Urban–Rural Environmental Regulation Convergence and Enterprise Export: Micro-Evidence from Chinese Timber Processing Industry

Abstract

Environmental regulations serve as a critical determinant of industrial competitiveness in the global market. Recent policy shifts have driven a gradual convergence of rural environmental standards with urban norms, fostering a dynamic landscape of “top-down competition” between urban and rural regulatory frameworks. While the economic consequences of regional regulatory disparities are well-documented, the specific impacts of this regulatory convergence remain insufficiently explored. To address this gap, this study constructs a novel index to measure the convergence of environmental regulations between urban districts and rural counties at the prefecture level. Utilizing an unbalanced panel dataset of 5600 county-level timber processing enterprises, the Heckman two-stage model is employed for empirical analysis. The results demonstrate that the convergence of urban and rural environmental regulations significantly enhances both the export probability and export intensity of county-level firms, with these effects exhibiting persistence and cumulative growth over time. These findings remain robust across a series of validation tests, including instrumental variable estimation, double machine learning, and alternative model specifications. Mechanism analysis reveals that regulatory convergence promotes exports primarily by improving access to green credit and enhancing peer quality within the industry. Furthermore, heterogeneity tests indicate that the positive effects are most pronounced for start-ups and firms in the decline stage, as well as for enterprises located in eastern China, those outside the Yangtze River Economic Belt, and those subject to minimal government intervention. This study provides critical micro-level evidence that helps enterprises navigate the evolving policy landscape and supports the formulation of strategies to boost export trade amidst the integration of environmental regulations.

1. Introduction

China has become one of the world’s leading producers and exporters of timber and wood products [1]. In recent years, China’s wood processing industry generated an output value exceeding RMB 2 trillion, with the trade value of wood and wood-based products reaching tens of billions of USD. Most wood processing enterprises are located in rural areas, where they extend their industrial chains through raw material procurement, thereby increasing farmers’ income and revitalizing rural industries [2]. However, the export potential of these enterprises is constrained by green trade barriers. On the one hand, under the “city-managed county” administrative framework, the wood processing industry is increasingly drawn away from urban areas. On the other hand, a lack of agglomeration effects results in low overall returns to scale within the industry. These challenges are becoming increasingly significant in the global market.

The convergence of environmental regulations between urban and rural areas may offer a potential solution. With the implementation of new environmental protection laws, enhanced inspections, and accountability mechanisms, ecological performance has become a factor in official promotions [3]. Consequently, the introduction of polluting enterprises into underdeveloped regions has declined [4], and the trend of “race to the bottom” in interregional environmental regulation has shown signs of easing. There is even evidence of stricter environmental regulations being applied in both urban and rural areas—a phenomenon referred to as the “race to the top [5].”

Meanwhile, county-level wood processing enterprises face inherent disadvantages in export competitiveness compared to their urban counterparts due to their geographic location and relatively small scale [6]. Thus, whether the convergence of environmental regulations will exacerbate these disadvantages or provide opportunities for improvement remains a critical and complex issue in the context of promoting new urbanization.

Existing literature has extensively discussed the economic impacts of environmental regulations, their effects on technological innovation [7,8], R&D activities [9], and their role in promoting exports. Research by Costantini et al. [10] found that environmental regulations can spur technological innovation, which offsets cost burdens and enhances export performance [11,12,13]. Conversely, Huang et al. [14] proposed a “U-shaped” relationship, where environmental regulations initially increase costs but, over time, lead to innovation that turns negative impacts into positive ones, boosting exports [15,16]. Recent studies have focused on provincial and district-level governance of environmental regulations, particularly the effects of the “race to the top” phenomenon [17,18]. Vogel et al. observed that the ecological performance evaluation mechanisms have driven a shift from a “race to the bottom” to a “race to the top,” with technological innovation mediating this shift [19,20,21,22,23]. Despite these valuable contributions, several gaps remain in the literature: First, there is a lack of research on the convergence of environmental regulation strictness between urban and rural areas and its specific effects on the exports of county-level timber processing enterprises [10,24,25]. Moreover, the transmission pathways of these effects remain underexplored. Second, previous studies primarily focus on provincial or district-level assessments, neglecting the differences in export performance between counties within municipalities and those in urban districts [25,26]. This makes it difficult to understand the evolution of urban–rural environmental regulation convergence and its impact on exports at the county level. Third, much of the existing research centers on large industrial enterprises, leaving a gap in understanding the dynamics of small and medium-sized enterprises (SMEs) at the county level, which make up the majority of timber processing enterprises [14,15]. This limits the generalizability and practical significance of the findings.

