Win-win of carbon and pollution co-mitigation in origin and destination regions: evidence from industrial enterprise relocation in China

INTRODUCTION

Environmental pollution and regional development inequality are common challenges faced by developing countries such as China, which are addressed by the United Nations’ Sustainable Development Goals (SDGs)^[1,2]. To tackle these challenges, China has implemented industrial relocation to optimize the spatial distribution of industries, establish a rational production division system, and promote regional coordinated development^[3]. Inland regions of China have actively embraced industries relocated from more developed coastal areas, seeking to create employment opportunities and foster economic development^[4-6]. In practice, such industrial relocation is often implemented in an organized manner through development zones and industrial transfer parks, where local governments provide land, infrastructure, and regulatory management to facilitate firm entry and industrial clustering, which may also shape environmental outcomes^[7]. A representative example is the “Replacing Old Capacity with New Capacity” policy implemented in the Beijing-Tianjin-Hebei region, which aims to utilize space freed up by relocating polluting industries to introduce new enterprises, promote industrial upgrading, and enhance regional efficiency^[8]. For instance, Shougang Group, a major state-owned steel company with a production capacity of about 8 million tons (equivalent to the UK’s national production in 2018), relocated from Beijing to Tangshan in 2010 to alleviate air pollution in Beijing^[9]. While this relocation presented destination regions with opportunities for economic development, it also posed challenges regarding environmental pollution^[8]. Given these dynamics, this paper aims to investigate the environmental impacts of industrial enterprise relocation (IER) on both the destination and origin regions. Specifically, does IER transform the destination regions into pollution havens?

A growing body of literature has examined the environmental impacts of industrial relocation, yet the findings remain highly contested^[10,11]. Much of the debate centers on the Pollution Haven Hypothesis, which suggests that pollution-intensive industries migrate to regions with laxer environmental regulations, thereby exacerbating local environmental degradation^[12], versus the Pollution Halo Hypothesis, which argues that relocated firms may introduce cleaner technologies and management practices, leading to environmental improvements^[10]. For origin cities, industrial transfer often reduces pollution sources, thereby lessening environmental pollution^[13,14]. In contrast, destination cities might experience increased environmental issues due to the incoming industries’ pollution^[13,15-18]. However, some relocated enterprises may leverage the opportunity to upgrade technologically and adopt greener practices, potentially leading to reduced pollution^[8,19-21] and beneficial technology and management spillovers^[12,22].

Recent studies further suggest that the environmental effects of industrial relocation critically depend on a range of influencing factors that shape firms’ relocation decisions and post-relocation behavior. On the one hand, environmental regulations and compliance costs can act as push factors, encouraging pollution-intensive firms to relocate and potentially generating pollution haven effects^[17,23]. On the other hand, relocation decisions are also influenced by broader economic and institutional factors, including differences in factor prices, industrial land policies, and place-based industrial transfer strategies adopted by local governments^[24]. Destination regions may attract relocating firms through industrial parks and policy incentives that facilitate agglomeration, while regional resource endowment and development stage condition firms’ technology choices and energy use after relocation^[7]. These considerations imply that industrial relocation is not merely a mechanical transfer of pollution, but a complex process in which firm behavior, regional characteristics, and policy environments jointly determine environmental outcomes. Consequently, relocation may worsen pollution when driven primarily by regulatory arbitrage, but may also generate co-benefits when accompanied by technological upgrading, process optimization, and knowledge spillovers^[23].

Despite these advances, important limitations remain in existing literature. Empirical research has predominantly adopted a macro industrial perspective, utilizing city and industrial-level economic data to construct indirect indicators for assessing industrial transfer, such as industrial location quotients, deviation-share measures, or multi-regional input-output tables^{[15,17,19,25]}. While these macro-level methods allow estimation of the relationship between inter-regional industrial scale changes and pollution intensity, they often fail to accurately capture the spatial dynamics of IER. Consequently, they may not effectively distinguish between the environmental impacts of relocating enterprises and those of local peers, potentially resulting in unreliable or biased findings^[26]. Furthermore, macro-level industrial transfers are driven by large-scale firm relocations, where enterprises directly influence both industrial transfers and environmental behaviors^[26]. Therefore, it is essential to evaluate the impacts of IER from a micro-level perspective to obtain accurate and actionable insights.

