countries because of the use of green technologies in their production activities (Mert and Caglar 2020). The results of studies on FDI have varied widely. Some scholars argue that FDI exacerbates pollution in host countries and note the pollution haven hypothesis in this regard (Wang et al. 2019). Other scholars argue that China’s agricultural technology spillovers have a pollution halo effect, i.e., they do not affect the local agricultural environment (Liu and Xu 2021), and that FDI has no effect on host country carbon emissions (Al-Mulali and Tang 2013). However, there are few studies on whether FDI affects the carbon emissions of construction enterprises. Therefore, this paper considers the PH hypothesis to study the effect of FDI on the carbon emissions of construction enterprises and its mechanism of action.

In summary, the PESTEL analytical model and the pollution halo hypothesis provide the theoretical basis for this paper in terms of constructing a model to assess the impact of the SCP on the carbon emissions of construction enterprises.

Research hypothesis

Impact of SCP on the carbon emissions of construction enterprises

The SCP is often defined as the primary means of achieving sustainable urban development and is therefore often advocated by governments (Martin et al. 2018). At the same time, it is the key to solving the problem of rapid urbanization (Datta 2015) and it represents a new way of thinking about urban space and future development goals (Hu et al. 2017). Existing studies on the evaluation of SCP have mostly focused on several aspects: first, the importance of implementing SCP is provided from the perspective of a low-carbon economy (Guo et al. 2022); second, the SCP is an important means to achieve sustainable urban development (Guo and Zhong 2022); third, the SCP can effectively improve the air quality of cities and thus curb the level of environmental pollution in cities (Gao and Yuan 2022). Construction enterprises, as an important part of urban development, improve the level of urban economic development but they also cause a large amount of environmental pollution (Li et al. 2022b). However, current studies on SCP evaluation tend to ignore the perspective of construction enterprises. Therefore, this paper starts from the perspective of the carbon emissions of construction enterprises to assess SCP. At the same time, the paper argues that the SCP can be used as a means to realize green development as well as carbon emission reduction in construction enterprises. Based on the above analysis, this paper proposes the following hypothesis:

Hypothesis 1: The SCP will curb the carbon emissions of construction enterprises.

Heterogeneity impact of the SCP on the carbon emissions of construction enterprises

Due to the vastness of China’s land, the effect of the SCP on carbon emission reduction in construction enterprises is affected by location, i.e., there is regional heterogeneity (Guo et al. 2022). The effect of the SCP on carbon emission reduction in China’s cities varies across different regions (Song et al. 2023), sizes (Jiang et al. 2021), and characteristics (Qian et al. 2021), resulting in different carbon emission reduction effects for different cities. Most studies have found that the SCP is effective in reducing carbon emissions in eastern cities, nonresource-based cities, and cities with higher levels of economic development (Wang et al. 2022). In this paper, the YREB is chosen as the study area because the region covers 11 provinces and cities, whereby the regional economic development is not balanced, and the industry types are different, so the implementation of the SCP may have large differences across different regions (Zhang et al. 2021). Therefore, this paper concludes that there is heterogeneity in the impact of the SCP on the carbon emissions of construction enterprises in the YREB across different river basins. Based on the above analysis, this paper proposes the following hypotheses:

Hypothesis 2: The SCP will curb the carbon emissions of construction enterprises in the upstream and downstream regions of the YREB.

Hypothesis 3: The SCP has a promoting effect on the carbon emissions of construction enterprises in the midstream of the YREB.

Mediating effect of FDI and R&D

Although the impact of FDI on carbon emissions has received scholarly attention, the relationship is controversial. Some studies suggest that FDI has a positive impact on China’s carbon emissions (Zhang and Zhang 2018), which supports the pollution halo hypothesis to a certain extent. However, the relationship between FDI and carbon emissions is not purely linear, and there are heterogeneous effects (Huang et al. 2022). In summary, FDI increases carbon emissions in countries and regions with lower levels of development and curbs carbon emissions in cities with strict environmental regulations, higher levels of economic development, and higher innovation levels. As the largest developing country, FDI in China has been an important economic source for enterprises, and thus it has an impact on the carbon emissions of such enterprises.

