Manufacturing Agglomeration and Corporate Environmental, Social, and Governance Performance

Abstract

Implementing environmental, social and governance (ESG) disclosure is critical for manufacturers’ sustainable development and high-quality growth. Amid manufacturing agglomeration, firms’ spatial concentration reshapes value creation and risk exposure, affecting ESG performance. Using 2010–2023 data from Chinese A-share listed manufacturers, this study empirically examines agglomeration’s impact on corporate ESG performance, based on heterogeneous firm and stakeholder theories. Results show agglomeration significantly improves ESG performance, via enhanced productivity (internal) and greater compliance pressure (external). Further analysis finds ESG performance mitigates adverse selection in agglomeration ecosystems, while cluster peer effects strengthen long-term ESG engagement, aligning with stakeholders’ demands for transparency and accountability. This enriches manufacturing agglomeration-ESG literature, guiding policymakers and firms in integrating sustainability into clustered development.

1. Introduction

The global push for sustainable development has elevated ESG principles as a core framework for balancing economic efficiency, social responsibility, and environmental protection [1]. ESG information disclosure, as a critical tool for communicating corporate sustainability efforts to stakeholders (e.g., investors, regulators, communities), directly addresses the information asymmetry between firms and their stakeholders [2]. However, significant gaps remain in disclosure practices, with firms varying widely in their willingness and capacity to report ESG performance, particularly in manufacturing, a sector pivotal to economic growth but also facing intense scrutiny over resource use and emissions.

Within the extensive theoretical and empirical research on corporate ESG performance, studies on its drivers are relatively scarce and can be broadly categorized into internal corporate factors and external environmental factors. Research on internal drivers primarily focuses on financial characteristics such as profitability [3], leverage [4], solvency [5], and non-financial indicators like organizational governance [6], digital transformation [7,8], and “directors and officers” (D&O) liability insurance [9]. The impact of these firm-specific factors on ESG performance decisions and quality has been widely discussed. External drivers are mainly studied based on environmental regulation [10], local government debt [11], tax reform [12], and investor attention [13]. Notably, the existing literature seldom considers the progressive convergence of characteristics and operational conditions among firms within the same industry when discussing corporate ESG performance. Facing similar external environments, firms exhibit a degree of “penguin huddle” behavior in their ESG performance decisions, fostering an atmosphere of mutual imitation and learning.

Manufacturing agglomeration, characterized by the spatial concentration of firms in the same industry, creates a unique ecosystem where stakeholder interactions are intensified. Drawing on stakeholder theory, agglomerated firms face heightened pressure from diverse stakeholders: local governments demand compliance with environmental regulations, peers set benchmarks for social responsibility, and investors prioritize governance transparency. This “stakeholder density” in agglomerations may reshape firms’ ESG strategies, yet existing literature has rarely explored this link. Manufacturing agglomeration also offers a natural laboratory to study ESG behavior. On one hand, agglomeration facilitates knowledge spillovers and resource sharing, potentially boosting productivity and enabling firms to invest in ESG [14]. On the other hand, clustered firms face stronger peer comparisons and regulatory oversight, increasing the cost of non-compliance with ESG expectations [15]. This study thus asks: How does manufacturing agglomeration influence corporate ESG information disclosure? And through what mechanisms?

Manufacturing agglomeration provides an internal space concentrating numerous firms from the same industry, offering a novel perspective for advancing the construction of domestic ESG performance standards and strengthening ESG regulatory policies. This paper focuses on manufacturing agglomeration to explore its impact on corporate ESG performance. On one hand, international economic development practices show that manufacturing development and the allocation of high-skilled labor are crucial for achieving economic leapfrogging [16]. Particularly under sustainable development strategy, as the primary sector for emission reduction, the ESG status of traditional high-pollution, high-emission industries like equipment manufacturing, steel and cement, and paper and printing requires more attention. On the other hand, building a modern economic system necessitates strategies inspired by the new development philosophy. Mutual influence and integration across different industries represent a new direction for industrial development [17]. Studying regional agglomeration effects and enhancing the agglomeration level of key industries will provide new analytical insights for promoting regional sustainable development.

This research contributes to the literature in three ways: (1) it integrates stakeholder theory into the analysis of agglomeration and ESG, explaining how spatial concentration amplifies stakeholder demands; (2) it identifies productivity and compliance pressure as dual mechanisms, bridging internal resource capacity and external stakeholder pressure; (3) it explores the role of ESG performance in mitigating agglomeration-related adverse selection, enriching understanding of sustainability in clustered ecosystems.

2. Theoretical Model and Research Hypotheses

As crucial micro-agents driving economic transformation, corporate performance in environmental protection, social responsibility, and governance is closely linked to high-quality regional economic development. Existing ESG literature predominantly concentrates on ESG ratings, ESG investment, and ESG impact effects, with relatively fewer studies on the drivers of corporate ESG performance. Addressing this gap, this paper focuses on the internal motivations and external environment influencing corporate ESG performance, utilizing the Melitz heterogeneous firm analytical framework and adopting manufacturing agglomeration as the entry point for specific analysis [18].

2.1. Theoretical Model Setup

Assume an economy consists of numerous heterogeneous economic agents (firms). The aggregate utility function, depicted by a CES function, can be expressed as:

$$U={({\displaystyle \int q{(w)}^{\rho }dw})}^{\frac{1}{\rho }}$$

(1)

where w represents different firms, q(w) denotes the product produced by each firm. If p(w) represents the price function faced by the firm, the revenue of each firm can be expressed as:

$$r(w)=q(w)\cdot p(w)$$

(2)

Total economic revenue is the integral of individual firm revenues:

$$R={\displaystyle \int r(w)dw}$$

(3)

Considering the utility maximization problem under the constraint of total revenue, the Lagrangian function is constructed as follows:

$$L={\left({\displaystyle \int q{(w)}^{\rho }dw}\right)}^{\frac{1}{\rho }}+\lambda \left(R-{\displaystyle \int q(w)p(w)dw}\right)$$

(4)

Solving this yields the firm’s demand function q(p(w)):

$$q(p(w))=p{(w)}^{\frac{1}{\rho -1}}{\left({\displaystyle \int p{(w)}^{\frac{\rho }{\rho -1}}dw}\right)}^{-\frac{1}{\rho }}{\left({\displaystyle \int q{(w)}^{\rho }dw}\right)}^{\frac{1}{\rho }}$$

(5)

Substituting this into the firm’s revenue function gives:

$$r(p(w))=p(w)q(p(w))=p{(w)}^{\frac{\rho }{\rho -1}}{\left({\displaystyle \int p{(w)}^{\frac{\rho }{\rho -1}}dw}\right)}^{-\frac{1}{\rho }}{\left({\displaystyle \int q{(w)}^{\rho }dw}\right)}^{\frac{1}{\rho }}$$

(6)

Next, consider the firm’s cost problem. Assume the representative firm’s labor demand in the industry satisfies:

$$l=f+{\displaystyle \frac{q}{\phi }}$$

(7)

where f represents the firm’s fixed cost, $\phi$ denotes the firm’s productivity. Therefore, firms with higher productivity have lower labor demands. Assuming production requires only a single input factor—labor—the cost c of a single firm can be expressed as $c=\omega ·l$, where $\omega$ is the wage. Free labor mobility implies wages are not determined by any single industry; here, the wage is generalized to 1, simplifying the model further. Thus, $c=l$, marginal cost $MC=1/\phi$, and the profit of a single firm $\pi \left(w\right)$ can be expressed as $\pi \left(w\right)=r-c=r-l$. The firm’s marginal revenue MR is as follows:

$$MR=p\left(1+{\displaystyle \frac{q}{p}}{\displaystyle \frac{\partial p}{\partial q}}\right)=p\left(1-{\displaystyle \frac{1}{\eta }}\right)$$

(8)

where η is the price elasticity of demand.