This study develops a theoretical and empirical model to identify the impact of IER on the environmental outcomes of both origin and destination cities using an industrial enterprises dataset in China. First, the analysis of the spatial and temporal characteristics of IER reveals distinct patterns. Southeast coastal and southwestern cities, along with provincial capital cities, are the main outflow cities of industrial enterprises, while the west, central, northeast, and non-provincial capital cities are the main inflow cities, which validates the industrial gradient transfer in China. Second, we assess the impacts of IER on the environmental outcomes of cities and find that IER can achieve a win-win situation for air quality improvements and carbon emissions reduction in both the destination and origin cities. Furthermore, we explore the mechanisms through which IER impacts environmental outcomes. We find that enterprise relocation promotes environmental benefits through technological innovation and changes in economic structure, without significantly affecting the regional economic scale. Moreover, our heterogeneity analysis indicates that the impacts of enterprise relocation are more pronounced in western regions and in areas with lower industrial development. Additionally, resource endowment weakens the environmental improvement in destination cities and enhances it in origin cities, validating the effectiveness of industrial gradient transfer in improving the environment. Lastly, while environmental regulations drive enterprises to move out, they are not the main factor attracting enterprise relocation.

This study contributes to the literature in the following aspects: First, we provide a nuanced understanding of IER from a micro perspective. Previous research predominantly examines industrial relocation from a macro perspective with industrial-level or region-level data, which cannot distinguish the effects of relocating enterprises from local enterprises entering or exiting the market, and also cannot precisely determine the direction of enterprise relocation. This can lead to inaccurate assessments of IER’s true environmental and economic impacts, resulting in potentially misguided policy recommendations. By utilizing a comprehensive database that reflects actual changes in enterprise locations, our study captures the trends and scale of IER in China more precisely, offering a detailed and reliable analysis of its real environmental impacts. Second, this study constructs theoretical and empirical models to explore the mechanisms through which IER impacts environmental outcomes. By focusing on innovation, structural, and scale effects, the research provides a nuanced understanding of how IER influences environmental performance and contributes to the application and development of the industrial gradient transfer theory. This analysis offers new insights into the complex dynamics of enterprise relocation. Third, our results expand the understanding of the Pollution Haven Hypothesis by showing that IER can lead to pollution reduction and carbon emission mitigation in both origin and destination regions, achieving win-win outcomes. This theoretical advancement provides a deeper understanding of how enterprise relocation can achieve sustainable development goals and offers valuable insights for both researchers and policymakers.

THEORETICAL HYPOTHESIS AND MODEL

According to the pollution haven hypothesis, the relocation of industrial enterprises could deteriorate the environmental performance in destination cities and improve the environmental performance in origin cities^[27].

For destination cities, with the influx of industrial enterprises that can lead to an increase in pollution sources, industrial production scale, and aggregation, thereby may heightening environmental pressures^[27,28]. After the relocation of enterprises with advanced production technologies and management experience, the destination cities may adopt these technologies and experience in the production control of local enterprises through “Learning by Doing”^[29]. Moreover, the introduction of new technologies can improve efficiency and reduce costs, prompting enterprises to invest more in innovation to gain profits and increase market share^[19]. If these technologies are used for pollution control, they will enhance environmental performance. However, if used to enhance production levels, they might contribute to increased energy consumption and a deteriorating environmental situation^[19,30].

In origin cities, the departure of industrial enterprises leads to a reduction in pollution sources and industrial scale, likely to improve local air quality. Additionally, enterprises that remain in their origin location will increase their investment in innovation to capture the market and increase their market share.

Hence, we propose the following hypothesis:

Hypothesis 1a. IER could deteriorate environmental situation in destination cities and mitigate air pollution in origin cities.

According to industrial gradient transfer theory, a country or region tends to transfer industries in which it has a relative disadvantage to countries or regions with a comparative advantage, aiming to achieve higher marginal returns on investment^[31].Relatively backward production capacities in developed regions could still be advanced and practical for less developed regions. Therefore, enterprises from developed regions bring new, environmentally friendly technologies and production methods, which can help to improve energy efficiency and environmental conditions^[8,32]. Moreover, enterprises from developed regions with better environmental management systems and higher energy efficiency have better environmental performance than local enterprises^[33]. Through technological spillovers, enterprises from developed regions can have an impact on the production pollution control capacity of the destination regions. Additionally, the influx of enterprises increases the demand for productive services such as logistics, finance, R&D, and information technology, improving the efficiency and quality of these services^[34]. As an intermediate input industry for industrial development, the productive service industry is knowledge-intensive, low-pollution, low-consumption, high-output, and high-employment. The concentration and development of productive service industry could reduce carbon emissions.