R&D is often studied as a key variable in current studies and is considered conducive to reducing the intensity of carbon emissions (Luan et al. 2019). In addition, Pan used R&D as an explanatory variable in the study of carbon emission reduction, and the results showed that the mandatory carbon emission reduction policy limited the expansion of the R&D scale of enterprises (Pan et al. 2021). At the same time, R&D is used as a mediating variable to verify the promotion effect of clean energy application on low-carbon development (Lin and Li 2022). However, few studies have used R&D as a mediating variable to explore the mediating effect of SCP on the carbon emissions of construction enterprises.

In summary, for China’s YREB, the SCP affects FDI and R&D levels toward curbing the carbon emissions of construction enterprises. Based on the above analysis, this paper proposes the following hypotheses:

Hypothesis 4: The SCP curbs the carbon emission level of construction enterprises through the mediating effect of reducing the level of FDI.

Hypothesis 5: The SCP has the mediating effect of increasing the R&D level to curb the carbon emission level of construction enterprises.

Moderating effect of new urbanization

In China, a single carbon emission reduction policy does not necessarily achieve the desired effect, and it is necessary to form a policy combination with corresponding policies to achieve the required effect (Zhao et al. 2020). New urbanization construction is a key step in the construction of SCP (Wang et al. 2023); by exploring the moderating effect of new urbanization construction in the process of influencing the carbon emissions of construction enterprises, it is realized that new urbanization construction is of great significance to the development of the SCP. Many studies have proven that new urbanization construction has a positive effect on the green development of enterprises (Xiao et al. 2023). In some of these studies, new urbanization construction is considered a new means to balance urbanization and carbon emissions (He and Liu 2022). However, there is no research that proves whether new urbanization construction can have a moderating effect between the SCP and the carbon emissions of construction enterprises. Therefore, this paper innovatively introduces the variable of new urbanization construction to explore its moderating effect on the influence of the SCP on the carbon emissions of construction enterprises. At the same time, this paper argues that new urbanization construction will positively regulate the inhibiting effect of the SCP on the carbon emissions of construction enterprises. Based on the above analysis, this paper proposes the following hypothesis:

Hypothesis 6: The construction of new urbanization can positively regulate the inhibiting effect of the SCP on the carbon emissions of construction enterprises.

In summary, Fig. 1 gives a model for assessing the impact of the SCP on the carbon emissions of construction enterprises.

Fig. 1: A model to assess the impact of smart city policies on the carbon emissions of construction enterprises.
figure 1

Yangtze River Economic Belt (YREB). First, we hypothesize the effect of smart city policy on the carbon emissions of construction enterprises; second, we hypothesize the effect of smart city policy on the heterogeneity of carbon emissions of construction enterprises; third, we hypothesize the mediating effect of FDI and R&D on the influence of smart city policy on the carbon emissions of construction enterprises; and finally, we hypothesize the moderating effect of new urbanization on the influence of smart city policy on the carbon emissions of construction enterprises.

Full size image

Methodology

Study area

The YREB refers to the economic circle along the Yangtze River, which covers 11 provinces and cities, including Shanghai, Jiangsu, Zhejiang, Anhui, Jiangxi, Hubei, Hunan, Chongqing, Sichuan, Yunnan and Guizhou, as shown in Fig. 2, covering an area of ~2,052,300 square kilometers, or 21.4% of the country, with a population and GDP of more than 40% of the country’s population and GDP. Meanwhile, the YREB is divided into upstream, midstream and downstream areas. Among these areas, the upstream region includes Yunnan, Sichuan, Guizhou, and Chongqing; the midstream region includes Hubei, Hunan, and Jiangxi provinces; and the downstream region includes Anhui, Jiangsu, Zhejiang, and Shanghai. This paper maps the carbon emission trends of construction enterprises in China’s YREB by calculating and collecting available data (Fig. 2).

Fig. 2: Carbon emissions of construction enterprises in the Yangtze River Economic Belt provinces.
figure 2

A shows provinces passing through the Yangtze River Economic Belt, B shows carbon emissions from construction enterprises in 2004, C shows carbon emissions from construction enterprises in 2012, and D shows carbon emissions from construction enterprises in 2021.

Full size image

Variables

Considering that this paper takes the carbon emissions of construction enterprises in the YREB as the research object, carbon emissions are chosen as the dependent variable in the regression model. At the same time, according to the model constructed in this paper to evaluate the impact of the SCP on the carbon emissions of construction enterprises, the number of patents is selected to measure political factors, the development of urban construction is selected to measure economic factors, the level of education is selected to measure sociocultural factors, R&D investment is selected to measure technological factors, the structure of the industry and the treatment of waste are selected to measure environmental factors, the construction of new towns and cities is selected to measure legal factors, and FDI is selected to test the pollution halo hypothesis. To test the pollution halo hypothesis, the dependent variables, core explanatory variables, control variables, mediating variables and moderating variables are as follows.