This study systematically investigates how the convergence of urban and rural environmental regulations affects the export of county-level wood-processing enterprises. Specifically, it (1) traces the evolution of regulatory convergence and develops a theoretical framework linking it to firm-level exports; (2) constructs a composite indicator system to measure both the degree of regulatory convergence and the export performances of county-level enterprises, and conducts a quantitative assessment using nationally representative data; (3) employs causal-identification strategies to mitigate endogeneity, estimates the export effect of regulatory convergence, and performs heterogeneity and robustness checks; and (4) evaluates the long-term sustainability of the effect and examines transmission channels such as green-credit access and quality upgrading among peer firms. The findings advance the theoretical understanding of how environmental-regulation convergence influences small- and medium-sized enterprise exports and provide policy guidance for promoting export growth in China’s county-level wood-processing sector. The remainder of the paper is organized as follows: Section 2 develops the research hypotheses and data analysis. Section 3 reports the empirical results. Section 4 discusses the underlying mechanisms and heterogeneity analyses. Section 5 concludes.

2. Research Hypotheses and Data Analysis

2.1. Research Hypotheses

2.1.1. The Direct Effects of Urban–Rural Environmental Regulation Convergence

The convergence of urban and rural environmental regulations creates a win-win situation for the export of county-level timber processing enterprises. According to Porter’s Innovation Compensation Effect and First-Mover Advantage Theory [27], stricter regulations prompt enterprises to innovate, yielding compensatory gains that offset regulatory costs. This innovation enhances competitiveness, turning environmental pressures into export opportunities. From the perspective of Resource Integration Theory [28], regulatory convergence drives enterprises to optimize resource allocation, apply technological innovations to production processes, reduce pollution, and enhance product differentiation. This reduces inefficiencies and fragmentation, improving market competitiveness and supporting exports. Additionally, Organizational Change Theory [29] suggests that the convergence of environmental regulations forces county-level timber processing enterprises to upgrade management structures, overcoming organizational inertia. By incorporating green practices into their operations, these enterprises reduce environmental management costs, transforming external pressures into first-mover advantages that promote export growth. Furthermore, Enterprise Capability Theory [30] posits that under the pressure of regulatory convergence, enterprises integrate sustainable innovations into core business strategies, improving brand value and expanding market share, thus stimulating export vitality.

Hypothesis H1:

The convergence of urban and rural environmental regulations has a significant positive impact on the exports of county-level timber processing enterprises.

2.1.2. Indirect Effects of Urban–Rural Environmental Regulation Convergence on County-Level Timber Processing Enterprises’ Exports

The convergence of urban and rural environmental regulations indirectly promotes the exports of county-level timber processing enterprises through green credit and peer quality effects [31]. Financing constraints are a key barrier to exports. The convergence of urban and rural environmental regulations strengthens green credit policies, encouraging enterprises to explore alternative funding sources. These policies promote the application of green production technologies, which alleviates financing difficulties and fosters export growth. Green credit enables enterprises to increase research and development investments, improve product quality, and enhance firms’ export performance [32]. By reducing production costs and increasing profit margins, green credit further stimulates export expansion.

Hypothesis H2:

The convergence of urban and rural environmental regulations increases the probability and intensity of export of county-level timber processing enterprises by easing green credit constraints.