Furthermore, when companies with advanced technology and management skills relocate to regions with stringent environmental laws and standards, the pollution halo hypothesis comes into play. As destination regions have increasingly emphasized that “enterprise relocation should not equate to pollution relocation”, enterprises leverage the opportunity to eliminate outdated production capacities, upgrade equipment, implement green technologies, and optimize production processes during the relocation process^[8]. Therefore, IER could lead to increased cleaner production technologies and management experience in destination cities^[35].

Based on the above analysis, we propose the following hypothesis:

Hypothesis 1b. IER could improve environmental performance in origin and destination cities.

Theoretical analysis of IER is shown in Figure 1.

Figure 1. Mechanisms of environmental impacts of IER. IER: Industrial enterprise relocation.

To substantiate the aforementioned theoretical hypothesis and link them with empirical results, we propose a simplified analytical framework designed to examine how enterprises of two distinct types, situated in the origin and destination regions, adopt varying production technologies under environmental regulation, such as a pollution fee. In this model, n_o enterprises in the origin region are in competition with n_d enterprises in the destination region. We assume that each enterprise possesses a primary technology for product manufacturing, characterized by a fixed unit production cost c and a quantifiable emission intensity α.

Enterprises may undergo technological adaptation upon relocation. For instance, if an enterprise enhances its production capacity in the destination region, the associated production cost c may escalate. Conversely, the adoption of greener technologies could reduce emission intensity α. The government imposes a pollution fee e for each unit of emissions, reflecting the regulatory stringency in the origin region, which is presumed to be more rigorous than in the destination region. Consequently, the introduction of IER results in a higher e in the origin region and a lower e in the destination region. Therefore, the total cost for unit capacity for each enterprise can then be articulated as c + αe.

Objectives. Enterprises in both the origin and destination regions aim to maximize their operating profits, producing q units of products at a market-clearing price p = A - b · (n_oq₀ + n_dq_d), moderated by a sensitivity parameter b > 0 and a potential market demand A > α + ae to preclude trivial outcomes. The profit maximization equations for origin and destination enterprises are as the following equation (2-1) and (2-2):

Therefore, we can express the environmental situation and economic situation in two regions. First, in our model, the pollution generated by enterprises in the origin and destination regions is directly related to their production levels and numbers of enterprises. Specifically, the environmental situation in origin region (I₀) and destination region (I_d) can be defined as I₀ = α · n_oq₀ and I_d = α · n_dq_d. Second, the economic situation is the total profit of all local enterprises, which can be shown as ₀ = n_oπ₀ in origin region and _d = n_dπ_d in destination region, respectively.

Equilibrium analysis. Utilizing the Cournot competition framework, we derive the market equilibrium outcomes, as delineated in Lemma 1.

Lemma 1. The optimal production capacities for enterprises in both the origin and destination regions can be shown as:

Lemma 1 delineates the pivotal role of several critical parameters - namely, the number of enterprises, production costs, emission intensities, and pollution fees - in determining the optimal production capacity for enterprises in both the origin and destination regions. The analysis reveals that there exists an inverse correlation between the optimal production capacity and the number of enterprises within each region. This inverse relationship is primarily driven by the heightened competitive pressures that emerge as the number of enterprises increases. Such competition dilutes individual market shares, leading to an elevation in the market-clearing price and consequently, a reduction in the equilibrium production level.

Moreover, the production cost, emission intensity, and pollution fees exert a negative influence on the optimal production capacity, underscoring the critical role of environmental regulations and the economic costs of pollution in shaping enterprise production strategies. This has significant implications for enterprise pollution migration: higher production costs and pollution fees discourage high production levels in regions with strict environmental regulations, incentivizing firms to adopt cleaner technologies or reduce output. Consequently, firms may relocate to regions with less stringent regulations to maintain higher production capacities without incurring significant costs. To mitigate this, policy harmonization across regions is crucial, preventing regulatory disparities from driving pollution migration. Additionally, promoting investments in clean technologies can help firms comply with environmental regulations, reducing the need to relocate.