  • Dependent variable: carbon emissions from the construction industry. This paper refers to the energy conversion coefficients in China’s General Rules for Calculating Comprehensive Energy Consumption (Standardization Administration & State Administration for Market Regulation, PRC 2020) and converts carbon emissions after standardizing the types of energy consumed by construction enterprises in each region.

  • Core explanatory variables: SCP pilot (DID). In this paper, the regional dummy variable, \(treated\), and the time dummy variable are set up. If the city is a pilot city, then treated takes the value of 1, and otherwise, it is 0. If the sample study time is after the pilot policy is carried out, then \(time\) takes the value of 1, and otherwise, it is 0. The core explanatory variable, DID, is the product of the regional dummy variable, \(treated\), and the time dummy variable, \(time\). The core explanatory variable, DID, is the product of the regional dummy variable, \(treated\), and the core explanatory variable, DID. The core explanatory variable DID is the product of the area dummy variable treated and the time dummy variable.

  • Control variables: considering that this paper takes the carbon emissions of construction enterprises as the dependent variable, after referring to the relevant literature, the paper selects the important variables that directly and indirectly affect the carbon emissions of construction enterprises as the control variables in the model based on the constructed theoretical framework model of the SCP evaluation; the specific indexes are as follows: (1) Number of patents: Yu and Zhang (2019) believe that patents can protect technological innovations, and therefore, the number of patents in each region is selected. In this paper, the number of patents in each region is chosen as a measure of the level of technological innovation, and thus is labeled PT; (2) industrial scale, this paper refers to Darnall et al.’s (2010) study and chooses the gross output value of the construction industry as an indicator of the industrial scale, hence it is labeled Acgdp; (3) urban construction and development, this paper refers to Yang and Chen’s (2010) study and chooses GDP per capita as an indicator of urban construction and development, thus it is labeled Cgdp; (4) education level, this paper refers to Mahalik et al.’s (2021) research, which analyses the impact of the level of education on the ecological environment and the impact of carbon emission intensity, so this paper chooses the number of students in higher education as a measure of the level of education, which is denoted Edu; and (5) waste treatment, this paper refers to Razzaq et al.’s (2021) study and chooses the efficiency of municipal industrial waste treatment as an indicator of waste treatment, and it is denoted FW.

  • Mediating variables: (1) scientific and technological research and development investment: this paper refers to Li et al. (2022a) and selects each region’s scientific and technological expenditure as a measure of scientific and technological research and development investment intensity, and thus it is labeled R&D; (2) foreign investment: this paper refers to foreign investment research as a measure of the corresponding indicator of technology introduction, hence it is labeled FDI (Fu et al. 2018).

  • Moderating variable: the interaction term of the SCP and new urbanization construction (Nurb). This paper refers to Lai’s (2022) study to adopt the pilot policy of new urbanization construction as a dummy variable. Therefore, the paper introduces the interaction term between the SCP and new urbanization construction as a moderator variable in the model to explore the moderating effect.

Data sources

The data in this paper mainly come from the China Statistical Yearbook, China Urban Statistical Yearbook and statistical yearbooks of provinces, and the number of patents in each region comes from the Patent Search Library of the State Intellectual Property Office and China Urban Statistical Yearbook. Due to the scientific nature and availability of data, this paper excludes some regions with substantial missing data. At the same time, because the implementation date of the SCP is 2012, to avoid the difference in results caused by the different times before and after the pilot node, this paper selects the data before and after the implementation node of 8 years and adds the latest data of official statistics. Finally, this paper selects panel data of 110 cities in 11 provincial administrative regions included in the YREB from 2004 to 2021, including 62 pilot cities and 48 nonpilot cities. Some of the missing data values are filled in by interpolation. To avoid the impact of extreme data values on the estimation results, the data are processed by taking the natural logarithm.

Methods

First, taking the carbon emissions state of construction enterprises as the research object, 110 cities in the YREB are selected as the study area. Second, a model is constructed to assess the impact of the SCP on the carbon emissions of construction enterprises, and this impact is analyzed by using the difference-in-differences (DID) method and the propensity score matching (PSM) method. Third, the mediating effect of R&D and FDI on the impact of the SCP on the carbon emissions of construction enterprises is explored through mediating effect analysis. Finally, through moderating effect analysis, the moderating effect of new urbanization construction on the impact of the SCP on the carbon emissions of construction enterprises is analyzed. Therefore, to clearly explain how the SCP affects the carbon emissions of construction enterprises and its mechanism, Fig. 3 gives the technical roadmap of this paper.