Convergence strengthens institutional pressures on enterprises, driving them to improve their product quality and environmental standards [33]. This leads to homogeneity among enterprises within a region, with newly established firms clustering with higher-quality peers. The improvement in group quality facilitates the spillover of green technologies, allowing enterprises to meet international environmental standards, reduce green trade barriers, and expand export markets [34]. This enhanced group quality also improves the brand image and overall competitiveness of county-level timber processing enterprises in international markets.

Hypothesis H3:

The convergence of urban and rural environmental regulations increases the probability and intensity of county-level timber processing enterprises’ export by enhancing group quality.

2.2. Data Sources

Using Python software (3.10.13) to retrieve environmental penalty case information for urban and rural areas from the “Institute of Public and Environmental Affairs (IPE) database”; The raw data on industrial sulfur dioxide emissions, nitrogen oxide emissions, industrial smoke emissions, and carbon emissions in “urban” and “rural” areas come from the “China Urban Statistical Yearbook” and the “China County Statistical Yearbook.” The three afore mentioned databases are matched to obtain the raw data to measure the convergence of urban and rural environmental regulations.

The financial and other characteristic data of timber processing enterprises come from the “National Tax Survey Database,” jointly collected by the Ministry of Finance and the State Administration of Taxation. The data has been updated until 2016, covering both large-scale and small and medium-sized enterprises, with large amounts of information and strong representativeness. To improve the data quality, the following adjustments were made: first, missing data for intermediate years in some enterprises were supplemented using interpolation methods; Second, samples that violated accounting principles or had obvious logical errors were excluded, such as those where fixed assets exceed total assets, total profit exceeds operating revenue, operating revenue and main business income are less than 0, research and development expenses exceed management expenses, year-end total assets are negative, or average annual number of employees and salary and bonus expenses are negative. Third, when data for export value, research and development expenses, management expenses, financial expenses, sales expenses, subsidy income, tax refund amounts, business tax and surcharges, advertising expenses, and net cash flow are missing, the value is assumed to be 0 [35]. The research subjects are clearly defined according to the “National Economic Industry Classification (GB/T 4754-2011)”, retaining enterprises classified under the categories of timber processing industry C201, artificial board manufacturing C202, wood products manufacturing C203, and wooden furniture manufacturing C211 [36]. In addition, the urban districts of each prefecture-level city are considered “urban” areas, while other counties (cities) within the same prefecture-level city are considered “rural” areas. The research subjects are county-level timber processing enterprises, i.e., those from the “rural” areas. Based on the first six digits of the taxpayer identification number of timber processing enterprises, the administrative district codes of the enterprise’s prefecture-level city and “urban” or “rural” location are identified. The timber processing enterprises’ financial and other characteristic data are then matched with the observational data on urban–rural environmental regulation convergence based on these administrative codes, excluding unmatched enterprise samples. After the above adjustments, a panel dataset of 5600 county-level timber processing enterprises from 2007 to 2016 was constructed, with 24,525 observations. The study period from 2007 to 2016 is primarily determined by data availability and consistency across multiple sources. Specifically, this time window captures a critical phase of China’s environmental governance transition, including the strengthening of environmental enforcement and the initial convergence of urban–rural regulatory standards. Therefore, while the data do not reflect the most recent industry conditions, the analysis is well suited to identify the underlying mechanisms through which urban–rural environmental regulation convergence affects firms’ export probability and export intensity. The geographic scope covers 900 counties in mainland China, with 1067 county-level timber processing enterprises involved in export activities. The sample is highly representative. Based on this, the impact of urban–rural environmental regulation convergence on county-level timber processing enterprises’ exports from 2007 to 2016 is examined, and the research conclusions are representative and reliable.

