The technology diffusion effect resulting from enhanced innovation capabilities is the primary driver of carbon emission improvements in administrative border regions. The “Dual pilot” cities optimize resource allocation by increasing innovation investment, aggregating innovation knowledge and talent elements, thereby promoting the “reduction in quantity and improvement in quality” of carbon emissions in border regions (Ryzhenkov, 2016).

Firstly, in terms of innovation investment, one of the key tasks of the “Dual pilot” policy is to promote technological innovation output by reasonably guiding the flow of fiscal funds and ensuring the supply of innovation elements. The central government also uses indicators such as the “proportion of local fiscal science and technology appropriation in local fiscal expenditure” and the “number of invention patents granted per million people” in “Dual pilot” cities as acceptance criteria. This compels local governments to establish special funds for the construction of ICP, increase support for science and technology funds, expand the scope of science and technology funds, and provide financial guarantees for the implementation of major technological breakthroughs and green production transformation projects. The emergence of numerous innovation funds and technological innovation achievements provides border regions with opportunities to acquire new technologies and methods, thereby promoting low-carbon development in these regions. Secondly, in the process of implementing the “Dual pilot” policy, pilot cities actively formulate policies to cultivate and attract innovative talent, thus fostering a talent advantage. The aggregation of innovative talent not only enhances the overall knowledge stock of cities but also effectively reduces the cost of knowledge and information transfer from central to border regions through the enhancement of human capital. This encourages the high-frequency exchanges, aiding polluting enterprises in border regions in better mastering market information and carbon reduction technologies, thereby effectively curbing carbon emissions in these regions. Thirdly, the “Dual pilot” policy accelerates the pace of new infrastructure deployment within jurisdictions, breaking down spatial and temporal barriers through the flow of information elements and effectively reducing information asymmetry in green technology innovation activities in border regions (Ke et al., 2017). In this context, on the one hand, the improvement of the innovation capacity in border regions and the diffusion of technology from central regions provide key momentum for energy conservation and emission reduction in polluting enterprises. Continuous innovation through green processes and products effectively optimizes the energy consumption structure and carbon resource utilization efficiency in border regions, fundamentally restraining carbon emissions. On the other hand, the combined effect of the “Dual pilot” policy and the “invisible hand” of the market can accelerate the reallocation of various elements from low-efficiency sectors to high-efficiency sectors, guiding the reallocation of capital, technology, and talent to high-tech industries in border regions (Lin and Monga, 2011), and continuously eliminating outdated production capacity. This, in turn, empowers the “reduction in quantity and improvement in quality” of carbon emissions in border regions by enhancing the overall efficiency of element utilization. Given this, we propose Hypothesis H2a:

H2a: The “Dual Pilot” innovation-driven policy can effectively achieve a reduction in scale of carbon emissions and an improvement in carbon emission efficiency in administrative border regions through the technology diffusion effect.

The “Dual pilot” cities primarily foster industrial and technological system innovation by establishing high-tech development zones, science parks, innovation bases, and implementing various policies. Local governments introduce target industries and related sectors for the ICP, achieving increasing returns to scale internally and economies of scale externally through the agglomeration of high-tech industries. The industrial agglomeration driven by these innovation zones also triggers a “circular cumulative” effect, accelerating innovation activities, knowledge spillovers, and technology diffusion. This process enhances energy efficiency in polluting enterprises while benefiting administrative boundary areas through the emergence of skilled labor, market demand, and information. These favorable conditions not only effectively reduce carbon abatement costs in the boundary areas but also motivate polluting enterprises to undertake technological upgrades and reshape production processes (Song and Wang, 2016). In this context, on one hand, innovation-driven policies, through institutional reforms, accelerate the integration of new resources and technologies while effectively reallocating existing factors of production through the market mechanism. Particularly, by guiding the flow of new production factors to boundary areas and recombining them with existing ones, these policies effectively match green technological progress, fostering the emergence of high-tech and knowledge-intensive industries represented by fintech and consumer technology. This leads to low-carbon development in boundary areas through industrial upgrading and rationalization (Cheng et al., 2019). Furthermore, the linkage and spillover effects of high-tech industries can create a “catfish effect” in traditional high-pollution industries (Lin et al., 2011), where intense market competition compels traditional extensive industries in boundary areas to accelerate their low-carbon transition. On the other hand, innovation-driven policies can significantly optimize the capital input-output structure by guiding the orderly migration of capital away from low-value-added “three high” industries towards greener emerging industries. Traditional pollution-intensive industries in administrative boundary areas, under policy pressure, may be compelled to adopt clean production practices, implement pollution control technologies, or exit entirely, thereby driving the low-carbon transformation of industries in these areas. Some “three high” enterprises may actively relocate under the “pollution haven” effect, optimizing the industrial structure in boundary areas through the exit of polluting firms. Therefore, the “Dual pilot” policy helps to harness structural optimization effects, effectively reducing carbon emissions and enhancing the quality of carbon resource consumption in administrative boundary areas. This paper proposes Hypothesis H2b:

H2b: The “Dual pilot” innovation-driven policy can effectively achieve a reduction in carbon emissions and an improvement in carbon efficiency in administrative boundary areas through structural optimization effects.