Fig. 3: Technology roadmap for the SCP affecting carbon emissions in construction enterprises.
figure 3

Yangtze River Economic Belt (YREB). First, a theoretical model framework for the evaluation of the SCP considering the carbon emissions of construction enterprises is constructed to screen seven relevant variables. Second, the DID and PSM methods are used to analyze the impact of the SCP on the carbon emissions of construction enterprises. Finally, the impact mechanism of the SCP is further analyzed using heterogeneity analysis, mediating effect analysis and moderating effect analysis.

Full size image

Difference-in-differences (DID)

Based on the DID method, the PSM-DID was proposed by Heckman to ensure the robustness of the findings by using the PSM method a second time to make the experimental and control groups as similar as possible in all characteristics (Heckman and Vytlacil 2001). Thus, the PSM-DID method was chosen to more accurately assess the effect of the SCP on the carbon emissions of construction enterprises.

The DID method is used to evaluate the SCP and the programs implemented at a particular point in time (Stuart et al. 2014). The DID method has the following advantages: (1) it ensures that endogeneity problems are avoided between the various sets of panel data; (2) the use of fixed effects estimation alleviates the problem of omitted variable bias to some extent; and (3) the DID model setting is more scientific than the traditional method. This paper focuses on assessing the impact of the SCP on the carbon emissions of construction enterprises, and hence it is essentially a policy evaluation study. At the same time, current studies have conducted policy evaluations of SCP through DID (Yao et al. 2020). Therefore, DID is chosen as the method of analysis in this paper. Specifically, the benchmark DID model setup is as follows:

$$Y_{it} = \alpha _0 + \alpha _1time \times treat + \varepsilon$$
(1)

where \(treat\) is the area dummy variable. If individual i is affected by policy implementation, then individual i belongs to the experimental group, corresponding to \(treat = 1\). If individual i is not affected by policy implementation, then individual i belongs to the control group, corresponding to \(treat = 0\), where \(time\) is the time dummy variable, \(time = 0\) before the year of policy implementation and \(time = 1\) after the year of policy implementation. \(time \times treat\) is the interaction term between the area dummy variable and the time dummy variable, and its coefficient α1 then reflects the net effect of policy implementation.

Propensity score matching (PSM)

The PSM method is a statistical method used for the observational processing of data from conducted experiments (Abadie 2005). In observational studies, bias and confounding variables often occur due to environmental influences and multifactorial interventions. The PSM method is designed to reduce these selective biases and to make the experimental and control group samples have similar characteristics so that a more reasonable comparison can be made between the experimental and control groups.

Let \(D_i = 1\) for experiments conducted and \(D_i = 0\) for no experiments conducted, while \(Y_i\) denotes the output variable tested. The counterfactual framework can then be represented as the following model:

$$Y_i = 1 - D_iY_i0 + D_iY_i$$
(2)

This model suggests that the outcome observed experimentally depends on the intervention state, i.e., the state of D. The average treatment effect for the treated is used to measure the average intervention effect for each area in the experimental state, i.e., it represents the difference between the observed outcome in area i in the experimental stage and its counterfactual and is called the standard estimate of the average intervention effect:

$$\begin{array}{l}ATT = E\left\{ {Y_i\left( 1 \right) - Y_i\left( 0 \right)|D = 1} \right\}\\ \qquad\;\, = E\left\{ {Y_i\left( 1 \right)|D = 1} \right\} - E\left\{ {Y_i\left( 0 \right)|D = 1} \right\}\end{array}$$
(3)

