https://orcid.org/0000-0001-7519-5357
Figures
Abstract
Tourism development is increasingly linked to both regional economic activity and environmental outcomes. However, empirical evidence remains limited on whether and through which channels proximity to major tourist attractions is associated with the environmental governance of industrial enterprises. Using enterprise-level data and geocoded locations of China’s national 5A tourist attractions in 2014, we examine the relationship between firms’ spatial proximity to 5A attractions and corporate environmental governance. We find that (1) proximity to 5A attractions is positively related to the environmental governance of nearby firms, and the results are robust across alternative specifications and a range of robustness checks, including IV-based estimates; (2) proximity is associated with lower operating costs, consistent with shared transportation infrastructure and environmental facilities, and with higher fixed-asset intensity that is linked to stronger environmental governance; and (3) the positive relationship is more pronounced for natural-landscape attractions and less pronounced in cities with higher innovation or marketization. These findings provide evidence consistent with positive environmental spillovers associated with tourism and underscore the potential value of coordinating tourism development strategies with industrial environmental governance policies.
Citation: Xiong X, Zhong C (2026) Spatial proximity to national 5A tourist attractions and corporate environmental governance: Evidence from China. PLoS One 21(3): e0345770. https://doi.org/10.1371/journal.pone.0345770
Editor: Simone Luongo, Pegaso Telematic University: Universita Telematica Pegaso, ITALY
Received: March 15, 2025; Accepted: March 10, 2026; Published: March 26, 2026
Copyright: © 2026 Xiong, Zhong. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the article and its Supporting information files.
Funding: This study was supported by the College of Humanities and Social Science Research Project, Jiangxi Province, China (Project no. JJ24111), the Research Project of Education Department, Jiangxi Province, China (Project no. SYH24226), the Key Research Project of Jiangxi Science and Technology Normal University, China (Project no. 2023XJGXSZZD01), the PhD Research Start-up Fund of Jiangxi Science and Technology Normal University, China (Project no. 2023BSQD09). There was no additional external funding received for this study. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
1. Introduction
Currently, tourism is broadly seen as a key impetus for economic growth in many countries [1]. Tourism refers to travel outside one’s usual environment for business, leisure, or other personal purposes, typically for less than one year, without being employed by resident entities in the visited place [2]. A number of studies have demonstrated the critical role of tourism in capital accumulation [3], poverty alleviation [4], and social welfare improvement [5]. These direct effects of tourism typically involve changes in “sales [6], employment [7], tax revenue [8], and income levels [9]” and are widely observed across tourism-related industries. Beyond its economic role, tourism is also linked to environmental outcomes [10]. Prior research reports mixed evidence, documenting both environmental pressures (e.g., emissions, resource use, waste, and pollution) and potential environmental benefits under certain conditions [11–13].
Tourism is widely recognized to impose substantial environmental pressures through higher carbon emissions, natural resource overuse, and increased solid waste and water pollution [14–16]. At the same time, evidence on tourism’s potential environmental benefits remains mixed and context-dependent. Some studies suggest that tourism can raise public awareness and strengthen policy attention to natural resource protection, encouraging local governments to adopt environmental measures [17,18]. Others find that tourism development may stimulate local investment in clean energy, waste management, and environmental infrastructure [19–21]. Finally, a strand of research proposes a non-linear pattern consistent with a tourism-environment Kuznets-type relationship, in which environmental quality may initially deteriorate but later improve as awareness and technology advance [22].
A growing firm-level literature suggests that environmental regulation reshapes corporate behavior, both by increasing compliance costs and, under some policy designs, by encouraging cleaner technologies and green innovation [23–25]. Beyond within-firm responses, recent evidence points to spatial interdependence in regulation and environmental governance. Firms can respond on location margins, including entry, exit, and relocation. Moreover, policy effects may spill over to nearby jurisdictions through market linkages and strategic interactions [26–30]. In parallel, work on place-based policies shows that geographically targeted interventions can shape local firm entry, capital accumulation, productivity, and agglomeration, highlighting that location is an important determinant of firm-level outcomes [31–34].
However, these effects are heterogeneous. Program design and institutional features can determine whether zones raise productivity or instead generate distortions and redistribution within targeted areas. More broadly, a growing literature emphasizes that place-based interventions may involve equity-efficiency tradeoffs and displacement risks. This makes it important to understand the mechanisms through which localized advantages translate into firm-level outcomes. Taken together, this literature suggests that firms’ exposure to localized amenities and policies increases with geographic proximity. Yet relatively little is known about whether and through which channels tourism-oriented place-based development, such as proximity to national 5A tourist attractions, is associated with firm-level environmental governance.
Corporate environmental governance has increasingly been studied as a firm-level response to local institutions, infrastructure, and place-based conditions. Yet national 5A tourist attractions have rarely been examined as spatial factors that may shape firms’ environmental governance through local spillovers. Building on this gap, our study examines how firms’ spatial proximity to these attractions is associated with corporate environmental governance. We provide suggestive evidence on potential channels through which tourism-related local conditions may be linked to firms’ environmental governance.
To collect evidence, this study is focused on China, which is a developing country that actively fosters its tourism industry. In recent years, China’s tourism sector has demonstrated notable achievements and has garnered substantial attention from the central government [35]. For this study, we employ a cross-sectional survey dataset taken from the Chinese Industrial Enterprise Database and national 5A Tourist Attractions during the year of 2014 to construct our empirical model, and we do this for two reasons. First, by 2014, China had established a large number of national 5A attractions across the country. These attractions are the highest-rated tourist destinations in China and span natural landscapes, historical monuments, and cultural sites.
We use 2014 partly because it predates the tightening of China’s environmental enforcement in 2015. This helps keep the regulatory environment relatively stable in our cross-sectional analysis. In particular, the revised Environmental Protection Law took effect in 2015 and signaled a shift toward stricter enforcement aimed at reducing pollution and improving environmental quality. Second, the China Industrial Enterprises Database is a survey-based dataset covering a broad set of industrial firms nationwide, which provides a large cross-sectional sample. We additionally report IV-based estimates as a supplementary analysis to assess the sensitivity of our results to endogeneity concerns. Because we use a cross section, the exclusion restriction cannot be fully verified.
The contributions of this paper include the following: First, we bridge the tourism-environment and firm-level environmental governance literatures by providing firm-level evidence on how proximity to major tourist attractions relates to corporate environmental governance. Existing tourism studies typically focus on destination-, city-, or regional-level environmental indicators [36–38], while firm-level research rarely considers tourism-related local development as a spatial determinant of environmental governance. We link enterprise microdata with geocoded national 5A attractions to examine firms’ responses to attraction-induced local conditions.
Second, we show that the estimated relationship is robust across alternative specifications and a set of endogeneity-related checks, including IV-based analysis. Third, we offer suggestive evidence on mechanisms and heterogeneity: proximity to 5A attractions is associated with lower operating costs, potentially reflecting shared transport infrastructure and environmental facilities, and with higher fixed-asset intensity that is linked to stronger environmental governance. More broadly, our results suggest that national tourist attractions, together with the public infrastructure and environmental services that accompany them, may affect nearby firms’ environmental governance beyond China.