Considering pricing equilibrium in a monopolistically competitive market, i.e., MC = MR:

$$p\left(1-{\displaystyle \frac{1}{\eta }}\right)={\displaystyle \frac{1}{\phi }}$$

(9)

Based on the properties of the CES utility function, the elasticity of substitution σ can be represented using the price elasticity of demand η, i.e., $\sigma ={\displaystyle \frac{1}{1-\rho }}\eta$. As the number of product varieties tends to infinity (ρ approaches 0), σ and η are approximately equal. At utility maximization, the price elasticity of demand η = 1, satisfying $\sigma ={\displaystyle \frac{1}{1-\rho }}$. At this point, price p(w) can be expressed as a function of firm productivity $\phi$: $p\left(\phi \right)={\displaystyle \frac{1}{\phi \rho }}$. The firm’s demand, revenue, and profit functions also satisfy the following relationships:

$${\displaystyle \frac{q({\phi }_1)}{q({\phi }_2)}}={\left({\displaystyle \frac{{\phi }_1}{{\phi }_2}}\right)}^{\sigma }$$

(10)

$${\displaystyle \frac{r({\phi }_1)}{r({\phi }_2)}}={\left({\displaystyle \frac{{\phi }_1}{{\phi }_2}}\right)}^{\sigma -1}$$

(11)

$$\pi (\phi )={\displaystyle \frac{r(\phi )}{\sigma }}-f$$

(12)

Assume there are N monopolistically competitive firms in the industry meeting the above conditions, each producing one substitutable good. Their productivity satisfies a random distribution $\mu \left(\phi \right)$, and defines the expected weighted average productivity for the entire industry as $\tilde{\phi }={\left({\displaystyle {\int }_0^{\infty }{\phi }^{\sigma -1}\mu (\phi )d\phi }\right)}^{\frac{1}{\sigma -1}}$.

The above process briefly describes the economic performance created by firms pursuing profit maximization during development. Countless rapidly growing firms not only effectively boost macroeconomic levels but also bring problems like market monopoly and environmental degradation. The government urges firms to proactively balance environmental, social, and governance interests alongside economic performance to create a better market environment for high-quality development and sustainable revitalization. Although ESG performance, as an integration of relevant non-financial information, offers a new approach for firms to contribute to sustainability, it must be noted that regulations do not currently mandate ESG performance in many regions of the world. Therefore, this paper assumes each firm undergoes a two-stage decision process during its operational cycle: In the first stage, the firm decides whether to enter the market; in the second stage, it decides whether to engage in ESG performance. Objectively, improving ESG performance levels can bring certain value gains to firms, reflected in aspects like access to external investment, rising stock prices, and stronger risk resilience. ESG performance helps enhance firm value, but only on a profitable basis can the level of ESG performance effectively improve the value relevance of earnings [19]. That is, only after a firm undergoes the first-stage decision and chooses to remain in the market might it consider the second-stage ESG performance decision in pursuit of enhanced firm value.

First, consider the first-stage firm decision. A rational firm chooses to enter a specific industry market and pay a fixed cost solely because it expects positive profits post-entry. The firm’s value function can be expressed as $\nu (\phi )=\max \left\{0,{\displaystyle \frac{1}{\delta }}\pi (\phi )\right\}$, where δ represents the probability of the firm facing an exogenous, irresistible shock leading to delisting. Combining the derived profit function, the free entry condition for firms is:

$$\pi ={\displaystyle \frac{\delta f}{1-G({\phi }^{*})}}$$

(13)

where $G\left(\phi \right)$ represents the productivity distribution function of all firms seeking market entry, ${\phi }^{*}$ is the critical productivity value at zero profit. If $g(\phi )$ denotes the probability density of the productivity distribution for all firms seeking entry, the relationship between $\mu \left(\phi \right)$ and $g(\phi )$ is $\mu (\phi )={\displaystyle \frac{g(\phi )}{1-G({\phi }^{*})}}$. Combining this with the expected weighted average productivity for the entire industry yields: $\tilde{\phi }={\left({\displaystyle {\int }_{{\phi }^{*}}^{\infty }{\phi }^{\sigma -1}\frac{g(\phi )}{1-G({\phi }^{*})}d\phi }\right)}^{\frac{1}{\sigma -1}}=\tilde{\phi }({\phi }^{*})$. By definition, $\pi ({\phi }^{*})={\displaystyle \frac{r({\phi }^{*})}{\sigma }}-f$, leading to the zero-profit condition (ZCP) at the firm production threshold:

$$\overline{\pi }={\displaystyle \frac{\prod }{N}}=\pi (\tilde{\phi })={\displaystyle \frac{r(\tilde{\phi })}{\sigma }}-f={\left({\displaystyle \frac{\tilde{\phi }}{{\phi }^{*}}}\right)}^{\sigma -1}\cdot f-f$$

(14)

The free entry market condition (FE):

$$\overline{\pi }={\displaystyle \frac{\delta f}{1-G({\phi }^{*})}}$$

(15)

Next, we focus on the number of firms N within an industry under the general equilibrium condition $\left(\overline{\pi },{\phi }^{*}\right)$. Assume total labor supply $L=L_{in}+L_{en}$, where $L_{in}$ represents labor hired by firms that successfully entered, and $L_{en}$ is labor hired by firms that paid the fixed cost but have not yet successfully entered. In other words, the fixed cost f can be measured in labor units, i.e., $N_{en}·f=\omega ·L_{en}=L_{en}$, where $N_{en}$ is the number of firms failing to enter successfully. Therefore, the sole cost for successfully entering firms is labor cost, also measurable in labor units: $R-\prod =\omega \cdot L_{in}=L_{in}$.