PSM-DID model

The SCP under study in this paper is considered based on the study area chosen for this paper as well as the identified study population. Therefore, this paper considers the SCP as a quasinatural experiment and uses the DID method to assess the impact of the SCP on the carbon emissions of construction enterprises. As seen from the previous section, the SCP was officially implemented in 2012, and several pilot cities were planned. In this paper, the pilot areas after the implementation of the policy are used as the experimental group, and the nonpilot areas are used as the control group. Therefore, to build a DID model, this paper first constructs two dummy variables, namely, the pilot area dummy variable and the pilot time dummy variable. The dummy variable is defined as 1 for the pilot region, 0 for the nonpilot region, 1 for the time in 2012 and afterward, and 0 for the time before 2012 (Eq. (4)):

$${\rm{CO}}_{2} = \alpha _0 + \alpha _1time_t \,*\, treat_i + \mathop {\sum}\limits_{j = 1}^N {\alpha _jcontrol_{ijt}} + \varepsilon _{it}$$
(4)

where i represents the region and t represents time. Where \({\rm{CO}}_{2}\) represents the carbon intensity of construction enterprises; \(treat_i\) represents the regional dummy variable, and the interaction term \(time_t \,*\, treat_i\) is the core explanatory variable, reflecting whether the corresponding policy has been implemented in the control group; and \(\varepsilon\) is the disturbance term.

$${\rm{CO}}_{2}^{PSM} = \alpha _0 + \alpha _1time_t \,*\, treat_i + \mathop {\sum}\limits_{j = 1}^N {\alpha _jcontrol_{ijt}} + \varepsilon _{it}$$
(5)

Mediating effect analysis

This paper refers to Zhou (2022) mediating effect test and uses the mediating effect model to analyze the roles of the two mediating variables, R&D and FDI, in the transmission mechanism, thus exploring the intrinsic mechanism of carbon emissions of construction enterprises in the pilot cities of the YREB, and using the following model settings:

In the first step, the effect of the SCP on the carbon emission level of construction enterprises in pilot cities is verified:

$${\rm{CO}}_{2} = \alpha _0 + \alpha _1time_{it} \times treat_{it} + \alpha _2Control + \varepsilon _{it}$$
(6)

In the second step, the effect of the SCP on R&D and FDI is verified:

$$Z = \beta _0 + \beta _1time_{it} \times treat_{it} + \beta _2Control + \varepsilon _{it}$$
(7)

In the third step, we verify the mediating effect of the SCP:

$${\rm{CO}}_{2} = \gamma _0 + \gamma _1treat_{it} \times time_{it} + \gamma _2Z + \gamma _3Control + \varepsilon _{it}$$
(8)

Z in the above equation denotes the mediating variable, which is R&D and FDI. Finally, this paper constructs a mediating effect analysis model, as shown in Eqs. (6)–(8).

Moderating effect analysis

This paper refers to the moderating effect test method in Feng and Wu (2022) and chooses new urbanization construction to measure the low-carbon constraints, substituting the low-carbon constraints into the regression model to analyze its regulatory effect on the SCP construction and the carbon emissions of construction enterprises. The construction of the moderating effect analysis model is shown in Eq. (9):

$$\begin{array}{l}{\rm{CO}}_{2} = \alpha _0 + \alpha _1time_{it} \times treat_{it} + \alpha _2time_{it1} \times treat_{it1}\\ \qquad\quad \,+\, \alpha _3time_{it2} \times treat_{it2} + \alpha _4Control + \varepsilon _{it}\end{array}$$
(9)

In the formula, \(time_{it1} \times treat_{it1}\) is 1, which means that the ith pilot city has implemented the new urbanization pilot in year t, otherwise it is 0; \(\alpha _2\) denotes the degree of influence of the new urbanization construction on the carbon emission level of the construction enterprises in the YREB; and \(time_{it2} \times treat_{it2}\) is 1, which means that the ith city has implemented both the SCP and the new urbanization construction in year t, otherwise it is 0. \(\alpha _3\) is the coefficient of the interaction term between the SCP and new urbanization construction, reflecting the effect of simultaneously influencing the carbon emission level of construction enterprises in the YREB through smart city construction and new urbanization construction.

Results

Benchmark regression results

Since the method chosen in this paper is DID, a parallel trend test should be carried out before the regression analysis, and the next regression analysis can be carried out only if the test result is passed. Meanwhile, to avoid multicollinearity, this paper refers to the practice of Ma (2023), which removes the dummy variables of the year of policy implementation. Finally, the results of the parallel trend test are shown in Fig. 4. As shown in Fig. 4, the carbon emissions of construction enterprises in the YREB did not have a large difference before the implementation of the policy, and the carbon emission level had a large change after the implementation of the policy, indicating that the research sample passed the parallel trend test.

Fig. 4: Results of the parallel trend test.
figure 4

Trends in variability (solid green line). Variability per period (solid blue dots). There was no significant difference between the experimental and control groups in the first 3 years of the policy implementation. In the 3 years following the implementation of the policy, there was a significant difference between the experimental and control groups.