Next, we will outline the environmental situation and economic situation in origin and destination regions.

The subsequent Proposition 1 and Proposition 2 will delineate the state of the environment and the economy in the origin and destination regions, providing a comprehensive framework for evaluating the multifaceted repercussions of IER.

Proposition 1. The impact of IER on the environmental situation can be shown as:

(a) $$ \frac{\partial I_{o}}{\partial c}<0, \quad \frac{\partial I_{d}}{\partial c}<0 $$

(b) If $$ 0<\alpha \leq \frac{A-c}{2 e}, \frac{\partial I_{o}}{\partial \alpha}>0, \frac{\partial I_{d}}{\partial \alpha}>0 ; $$ otherwise, $$ \frac{\partial I_{o}}{\partial \alpha}<0, \frac{\partial I_{d}}{\partial \alpha}<0; $$

(c) $$ \frac{\partial I_{o}}{\partial e}<0, \quad \frac{\partial I_{d}}{\partial e}<0; $$

(d) $$ \frac{\partial I_{o}}{\partial n_{o}}<0, \quad \frac{\partial I_{d}}{\partial n_{d}}<0. $$

Proposition 1 elucidates the intricate dynamics between economic factors and environmental regulations, highlighting how the environmental situation in both regions is influenced by IER. The environmental intuition behind Proposition 1(a) suggests that an escalation in production costs precipitates a diminution in environmental impact across regions. This phenomenon arises as heightened production costs diminish the economic feasibility of sustaining elevated production levels, resulting in reduced emissions and improved environmental conditions. This indicates that as production costs increase, enterprises are less likely to produce at high levels, thus reducing pollution emissions in both the origin and destination regions. The following empirical findings will support this, showing that IER significantly reduces CO₂ emissions and PM_2.5 concentrations, achieving environmental improvement. Furthermore, the impacts of enterprise relocation are more pronounced in western regions and in areas with lower industrial development, where higher production costs lead to more significant environmental benefits.

The impact of emission intensity on the environmental situation in Proposition 1(b) is contingent upon its relative magnitude to the other economic parameters. If the emission intensity is at the lower level, an increase in α results in a higher environmental impact in both regions. This counterintuitive result can occur because lower emission intensity of the present technology might not be sufficient to incentivize enterprises to adopt cleaner technologies or significantly reduce production. Instead, enterprises may continue their operations at similar levels, leading to higher overall emissions. Conversely, if the emission intensity surpasses a critical threshold, it serves as a potent impetus for enterprises to diminish emissions, either through the adoption of cleaner production methodologies or by reducing operational scale, thereby enhancing the environmental improvement in both regions. This will be consistent with the following empirical findings, where IER promotes environmental benefits through technological innovation and changes in economic structure, validating the effectiveness of industrial gradient transfer in improving the environment.

Proposition 1(c) indicates that pollution fees serve as a direct economic lever for curbing emissions. Higher fees raise the marginal cost of production for polluting enterprises, incentivizing them to either reduce their pollution levels through cleaner technologies or to relocate to regions with lower fees. This mechanism effectively mitigates the overall environmental impact by aligning economic incentives with environmental objectives, achieving a win-win situation in both origin and destination regions. However, we need to clarify that although environmental regulations in the theoretical model encourage companies to relocate and improve the environment, whether they are the main factors attracting companies to relocate still needs to be further verified by empirical models.

Proposition 1(d) highlights that an increase in the number of enterprises in both the origin and destination regions leads to a reduction in the environmental impact. The underlying mechanism can be attributed to reduced aggregate production and emissions when the enterprise count diminishes, which aligns with the broader goals of IER to mitigate environmental degradation.