The remainder of this paper is organized as follows. Section 2 develops the hypotheses. Section 3 describes the data and empirical strategy. Section 4 presents the results. Section 5 explores mechanisms and additional robustness checks. Section 6 concludes.
2. Theoretical analysis and hypothesis development
2.1. Proximity to 5A tourist attractions, shared infrastructure, and firms’ operating costs
The theory of the geographical advantage of tourism, which is a core concept in tourism geography and regional economics, explains the spatial distribution of tourism resources, the formation of tourist destinations, and the locational choices that support tourism industry development [39]. The theory proposes that tourist attractions exert significant economic spillovers on their surrounding areas [40,41]. China classifies tourist attractions into five quality tiers, from A to AAAAA, under a centralized evaluation system. National 5A attractions represent the highest tier and are designated by the central government based on strict criteria. One key requirement is accessibility. In practice, 5A attractions are expected to be well connected by major transport modes such as air travel, high-speed rail, expressways, or high-standard waterways. The designation also imposes minimum thresholds for scale and visitation. Each attraction must cover at least three square kilometres and receive at least 600,000 visitors annually, including at least 50,000 overseas visitors. The attraction must have been upgraded from 4A after having been maintained at 4A for at least three consecutive years [42].
Therefore, 5A tourist attractions have a significant spillover effect on nearby industrial enterprises. The planning and construction of 5A-level tourist attractions are often accompanied by substantial local government investments in the regional infrastructure, including upgrades to critical environmental facilities such as transportation networks, water and electricity supply, and wastewater treatment systems. These infrastructure expansions not only support the tourism industry but also create favourable external conditions for nearby non-tourism businesses. For instance, industrial enterprises can comply with discharge standards through the use of centralized wastewater treatment systems, thereby avoiding the high costs associated with the construction and operation of individual facilities [43,44]. Such a shared infrastructure significantly reduces the production, transportation, and environmental governance costs for businesses, particularly for industrial enterprises. Consequently, the infrastructure improvements associated with tourist attractions serve as an important mechanism for lowering operational costs for nearby industrial firms. In summary, we hypothesize that proximity to 5A tourist attractions may be linked to lower operating costs for nearby industrial enterprises.
2.2. Proximity to 5A tourist attractions and firms’ environmental governance strategies
We hypothesize that proximity to 5A tourist attractions is associated with stronger environmental governance among nearby firms, and we explore two candidate mechanisms. First, the development and operation of such attractions not only highlight the overall appeal of the regional tourism industry but are also closely tied to local government performance evaluations and economic development objectives [45]. To safeguard the stability and sustainability of the surrounding ecological environment, local environmental authorities often impose stricter and more frequent regulations on nearby enterprises to mitigate the adverse effects of pollution on tourism brands. This differentiated regulatory approach creates a form of “selective enforcement,” making those industrial enterprises located in proximity to scenic areas more likely targets of environmental law enforcement [46]. Under the combined influence of governmental “window guidance” and the pressure of public opinion, firms tend to proactively improve their environmental governance as a means of avoiding regulatory risks and reputational damage, even in the absence of strong intrinsic motivation [47]. Consequently, we hypothesize that tourist attractions are associated with spillovers in environmental regulation, which may provide an external incentive for nearby non-tourism firms to improve their environmental performance.
Second, the construction and operation of 5A-level tourist attractions result in a continuous influx of visitors, which substantially increases the levels of public attention and media exposure in the region, thereby elevating societal expectations regarding the quality of the surrounding ecological environment. In this highly visible context, pollution problems are more likely to attract public scrutiny and complaints, particularly as monitoring by local residents, tourists, and environmental organizations is becoming more frequent. This heightened public oversight constitutes an “informal institutional constraint” that generates persistent reputational pressure and social accountability risks for nearby businesses [48]. To maintain legitimacy and avoid reputational damage, firms often proactively strengthen their environmental governance and adopt more standardized environmental practices to procure the “social licence” necessary for sustained operations [49]. Thus, tourist attractions can shape the local social environment, which may affect not only tourism-related industries but also the environmental governance of nearby industrial enterprises through social pressure.
However, environmental governance is inherently capital intensive [50–52]. Whether it involves the purchase of wastewater treatment equipment or the adoption of clean production processes, firms need to strengthen their environmental governance capacities through substantial fixed asset investment. An increase in the fixed asset ratio, particularly for environmental protection-related assets, therefore serves as a tangible indicator of improved corporate environmental governance. From the perspective of corporate behavioural drivers, Porter’s hypothesis and its extended theoretical framework suggest that, rather than undermining competitiveness, moderate environmental regulation can stimulate technological upgrades in production and governance through “incentive–innovation” mechanisms [53,54]. In such cases, firms may respond by increasing capital investment in environmental governance to meet compliance requirements or protect reputation, which may improve their environmental performance.
Building on this logic, we hypothesize that proximity to 5A tourist attractions is associated with lower operating costs for nearby industrial enterprises. Lower operating costs may relax constraints and be linked to higher fixed-asset investment, including investment in environmental equipment, which is in turn associated with stronger environmental governance. Based on this reasoning, we propose the following hypotheses.
Hypothesis 1: Proximity to 5A tourist attractions is positively related to corporate environmental governance among nearby industrial enterprises.
Hypothesis 2: Proximity to 5A tourist attractions is negatively related to operating costs and positively related to fixed-asset intensity.
Hypothesis 3: Fixed-asset intensity is positively related to corporate environmental governance, and it partially accounts for the relationship between proximity to 5A tourist attractions and corporate environmental governance.
2.3. Heterogeneity analysis
We suggest that the nature of 5A tourist attractions (e.g., natural vs. cultural landscapes) and the differing levels of technological innovation and business environments in their host cities may condition how proximity to 5A tourist attractions relates to firms’ environmental governance. Accordingly, we examine whether attraction characteristics and local institutional environments are associated with heterogeneity in the relationship between proximity to 5A tourist attractions and firms’ environmental governance.
2.3.1. Nature of 5A tourist attractions.
China’s 5A tourist attractions fall into two categories: natural and cultural landscapes. Natural landscapes form around natural resources, such as the Jade Dragon Snow Mountain in Yunnan Province, and are often located far from urban centres. Attractions located farther from central cities may be less exposed to urban innovation ecosystems and high-quality business environments. This may limit nearby firms’ access to frontier technologies and constrain their ability to upgrade environmental governance through technology-intensive approaches.
In contrast, cultural landscapes emerge from human activities and civilization, and they feature historical or artistic traits (e.g., the Forbidden City Museum and the Old Summer Palace). Typically situated in central cities related to specific historical events, these sites expose the surrounding industrial enterprises to advanced transportation systems as well as to higher levels of technological innovation and business environments. Consequently, higher levels of urban innovation or marketization may relax technological and institutional constraints faced by nearby firms. This may strengthen their ability to adopt cleaner production upgrades and improve environmental governance. We thus propose the following hypothesis:
Hypothesis 4: The positive relationship between proximity to 5A tourist attractions and firms’ environmental governance is more pronounced when the attraction is a natural landscape.