Thus, $L=L_{in}+L_{en}=R-\prod +N_{en}\cdot f=R-\prod +{\displaystyle \frac{\delta Nf}{1-G({\phi }^{*})}}=R-\prod +N\cdot \overline{\pi }=R$, and the number of firms N can be expressed as:

$$N={\displaystyle \frac{R}{\overline{r}}}={\displaystyle \frac{L}{r(\tilde{\phi })}}={\displaystyle \frac{L}{\sigma \left[\overline{\pi }+f\right]}}$$

(16)

Next, we examine the second-stage firm decision: whether to engage in ESG performance. To simplify the analysis, we treat a firm’s post-ESG performance as operating in a distinct market, with profits calculated separately. Such a firm will face a new price-demand function and productivity distribution. The derivation process is similar to the first stage; subscripts 1 and 2 denote the original and ESG-related components of the value function, respectively. Similarly, a rational firm chooses to maintain ESG performance only if subsequent operations yield positive profits. At this point, the firm’s revenue and profit functions can be expressed as:

$$r_1(\phi )={(\phi \rho )}^{{\displaystyle \frac{-\rho }{\rho -1}}}{\left({\displaystyle \int p{(w)}^{\frac{\rho }{\rho -1}}dw}\right)}^{-{\displaystyle \frac{1}{\rho }}}{\left({\displaystyle \int q{(w)}^{\rho }dw}\right)}^{{\displaystyle \frac{1}{\rho }}}$$

(17)

$$r_2(\phi )={\xi }^{{\displaystyle \frac{\rho }{\rho -1}}}{(\phi \rho )}^{{\displaystyle \frac{-\rho }{\rho -1}}}{\left({\displaystyle \int p{(w)}^{\frac{\rho }{\rho -1}}dw}\right)}^{-{\displaystyle \frac{1}{\rho }}}{\left({\displaystyle \int q{(w)}^{\rho }dw}\right)}^{{\displaystyle \frac{1}{\rho }}}={\xi }^{{\displaystyle \frac{\rho }{\rho -1}}}{r}_1(\phi )$$

(18)

$${\pi }_1(\phi )={\displaystyle \frac{r_1(\phi )}{\sigma }}-f_1$$

(19)

$${\pi }_2(\phi )={\displaystyle \frac{r_2(\phi )}{\sigma }}-f_2$$

(20)

where $\xi$ is a price transformation coefficient (To simplify the analysis, it is assumed that the price function faced by a firm following ESG disclosure is related to the original price function, with a direct conversion factor existing between them), satisfying the equation $p_2\left(\phi \right)=\xi p_1\left(\phi \right)$. Simple derivation shows that when $f_2·{\xi }^{\sigma -1}>f_1$, we have ${\pi }_2<0<{\pi }_1$. Here, the firm gains positive profit from its original production but incurs a loss from ESG performance, leading the firm to continue original production and forgo ESG performance. The critical productivity satisfies:

$${\phi }_2^{*}={\phi }_1^{*}\cdot \xi {\left({\displaystyle \frac{f_2}{f_1}}\right)}^{{\displaystyle \frac{1}{\sigma -1}}}$$

(21)

Assume the proportion of firms choosing ESG performance is $\theta$. The equilibrium solution $\left(\overline{\pi },{\phi }^{*}\right)$ for the second-stage ESG performance decision can be expressed as follows:

$$\overline{\pi }={\pi }_1(\widetilde{{\phi }_1})+\theta {\pi }_2(\widetilde{{\phi }_2})=\left[{\left({\displaystyle \frac{\widetilde{{\phi }_1}}{{\phi }_1^{*}}}\right)}^{\sigma -1}-1\right]\cdot f_1+\theta \left[{\left({\displaystyle \frac{\widetilde{{\phi }_2}}{{\phi }_2^{*}}}\right)}^{\sigma -1}-1\right]\cdot f_2$$

(22)

Under the new equilibrium, the number of firms N can be calculated using the following:

$$N={\displaystyle \frac{L}{\sigma \left[f_1+\theta f_2+{\pi }_1(\widetilde{{\phi }_1})+\theta {\pi }_2(\widetilde{{\phi }_2})\right]}}$$

(23)

2.2. Research Hypotheses

From the model’s equilibrium solutions, the solution involving second-stage ESG performance contains more unknown parameters. This signifies that ESG, as a new domain for firm development, opens a novel path for value creation but also entails relatively high market risk. Equations (15) and (21) indicate that firms with productivity below the critical value ${\phi }_1^{*}$ will exit the market; firms with productivity in the interval $({\phi }_1^{*},{\phi }_2^{*})$ will remain in the market but refrain from ESG performance; firms with productivity above the critical value ${\phi }_2^{*}$ will remain in the market and maintain ESG performance. Thus, firm productivity is the primary internal factor influencing the ESG performance decision. Observing Equations (16) and (23), it is evident that in an economy with N firms, higher total labor L leads to higher profits for the representative firm. In other words, a higher degree of labor agglomeration $L/N$ in an industry corresponds to higher profit levels $r\left(\tilde{\phi }\right)$ for its representative firms and a greater likelihood $\left(\tilde{\phi }{>\phi }^{*}\right)$ of engaging in ESG performance in the second stage. The Melitz model assumes only labor as a single input factor; thus, labor agglomeration represents the concentration of production factors, i.e., the level of agglomeration. In a more general economic reality, industries with higher agglomeration levels exhibit more frequent comparison and competition among firms. ESG performance decisions exhibit stimulating and radiating effects. Heterogeneous firms facing similar external environments imitate and learn from each other to some extent, imbuing corporate ESG performance decisions with group motivation.

It is noteworthy that ESG performance or ESG ratings integrate two interrelated but distinct dimensions: (1) a firm’s actual ESG performance and (2) the transparency of ESG information disclosure. While disclosure serves as a critical input for rating agencies to assess and quantify performance—without sufficient disclosure, even strong ESG practices may be underrepresented in ratings—ESG ratings are not equivalent to “ESG disclosure” alone. Ratings further incorporate third-party verification, industry benchmarking, and qualitative evaluations of practice materiality, which extend beyond the mere provision of information. This distinction is pivotal for our theoretical reasoning: when we refer to “ESG ratings” or “ESG performance” in the following analysis, we acknowledge the role of disclosure as a complementary component, but center our arguments on the holistic assessment of firms’ ESG outcomes (performance + disclosure quality) captured by ratings, rather than reducing the construct to disclosure alone. This alignment between terminology and measurement ensures that our theoretical mechanisms (e.g., agglomeration-driven ESG improvements) directly map to the empirical variable we test, avoiding conceptual ambiguity.

Based on the above, this paper proposes Hypothesis 1 concerning the impact of manufacturing agglomeration on corporate ESG performance:

Hypothesis 1.

Manufacturing agglomeration significantly improves corporate ESG performance and rating.

Regarding the mechanism through which manufacturing agglomeration influences ESG performance, this paper explains it from both internal motivations and external environmental perspectives. From an internal perspective, manufacturing agglomeration can enhance firm productivity [14], making ESG performance profitable. This “profit-seeking” motivation manifests not only in quantifiable economic performance indicators like improved investment efficiency [20], enhanced firm performance [21], alleviated financing constraints [22], and increased firm value [23], but also in “additional value-added” aspects such as institutional investor preference [24], media attention [25], and access to trade credit [26]. Higher industry agglomeration leads to higher average firm productivity within the industry, enabling more firms to engage in ESG performance and capture the “extra value” created by ESG. Based on this, we propose:

Hypothesis 2.

Manufacturing agglomeration promotes corporate ESG performance and rating by enhancing firm productivity, thereby reducing the opportunity cost of ESG practice and high-quality disclosure.