2.3. Variable Selection

The core explanatory variable in this study is the convergence of urban–rural environmental regulations. The convergence of urban and rural environmental regulations is measured at the district-level city level from 2007 to 2016. In this context, the district within each city is designated as “urban,” and the surrounding counties or county-level cities within the same district are classified as “rural.” Data for this measurement is drawn from the “Institute of Public and Environmental Affairs (IPE) database” and the “China Legal Digital Library.” The following steps outline the procedure for calculating urban–rural environmental regulation convergence. Given that environmental regulation stringency is not directly observable at the subnational level, we employ outcome-based indicators to proxy the effective intensity of environmental regulation. Specifically, we collect county and district level indicators within each prefecture-level city, including the number of environmental penalty cases per unit of GDP, pollutant emissions per unit of GDP (SO₂, nitrogen oxides, chemical oxygen demand, and ammonia nitrogen), and carbon emissions per unit of GDP. These indicators reflect regulatory enforcement and compliance outcomes. The entropy weight method is then applied to calculate a comprehensive environmental regulation evaluation score for each county or district, yielding a single index ranging from 0 and 1, with higher values indicating stronger effective environmental regulation. Urban environmental regulation intensity is calculated as a GDP-weighted average of the composite indices across municipal districts, while rural environmental regulation intensity is computed analogously using counties’ GDP shares. Urban–rural environmental regulation convergence is subsequently measured as the ratio of the weighted rural regulation index to the weighted urban regulation index for each prefecture-level city in each year. A higher ratio indicates a smaller gap between urban and rural regulatory outcomes and thus a higher degree of convergence.

Enterprise export probability and export intensity is assessed through both the extensive margin and the intensive margin. The extensive margin is represented by a dummy variable that indicates whether a firm engages in exports. If the total export value is greater than 0, it is coded as 1; otherwise, it is coded as 0. The intensive margin is measured using export intensity, defined as the ratio of a firm’s total export value to its total sales value, following standard practice in the international trade literature [24].

To explore the transmission mechanisms through which urban–rural environmental regulation convergence affects export probability and export intensity, we focus on three channels: green credit, industry aggregation, and group quality. Green Credit, Existing literature often measures green credit using dummy variables based on the announcement points of green credit policies, the proportion of loans allocated to environmental projects, or the interest expenditures for high-energy-consuming industries [37]. To capture the intensity of green credit changes over time, this study follows the methodology of Yin et al. [38], measuring green credit as the ratio of loans for environmental projects to total regional credit. Industry Aggregation and Group Quality, the quality of an enterprise is measured through its size, with fixed effects incorporated to account for variations in size [31]. Group quality is assessed based on the quality of other enterprises within the same group, specifically by the proportion of newly established companies relative to the total number of new individual businesses.

A set of control variables is included to account for firm-level characteristics and financial conditions that may influence export probability and export intensity. These include human capital, enterprise size, capital intensity, and green total factor productivity, as shown in

Table 1. Variables and definitions.

Variable Type	Variable Name	Notation	Variable Interpretation
explained variable	Exported or not	Eo	Takes the value of 1 when the firm’s exports are greater than 0 and 0 otherwise.
	export intensity (intensive margin)	Ex	Total exports/total sales
explanatory variable	Integration of urban and rural environmental regulation	Er	County Unit Industrial Sulfur Dioxide Emissions
			County Unit Industrial NOx Emissions
			County Unit Industrial Soot Emissions
			County Carbon Emissions
			Number of environmental penalties in the county
control variable	Level of human capital	Hl	Total wages divided by the number of employees
	Capital intensity	Ci	Fixed assets per employee
	Enterprise size	Es	(Logarithmic value of (number of employees + 1)
	Earnings on assets	Ro	Total Profit/Total Assets
	Capital structure	Al	Debt-to-asset ratio at the beginning of the year
	Lending capacity	Lc	Finance costs/total assets
	Tax burden	Tb	(Logarithmic value of (sales tax and surcharge + 1)
	Total assets	Ta	(Logarithmic value of (total assets + 1)
	Government support	Sg	(Logarithmic value of (subsidy income + amount of tax rebate received + 1)
	Business costs	Oc	Ratio of operating costs to operating income
	Sales expense	Sc	Ratio of selling expenses to operating income
	Overhead	Mc	Ratio of administrative expenses to operating income
	Net cash flow	Nc	(Logarithm of (net cash flow + 1)
	Green total factor productivity	Gt	Measured using a non-radial SBM-ML index
Mechanism variables	Green credit	Gc	Total credits for environmental projects/total credits
	Cohort quality	Nc	Number of newly created companies/Number of newly created self-employed households