2.3.2. City innovation ability.
To further motivate Hypothesis 4, we consider how the host city’s innovation environment may shape firms’ environmental governance responses. When local innovation capacity is higher, firms may face fewer constraints in upgrading environmental governance through technology-intensive and cleaner production approaches, partly by adopting and adapting existing technologies rather than undertaking frontier innovation. Relative to original innovation, technology adoption and imitation can reduce adjustment costs and implementation risks in environmental management [55]. Such upgrades may deliver both environmental improvements and private returns by improving efficiency and reducing regulatory and reputational risks. Based on this reasoning, we propose the following supplementary hypothesis:
Hypothesis 4.1: The positive relationship between proximity to 5A tourist attractions and firms’ environmental governance is less pronounced in cities with higher levels of innovation.
2.3.3. City marketization.
We also collect city-level measures of marketization to further examine Hypothesis 4. In cities with higher marketization, credit allocation tends to rely more on market-based mechanisms, which may reduce firms’ incentives to engage in non-market activities to obtain financing [56,57]. A more market-oriented environment may also facilitate technology adoption and cleaner production upgrades by improving access to finance, inputs, and competitive pressure. As a result, the relationship between proximity to 5A tourist attractions and firms’ environmental governance may differ across cities with different marketization levels. Based on this reasoning, we propose the following hypothesis.
Hypothesis 4.2: The positive relationship between proximity to 5A tourist attractions and firms’ environmental governance is less pronounced in cities with higher marketization.
The theoretical analysis framework used in this study is presented in Fig 1.
3. Sample data and empirical identification
3.1. Data
3.1.1. Sample.
This study did not involve human participants or animal subjects. Ethical approval and informed consent were not required as the research used publicly available corporate data. To examine the relationship between firms’ spatial proximity to China’s national 5A tourist attractions and corporate environmental governance, we use the 2014 survey wave of the China Industrial Enterprises Database. This database covers the industrial enterprises in 31 provinces across mainland China, with each reporting sales of above 5 million yuan. The database includes various enterprise types, such as state-owned, privately owned, and foreign-invested firms. We also retrieve data on the total volume of wastewater treatment from the database and then match it to the sampled enterprises. Additionally, we clean the dataset by removing anomalies, including any enterprises with zero registered capital and those reporting an average annual workforce of zero.
3.2. Main variables
3.2.1. Shortest geographical distance from an industrial enterprise to a 5A tourist attraction.
To construct our key variable of interest, firms’ proximity to 5A tourist attractions, we proceed as follows: First, we gather the latitude and longitude of all 5A tourist attractions in China during the years of 2014 and 2023, along with the geographical coordinates of each industrial enterprise. Next, we use R software to calculate the geographical distance between each industrial enterprise and the nearest 5A tourist attraction. Finally, we measure a firm’s proximity to 5A tourist attractions using the shortest distance from the firm to the nearest 5A attraction (Dis2014, Dis2023). This approach follows the related literature, which suggests that the influence of tourist amenities on nearby firms and local conditions attenuates with distance [58]. Therefore, we treat firms located closer to 5A tourist attractions as having greater exposure to tourism-related local conditions. The geographic coordinates of China’s 5A attractions are obtained from the official website of the Ministry of Culture and Tourism.
3.2.2. Environmental governance of enterprises.
We measure enterprises’ environmental governance (CEG) through the industrial wastewater treatment rate, which is calculated as the total treated industrial wastewater divided by the firm’s total fixed assets. This indicator is especially important because wastewater, compared with other pollutants, is more easily detected by both environmental regulators and non-governmental organizations. Therefore, the wastewater treatment rate offers a more accurate reflection of the environmental governance of industrial enterprises (industrial enterprise wastewater treatment data are taken from the Chinese Ministry of Ecology and Environment, and they reflect information on environmental pollution and governance, industrial enterprise emissions and environmental management).
3.2.3. Control variables.
Drawing on the literature, we adopt several control variables [59,60]. We define Size as the natural log of the average annual number of employees, lnAge as the natural log of the enterprise’s years in operation and Debt as the natural log of total long-term liabilities. Nature is a dummy variable indicating the enterprise’s ownership type, and it takes a value of 1 if state-owned capital accounts for at least 50% and 0 otherwise. The enterprise’s ownership type is a critical influencing factor on its environmental governance. The literature indicates that in China, state-owned enterprises typically exhibit stronger environmental governance than private enterprises do, owing to more stringent environmental regulations. Subsidy is the enterprise’s subsidy income rate, which is calculated by dividing its annual subsidy income by its total paid-in capital. Lighting represents the nighttime light intensity in the enterprise’s city, which indirectly reflects human activity and serves as a proxy for the city’s economic development level [61]. Pline indicates whether the enterprise is located in a county that borders a provincial boundary, and it is based on the enterprise’s geographical coordinates. Our assumption is that those enterprises located closer to provincial boundaries typically operate under less developed economic conditions. Coast identifies whether an enterprise is situated in a coastal county (based on latitude and longitude), and it indicates an advantage in import-export trade (nighttime light data are obtained from the DMSP-OLS dataset provided by NOAA’s National Geophysical Data Center (https://ngdc.noaa.gov/eog/dmsp/downloadV4composites.html)).
Table 1 shows the input-output indicators. A correlation coefficient test was conducted on all the variables listed inTable 2, revealing significant correlations among most of them.
3.3. Empirical identifications
3.3.1. Benchmark model.
To examine how proximity to 5A tourist attractions is associated with enterprises’ environmental governance, we employ the following empirical model:
where
the coefficient of interest;
and
denote the firm and
, respectively;
is a vector of the control variables; and
is the error term. On the basis of our theoretical framework, we expect
<0, implying that a shorter distance (i.e., greater proximity) to the nearest 5A tourist attraction is associated with stronger environmental governance among nearby firms.
3.3.2. Potential mechanisms linking proximity to 5A tourist attractions and corporate environmental governance.
Considering the proposed mechanisms, we examine whether operating costs and fixed-asset intensity are potential channels linking firms’ proximity to 5A tourist attractions and corporate environmental governance.
In Equation (3), we explore a mediation-type relationship in which proximity to 5A tourist attractions is associated with operating costs and fixed-asset intensity, which may help account for the association with corporate environmental governance. The cost rate (Cost) is defined as business operating costs divided by total paid-in capital, and fixed-asset intensity (CFI) is defined as total fixed assets divided by total paid-in capital.
To test for heterogeneity, we define three dummy variables as follows: if the 5A tourist attraction is natural and 0 otherwise (see Section 5.4.1);
if the enterprise’s city has a high level of innovation and 0 otherwise (see Section 5.4.2); and
if the city has a high marketization level and 0 otherwise (see Section 5.4.3).
4. Empirical results and discussion
4.1. Benchmark regression results
We measure firms’ exposure to 5A tourist attractions using the distance to the nearest 5A attraction. This choice implicitly assigns the largest weight to the closest attraction. In practice, firms may also be exposed to multiple 5A attractions, which motivates robustness checks based on alternative exposure measures. To address this, we additionally control for the median distance to all other 5A attractions (excluding the nearest) in our baseline regressions (Dis2014_med, Dis2023_med). To mitigate the influence of outliers, we winsorize the main continuous variables at the 1st and 99th percentiles (p1-p99), while leaving dummy variables unchanged.