From the perspective of the firm’s external environment, manufacturing agglomeration facilitates information exchange among firms within the industry. When ESG performance creates value and the average level of ESG performance within the industry is high, firms that do not yet disclose ESG will pay greater attention to ESG-related initiatives to avoid lagging behind. They leverage the reduced communication costs brought by agglomeration to achieve ESG value gains at lower costs. Dynamic competition theory posits that interactions within a competitive environment prompt firms to react to competitors’ actions to prevent the establishment of competitive barriers and the loss of their own advantages. In highly agglomerated industries, the flow of information, personnel, and management practices among firms is more convenient. When the overall industry ESG performance level is high or trending upward, firms’ cognitive legitimacy and normative pressure increase simultaneously. Amidst rising public concern over environmental issues and sustainability, conducting reasonable and effective ESG performance has become a strong demand from the public and local regulators. Current external pressures stemming from government regulation, market environment, social opinion, and media supervision compel firms to contribute to environmental protection and social responsibility, forcing them to adopt ESG performance as a means of “harm avoidance.” Based on this, we propose:

Hypothesis 3.

Manufacturing agglomeration forces firms to engage in ESG performance under increasing environmental compliance costs by promoting inter-firm information exchange and enhancing cognitive legitimacy.

3. Research Design

3.1. Data and Sample

The core explanatory variable is manufacturing agglomeration, measured using the Location Quotient (LQ) indicator. Taking city-level manufacturing as an example, the calculated LQ reflects the relative level of that city’s overall manufacturing sector nationally. The formula for the location quotient is as follows:

$$L{Q}_i={\displaystyle \frac{m_i/n_i}{\mu }}$$

(24)

where ${LQ}_i$ denotes the manufacturing location quotient of region i, $m_i$ represents the manufacturing employment in region i, $n_i$ is the total employment in region i, and $\mu$ is the ratio of national manufacturing employment to total national employment. The relevant employment statistics are sourced from the China City Statistical Yearbook and the China Statistical Yearbook.

The dependent variable is corporate ESG performance, sourced from the Shanghai Huazheng Index ESG rating. For robustness checks, Bloomberg’s ESG rating was also used as a substitute variable. Values were averaged quarterly, and ratings were assigned values from 1 (lowest) to 9 (highest). Due to data gaps for some firms in 2009 following the 2008 subprime crisis, the sample period is 2010–2023. The initial sample comprised all manufacturing firms among listed enterprises. Following common practice in firm-level studies, the sample data underwent preliminary cleaning and organization: (1) exclude firms that underwent IPO during the study period; (2) exclude firms with PT, ST, or *ST status during the study period; (3) exclude firms with discontinuous information regarding industry or entity changes during the study period; (4) exclude firms with severely missing data; (5) apply 95% winsorization to the sample; (6) exclude samples with fewer than 4 quarters of ESG reports (survival period less than one year). The final sample consists of 1698 listed firms.

3.2. Model Specification

To better address omitted variable bias, this paper incorporates numerous control variables. The specific model is as follows:

$$ES{G}_{it}={\beta }_0+{\beta }_1\cdot Agglomeratio{n}_{it}+{\beta }_2\cdot X+{\lambda }_i+{\gamma }_t+{\epsilon }_{it}$$

(25)

where ${ESG}_{it}$ is corporate ESG performance, ${Agglomeration}_{it}$ is the manufacturing agglomeration level in region i in year t, X is the set of control variables (detailed below), ${\lambda }_i$ represents firm fixed effects, ${\gamma }_t$ represents year fixed effects, and ${\epsilon }_{it}$ is the random error term. Based on the model derivation, higher regional manufacturing agglomeration is expected to improve corporate ESG performance; the expected sign of ${\beta }_1$ is positive.

For control variables, existing ESG research and relevant theory suggest that factors influencing corporate ESG performance can be categorized into external and internal factors. Internal factors are mostly firm-level variables. Drawing on studies by Giese et al. and Maji and Lohia [27,28], this paper controls for variables at both financial and non-financial levels. External factors influencing corporate ESG primarily stem from national policies, industry attributes, and investor attention. Within this paper’s corporate ESG decision model, the key external controls are the probability of exogenous shocks causing delisting and the substitutability elasticity of the firm’s product in the market. The former can be viewed as the aggregate effect of external influences, i.e., the combined effect of national policies, industry attributes, and institutional investor attention. The latter depends mainly on industry attributes and competition. As this study targets manufacturing firms, industry attributes exhibit relatively smaller variation compared to studies covering all industries. Industry fixed effects can capture most of these differences; results using industry fixed effects are also reported in subsequent empirical tests.

After screening and organization, 11 control variables were included in the baseline regression: 3 at the city level (Environmental Regulation, Local Government Debt, Supply Chain Innovation) and 8 at the firm level (4 financial indicators and 4 firm characteristic indicators). Given the association between ESG and CSR (Corporate Social Responsibility) investing, CSR performance was added as an additional control variable in subsequent robustness checks. In addition, in the subsequent analysis, corporate profits will be used to test the adverse selection effect of manufacturing agglomeration. Corporate profits are measured by the net profits of listed companies. Data sources for controls: Environmental regulation data was measured using the frequency of environmental and pollution-related keywords in city government work reports, obtained via Python 3.13 text segmentation of reports. Local government debt data used local government debt balances, sourced from the China Statistical Yearbook and city statistical bureaus. Supply Chain Innovation: Inspired by Xu et al., who suggest supply chains may influence ESG performance [29], this paper controlled for supply chain factors. Based on the list of National Supply Chain Innovation and Application Demonstration Cities published by the Ministry of Commerce and seven other departments, cities were assigned a value of 1 in the year of approval and subsequent years, and 0 otherwise. Firm-level controls include the following: Total Market Value, Return on Assets (ROA), Asset Turnover Ratio (ATO), Profit Margin, Board Size, Ownership (State-owned or not), Firm Age, Executive Compensation. The specific indicators are shown in the following

Table 1. Variable Description.

Variable	Indicator Description	Data Source
ESG	Average ESG score at the end of the quarter of the enterprise	Huazheng ESG report and Bloomberg ESG report
Agglomeration	Location Quotient calculated by the number of manufacturing employees	China Urban Statistical Yearbook
Environmental regulation	Logarithm of the sum of word frequency of environmental keywords in government report	Python government report keyword extraction, keyword lexicon: environmental protection, pollution, energy consumption, emission reduction, sewage, ecological, green, low carbon, air, chemical oxygen demand, sulfur dioxide, carbon dioxide, PM10, PM2.5
Local government debt	Logarithm of local government debt balance	China Urban Statistical Yearbook
Supply Chain Innovation	Dummy variable indicating whether the city is a national-level demonstration city for supply chain innovation and application	Matching cities with the officially published list; 1 if yes, 0 otherwise
Total Market Value	Logarithm of the total market value of listed	CSMAR database
Return on Assets (ROA)	Net profit divided by average total assets	CSMAR database
Asset Turnover Ratio (ATO)	Operating revenue divided by average total assets	CSMAR database
Profit Margin	Net profit divided by operating revenue	CSMAR database
Board Size	Logarithm of the number of board members	CSMAR database
Ownership	Dummy variable indicating state-owned enterprises	CSMAR database; 1 for state-owned, 0 otherwise
Firm Age	Logarithm of (current year − establishment year + 1)	CSMAR database
Executive Compensation	Logarithm of the total annual salary of directors, supervisors, and senior management	CSMAR database
Firm Profit (FP)	Logarithm of the net profit of the enterprise	CSMAR database

.

4. Empirical Results

Based on the theoretical analysis of the corporate ESG decision model and corresponding data variables, this section empirically tests the impact of manufacturing agglomeration on corporate ESG performance. Robustness checks and potential endogeneity issues are further addressed by replacing indicator variables, controlling for special samples, and accounting for specific events.