Proposition 2. The impact of IER on the economic situation can be shown as:

(a) If $$ 0<c \leq \frac{A-\alpha e}{2}, \frac{\partial \Pi_{o}}{\partial c}<0, \frac{\partial \Pi_{d}}{\partial c}<0 ; $$ otherwise, $$ \frac{\partial \Pi_{o}}{\partial c}>0, \frac{\partial \Pi_{d}}{\partial c}>0 ; $$

(b) If $$ 0<\alpha \leq \frac{A-c}{2 e}, \frac{\partial \Pi_{o}}{\partial \alpha}<0, \quad \frac{\partial \Pi_{d}}{\partial \alpha}<0 ; $$ otherwise, $$ \frac{\partial \Pi_{o}}{\partial \alpha}>0, \quad \frac{\partial \Pi_{d}}{\partial \alpha}>0 ; $$

(c) If $$ 0<e \leq \frac{A-c}{2 \alpha}, \frac{\partial \Pi_{o}}{\partial e}<0, \frac{\partial \Pi_{d}}{\partial e}<0 ; $$ otherwise, $$ \frac{\partial \Pi_{o}}{\partial e}>0, \frac{\partial \Pi_{d}}{\partial e}>0 ; $$

(d) The number of enterprises in both origin and destination regions cannot affect the economic situations.

Proposition 2 shows that the impact of IER on economic situation is different from environmental situations. Proposition 2(a) asserts that the impact of production costs c on the economic situation of two regions. When production costs are low, an increase in production costs negatively affects the economic situation in both regions. This is because lower production costs enable enterprises to operate more profitably, and any increase can significantly reduce their margins, leading to lower overall profitability. However, when production costs are high, enterprises are already operating under tight margins, and further increases can lead to cost efficiencies and technological innovations that may improve profitability.

Proposition 2(b) indicates that the impact of emission intensity on the economic situation is contingent upon its relative magnitude to other economic parameters. Low emission intensity fails to drive significant behavioral changes, raising operational costs without altering production processes. However, high emission intensity pressures enterprises to adopt cleaner technologies or relocate, leading to augmented profits from cost savings on emissions or pollution fees. Proposition 2(a) and (b) will be confirmed by empirical results that IER drives enterprises to adopt cleaner technologies, enhancing profitability, especially in regions with lower industrial development.

Proposition 2(c) suggests that the impact of pollution fees depends on their relative magnitude. Low pollution fees raise costs without incentivizing behavior change, while high pollution fees motivate cleaner technologies or relocation, reducing emissions, thereby enhancing the profitability of each enterprise. Empirical results will demonstrate that higher pollution fees drive economic improvements through innovation and structural changes.

Proposition 2(d) asserts that the quantity of enterprises in both regions does not independently influence the economic outcomes. This seemingly counterintuitive outcome may be rationalized by the assumption that market conditions or external factors may neutralize the impact of enterprise count, rendering the aggregate economic results in both regions agnostic to the number of individual enterprises. Empirical results will also show that IER promotes environmental benefits without significantly impacting economic stability, regardless of enterprise count.

DATA AND EMPIRICAL STRATEGY

RESULTS

Spatial and temporal characteristics of IER

The temporal characteristics of IER exhibit a fluctuating and irregular pattern, varying across different stages of China’s economic and social development [Supplementary Figure 1]. A significant increase in IER occurred in 2008-2009, followed by a decline to pre-financial crisis levels in 2010, and another increase in 2011. This pattern reflects the far-reaching impact of the 2008 global financial crisis on the world economy, including China. In response to the crisis, the Chinese government implemented a series of economic stimulus programs, including large-scale infrastructure construction and policies to expand domestic demand. While these measures stimulated the recovery of the domestic economy to some extent, they also intensified competitive pressures among enterprises, prompting some to seek more favorable production and operating environments. By 2010, China’s economy had begun to gradually recover and maintain steady growth. In a stable economic environment, companies may prefer to stabilize their existing production bases rather than relocate to reduce uncertainty and relocation costs. Furthermore, China started to establish the National Industrial Undertaking Demonstration Zone in 2010 [Supplementary Table 2], leading to an increase in enterprise relocations in 2011.

From a spatial perspective, IER is more frequent in the central, western, and northeastern cities [Figure 2]. Southeast coastal and provincial capital cities are the main outflow cities of industrial enterprises, while west, central, northeast, and non-provincial capital cities are the main inflow cities. The Chinese government has implemented strategies such as the Western Development Strategy, the revitalization of Northeast China, and the Strategy for the Rise of Central China to promote balanced regional economic development. These initiatives effectively promote the flow of capital, labor, and other production factors, facilitating enterprise relocation.