The “Dual pilot” cities can leverage the effect of policy overlap to provide policy support for carbon emission reduction within administrative boundary areas. Institutional factors, especially in China’s transitional period, significantly influence urban development and should not be underestimated (Berman and Bui, 2001). Among these factors, the institutional environment and government functions serve as key drivers for low-carbon development, playing a crucial role in promoting the “decrease in quantity, increase in quality” of carbon emissions in boundary areas. By integrating the two pilot reform policies, the innovative city pilot program optimizes the urban innovation environment, accelerates industrial upgrading, and facilitates the spillover of urban innovation capabilities. Meanwhile, the NIDZ aims to provide flexible and reliable institutional convenience for enterprises engaging in technological innovation by transforming government functions. If an innovative pilot city also benefits from the NIDZ policy, the resulting reduction in institutional transaction costs and further refinement of market competition mechanisms can provide substantial policy support for enterprises in boundary areas to access more clean production factors. For instance, Shenzhen’s “Plan” emphasizes deepening the reform of administrative management systems, streamlining administrative levels, and establishing an efficient service-oriented government. The simplification of administrative approval processes not only enhances the efficiency of administrative approvals and environmental regulation by the government but also lowers the entry barriers and non-productive costs for enterprises, thereby encouraging high-tech enterprises to cluster in boundary areas. The enhanced empowerment of factors such as new technologies, new knowledge, and human capital further improves the potential for a “decrease in quantity and increase in quality” of carbon emissions in boundary areas, achieving a “multiplier effect”. Similarly, the inhibitory effect of the NIDZ on carbon emissions in boundary areas will also benefit from the ICP policy. Moreover, Dual pilot cities are typically accompanied by a series of supportive policies, such as tax incentives, R&D subsidies, credit facilitation, and land use preferences (Alder et al., 2016). The institutional system optimization effect resulting from policy overlap not only provides effective financial support for the implementation of clean production activities and green technology innovation in boundary areas, thereby incentivizing industries in these areas to achieve low-carbon transformation, but also helps reduce the uncertainties caused by a lack of information and knowledge, and enhances their resilience to such uncertainties. Additionally, the advancement of various laws, regulations, environmental policies, and other support tools conducive to green and low-carbon development in Dual pilot cities directly or indirectly creates favorable policy conditions for the low-carbon transformation of boundary areas, facilitating the “decrease in quantity and increase in quality” of carbon emissions in these areas. Accordingly, this paper proposes Hypothesis H2c:

H2c: The “Dual pilot” policy driven by innovation can effectively reduce the scale of carbon emission and improve carbon emission efficiency in administrative boundary areas through the effect of policy overlap.

The variable relationship diagram of this paper is shown in Fig. 1.

Fig. 1.

Relationship of variables.

To address potential endogeneity issues stemming from omitted variables, measurement errors, or reverse causality, we further employ an instrumental variable (IV) approach. Following the methodology of Dong et al. (2021a), who used historical data as instruments to mitigate endogeneity concerns, we select the status of a city as a treaty port between 1840 and 1930 as our instrumental variable. This choice is justified for two reasons. First, treaty ports were pivotal locations for the dissemination of Western modern culture, ideas, and ideologies in China. These areas also experienced rapid economic development following the reform and opening-up period. This historical context provides a strong foundation for the implementation of pilot policies and serves as a model for subsequent policy expansion, thereby satisfying the relevance condition of the instrumental variable. Second, the status of a city as a treaty port is a historical variable that is unlikely to directly influence carbon emissions within administrative boundary areas, thus meeting the exogeneity condition. The regression results are presented in Table 3 and indicate that our main findings remain robust even after accounting for potential endogeneity. Specifically, the “Dual pilot” policy has effectively promoted “qualitative improvement with quantitative reduction” in carbon emissions within administrative boundary areas. Additionally, we conducted tests for weak instruments, and the results confirm that weak instrument issues are not a concern.