Table 3 presents our baseline regression results. The results presented in Columns (1) indicate that firms closer to a 5A tourist attraction exhibit stronger environmental governance. After controlling for Dis2014_med, column (2) shows that the coefficient of Dis2014 remain negative and statistically significant, reducing concerns that the estimate is driven by the spatial distribution of other attractions. Column (3) shows that, even using Conley-HAC standard errors with a 100-km cutoff, the coefficient of Dis2014 remains significantly negative (The 100-km bandwidth spans within- and near-county interactions, thereby mitigating the risk of overstated significance due to spatially correlated shocks). Column (4) shows that, under a stricter specification with both county (company’s county-level location) and industry (finer-grained segmentation of industrial sectors) fixed effects, the negative coefficient of Dis2014 remains robust and is slightly larger in magnitude. The results displayed in Columns (5)–(6) show that the results are robust to alternative specifications that replace the Dis2014 with Dis2023. Overall, the evidence supports Hypothesis 1.
To further assess the robustness of the results of baseline, we construct an alternative measure based on the 2010 geographic distribution of 5A attractions (Dis2010). Columns (1)–(2) of Table 4 report negative and statistically significant coefficients on Dis2010, suggesting that the negative association between distance to 5A attractions and firms’ environmental governance is not driven by the attraction layout in any single year. Second, we construct Asset_S (enterprise accumulated depreciation divided by total capital) as a placebo variable with no direct link to tourism demand. Columns (3)–(4) indicate that the coefficients of Dis2014 are small and statistically insignificant, alleviating concerns about spurious spatial-gradient correlation. Third, given substantial missingness in the 2014 CEG for industrial enterprises (18,411 non-missing observations; 93% of the sample missing), we first visualize that observed firms are geographically dispersed with no systematic clustering (Fig 2), and then impute missing CEG values as zero to obtain a conservative lower bound and estimate a logit selection model to assess robustness. Columns (5)–(6) yield estimates for Dis2014 that are similar to the baseline results, suggesting that the association between Dis2014 and CEG is not particularly sensitive to the sample restriction. Table 3 reports baseline coefficients on Dis2014 of approximately −0.020 to −0.046. Consistent with these estimates, a 10-km increase in distance is associated with a 0.20–0.46 unit decrease in CEG. After introducing finer fixed effects (county and industry), the estimated magnitude increases slightly, and the inference remains in line with the spatially robust results based on Conley standard errors. Overall, results using an alternative distance measure (Dis2010), a placebo variable (Asset_S), and the two-step checks yield a similar pattern, suggesting that the main findings are qualitatively unchanged across alternative measures and sample-processing choices.
The administrative boundary data are from the National Catalogue of Geographic Information of China (http://www.webmap.cn/), which is in the public domain).
4.1.1. Robustness testing.
Owing to data availability constraints, we were unable to obtain detailed tourism industry data for the areas surrounding each sampled industrial enterprise. Therefore, we selected the shortest geographical distance from industrial enterprises to the nearest 5A-level tourist attraction as the core explanatory variable. In practice, tourism activities are primarily concentrated in the tertiary sector, which creates a potential mismatch between the independent and dependent variables used in this study. To address this issue, we construct TR as the share of accommodation and food service workers in a city’s tertiary sector, measured by employment in accommodation and food services divided by total urban employment. We then use TR as an alternative measure of local tourism activity and re-estimate the baseline regression of CEG on this proxy as a robustness check.
We measure a firm’s environmental governance as wastewater treatment volume scaled by total fixed assets, reflecting treatment activity conditional on capital equipment. Nonetheless, data limitations imply that this measure may imperfectly reflect firms’ overall environmental governance. To address this concern, we consider two alternative measures as robustness checks. First, using the 2014 China Enterprise Patent Database, we match each sample firm to its total number of granted green patents and construct Patent, defined as the natural logarithm of this total. This measure captures a firm’s environmental governance from a technological-innovation perspective. Second, we construct CEG_S by dividing wastewater treatment volume by total industrial output, providing an output-adjusted measure of wastewater treatment intensity.
Columns (1)–(2) of Table 5 show that replacing the core independent variable with TR yields a positive and statistically significant association between accommodation and food service employment and firms’ environmental governance (CEG). These findings are consistent with the benchmark regression, suggesting that a larger accommodation and food services employment share (a proxy for local tourism activity) is positively associated with corporate environmental governance. Columns (3)–(6) further show that the coefficient on distance to the nearest 5A tourist attraction remains negative and statistically significant for both alternative outcomes: green patenting and wastewater treatment per unit of industrial output. The results are broadly consistent across specifications.
4.1.2. Further discussion on the benchmark regression results.
The benchmark regression yields a negative and statistically significant coefficient on Dis2014, implying that greater proximity to a 5A tourist attraction is associated with stronger environmental governance. However, this does not directly imply that “closer companies have stronger CEGs.” To strengthen the logical rigor of our conclusions and their empirical support, we conduct two supplementary analyses. First, we perform a distance-group regression by dividing enterprises into three groups on the basis of their proximity to the nearest 5A tourist attraction. As shown in columns (1)–(2) of Table 6, the coefficient on Dis2014 is negative and statistically significant for the short-distance group, whereas the corresponding coefficients for the medium- and long-distance groups are also negative but not statistically significant. These results are consistent with the view that proximity to 5A tourist attractions is associated with stronger environmental governance. Second, we visualize group means by classifying firms into three distance bins (5–15 km, 15–20 km, and >20 km) and plotting average CEG with standard error bars (Fig 3). The figure shows that firms in the closest bin have higher average CEG than those farther away. Although the differences are modest, they are statistically significant, and the overall pattern is consistent with the benchmark regression results.
To assess the robustness of the baseline model, we further plot distance-binned boxplots for the key control variables (Size, Debt, Subsidy, Nature, Pline) to evaluate covariate balance. Figs 4, 5, 6, 7, 8 show that, for each control variable, the distance-binned boxplot means vary only slightly with distance to the 5A tourist attraction. The two groups closest to the attraction are similar in levels, with only minor differences, indicating that sample-selection bias is unlikely in the baseline regression. Then, we regress CEG on the full set of controls and fixed effects, obtain the residuals, and plot a binscatter with a local polynomial fit using these residuals. Fig 9 shows a clear downward pattern in residualized CEG as distance increases, indicating that, after conditioning on covariates and fixed effects, CEG remains negatively associated with distance. Finally, we split firms into “near” and “far” groups based on whether their distance to the nearest 5A tourist attraction is below or above the sample median, and assess common support using a distance histogram. Fig 10 shows substantial overlap between the two distributions around the median, with only limited mass in the tails, suggesting reasonable common support and a similar distance range across the two groups.
4.1.3. Placebo tests.
Furthermore, we conduct placebo tests by repeating the regression 1,000 times using randomly generated (placebo) values of Dis2014 and Dis2023. The distribution of placebo coefficients is centered around zero and is rarely statistically significant, indicating that the baseline estimates are unlikely to be driven by chance (Figs 11 and 12). Overall, the placebo evidence is consistent with Hypothesis 1.