4.1. Baseline Regression

As shown in

Table 2. Regression Results of Manufacturing Agglomeration and Corporate ESG Performance.

	(1) ESG	(2) ESG
Agglomeration	0.0907 *** (0.0137)	0.0876 *** (0.0276)
Agglomeration2		0.1134 (0.1058)
Total Market Value		0.1078 (0.0996)
ROA		0.2378 * (0.1248)
ATO		0.0799 ** (0.0388)
Profit Margin		0.0001 (0.0008)
Board Size		−0.0255 ** (0.0089)
Ownership		0.2756 ** (0.1039)
Firm Age		−0.0289 (0.0227)
Executive Compensation		−0.5611 * (0.3369)
Environmental Regulation		−0.1552 ** (0.0614)
Local Government Debt		−0.8131 (0.7789)
Supply Chain Innovation		0.2015 *** (0.0415)
_Cons	1.8669 *** (0.0347)	2.1423 *** (0.0718)
Year FE	Yes	Yes
Firm FE	Yes	Yes
N	9587	9587
Adj. R2	0.0165	0.0189

Notes: two-way clustering standard errors (firm × year) in parentheses; *** p < 0.01, ** p < 0.05, * p < 0.10.

below, to ensure robustness and reliability of the results, this paper employs a two-way fixed effects model (firm and year). Results with added controls are also reported. The results indicate that regional manufacturing agglomeration significantly affects corporate ESG performance at the 5% level. Specifically, higher overall regional manufacturing agglomeration leads to better ESG performance among manufacturing firms within the region, supporting Hypothesis 1. For manufacturing firms facing external environmental regulations and information asymmetry in capital markets, the atmosphere of mutual imitation fostered by manufacturing agglomeration leads firms within the region to exhibit similar behavioral motivations in ESG performance decisions. Specifically, regions with higher manufacturing agglomeration offer lower opportunity costs for ESG performance, and Sun et al. indicate that densely distributed firms typically face higher environmental compliance costs, compelling firms to devote more effort to ESG performance to avoid losing competitive advantages [21]. Against the backdrop of intensifying global climate change and the pursuit of high-quality economic development, external demands for corporate sustainability and green development are increasing, particularly in regions with manufacturing agglomerations. However, for firms, achieving low-carbon transformation and disclosing ESG information entail increased economic costs. Therefore, subsequent sections will discuss the challenges faced by firms within agglomerated spaces.

4.2. Robustness Checks

To further enhance the robustness of the conclusions, this paper sequentially replaced measurement indicators, added key control variables, controlled for different fixed effects and clustered standard errors, and used different subsamples for further testing. Results are shown in

Table 3. Robustness Checks Results.

	(1) Replace Dep. Var. ESG_PB	(2) Control Var. CSR ESG	(3) Different FE ESG	(4) Industry Clustered SE ESG	(5) Exclude Samples (HK Listed) ESG	(6) Exclude Samples (IPO < 3 Years) ESG
Agglomeration	0.0933 ** (0.0458)	0.0959 *** (0.0263)	0.0967 *** (0.0301)	0.0967 *** (0.0422)	0.0323 * (0.0165)	0.0163 *** (0.0323)
CSR		0.6980 ** (0.9265)
_Cons	1.8961 *** (0.6607)	1.7696 *** (0.1118)	1.8178 *** (0.2121)	1.7067 *** (0.3281)	1.2200 *** (0.9156)	1.8235 *** (0.0461)
CV	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes
Firm FE	Yes	Yes		Yes	Yes	Yes
Industry FE			Yes
City FE			Yes
N	9587	9587	9587	9587	8626	8055
Adj.R2	0.0189	0.0257	0.0213	0.0221	0.0152	0.0236

*** p < 0.01, ** p < 0.05, * p < 0.10.

. To avoid subjectivity in the choice of the core dependent variable measure and potential bias from measurement error, Bloomberg ESG data (ESG_PB) replaced Huazheng ESG data. Column (1) in Table 2 shows that the conclusion that manufacturing agglomeration significantly enhances manufacturing firms’ ESG performance remains valid. Column (2) reports results after adding a key control variable: Corporate Social Responsibility (CSR). Recognizing the overlap between ESG and CSR in scope and boundaries, CSR performance was added as a control. CSR data come from hexun and RKS CSR rating data, with a full score of 100 points. The higher the score, the better the performance of social responsibility; the conclusion holds. Columns (3) and (4) report results controlling for city-industry fixed effects and clustering standard errors at the industry level, respectively. Given that firm financial data can be volatile post-IPO, samples within 3 years of IPO were excluded based on the study window; results are in column (5). Considering that firms listed overseas (mainly Hong Kong) might alter ESG decisions due to differing investor philosophies, samples dually listed in Hong Kong were excluded; results are in column (6). These test results are largely consistent with the main effect, indicating robust findings supporting Hypothesis 1: Higher regional manufacturing agglomeration leads to better ESG performance among manufacturing firms.

4.3. Endogeneity Treatment

Considering the research object and model, potential sources of endogeneity include: First, sample selection bias. Since firms’ decisions to disclose ESG are conditional on their own characteristics, using only disclosing firms would lead to biased estimates. Heckman’s two-step method was employed: Stage 1 used a binary Probit regression with whether a listed firm discloses ESG as the dependent variable, including variables potentially influencing the decision; Stage 2 incorporated the Inverse Mills Ratio (IMR) from Stage 1 into the OLS model to correct sample bias. Second, omitted variable bias. Unobserved heterogeneity might affect ESG performance. Instrumental Variable (IV) estimation was used. Topographic relief and whether a railway passed through the area in 1933 were chosen as IVs for agglomeration. These IVs are highly random yet plausibly influence agglomeration formation, satisfying relevance and exogeneity requirements. To avoid inconsistency from weak instruments, Durbin-Wu-Hausman (DWH) tests were conducted; results indicated no weak instrument problem. To enhance the reliability of causal inference, we conducted a placebo test. We selected the firm registration location as the placebo outcome variable, measured using the city code of the firm’s headquarters. Since this variable is determined by historical factors, it should not, in theory, be directly influenced by the current level of regional manufacturing agglomeration. If the core results were driven by certain unobserved regional confounding factors, these same factors might also affect the firm registration location. However, the estimated effect of manufacturing agglomeration on the firm registration location was statistically insignificant (see Appendix A 

Table A1. Placebo Test for Instrument Variable.

	Placebo Test: Firm Registration Location
Agglomeration	−0.0070 (0.0322)
CV	YES
Year FE	YES
Firm FE	YES
Stock-Yogo critical values (10% maximal IV size)	16.38
Cragg-Donald Wald F	39.056
Kleibergen–Paap rk Wald F	0.513

for detailed results). This finding provides further support for our main conclusion that manufacturing agglomeration has a specific causal effect on corporate ESG performance. Third, simultaneity bias. Manufacturing agglomeration and ESG performance might co-evolve, and dynamic links over time could bias results. System GMM estimation was employed. In this model, L.ESG represents the one-period lagged term of the core explanatory variable, ESG. Consistent with the baseline regression, it is measured as the mean value at the end of the quarter of the following year. In the system GMM estimation, we carefully specified the instrumental variables to address endogeneity and ensure estimation efficiency. Specifically, we used lagged terms of all endogenous variables (including ESG, manufacturing agglomeration, and relevant control variables) as instruments for themselves. To ensure the exogeneity of the instruments, we restricted the lag depth, employing only the second to fourth lags (t-2 to t-4) as instruments and avoiding the use of the first lag (t-1) to prevent potential correlation. Furthermore, to mitigate bias from an excessive number of instruments, particularly in finite samples, we collapsed the instrument matrix. This approach effectively reduces the number of instruments and enhances the power of the Hansen test for overidentifying restrictions. Endogeneity treatment results are presented in

Table 4. Testing for Endogeneity.