During the period examined in this study, related environmental policies could potentially bias the estimated results. To address this concern, we systematically exclude the impact of policies that might affect carbon emissions within administrative boundary areas. These policies include: (i) the “Five Provinces and Eight Cities” low-carbon city pilot policy implemented by the National Development and Reform Commission in July 2010; (ii) the “Air Pollution Prevention and Control Action Plan” introduced by the State Council in 2013, targeting 57 key cities; (iii) the Environmental Protection Tax Law of the People’s Republic of China, which came into effect on January 1, 2018, with 12 provinces, including Hebei, raising their tax rates following the transition from fees to taxes; (iv) the sulfur dioxide emission trading pilot program initiated by the central government in 2007 and implemented in nine provinces, including Hebei and Shanxi, as well as in the cities of Tianjin and Chongqing. Considering the common sources of pollution, we conducted robustness checks by excluding the samples from these four policy pilots individually. The regression results, reported in Columns (1) and (2) of Table 4, show that after excluding the interference from other policies, the “Dual pilot” policy continues to suppress the scale of carbon emission and enhance carbon emission efficiency in boundary areas, with robust results.

Although the staggered DID method applied earlier provides useful insights, the non-random nature of the ICP and NIDZ means that they are not strict natural experiments, and thus there remains a potential for selection bias. To address this concern, we further employ a multiple-period Propensity Score Matching-Difference-in-Differences (PSM-DID) model for robustness checks. The regression results are presented in Columns (3) and (4) of Table 4. The findings reveal that the regression coefficient for CESC remains significantly negative, while the coefficient for CEEF in boundary towns remains significantly positive, consistent with the baseline regression results. This indicates that the effect of the “Dual pilot” policy in promoting “qualitative improvement with quantitative reduction” in carbon emissions within administrative boundary areas is robust.

To test the robustness of our findings, we conducted a placebo test by randomly reassigning the treatment and control groups. Specifically, we used computer-generated randomization to create new treatment groups and repeated this process 1000 times. We then applied a staggered Difference-in-Differences (DID) regression to each set of randomly assigned groups. The results are presented in Fig. 4. Finally, we compared the p-values obtained from the placebo tests with those from the actual treatment effect reported in Column (2) of Table 2. We found that the p-values from the placebo tests were not significant, whereas the p-values from the actual treatment effect remained significant.

Fig. 4.

Placebo test.

Administrative boundary areas often suffer from inadequate influx of new technologies and information due to information isolation and low levels of market integration. Under the influence of the “Dual pilot” policy, which includes the ICP and the NIDZ, the central cities have experienced a gradual concentration of innovation investments, cutting-edge knowledge, and innovative talent, as well as improvements in information infrastructure. These changes have led to significant technology spillover effects in administrative boundary areas, stimulating pollution-intensive firms in these regions to enhance their carbon reduction capabilities through process improvements and upgrades to pollution control equipment. Moreover, the technological advancements in carbon emission management and related equipment have also contributed to further reductions in carbon reduction costs in boundary areas, thus promoting “qualitative improvement with quantitative reduction” in carbon emissions. We test this mechanism in the paper, and the mediation test results are reported in Table 5, columns (1) and (4). The findings indicate that the “Dual pilot” policy significantly promotes technology diffusion. Specifically, increased technology diffusion further suppresses the scale of carbon emissions and enhances carbon emission efficiency in boundary areas. This demonstrates that the implementation of the “Dual pilot” policy effectively facilitates the spread of advanced technologies to administrative boundary areas, thereby leading to the replacement and integration of new production factors with traditional ones. Consequently, through the emergence of new business models and the transformation of high-carbon industries, the policy significantly impacts carbon emissions in these regions, confirming Hypothesis H2a.

The results of the mediation analysis for the structural optimization effect are reported in Columns (2) and (5) of Table 5. Following the implementation of the “Dual pilot” policy for ICP and NIDZ, we observe an increase in the share of the tertiary sector, a significant reduction in the scale of carbon emissions, and a notable improvement in carbon emission efficiency in administrative boundary areas. These findings suggest that the innovation-driven “Dual pilot” policy has, to some extent, hindered the development of traditional high-energy-consuming industries, thereby leading to a notable increase in the proportion of the tertiary sector—an important indicator of industrial structure upgrading. Additionally, the wider application of clean energy technologies has facilitated the substitution of fossil fuels, such as coal, with clean energy sources like wind and solar power. This substitution has effectively alleviated the path dependence on traditional energy sources in administrative boundary areas and driven pollution-intensive firms to accelerate their low-carbon transformation. Furthermore, the implementation of the “Dual pilot” policy has also pressured local governments to strengthen environmental regulation, thereby transforming administrative boundary areas from carbon emission havens into more regulated zones. The tightening of environmental constraints has led to the exit of some high-energy-consuming firms, thereby significantly reducing the scale of carbon emissions and improving carbon emission efficiency through the structural optimization effect. Therefore, hypothesis H2b is supported.