Full size image

In this paper, the DID method is used to investigate the impact of the SCP on the carbon emissions of construction enterprises in the pilot cities of the YREB, and the regression results are shown in Table 1. As shown in column (1) of Table 1, the coefficient of the core explanatory variable DID is significantly negative without adding control variables, indicating that the SCP significantly curbs the carbon emission level of the construction enterprises in the pilot cities of the YREB. In addition, according to columns (2)–(6), the regression results of the core explanatory variable are still significantly negative under the conditions of controlling the variables of urban economic development, technological innovation, industrial structure, education level and waste disposal, and this further illustrates the validity of the conclusion.

Table 1 Benchmark regression results.
Full size table

PSM-DID regression analysis results

To ensure that the samples in the experimental and control groups have similar characteristics, this paper not only overcomes the systematic bias that may exist in the treatment and control groups through the PSM method but also reduces the DID estimation bias. In addition, after processing the study samples, this paper verifies the validity of the base regression results by conducting the DID regression again.

First, this paper adopts the 1:1 neighbor matching method to process the sample data and obtain the propensity score values; second, the obtained propensity score values are analyzed to retain the samples with similar characteristics in the treatment and control groups to reduce the systematic bias; last, the PSM-processed sample data are examined for balancing, and Fig. 5 shows that the matched sample data are well balanced so that the next regression analysis can be carried out.

Fig. 5: Balance test results.
figure 5

No significant difference points (x = 0 gray solid line). The unmatched sample data (black dots) were matched to give the matched sample data (x markers).

Full size image

As shown in Table 2, this paper performs the DID regressions a second time on the processed research sample. Specifically, the regression results in column (1) of Table 2 do not include control variables, and the regression results in column (2) control five variables. As shown in Table 2, the core explanatory variable is significantly negative at the 1% level for both cases, i.e., the implementation of the SCP significantly reduces the carbon emissions of construction enterprises in the pilot cities of the YREB. Meanwhile, comparing the regression results obtained after the PSM treatment with the baseline regression results, it is found that the coefficient of the core explanatory variable did not change significantly, i.e., it proves that the research results have good stability. This result verifies hypothesis 1. Therefore, further heterogeneity analysis will be carried out to explain the reason behind this result.

Table 2 PSM-DID regression results.
Full size table

Heterogeneity analysis results

From the regression results of the PSM-DID, it can be observed that the implementation of the SCP significantly reduces the carbon emissions of the construction enterprises in the pilot cities of the YREB. However, based on the hypotheses proposed, it is argued that there is heterogeneity in the SCP on the carbon emissions of construction enterprises in the YREB. Therefore, the YREB was divided into three regions: upstream, midstream and downstream (Network of the Development of the Yangtze River Economic Belt 2019). This paper conducts a heterogeneity test for each region.

The test results are shown in Table 3, where columns (1)–(3) indicate the regression results for the upstream, midstream and downstream regions, respectively. In addition, the regression results in columns (1) and (3) indicate that the SCP has a significant inhibitory effect on the carbon emission level of the pilot cities in the upstream and downstream regions of the YREB, reducing the level of carbon emission by 18.04% and 4.49%, respectively. The regression result in column (2) indicates that the SCP has a significant contribution to the carbon emission level of the construction enterprises in the pilot cities in the midstream of the YREB, increasing the carbon emission level by 14.55%. This result passes hypothesis 2 and hypothesis 3.

Table 3 Heterogeneity analysis results.
Full size table

Mediating effect analysis results

The results of the mediating effect analysis are shown in Table 4. Column (1) shows the regression results of formula (6), indicating that the SCP has a significant contribution to the carbon emission level of construction enterprises in the pilot cities of the YREB. Columns (3) and (5) show the regression results of formula (7), indicating that the coefficients of FDI are significantly negative, and the coefficients of R&D are significantly positive, indicating that the SCP significantly reduces the pilot city’s FDI level while it significantly increases the level of R&D in the pilot city. Columns (2) and (4) are the regression results of formula (8), and both the FDI and R&D coefficients are significantly positive, indicating that there is a mediating effect in the model, i.e., the SCP curbs the FDI level of the pilot city and it improves the level of R&D in the region, thus leading to an increase in the level of carbon emissions of the construction enterprises in the region. This result passes hypotheses 4 and 5 as proposed above.