. Average wage is used as a proxy for human capital due to data limitations. These variables capture the impact of labor, capital, and technology on export probability and export intensity. Additional financial variables include total assets, net cash flow, operating costs, management expenses, sales expenses, and return on assets. These reflect the financial health of enterprises, including their cash-generating ability, cost control, and operational management capacity. Capital structure, loan capacity, tax burden, and government support are also considered to examine the efficiency of financial structures, financing, and the influence of institutional costs (such as taxes) on export probability and export intensity [39].

2.4. Research Methods

2.4.1. Benchmark Regression Model

Exporting and non-exporting enterprises face systematically different conditions and treating them as homogeneous may lead to biased estimation. To test the impact of urban–rural environmental convergence on the export probability and export intensity of county-level timber processing enterprises, based on the above theoretical analysis and combined with the studies of Heckman (1979) [40] and Du and Li (2020) [25], a two-stage selection model is established. The first stage is the export participation regression model for county-level timber processing enterprises:

$${E\mathrm{o}}_{it}={\partial }_0+{\partial }_1{Er}_{it}+{\partial }_2{CTR}_{it}+{Eo}_{it-1+}{\mu }_t+{\nu }_i+{\epsilon }_{it}$$

(1)

In the first stage, the Probit estimation method is used to estimate the relationship between urban–rural environmental convergence and the marginal expansion of county-level timber processing enterprises’ exports. Eo_it is the dummy variable reflecting whether county-level timber processing enterprises export, Er_it is the degree of urban–rural environmental convergence, and its coefficient directly reflects the effect of urban–rural environmental convergence on whether county-level timber processing enterprises export. CTR_it is the set of control variables affecting the export decision, µ_t is the time fixed effect, v_i is the individual fixed effect, and ε_it is the random disturbance term in the first stage.

In addition, at least one variable in the export participation equation cannot be included in the export intensity equation. Following Lu J’s [41] study, the lagged term of whether the enterprise exports, Eo_it−1, is added in Equation (1). This is because prior export history helps enterprises overcome sunk costs and supports continued export activity. The second stage is the export intensity regression model for county-level timber processing enterprises:

$${E\mathrm{x}}_{i\mathrm{t}}={\beta }_0+{\beta }_1{Er}_{it}+{\beta }_2{CTR}_{it}+{\mu }_t+{\nu }_i+{\xi }_{it}+\psi \hat{\lambda }$$

(2)

Ex_it reflects the export intensity margin of county-level timber processing enterprises, i.e., the ratio of the total export value of enterprise i in year t to its total sales. CTR_it is a set of control variables, and similarly, fixed time effects (µ_t) and individual effects (v_i), with ξ_it being the random disturbance term of the second stage. Due to the correlation between the two-stage regressions, it is assumed that ε_it and ξ_it follow a normal distribution. When their correlation coefficient is not zero, it indicates the presence of selection bias in the equation. At this point, the Heckman two-step method can be appropriately applied for correction. The principle is as follows: estimate Equation (1): $\mathrm{P}({E\mathrm{o}}_{i\mathrm{t}}=1)=\varphi ({\partial }_0+{\partial }_1{Er}_{it}+{\partial }_2{CTR}_{it}+{\mu }_t+{\nu }_i+{\epsilon }_{it})$.

After calculating, the inverse Mills ratio (λ) obtained above is added as a control variable to estimate Equation (2) to correct the bias issue.