Figure 2. Spatial characteristics of IER. The base map is sourced from the Ministry of Natural Resources of the People’s Republic of China (http://bzdt.ch.mnr.gov.cn/), approved under No. GS(2019)1822. It is used solely for academic research purposes, and the figure was generated using ArcGIS. Figure 2A displays industrial enterprises moving in. Figure 2B shows industrial enterprises moving out. Figure 2C represents net relocation (industrial enterprises moving in minus industrial enterprises moving out). IER: Industrial enterprise relocation.

MECHANISM ANALYSIS

The baseline estimates suggest that IER contributes to a win-win situation in environmental improvement in both destination and origin cities. This section further explores the mechanisms through three channels: technological effects, structural effects, and scale effects.

FURTHER DISCUSSION: THE ROLE OF ENVIRONMENTAL REGULATION

Existing research has demonstrated the effectiveness of environmental policies in agglomeration economies and enterprise relocation^[48]. This section further explores the role of environmental regulation on IER. We use the frequency of environmental-related terms in government documents as a proxy for the government’s emphasis on environmental protection and the pollution fee per pollution equivalent as a proxy for enterprise pollution costs.

The results indicate that environmental regulation significantly influences enterprises moving out [Table 7]. Increases in pollution fees and the importance placed on environmental issues by the government drive enterprises to move out, validating the pollution haven hypothesis. Specifically, for every unit increase in pollution fees and governmental emphasis on environmental issues, enterprise relocation increases by 21.1% and 1.1%, respectively [Columns (3) and (4)]. However, environmental regulation does not significantly impact the attraction of enterprises to new locations [Columns (1) and (2)], suggesting that the relaxation of environmental regulations is not a primary factor in attracting enterprise relocation.

CONCLUSION

This study provides evidence that IER generates simultaneous reductions in carbon emissions and air pollution in both origin and destination cities in China, indicating a clear win-win environmental outcome. These findings challenge the conventional view that industrial relocation necessarily creates pollution havens in receiving regions. Instead, our results show that relocation can promote environmental improvement through technological innovation and industrial structure upgrading, without undermining local economic output. The environmental benefits are more pronounced in western regions and in cities with lower levels of industrialization, while resource endowment weakens the carbon-reduction effect in destination cities but strengthens it in origin cities, providing empirical support for the mechanism of industrial gradient transfer. In addition, stricter environmental regulations play an important role in driving enterprises to relocate from their original locations. These results suggest that the key policy challenge is not whether industrial relocation occurs, but how it is guided, governed, and complemented by innovation and environmental regulation to ensure sustainable regional development.

The results carry important policy implications for coordinated regional development. First, industrial relocation should be guided in an orderly manner under stringent environmental regulations to prevent the inflow of highly polluting activities into less developed regions. Policies that encourage green technological innovation and industrial upgrading can help ensure that environmental improvements do not come at the expense of economic performance. Second, policymakers should implement strategies that minimize negative impacts while optimizing resource allocation for coordinated economic development. Western and less industrialized cities should receive targeted institutional and infrastructural support, while more developed regions should maintain high entry standards and avoid transferring pollution-intensive production. In addition, achieving a sustainable “win-win” outcome may require the establishment of cross-regional coordination and compensation mechanisms. If western regions contribute to national carbon mitigation by accommodating industrial transfers, eastern regions, which benefit from improved environmental quality and industrial upgrading, could provide financial compensation, technology transfer, or capacity-building support to help destination regions strengthen environmental governance and maintain ecosystem service functions. Such mechanisms could enhance policy fairness, reduce long-term environmental risks, and promote more balanced and sustainable regional development.

Despite providing valuable insights, this study has several limitations. Identifying all relocating enterprises in China remains challenging, and data constraints may limit the representativeness of the sample. Future research could compile more comprehensive firm-level datasets to better capture relocation dynamics and enterprise characteristics. A full life cycle assessment would be required to comprehensively evaluate the embodied carbon and ecological footprints generated during the relocation process. In addition, while green technology innovation is used as a proxy for technological upgrading, future studies that link innovation outcomes to the actual adoption of pollution control equipment or cleaner production processes would allow a more direct assessment of improvements in eco-efficiency.