	(1) Heckman Stage II (OLS)	(2) IV: Topographic Relief	(3) IV: Railway Operation	(4) System GMM
Agglomeration	0.0816 ** (0.0248)	0.0540 ** (0.1166)	0.1025 *** (0.4197)	0.0932 *** (0.0281)
L.ESG				0.9321 *** (0.0117)
IMR	1.3315 ** (0.2983)
CV	Yes	Yes	Yes
Year FE	Yes	Yes	Yes
Firm FE	Yes	Yes	Yes
Stock-Yogo critical values (10% maximal IV size)		16.38	16.38
Cragg-Donald Wald F		407.32	32.00
Kleibergen–Paap rk Wald F		1493.58	30.18
first-stage R2		0.0147	0.0150
	χ2 (2) = 287.213 (p = 0.00) F statics = 30.183 (p = 0.00)	DWH = 424.92 (p = 0.00)	DWH = 28.93 (p = 0.00)	AR (1) = 0.007 AR (2) = 0.258 Sargen test (prob > chi2) = 0.172 Hansen test (prob > chi2) = 0.224 Number of instruments = 34

*** p < 0.01, ** p < 0.05.

and also in Supplementary Materials.

5. Further Analysis: Adverse Selection and Peer Effects

In the theoretical model, the value function $\nu (\phi )=\max \left\{0,{\displaystyle \frac{1}{\delta }}\pi (\phi )\right\}$ is the decisive variable for ESG decisions. Rational firms will only maintain ESG performance as a strategic decision if it is profitable, yielding positive profits. At this point, the factors influencing the value function are solely the internal heterogeneity source—productivity—and the external environmental factor forcing delisting—compliance pressure. Therefore, the mechanism discussion focuses on firm productivity differences and external compliance pressure within the process of manufacturing agglomeration, affecting ESG performance.

5.1. Internal Factor: Firm Productivity

To investigate the role of firm productivity in this relationship, a mediation effect model was employed.

Table 5. Mediation Effect Test of Production Efficiency.

	(1) TFP	(2) ESG
Agglomeration	0.0698 *** (0.0077)	0.0864 *** (0.0272)
TFP		0.0794 * (0.0453)
CV	Yes	Yes
Year FE	Yes	Yes
Firm FE	Yes	Yes
N	8978	8978
Wald	189.11 ***	521.78 ***

Notes: Sobel test confirmed the mediation effect (Z = −4.011, p < 0.01). *** p < 0.01, * p < 0.10.

shows the relationship between manufacturing agglomeration and firm productivity (TFP measured by OP method). The OP method is as follows:

$$\ln Y={\beta }_0+{\beta }_K\ln K+{\beta }_L\ln L+{\beta }_M\ln M+{\beta }_{A}Age+{\beta }_{S}SOE+\gamma +\delta +\phi +\epsilon$$

where Y denotes sales revenue, L denotes labor input measured by the number of employees, K denotes capital input measured by the book value of fixed assets, and M denotes intermediate input. Following the distribution method, M is calculated as sales revenue minus value-added. Value-added is the sum of four components: depreciation, compensation of employees, net taxes on production, and operating surplus. These are, respectively, measured using the following financial statement items from listed companies: depreciation of fixed assets, cash paid to and on behalf of employees, taxes and surcharges, and operating profit. Age represents firm age, and SOE is a dummy variable for ownership type. Firm-level data are primarily sourced from the CSMAR database, supplemented with annual reports where necessary. $\gamma ,\delta ,\phi$ represent time, region, and industry fixed effects, respectively, and $\epsilon$ is the random error term. To accurately reflect the contribution of input factors to economic growth, all nominal variables are deflated using the corresponding price indices (CPI) and converted into real values based on the year 2000, and ESG performance. Column (1) shows that agglomeration enhances firm TFP; agglomeration spillovers allow firms within clusters to share knowledge, technology, talent, etc., improving productivity. Column (2) further shows that higher productivity improves ESG performance. Productivity gains enhance internal governance and social responsibility fulfillment, and the capacity for high-quality ESG information disclosure, thereby boosting ESG performance.

It is noteworthy that, assuming consistent external environments, the primary factor firms consider in ESG performance decisions is their own productivity. When numerous firms with varying productivity coexist within an agglomeration, the “benefits” derived from agglomeration also differ. For instance, laggard firms often enjoy more spillovers within clusters, while industry leaders gain relatively less, i.e., an externality imbalance issue. Even assuming all firms receive average external benefits, leading firms contribute more to cluster externalities (e.g., technology, human capital, training, supply chains). Thus, the average benefit remains low relative to their contribution, making laggard firms the primary beneficiaries. This partly explains why large firms prefer relocating over joining existing clusters, while clusters aggregate relatively average SMEs—the adverse selection problem in agglomeration. We discuss this issue below.

Based on signaling theory in microeconomics, ESG performance itself can be viewed as a signal transmission, helping maintain dynamic balance in externalities and mitigating adverse selection to some extent. Firms with better ESG performance attract more investor attention and build a better public image by assuming more social responsibility. This positive signal release and transmission theoretically allows firms to gain implicit “added value,” compensating for losses due to externality imbalance and sustaining the “Agglomeration → Productivity → ESG Performance” mechanism chain. To further explore this potential adverse selection stemming from productivity differences, quantile regressions at the 25%, 50%, and 75% levels were conducted separately for manufacturing firm samples that did and did not disclose ESG, estimating firm profits (RE, measured by total profit) at different productivity levels. As shown in

Table 6. Quantile Test Results on Adverse Selection in Manufacturing Agglomeration.

	(1) 25th Quantile Discloser FP	(2) 25th Quantile Non-Discloser FP	(3) 50th Quantile Discloser FP	(4) 50th Quantile Non-Discloser FP	(5) 75th Quantile Discloser FP	(6) 75th Quantile Non-Discloser FP
Agglomeration	0.0319 ** (0.0187)	0.0411 *** (0.0105)	0.0451 *** (0.0137)	0.0398 ** (0.0133)	0.0288 ** (0.0135)	0.0363 *** (0.0091)
CV	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes
Firm FE	Yes	Yes	Yes	Yes	Yes	Yes
N	9587	6712	9587	6712	9587	6712
Adj.R2	0.0219	0.0192	0.0211	0.0192	0.0281	0.0247

Note: Sample loss due to some manufacturing listed firms that did not publicly disclose ESG information for comparative analysis. *** p < 0.01, ** p < 0.05.