Table 4 Mediating effect analysis results.
Full size table

Moderating effect analysis results

The results of the moderating effect analysis are shown in Table 5. After the introduction of the moderating variables, the coefficients of the core explanatory variables of the SCP did not change significantly, indicating that the impact of the SCP on the carbon emissions of construction enterprises in the YREB has not been affected by the impact of the construction of new urbanization, which further tests the robustness of the conclusions. Meanwhile, with the new urbanization construction, the coefficient of the interaction term between the SCP and the new urbanization construction is −0.096, indicating that there is a significant positive moderating effect of the new urbanization construction in the relationship of the SCP affecting the carbon emission level of the construction enterprises, i.e., the joint implementation of the new urbanization construction and the SCP enhances the inhibition of the carbon emission level of the construction enterprises in the pilot cities. Therefore, this paper concludes that a single SCP weakens the carbon emissions of construction enterprises, but the combination of other environmental governance policies may play a key role in the governance of the carbon emissions of construction enterprises. This result verifies Hypothesis 6.

Table 5 Moderating effect analysis results.
Full size table

Discussion

This paper constructs a model to assess the impact of the SCP on the carbon emissions of construction enterprises through the lens of the PESTEL model and the pollution halo hypothesis and employs the DID method to assess the impact of the SCP on the carbon emission intensity of construction enterprises in pilot cities. Meanwhile, the paper adopts the PSM method to build a new PSM-DID model to verify the robustness of the results. Finally, the mediating effect analysis and moderating effect analysis are used to determine the mechanism of how the SCP affects the carbon emissions of construction enterprises.

The direct impact of SCP

This paper finds that the SCP significantly curbs the carbon emission level of construction enterprises in the pilot cities of the YREB. As a new means to ensure the healthy and harmonious development of cities, the SCP is of great significance to the green and low-carbon development of the region. At the same time, the SCP is often considered to promote the green and low-carbon development of cities (Cheng et al. 2022), significantly reduce the carbon emission intensity of enterprises (Liu et al. 2022), and curb per capita carbon emissions (Guo et al. 2022). However, in contrast to previous studies, this paper chooses the perspective of construction enterprises and analyses the impact of the SCP on the carbon emissions of construction enterprises based on the study area of the YREB and concludes that it can significantly curb the carbon emissions of construction enterprises; this finding, to a certain extent, supports the viewpoints of previous studies that the SCP will promote urban carbon emission reduction.

Heterogeneity impact of SCP

There is heterogeneity in the impact of the SCP on the carbon emission levels of construction enterprises in the pilot cities of the YREB, i.e., it curbs the carbon emission levels of construction enterprises in the pilot cities in the upstream and downstream regions of the YREB and significantly promotes the carbon emission levels in the midstream region. First, current studies suggest that when the SCP is implemented, it promotes industrial upgrading downstream of the Yangtze River, where the level of industrial transformation is higher (Song et al. 2023). At the same time, stricter environmental regulations cause some highly polluting enterprises in the downstream region to accelerate their transformation or prompt them to move to areas with less stringent environmental regulations (Dou and Han 2019). Second, according to the environmental Kuznets curve, there is an inverted U-shaped relationship between environmental pollution and the level of economic development. Therefore, research suggests that the reason for the increase in carbon emissions from construction enterprises in the pilot cities in the midstream of the Yangtze River is that the midstream of the Yangtze River is in the midstream of economic development, implying a higher level of environmental pollution in the region (Wu et al. 2022). At the same time, the transfer of high-polluting enterprises in the downstream region and the low level of green development in the midstream region of the Yangtze River (Cui et al. 2021) lead to an increase in carbon emissions from construction enterprises in the pilot cities in the midstream region. Finally, the upstream region of the YREB has a lower level of economic development than other regions and a single industrial structure with a lower level of industrialization, leading to a lower level of environmental pollution in the region (Chen et al. 2021). When the SCP is implemented, it can further reduce the carbon emission level of the construction enterprises in the pilot cities of the region. Therefore, this paper supports the viewpoints of previous studies by analyzing the heterogeneity of the impact of the SCP on the carbon emissions of construction enterprises in the YREB.

Mediating impacts of SCP

R&D and FDI are important transmission mechanisms through which the SCP promotes the carbon emission levels of construction enterprises in the pilot cities in the YREB. Specifically, the SCP increases the carbon emission levels of construction enterprises in pilot cities by suppressing FDI and increasing R&D levels in pilot cities. Previous studies on the impact of FDI on carbon emissions have found that FDI significantly contributes to China’s carbon emission levels (Song et al. 2021) and significantly accelerates such emissions (Zheng et al. 2022), while R&D is considered the key to reducing domestic carbon intensity (Huang et al. 2020). Therefore, this paper analyses the impact of FDI and R&D on the carbon emission level of construction enterprises through the mediating effect, which supports the viewpoints of previous studies.