4.2. Instrumental variable (IV) estimation
4.2.1 IV robustness checks: baseline and channel results.
During the matching process, we found that many industrial enterprises reported zero or missing wastewater discharge data. This may reflect non-reporting by the surveying agency or production processes that generate little or no wastewater. Because such missingness may be non-random, our baseline estimates could be affected by sample selection and other unobserved confounders. To further assess whether our baseline association is robust to potential endogeneity, we implement an instrumental-variables (IV) estimation as a robustness check. We use terrain slope at the city level as an instrument. Terrain is an important determinant of tourism resource endowments. Cities with steeper or more rugged terrain are more likely to host prominent scenic resources and sites that are subsequently designated as 5A tourist attractions. Accordingly, city-level slope should be correlated with firms’ proximity to 5A tourist attractions.
Because our analysis relies on cross-sectional data, the IV strategy cannot deliver definitive causal identification. We therefore use city-level terrain slope in a 2SLS framework as a robustness check for potential endogeneity. The exclusion restriction may be violated if slope directly affects environmental governance through drainage conditions, wastewater dispersion, infrastructure and abatement costs, or industrial siting patterns. Accordingly, we interpret the IV estimates as suggestive rather than conclusive.
Table 7 reports the 2SLS estimates. Columns (1)–(2) show that city-level terrain slope is statistically associated with Dis2014 and Dis2023, which is consistent with the relevance condition. Columns (3)–(4) report an informal direct-effect check by regressing CEG on slope while conditioning on the distance measures. The slope coefficients are small and statistically indistinguishable from zero, suggesting that slope does not exhibit a residual association with CEG once distance is included. We emphasize, however, that this exercise is not a formal test of the exclusion restriction, and slope may still affect firms’ environmental governance through other channels discussed above. Columns (5)–(6) present the second-stage estimates. The coefficients on the proximity measures remain qualitatively similar to the baseline association, suggesting that firms located closer to 5A attractions tend to exhibit stronger environmental governance on average. The IV results are broadly in line with our baseline findings for Hypothesis 1.
We further examine a potential channel by estimating the 2SLS specification with the fixed-asset ratio (CFI) as the outcome. Our hypothesis is that firms located closer to 5A attractions may exhibit higher fixed-asset intensity, which is consistent with stronger environmental governance. Table 8 reports the results. Columns (5)–(6) present the second-stage 2SLS estimates. The coefficient on Dis2014 remains statistically significant and indicates that firms closer to 5A attractions tend to have higher fixed-asset ratios on average. Taken together, these results are consistent with the channel hypothesis that proximity to 5A attractions is associated with higher fixed-asset intensity.
4.2.2 Sensitivity checks and OLS–IV comparisons.
Table 9 reports additional sensitivity checks for the 2SLS specification. Columns (1)–(4) show that city-level slope remains a strong predictor of Dis2014 in the first stage. The slope coefficient is statistically significant, and the first-stage F-statistics range from 21 to 23, mitigating weak-instrument concerns. The estimated relationship is stable when we add a set of geographic controls (Inside and Touch are constructed from administrative-boundary adjacency, where Inside captures proximity to provincial borders, the coastline, or national borders, and Touch captures proximity to land borders. North is defined using a standard north-south division of China (e.g., the Qinling–Huaihe line). Industry fixed effects are defined at the [2-digit] industry level based on firms’ primary industry codes).Specifically, Inside is an indicator for counties adjacent to a provincial boundary, the coastline, or a national border; Touch indicates whether a county is adjacent to a land border; and North indicates whether a county is located in northern China. Column (4) further includes industry fixed effects. Taken together, these results suggest that the first-stage association is not mechanically driven by border attributes, north-south differences, or industry composition. Column (5) reports the second-stage estimate with these additional controls, and the coefficient on Dis2014 remains in the same direction as in the baseline models.
These checks do not validate the exclusion restriction, but they help assess whether the IV results are sensitive to observable geographic and industry-related channels through which slope could directly affect environmental governance. With these controls, the IV estimates remain broadly consistent with the baseline association, and we therefore interpret them as suggestive rather than definitive.
Table 10 compares OLS and 2SLS estimates using both the raw CEG measure and its standardized counterpart. Using the raw outcome, the 2SLS coefficient on Dis2014 is −5.292 (Table 7 and Table 10, Column (2)), whereas the corresponding OLS coefficient is −0.022 (Table 10, Column (1)). The first-stage F-statistic is 22.37, mitigating weak-instrument concerns. We interpret the 2SLS magnitude cautiously given the cross-sectional design and the possibility of alternative channels linking slope to CEG, as discussed above.
Because CEG is defined as wastewater treatment volume scaled by total fixed assets, its level can be sensitive to rescaling (mean = 0.181; SD = 0.749). Consistent with this, re-estimating the model using standardized CEG yields a statistically significant effect with the same sign (CEG_std: −7.062; Table 10, Column (4)). Finally, even under standard IV assumptions, 2SLS identifies a local relationship for compliers; with heterogeneous effects, its magnitude need not coincide with the OLS estimate [62].
5. Mechanism testing
5.1. Mediation analysis via operating costs and fixed-asset intensity
To explore potential channels through which proximity to 5A attractions is associated with firms’ environmental governance, we implement a causal mediation analysis following the framework in [63,64]. Using Equations (2)–(3), we decompose the total association into an indirect component operating through the mediator (Cost or CFI) and a direct component, and report the average causal mediation effect (ACME). In practice, we estimate the mediator and outcome models specified in Equations (2)–(3) and compute ACME based on simulated counterfactual predictions [65]. Given the cross-sectional design, these mediation results should be interpreted as suggestive evidence conditional on the identifying assumptions of the mediation framework, rather than definitive causal mechanisms.
5.2. Addressing measurement and confounding concerns
Due to data limitations, our operating-cost measure (Cost) aggregates multiple expense categories and may not cleanly capture transportation-related cost changes. We therefore complement the mechanism discussion with two proxies more directly linked to traffic accessibility: lnSco, defined as the logarithm of firms’ main-business product sales, and lnHwy, defined as the logarithm of the distance from each firm to its nearest primary highway toll station. We obtain the coordinates of primary highway toll stations from China’s 2014 National Toll Road Statistical Bulletin and compute the great-circle distance between each firm and its nearest primary toll station using R. The resulting distance is log-transformed to construct lnHwy. The Bulletin reports 1,665 mainline toll stations nationwide by end-2014, including 442 on primary highways.
These auxiliary regressions help assess whether the baseline patterns are consistent with a traffic-accessibility interpretation. We also note that lower operating costs do not necessarily capture firms’ governance incentives; instead, they may be related to governance capacity through capital expenditures. Accordingly, we examine the fixed asset-to-asset ratio (CFI) as an additional channel outcome. To reduce confounding from production expansion, we control for firms’ operating scale and investment intensity by including lnRoi (returns on investment), lnInt (inventories), and lnSrn (main-business revenue) in the regressions.