, columns (2), (4), and (6) (non-disclosers) show significantly positive coefficients for Agg at all three quantiles (1% level), indicating a significant positive effect of agglomeration on firm profits. Furthermore, the results show that higher productivity (higher quantile) corresponds to a smaller regression coefficient for Agg, signifying relatively lower gains from agglomeration. This suggests heterogeneous effects of agglomeration on profits across productivity levels, with a decreasing trend in coefficients across quantiles—lower productivity firms benefit more, indicating externality imbalance potentially leading to adverse selection. In contrast, columns (1), (3), and (5) (disclosers) show no significant difference in regression coefficients; coefficients for higher and lower quantiles are similar, without a decreasing trend. This indicates that the externality imbalance caused by productivity differences is mitigated to some extent. The path “Agglomeration → Productivity → Profit” is not interrupted by potential adverse selection. Therefore, under consistent external environments, firms of varying productivity can benefit from agglomeration, leading to greater investment in ESG decisions and ultimately impacting ESG performance.

5.2. External Factor: Compliance Pressure

Existing research finds that industry characteristics significantly influence corporate environmental [30], social [31], and governance [32,33] performance. That is, differences in environmental sensitivity led firms to react differently to external shocks (e.g., carbon emissions trading pilot policies), resulting in varying stakeholder expectations regarding ESG performance and further affecting subsequent ESG decisions [34]. Accordingly, this paper divides sample firms into high and low environmental sensitivity industries based on whether they belong to heavily polluting industries (Under the Guidelines for Industry Classification of Listed Companies (revised 2012) issued by the China Securities Regulatory Commission (CSRC) and the Listed Companies Environmental Verification Industry Categorization Directory issued by the Ministry of Environmental Protection, the following manufacturing sub-sectors, identified by their two-digit industry codes, are defined as heavy-polluting industries: C15, C17, C18, C19, C22, C25, C26, C27, C28, C29, C31, C32).

Table 7. Environmental Sensitivity Test of Corporate ESG Performance in three dimensions.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
	High Sensitivity E	Low Sensitivity E	High Sensitivity S	Low Sensitivity S	High Sensitivity G	Low Sensitivity G	High Sensitivity ESG	Low Sensitivity ESG
Agglomeration	0.0678 * (0.0341)	0.0749 (0.0353)	0.0554 ** (0.0147)	0.0402 ** (0.0210)	0.0322 ** (0.0087)	0.0181 * (0.0096)	0.0904 *** (0.0209)	0.0871 * (0.0530)
CV	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Firm FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
N	4087	5418	4087	5418	4087	5418	4087	5418
Adj.R2	0.0137	0.0540	0.0214	0.0965	0.0790	0.1110	0.0545	0.1195

Note: Regression sample is split into two groups based on whether the firm belongs to a high-pollution industry; the total sample size remains largely unchanged. *** p < 0.01, ** p < 0.05, * p < 0.10.

tests whether the impact of manufacturing agglomeration on ESG performance differs across environmental sensitivity. Results show significantly positive coefficients for Agg on the ESG performance and its three dimensions (E, S, G) for high-sensitivity firms, where the scores for the individual E, S, and G components are represented by their respective standardized rating metrics, typically obtained from Huazheng and Bloomberg ESG ratings. Moreover, the effect is more pronounced for high-sensitivity firms compared to low-sensitivity firms. This indicates that high-sensitivity industry firms face stronger public and stakeholder scrutiny regarding environmental protection, social responsibility, and governance, making their ESG performance more susceptible to external shocks. Facing greater compliance pressure, firms in heavily polluting industries demonstrate more significant improvements in ESG performance.

The preceding analysis establishes that corporate ESG performance is demonstrably sensitive to exogenous shocks, translating at the firm level into heightened compliance pressure. From an economic perspective grounded in industrial agglomeration theory, the genesis of this pressure is multifaceted. Firstly, within concentrated manufacturing clusters, particularly those housing pollution-intensive sectors, regulatory authorities may intensify environmental oversight under certain fiscal conditions [35], thereby amplifying external compliance requirements for all cluster participants. This regulatory channel typically materializes through stricter environmental mandates and elevated compliance costs [15], often stemming from the credible threat of substantial financial penalties. Secondly, the competitive microclimate inherent within agglomerations fosters strategic interdependence. Firms actively monitor and react to competitor actions to mitigate potential competitive disadvantages. In this context, compliance pressure manifests as imitation costs—firms experience economic incentives to emulate the observable ESG behaviors of proximate peers, reflecting a distinct peer effect. Given the significant potential influence of both pathways on ESG performance, this study formally incorporates compliance costs and peer effects as mediating variables within the theoretical framework linking manufacturing agglomeration to ESG outcomes. Specifically, compliance costs are empirically proxied by firm expenditures on Directors’ and Officers’ (D&O) liability insurance. D&O insurance serves as a risk management instrument [36], indemnifying executives against legal costs and liabilities arising from alleged negligence or misconduct during their duties. Economic research supports the interpretation that D&O insurance premiums reflect anticipated compliance costs. Higher premiums signal heightened corporate perception of exposure to external risks, primarily encompassing credit, market, and operational vulnerabilities, thereby serving as a measurable indicator of the cost burden associated with regulatory adherence and litigation risk mitigation. Concurrently, the peer effect is quantified employing the methodology established by Manski [37], operationalized as the mean ESG performance score of other firms within the same narrowly defined industry classification.

Empirical results, detailed in

Table 8. Impact Mechanism Test Results of External Compliance Pressure on Corporate ESG Performance.

	(1)	(2)	(3)	(4)	(5)
	Compliance Cost	ESG	Peer Effect	ESG	ESG
Agglomeration	0.0713 *** (0.0078)	0.0858 *** (0.0271)	0.0053 ** (0.0019)	0.0886 *** (0.0198)	0.0659 *** (0.0177)
Compliance Cost		0.0858 (0.0538)			0.0080 (0.1717)
Peer Effect				0.3901 * (0.2411)	0.3013 ** (0.1384)
CV	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes
Firm FE	Yes	Yes	Yes	Yes	Yes
N	8978	8978	9074	9074	8826
Adj.R2	0.7882	0.0146	0.7795	0.0188	0.0165

Note: Some samples were excluded due to the inability to match data on compliance costs. *** p < 0.01, ** p < 0.05, * p < 0.10.

, provide robust evidence supporting the hypothesized mediation effects. Columns (1) and (2) demonstrate a statistically significant mediating role for compliance costs, while columns (3) and (4) confirm the significant mediating influence of peer effects. The findings collectively indicate that manufacturing agglomeration positively influences corporate ESG performance through dual channels: by elevating the direct and indirect costs associated with compliance and by intensifying the economic incentives for behavioral convergence driven by peer observation. This empirical validation strongly supports Hypothesis 3, confirming that agglomerative forces increase external compliance pressure, compelling firms towards enhanced ESG performance. Further analysis in column (5) reveals that within contemporary manufacturing clusters, the predominant source of compliance pressure originates from peer effects, suggesting that inter-firm imitation behavior is a particularly salient economic driver. It is pertinent to note that, for the purposes of this economic analysis, the precise distinction between pure imitation and intentional learning is less critical than the observable outcome of strategic convergence. Given the inherent uniqueness and complexity of ESG-related information, both processes can lead to similar patterns of disclosure alignment, potentially introducing biases into the disclosed information itself. From a dynamic perspective, ESG development represents a long-term investment commitment for firms. For entities already prioritizing ESG, the positive feedback loop generated by improved ratings creates sustained economic incentives for progressively more proactive disclosure and substantive implementation. Conversely, for firms initially neglecting ESG considerations, the escalating external compliance pressure—whether regulatory or peer-driven—functions as a mechanism to progressively internalize the associated economic calculus. This includes recognizing the potential opportunity costs foregone and the tangible constraints imposed by market and regulatory environments. Ultimately, this heightened awareness of the material risks, reputational effects, and future business viability implications stemming from ESG neglect provides a compelling economic rationale for firms to integrate ESG considerations into their core strategic planning and operational value chains.