The results of the mediation analysis for the policy overlap effect are reported in Columns (3) and (6) of Table 5. Following the implementation of the “Dual pilot” policy for ICP and NIDZ, the intensive release of related supporting policies has created significant overlap effects. This has led to a substantial reduction in the scale of carbon emissions and a notable improvement in carbon emission efficiency in administrative boundary areas. These findings indicate that “Dual pilot” cities are more conducive to enhancing local governments’ strategic guidance on urban low-carbon development. On one hand, policy overlap effects can be effectively leveraged through streamlining administration and optimizing government service processes. On the other hand, financial incentives such as subsidies, tax breaks, and land use discounts provide critical funding support, achieving a “small investment with significant returns” effect and ensuring effective financial backing for low-carbon development in boundary areas. Moreover, environmental regulation policies and other supporting tools are more readily advanced in “Dual pilot” cities, directly or indirectly promoting “qualitative improvement with quantitative reduction” in carbon emissions in administrative boundary areas. Therefore, Hypothesis H2c is supported. (Results for the dependent variable, carbon emission efficiency, were retrieved and found to be consistent with the requirements.)

Although both the Innovative City Pilot Policies (ICP) and the National Innovation Demonstration Zones (NIDZ) are national pilot policies, the sequence of policy implementation differs among “Dual pilot” cities. To further investigate how the sequence of the “Dual pilot” policy affects the “quantity” and “quality” of carbon emissions in administrative boundary areas, we categorize the sample based on the policy implementation sequence. One group consists of cities that first established NIDZ before becoming “Dual pilot” cities, reflecting the dual impact of first becoming the NIDZ and then the ICP on boundary area carbon emissions. The other group includes cities that first became Innovative Cities before being designated as NIDZ, to examine the dual impact of this sequence on boundary area carbon emissions. The regression results are presented in Columns (1) and (2) of Table 6. These results indicate that becoming the NIDZ before being designated a “Dual pilot” city does not significantly affect the scale or efficiency of carbon emissions in administrative boundary areas. Columns (3) and (4) reflect the impact of the sequence where cities first pilot as ICP and then as NIDZ. This sequence significantly promotes the “quantitative reduction and qualitative improvement” of carbon emissions in boundary areas. This may be due to several factors. First, cities that first implemented the ICP policy had a pilot period of up to eight years, whereas cities that first implemented the NIDZ policy had a shorter period of 1 to 3 years. The shorter pilot period limited the accumulation of technology and information, thus making the short-term effects on “quantitative reduction and qualitative improvement” less evident. Second, compared to the NIDZ, the ICP focuses more on top-level strategic guidance and covers various aspects such as development models, management systems, and cultural institutions. Implementing the ICP policy first can provide a better institutional environment and innovation foundation for the NIDZ, thereby enhancing the policy’s effectiveness through a “one plus one is greater than two” effect and promoting low-carbon development in boundary areas.

Resource endowments are a key factor influencing urban development and may result in heterogeneous effects of the “Dual pilot” policy. To examine the heterogeneity arising from differences in urban classification due to initial resource endowments, this study utilizes the list of resource-oriented cities released by the State Council in 2013. A dummy variable is constructed, coded as 1 for resource-based cities and 0 for non-resource-based cities. The results are reported in Table 7. We find that the “Dual pilot” policy induces a “quantitative reduction and qualitative improvement” effect on carbon emissions in the boundary areas of non-resource-based cities. This may be attributed to the fact that resource-based cities primarily rely on natural resource endowments for development, which often results in a singular economic structure that is susceptible to the “resource curse” and severe path dependence. Moreover, resource-based industries tend to concentrate in administrative boundary areas, and local governments’ protectionist behaviors aimed at economic growth can impede the flow of factors and low-carbon development. Consequently, it is challenging for the current “Dual pilot” policy to exhibit its carbon reduction effects in the boundary areas of resource-based cities. In contrast, non-resource-based cities generally exhibit a more diversified industrial structure and a stronger innovation foundation. These characteristics facilitate the flow of innovation elements between regions. Additionally, local governments in these cities are also better positioned to quickly adapt to trends in high-quality economic development. Consequently, these factors enhance the “quantitative reduction and qualitative improvement” effect of the “Dual pilot” policy on carbon emissions in boundary areas.

Compared with inland cities, coastal cities generally exhibit higher levels of economic development, innovation infrastructure, and market orientation, making them more likely to actively promote low-carbon development in boundary areas under the “Dual pilot” policy. To investigate this heterogeneity, we classify cities based on whether they are coastal, assigning a value of 1 for coastal cities and 0 otherwise, and conduct the analysis accordingly. The regression results are reported in Table 8. We find that the “Dual pilot” policy has a significantly stronger effect on the “quantitative reduction and qualitative improvement” of carbon emissions in coastal cities than in inland cities. This finding may be attributed to the more advanced economic conditions, faster flow of innovation factors, and more comprehensive supporting policies in coastal regions. Alternatively, it could also reflect selective enforcement and relaxed regulations in less developed inland provinces aimed at achieving economic growth.