Columns (1)–(2) of Table 11 report auxiliary regressions using two proxies for traffic accessibility. Dis2014 is positively associated with lnHwy and lnSco, indicating that firms located closer to 5A attractions tend to be closer to primary highway toll stations and exhibit lower transport-related frictions as captured by these measures. Taken together, these patterns are consistent with a traffic-accessibility interpretation in which areas around 5A attractions are better connected.
Columns (3)–(5) examine fixed-asset intensity (CFI). The coefficient on Dis2014 remains statistically significant after controlling for proxies of production expansion (lnRoi, lnInt, and lnSrn). Given that Dis2014 is a distance measure, the negative coefficient implies that firms closer to 5A attractions tend to have higher fixed-asset intensity on average, and this relationship is not mechanically driven by differences in expansion-related firm characteristics.
In our framework, fixed-asset intensity is not meant to capture governance intentions directly. Instead, it proxies for firms’ capacity to undertake environmental governance through capital expenditures. Under stronger institutional pressure and reputational concerns, firms may upgrade or expand physical assets that facilitate compliance and abatement. Therefore, while fixed assets are a broad category, the observed association provides suggestive support for the capacity-based channel in areas near 5A attractions.
5.3. Results of mediation analysis
Panel A of Table 12 reports exploratory mediation and decomposition estimates. For Dis2014, the total effect on CFI is negative and statistically significant (−0.129, p < 0.01), indicating that firms located farther from 5A attractions tend to exhibit lower fixed-asset intensity on average. A similar pattern holds for Dis2023 (total effect = −0.124, p < 0.01).
The decomposition further shows a positive indirect component operating through Cost (ACME = 0.012 for Dis2014 and 0.021 for Dis2023, both p < 0.01), while the direct component remains negative (−0.142 and −0.145, respectively). This sign reversal implies an “inconsistent” (competitive) mediation pattern: the cost-related channel offsets part of the negative total association rather than reinforcing it. Consistent with this, the share mediated is negative (−9.0% for Dis2014 and −16.6% for Dis2023), suggesting that the Cost pathway partially counteracts the overall negative association between distance and fixed-asset intensity.
The sign reversal in the benchmark mediation results may reflect that Cost partly captures firm scale or capital intensity. Larger firms tend to incur higher operating costs but may also have stronger investment capacity, which can mechanically generate a positive indirect component. To reduce this scale-related confounding, we add paid-in capital as a control in the mediation models. Paid-in capital is measured as the logarithm of a firm’s annual paid-in capital (lnCap).
Panel B of Table 12 shows that, after controlling for lnCap, the estimated indirect component through Cost (ACME) becomes negative (−0.012, p < 0.05 for Dis2014; −0.002, p < 0.10 for Dis2023). This pattern is consistent with a cost-based channel in which greater distance is associated with higher operating costs and, in turn, lower fixed-asset intensity. The implied mediated share is sizeable (52.4% for Dis2014 and 18.9% for Dis2023). Interpreted with caution, the mediation results suggest that the Cost pathway may account for a non-trivial portion of the overall association highlighted in Hypothesis 2.
To explore whether fixed-asset intensity (CFI) is a potential channel linking proximity to 5A attractions and environmental governance (CEG), Panel C of Table 12 reports mediation estimates with CFI as the mediator and CEG as the outcome. For Dis2014, the estimated indirect component through CFI is negative and statistically significant (ACME = −0.003, p < 0.01), consistent with the idea that firms farther from 5A attractions tend to have lower fixed-asset intensity, which is in turn associated with weaker wastewater treatment capacity. The direct component is also negative (−0.020, p < 0.10), and the total effect is negative and statistically significant (−0.025, p < 0.05). The implied mediated share is about 11.5%. Overall, the pattern is consistent with CFI being one observable correlate of the relationship between proximity to 5A attractions and CEG, and it provides supportive but not definitive evidence for Hypothesis 3.
5.4. Heterogeneity analysis
5.4.1. Heterogeneity by attraction type.
We examine whether the relationship between proximity to 5A attractions and firms’ environmental governance varies with the type of attraction. We classify 5A attractions as natural or cultural and define a dummy variable, Lands, which equals one for natural attractions and zero otherwise, based on the descriptions provided on the official website of the Ministry of Culture and Tourism of China.
We estimate Equation (4) by interacting the distance measure with Lands. Consistent with our baseline IV strategy, we implement a 2SLS specification in which we instrument both Dis2014 and the interaction term Dis2014 × Lands using the terrain-slope instrument and its interaction with Lands (Slop and Slop×Lands). Columns (1)–(2) of Table 13 report the first-stage results, which indicate that the instruments are correlated with the corresponding endogenous regressors. Column (3) reports the second-stage estimates. To make the heterogeneity more transparent, Fig 13 plots the implied relative predicted CEG-distance profiles separately for natural and cultural attractions (normalized to zero at distance = 0 within each group). The coefficient on Dis2014 × Lands is statistically significant, suggesting that the association between distance to 5A attractions and CEG differs between natural and cultural attractions. In particular, the estimates imply that the relationship between proximity to 5A attractions and CEG is stronger in counties with natural 5A attractions than in counties with cultural attractions.
Given the cross-sectional design and the potential threats to the exclusion restriction, we interpret this heterogeneity pattern as suggestive. One possible interpretation is that natural attractions are more likely to be accompanied by environmental-amenity investments and regulatory attention, which may be associated with greater environmental compliance efforts and governance capacity among nearby firms. This evidence is broadly consistent with Hypothesis 4. We next examine auxiliary Hypotheses 4.1 and 4.2.
5.4.2. Heterogeneity by urban innovation.
We examine heterogeneity by urban innovation intensity. We define Innovation as an indicator equal to one if a city’s innovation expenditure to GDP ratio in 2014 (innovation expenditure divided by GDP) is at or above the sample median across all cities, and zero otherwise. We then estimate Equation (4) by interacting Dis2014 with Innovation.
Following the approach in Table 13, we implement a 2SLS specification in which both Dis2014 and Dis2014 × Innovation are instrumented using the terrain-slope instrument and its interaction with Innovation (Slop and Slop×Innovation). Table 14 reports the results. Column (3) shows that the coefficient on Dis2014 × Innovation is statistically significant, implying that the relationship between proximity to 5A attractions and CEG differs across high- and low-innovation cities. The estimates suggest that this relationship is weaker in more innovative cities. Fig 14 provides a complementary visualization of this pattern. Normalizing each group’s predicted CEG to zero at zero distance, the figure shows that residualized CEG declines with distance in both groups, but the decline is less steep in high-innovation cities, consistent with the positive interaction term. One possible interpretation is that firms in more innovative cities rely more on technology-based upgrading and process improvements. Overall, the evidence is consistent with Hypothesis 4.1.
5.4.3. Heterogeneity by marketization.
We examine heterogeneity by the degree of city marketization. We define Market as an indicator equal to one if a city’s 2014 marketization index is at or above the median across all cities in the sample, and zero otherwise. The marketization index is taken from the Regional Marketization Report (2020). We then estimate Equation (4) by interacting Dis2014 with Market.