Moderating impacts of the SCP

The construction of new urbanization will promote the inhibiting effect of the SCP on the carbon emissions of construction enterprises. The results of the moderating effect analysis in terms of the introduction of new urbanization construction show that new urbanization construction has a positive moderating effect on the carbon emission reduction effect of the SCP on construction enterprises. Previous studies have concluded that the relationship between new urbanization and carbon emissions is not purely linear and that new urbanization has a contradictory effect on CO2 emissions and significant spatial heterogeneity (Wang et al. 2019), but the coordination of new urbanization construction and carbon emission reduction in China has been significantly improved (Jiang et al. 2022). This paper innovatively introduces the interaction term between the SCP and new urbanization construction as a moderating variable to verify the moderating effect of new urbanization construction, which enriches the current research on policy mixes.

Conclusions and implications

Conclusions

This paper is based on a model assessing the impact of the SCP on the carbon emissions of construction enterprises and clarifies the mechanism of the role of the SCP in influencing the carbon emissions of construction enterprises through the DID method, PSM method, mediating effect analysis, and moderating effect analysis. The main research conclusions are as follows:

  • The SCP significantly curbs the carbon emission level of construction enterprises in the pilot cities of the YREB and reduces the carbon emission of construction enterprises in the pilot cities by 4.4%.

  • There is heterogeneity in the effect of the SCP on the carbon emission levels of construction enterprises in the pilot cities of the YREB, i.e., it curbs the carbon emission levels of construction enterprises in the pilot cities in the upstream and downstream regions of the YREB and significantly promotes the carbon emission levels of those in the midstream region.

  • R&D and FDI are important transmission mechanisms for the SCP to promote the carbon emission levels of construction enterprises in the pilot cities of the YREB, i.e., the SCP curbs the FDI level of the pilot cities and increases the regional R&D level, thus leading to an increase in the carbon emission levels of construction enterprises in the region.

  • The results of the moderating effect analysis in terms of the introduction of new urbanization construction show that new urbanization construction has a positive moderating effect on the carbon emission reduction of the SCP on construction enterprises. Specifically, with the development of new urbanization construction, the level of suppression of the carbon emissions of construction enterprises in the YREB by the SCP increased from 4.4 to 9.6%.

Implications

The conclusions of this paper not only clarify the impact and mechanism of the SCP on the carbon emissions of construction enterprises in the YREB but also have certain significance for the government to improve the SCP and it is also important for construction enterprise managers to make strategic plans for carbon emission reduction.

  • When the government or enterprise managers make decisions to curb the carbon emissions of construction enterprises, they can learn from the management style of the SCP pilot cities to improve the level of carbon emission reduction in construction enterprises.

  • The results of the study show that there is a heterogeneous effect of the SCP on the carbon emissions of construction enterprises in the YREB. Therefore, when the government implements SCP, it should consider the differences between implementation regions and implement SCP according to local conditions.

  • FDI and R&D investment were found to have significant mediating effects in the process of SCP affecting the carbon emissions of construction enterprises in the YREB. Therefore, the government should increase the level of R&D investment and fully improve the structure of FDI when implementing the SCP to ensure an ameliorative effect on the carbon emission problem of construction enterprises.

  • A single policy may have a limited effect on solving carbon emissions. Therefore, the government should improve the corresponding supporting policies to form a policy mix to promote low-carbon development as well as green development.

Research limitations and future research directions

This paper, like previous studies, has certain limitations. First, in terms of data selection, this paper only selects a panel of provincial-level administrative districts in the main channel of the YREB from 2004 to 2021 due to the availability of data. Second, this paper only explores the impact of SCP on the carbon emissions of construction enterprises, and future researchers may consider establishing a comprehensive system that includes multiple policies for analysis. Finally, many factors influence the carbon emissions of construction enterprises, such as the government’s determination to govern and people’s willingness to do so. This paper has not yet examined those factors that are difficult to quantify, which provides an opportunity for researchers to further investigate the impact and mechanisms of these factors on the carbon emissions of construction enterprises in countries or regions worldwide using computer simulations and other tools.