6. Discussion

6.1. Summary

This study proposes three interrelated hypotheses to systematically examine the relationship between manufacturing agglomeration and corporate ESG performance, with empirical results not only validating these hypotheses but also uncovering the nuanced mechanisms through which spatial clustering reshapes firms’ sustainable behaviors.

First, regarding Hypothesis 1 (manufacturing agglomeration as a direct driver of ESG performance), empirical analysis confirms a statistically significant positive impact of agglomeration on ESG performance after controlling for firm-level and city-level confounders; this conclusion is further reinforced by robustness checks and endogeneity treatments that rule out reverse causality and omitted variable bias. Theoretically, this aligns with stakeholder theory: agglomeration intensifies interactions between firms and diverse stakeholders, creating a “stakeholder density” that elevates transparency demands—clustered firms face more concentrated oversight than dispersed ones, reducing information asymmetry and turning ESG performance from an “option” into a “strategic necessity,” while resource sharing lowers ESG implementation costs to further incentivize disclosure.

Second, for Hypothesis 2 (productivity as an internal mediator), empirical tests show that agglomeration significantly boosts total factor productivity (TFP), which in turn improves ESG performance. Additional analysis adds context that, among non-ESG disclosers, the positive impact of agglomeration on profits weakens with higher productivity—low-productivity firms benefit more from spillovers like technology diffusion, while high-productivity firms face “externality imbalance”. This pattern reveals a key insight: ESG performance acts as a “corrective mechanism” for agglomeration-induced adverse selection, as it helps high-productivity disclosers gain implicit value to offset externality losses, sustaining the “Agglomeration → Productivity → ESG” channel. Practically, this implies that agglomeration policies focusing solely on productivity need to be paired with ESG performance incentives to prevent high-productivity firms from exiting clusters.

Third, concerning Hypothesis 3 (compliance pressure and peer effects as external mediators), heterogeneity analysis indicates the “agglomeration-ESG” link is stronger for firms in high environmental sensitivity industries due to stricter regulatory oversight and public scrutiny, and mechanism tests confirm compliance costs and peer effects both mediate this relationship, with peer effects playing a more dominant role. Analytically, this underscores the role of “competitive mimicry” in clustered ecosystems: firms closely observe peers’ ESG actions, and non-disclosers risk losing market share or legitimacy if competitors use ESG performance to attract investors or avoid regulatory penalties—especially in high-sensitivity industries, where one firm’s poor ESG performance can damage the entire cluster’s reputation, turning ESG from an “individual choice” into a “collective norm”. Notably, the dominance of peer effects suggests policymakers can leverage social learning in agglomerations to drive voluntary disclosure, complementing top-down regulation.

6.2. Policy Implications

Based on the empirical findings, this study derives three targeted policy recommendations to integrate ESG into manufacturing agglomeration and promote sustainable industrial development.

First, leverage agglomeration to standardize ESG regulation. Manufacturing agglomeration concentrates firms with similar industrial attributes, creating a “natural regulatory zone” for ESG oversight. Policymakers should design cluster-specific ESG performance frameworks—for example, developing unified reporting standards for high-polluting agglomerations and establishing regional ESG information platforms to streamline data collection and verification. This not only reduces regulatory costs but also ensures consistency in ESG metrics, facilitating cross-firm and cross-region comparisons. Additionally, local governments can link agglomeration support policies to ESG performance, incentivizing clusters to adopt higher sustainability standards.

Second, strengthen productivity-driven ESG incentives for firms. Given that productivity is a core internal driver of ESG performance, firms should prioritize technological innovation and efficiency gains to reduce the opportunity cost of ESG investment. Policymakers can support this by offering R&D subsidies or low-interest green loans to agglomerated firms that integrate ESG into their innovation strategies. For small and medium-sized enterprises—which often face resource constraints—governments can establish ESG capacity-building programs within clusters, leveraging agglomeration spillovers to disseminate best practices.

Third, mitigate peer effect risks and prevent greenwashing. While peer effects promote ESG adoption, they may also lead to homogenized disclosure or greenwashing. To address this, regulators should mandate “materiality-based” ESG performance, requiring firms to report industry-specific, impact-relevant metrics rather than generic information. Additionally, governments can introduce post-disclosure verification mechanisms—such as third-party audits for high-agglomeration regions—and impose penalties for misleading ESG claims. For clusters with strong peer effects, policymakers can further encourage “ESG leadership programs,” where leading firms share actionable sustainability strategies to drive substantive improvements rather than superficial imitation.

6.3. Research Limitations and Future Directions

While the findings above contribute to understanding the link between manufacturing agglomeration and corporate ESG performance, this study is not without limitations, which also point to avenues for future research.

First, cross-region comparative analysis is not feasible, as the empirical work focuses solely on Chinese A-share listed manufacturers due to three key data and institutional constraints: ESG performance practices and regulatory frameworks vary drastically across countries, making direct cross-country comparisons of the agglomeration-ESG relationship unreliable without harmonized data; manufacturing agglomeration in China is largely government-led while that in developed economies is more market-driven, creating fundamental differences in the mechanisms linking agglomeration to ESG that complicate causal identification across regions; and granular data on agglomeration and control variables are not consistently available for other areas firms, even with global ESG ratings from sources like Bloomberg or MSCI, precluding a balanced cross-country sample. Second, the study simplifies ESG measurement by using aggregate ratings as the dependent variable, which may obscure heterogeneity across the environmental (E), social (S), and governance (G) dimensions—for instance, agglomeration might have a stronger impact on emission reductions (E) than labor welfare (S), a distinction lost in aggregate scores. Third, the analysis of peer effects is static, measuring them as industry peers’ average ESG performance at a given time without capturing how these effects evolve dynamically, which limits understanding of how inter-firm interactions shape long-term ESG engagement.

To address these limitations, future research can pursue three directions: as global ESG standards become more widespread, constructing cross-country samples of agglomerated manufacturers to compare the agglomeration-ESG relationship across institutional contexts could clarify how institutional factors moderate the roles of productivity and compliance pressure; using firm-level survey data or case studies to explore the micro-foundations of ESG decisions in agglomerations—such as how managers perceive peer pressure or how board characteristics like independent directors interact with agglomeration to influence ESG performance—could refine understanding of the “black box” between agglomeration and ESG; and applying spatial econometric models to capture spillover effects between adjacent agglomerations (e.g., whether a high-ESG cluster influences neighboring ones) and DID designs to examine how exogenous shocks alter the agglomeration-ESG relationship over time could provide deeper insights into policy effectiveness.