Following Tables 13, 14, we estimate a 2SLS specification that instruments both Dis2014 and Dis2014 × Market using terrain slope and its interaction with Market. Table 15 reports a positive and statistically significant coefficient on Dis2014 × Market, implying that the distance-CEG association is weaker in high-marketization cities. Fig 15 visualizes this pattern by plotting relative predicted CEG against distance separately for high- and low-marketization cities, with each group normalized to zero at distance = 0. The high-marketization profile is flatter, consistent with the interaction estimate. One interpretation is that more market-oriented environments improve access to external finance and strengthen market-based incentives, facilitating technology upgrading and cleaner production. Prior studies likewise document that financing constraints can impede firms’ adoption of environmental technologies [66,67]. Overall, the evidence is consistent with Hypothesis 4.2.
6. Conclusion and limitations
6.1 Conclusion
This study examines the relationship between firms’ proximity to nationally designated 5A tourist attractions and corporate environmental governance (CEG) among industrial enterprises. Using 2014 cross-sectional data, we find a robust association in which firms located closer to 5A attractions tend to exhibit stronger CEG outcomes. The auxiliary evidence is consistent with a shared-infrastructure interpretation: proximity is associated with proxies related to traffic accessibility and operating conditions, and firms closer to attractions also tend to display higher fixed-asset intensity, which may reflect greater capacity for compliance-related investment.
Policy implications are therefore conditional. If proximity to major attractions is associated with stronger environmental governance through shared public goods, including transport connectivity and environmental service provision, then policymakers might consider coordinating tourism planning with industrial environmental management. Some aspects of the infrastructure-spillover channel may generalize to other contexts with major tourist destinations or place-based investments. Other aspects are setting-specific, including the 5A designation system and the administrative attention attached to it.
6.2 Limitations
Several limitations matter for interpretation. First, the cross-sectional design cannot rule out unobserved confounders, so the estimates should be interpreted as associations rather than definitive causal effects. Second, the IV strategy may violate the exclusion restriction because terrain slope can affect environmental outcomes through drainage, infrastructure costs, and industrial siting, among other channels; IV results are best viewed as a robustness exercise. Third, our CEG measure captures wastewater-treatment intensity and may not reflect broader environmental management. Fourth, missingness in wastewater variables and the composite nature of the cost proxy raise sample-selection and measurement concerns. Future work using panel data, richer environmental outcomes, and alternative settings would help assess robustness and external validity.
References
- 1. Zhang P, Yu H, Xu L, Guo W, Shen M. Synergistic relationship or not? Understanding the resilience and efficiency of the tourism economy: evidence from Hainan Province, China. Environ Dev Sustain. 2023;1–25. pmid:36624731
- 2.
United Nations. Recommendations on Statistics of International Migration. 1998;58(1).
- 3. Chaitanya AVK, Swain SK. Economic leakages in tourism: A comprehensive review of theoretical and empirical perspectives. Tour Econ. 2024;30(5):1306–23.
- 4. Deller S. Rural poverty, tourism and spatial heterogeneity. Annals of Tourism Research. 2010;37(1):180–205.
- 5. Lee S. Income inequality in tourism services-dependent counties. Curr Issues Tour. 2009;12(1):33–45.
- 6. Yiu CY. The impacts of shopping tourism on retail sales and rents: lessons from the covid-19 quasi-experiment of Hong Kong. J Risk Financial Manag. 2023;16(6):301.
- 7. Bolukoglu A, Gozukucuk T. Tourism development and women employment: A study on the European union countries. Tour Econ. 2024;30(2):442–56.
- 8. Mohan P, Strobl E. Tourism and tax revenue: evidence from stay-over tourists in the Eastern Caribbean. Curr Issues Tour. 2024;27(8):1185–207.
- 9. Fang J, Gozgor G, Paramati SR. The impact of tourism growth on income inequality: Evidence from developing and developed economies. Tour Econ. 2021;27(8):1669–91.
- 10. Buckley R. Tourism and environment. Annu Rev Environ Resour. 2011;36(1):397–416.
- 11. Zhao L, Li Q, Sheng YC. Environmental policies as engines for tourism eco-efficiency: quasi-experimental evidence. J Sustain Tour. 2025;33(9):1–21.
- 12. Ehigiamusoe KU. Tourism, growth and environment: Analysis of non-linear and moderating effects. J Sustain Tour. 2020;28(8):1174–92.
- 13. Koçak E, Ulucak R, Ulucak ZŞ. The impact of tourism developments on CO2 emissions: An advanced panel data estimation. Tour Manag Perspect. 2020;33:100611.
- 14. DeMaagd N, Fuleky PF, Burnett K. Tourism water use during the COVID-19 shutdown: A natural experiment in Hawai ‘i. Ann Tour Res. 2022;97:103475.
- 15. Radwan HRI, Jones E, Minoli D. Solid waste management in small hotels: a comparison of green and non-green small hotels in Wales. J Sustain Tour. 2012;20(4):533–50.
- 16. Gössling S. The consequences of tourism for sustainable water use on a tropical island: Zanzibar, Tanzania. J Environ Manage. 2001;61(2):179–91. pmid:11381774
- 17. Peters A, Fuchs M. A relational exploration of tourists’ environmental values and their perception of restrictions in protected nature. J Sustain Tour. 2025;33(1):28–45.
- 18. Xu S, Hu Y. How do residents respond to negative environmental impacts from tourism? The role of community participation in empowering residents’ environmentally responsible behavior. Int J Tour Res. 2021;23(6):1099–111.
- 19. Awan A, Hussain K, Zafar M, et al. Is tourism conducive to green growth? An investigation of the role of tourism-based economic expansion in the nexus of leisure tourism and business tourism. J Hosp Tour Insights. 2025;8(1):120–37.
- 20. Juvan E, Grün B, Dolnicar S. Waste production patterns in hotels and restaurants: An intra-sectoral segmentation approach. Ann Tour Res Empir Insights. 2023;4(1):100090.
- 21. Wei L, Ullah S. International tourism, digital infrastructure, and CO2 emissions: fresh evidence from panel quantile regression approach. Environ Sci Pollut Res Int. 2022;29(24):36273–80. pmid:35060044
- 22. Shah SAR, Balsalobre-Lorente D, Radulescu M. Revising the tourism-induced environment Kuznets curve hypothesis in top 8 Asian economies: the role of ICT and renewable energy consumption. J Hosp Tour Technol. 2025;16(1):1–32.
- 23. Li X, Du K, Ouyang X, Liu L. Does more stringent environmental regulation induce firms’ innovation? Evidence from the 11th Five-year plan in China. Energy Econ. 2022;112:106110.
- 24. Wang Y, Yang Y, Fu C, Fan Z, Zhou X. Environmental regulation, environmental responsibility, and green technology innovation: Empirical research from China. PLoS One. 2021;16(9):e0257670. pmid:34551024
- 25. He L, Wang M. Environmental regulation and green innovation of polluting firms in China. PLoS One. 2023;18(3):e0281303. pmid:36893093
- 26. Xie T, Yuan Y. Go with the wind: Spatial impacts of environmental regulations on economic activities in China. J Dev Econ. 2023;164:103139.
- 27. He Y, Zhang X, Xie Q. Environmental regulation and carbon emission efficiency: Evidence from pollution levy standards adjustment in China. PLoS One. 2024;19(2):e0296642. pmid:38300962
- 28. Wu J, Segerson K, Wang C. Is environmental regulation the answer to pollution problems in urbanizing economies?. J Environ Econ Manage. 2023;117:102754.
- 29. Wang J, Song Z, Zhang Y, Hussain RY. Can low-carbon pilot policies improve the efficiency of urban carbon emissions?--A quasi-natural experiment based on 282 prefecture-level cities across China. PLoS One. 2023;18(2):e0282109. pmid:36827253
- 30. Chen S, Graff-Zivin J, Wang H, Xiong J. Combating cross-border externalities: Evidence from China’s inter-provincial ecological compensation initiatives. J Public Econ. 2025;252:105495.
- 31. Görg H, Mulyukova A. Place-based policies and firm performance: Evidence from special economic zones in India. Eur Econ Rev. 2024;165:104752.
- 32. Tian W, Wang Z, Zhang Q. Land allocation and industrial agglomeration: Evidence from the 2007 reform in China. J Dev Econ. 2024;171:103351.
- 33. Tian X, Xu J. Do place-based policies promote local innovation and entrepreneurship? Rev Finance. 2022;26(3):595–635.
- 34. Liu J, Li N, Du X. Did place-based industrial policy promote regional economic growth? Evidence from China. PLoS One. 2023;18(4):e0283688.
- 35.
Nawaz A, Shakeel M. Tourism development and environmental sustainability in the European Union: moderating role of institutional governance. Green Low-Carbon Econ. 2025;3(1):96–104. https://doi.org/10.47852/bonviewGLCE32021749
- 36. Dobruszkes F, Mondou V, Ghedira A. Assessing the impacts of aviation liberalisation on tourism: Some methodological considerations derived from the Moroccan and Tunisian cases. J Transp Geogr. 2016;50:115–27.
- 37. Fu T, Chang D, Miao C. Fuel regulation in a developing country: An interventional perspective. Energy Econ. 2022;113:106031.
- 38. Lu Y. Transforming China’s tourism industry: the impact of industrial integration on quality, performance, and productivity. J Knowl Econ. 2024;15(4):18116–53.
- 39. Zhang N, Ren R, Zhang Q. Air pollution and tourism development: An interplay. Ann Tour Res. 2020;85:103032.
- 40. Fethi S, Senyucel E. The role of tourism development on CO2 emission reduction in an extended version of the environmental Kuznets curve: evidence from top 50 tourist destination countries. Environ Dev Sustain. 2021;23(2):1499–524.
- 41. Tong Y, Zhang R, He B. The Carbon Emission Reduction Effect of Tourism Economy and Its Formation Mechanism: An Empirical Study of China’s 92 Tourism-Dependent Cities. Int J Environ Res Public Health. 2022;19(3):1824. pmid:35162847
- 42. Liu H, Gao C, Tsai H. Spatial spillover and determinants of tourism efficiency: A low carbon emission perspective. Tour Econ. 2024;30(3):543–66.
- 43. Wang J. Evaluating the impact of high-speed rail on local tourism economy in China: A quasi-natural experiment. J Comput Methods Sci Eng. 2025;25(3):2459–70.
- 44. Li S, Lv Z. Do spatial spillovers matter? Estimating the impact of tourism development on CO2 emissions. Environ Sci Pollut Res Int. 2021. pmid:33630260
- 45. Lin YX, Lin BS, Chen MH. 5A Tourist attractions and China’s regional tourism growth. Asia Pac J Tour Res. 2020;25(5):524–40.
- 46. Hernández-Chover V, Bellver-Domingo Á, Hernández-Sancho F. Efficiency of wastewater treatment facilities: The influence of scale economies. J Environ Manage. 2018;228:77–84. pmid:30212677
- 47. Hu W, Guo Y, Tian J. Eco-efficiency of centralized wastewater treatment plants in industrial parks: A slack-based data envelopment analysis. Resour Conserv Recycl. 2019;141:176–86.
- 48. Liu A, Qin Y, Lin VS. Local leaders’ promotion pressure and tourism development: Evidence from China. J Travel Res. 2025;64(6):1464–79.
- 49. Van Rooij B, Lo CWH. Fragile convergence: Understanding variation in the enforcement of China’s industrial pollution law. Law Policy. 2010;32(1):14–37.
- 50. Tan X, Liu H. Corporate environmental performance under the pressure of government environmental information disclosure. Financ Res Lett. 2024;69:106152.
- 51. Tao Y, Wang D, Ye Y. The role of public environmental concern on corporate social responsibility: Evidence from search index of web users. Energy Econ. 2023;126:107041.
- 52. Zhang J, Zhang L, Zhang M. Media pressure, internal control, and corporate environmental information disclosure. Financ Res Lett. 2024;63:105369.
- 53. Farooq U, Wen J, Tabash MI. Environmental regulations and capital investment: Does green innovation allow to grow? Int Rev Econ Finance. 2024;89:878–93.
- 54. Dang TV, Wang Y, Wang Z. The role of financial constraints in firm investment under pollution abatement regulation. J Corp Finance. 2022;76:102252.
- 55. Xiao H. Environmental regulation and firm capital structure dynamics. Econ Anal Policy. 2022;76:770–87.
- 56. Qiu LD, Zhou M, Wei X. Regulation, innovation, and firm selection: The porter hypothesis under monopolistic competition. J Environ Econ Manage. 2018;92:638–58.
- 57. Van Leeuwen G, Mohnen P. Revisiting the Porter hypothesis: an empirical analysis of green innovation for the Netherlands. Econ Innov New Technol. 2017;26(1-2):63–77.
- 58. Zeng H, Huang Z, Zhou Q. Corporate environmental governance strategies under the dual supervision of the government and the public. Bus Soc. 2023;62(4):860–907.
- 59. Fu T, Jian Z. Corruption pays off: How environmental regulations promote corporate innovation in a developing country. Ecol Econ. 2021;183:106969.
- 60. Xiong X, Chen M, Chen C. Corporate governance matters: How do reduced state-owned shares lower firms’ market power in a developing country? Int Rev Econ Finance. 2024;93:211–24.
- 61. Chen M, Chen C. Financial constraints alleviation: why does state-owned share reduction in China promote firm performance? Financ Res Lett. 2023;55:103824.
- 62. Fu T, Yang Y, Zhang H. Risk culture as a blessing in tourism development: long-run effects of epidemic disasters. Ann Tour Res. 2023;103:103682.
- 63. Ding J, Lu Z, Yu CH. Environmental information disclosure and firms’ green innovation: Evidence from China. Int Rev Econ Finance. 2022;81:147–59.
- 64. Henderson JV, Storeygard A, Weil DN. Measuring economic growth from outer space. Am Econ Rev. 2012;102(2):994–1028. pmid:25067841
- 65. Abadie A. Semiparametric instrumental variable estimation of treatment response models. J Econom. 2003;113(2):231–63.
- 66. Imai K, Keele L, Yamamoto T. Identification, inference and sensitivity analysis for causal mediation effects. Statist Sci. 2010;25(1):51–71.
- 67. Hicks R, Tingley D. Causal mediation analysis. Stata J. 2011;11(4):605–19.