2.4.2. Mechanism Impact Testing Model

After identifying the total effect of the convergence of urban and rural environmental regulations on the export of county-level timber processing enterprises, a mediation model is introduced to further explore the impact path. Based on the research of Wen and Ye [42], the following Formulas (3)–(5) are established for testing. In Formulas (3) and (4), δ₁ and Γ₁ are the impact coefficients of the convergence of urban and rural environmental regulations on green credit and peer quality, respectively. Combined with θ₁ and θ₂ in Formula (5), δ₁ × θ₁ and Γ₁ × θ₂ represent the indirect effects of ${Er}_{\mathrm{it}}$.

$${\mathrm{Gc}}_{i\mathrm{t}}={\delta }_0+{\delta }_1{Er}_{it}+{\delta }_2{CTR}_{it}+{\mu }_t+{\nu }_i+{\xi }_{it}+\psi \hat{\lambda }$$

(3)

$${\mathrm{Nc}}_{i\mathrm{t}}={\Gamma }_0+{\Gamma }_1{Er}_{it}+{\Gamma }_2{CTR}_{it}+{\mu }_t+{\nu }_i+{\xi }_{it}+\psi \hat{\lambda }$$

(4)

$${E\mathrm{x}}_{i\mathrm{t}}={\sigma }_0+{\sigma }_1{Er}_{it}+{\sigma }_2{CTR}_{it}+{\vartheta }_1{Gc}_{it}+{\vartheta }_2{Nc}_{it}+{\mu }_t+{\nu }_i+{\xi }_{it}+\psi \hat{\lambda }$$

(5)

Taking green credit as an example, the mainstream methods for testing mediation effects mainly include three approaches: the first is to use the three-step method, fitting Equations (1), (3) and (5), and sequentially testing the significance of β₂, δ₂, and θ₁. If all are significant, it indicates that green credit has a mediation effect. The second method is under the premise that β₂ is significant, and at least one of δ₂ or θ₁ is insignificant. In this case, the Sobel test is used to check the significance of the product of the coefficients δ₂ × θ₁ (H0: δ₂ × θ₁ = 0). The test statistic is calculated as follows:

$$Z={\displaystyle \frac{{\delta }_2\times {\theta }_1}{\sqrt{{\theta }_1^2\times S_{{\delta }_2}^2+{\delta }_2^2\times S_{{\theta }_1}^2}}}$$

(6)

where $S_{{\delta }_2}$ and $S_{{\theta }_1}$ denote the standard errors of δ₂ and θ₁, respectively. When β₂ is insignificant, the bootstrap method is applied to assess the significance of the indirect effect based on repeated resampling.

3. Results Analysis

3.1. Typical Case Analysis

Figure 1 illustrates the spatiotemporal evolution of urban–rural environmental regulation convergence during 2007–2016: the convergence index exhibited a monotonic upward trend, with regional disparities narrowing substantially and low-convergence areas diminishing markedly. Spatially, high-convergence regions were concentrated in eastern coastal provinces, aligning with the gradient distribution of institutional capacity and marketization levels. Although central and western regions maintained lower absolute convergence levels, their rapid improvement rates demonstrated a notable catching-up effect. The evolution displayed distinct phase characteristics: convergence progressed gradually from 2007–2010 but accelerated sharply after 2013. This acceleration likely reflects intensified ecological supervision and the incremental effectiveness of environmental policies.

Panels (a)–(d) show the provincial-level values of urban–rural environmental regulation convergence in 2007, 2010, 2013, and 2016, respectively. Darker shading indicates a higher degree of convergence, while lighter shading represents lower levels. Provinces with missing data are shown in hatched patterns. Provincial boundaries are delineated for reference.

Table 2. Descriptive statistics of variables.

	County Wood Processors (5600)			County Wood Processing Exporters (1067)			County Wood Processing Non-Exporters (4533)
Variant	Mean	Min	Max	Mean	Min	Max	Mean	Min	Max
Eo	0.224	0	1	1	1	1	0	0	0
Ex	0.140	0	1	0.626	0	1	0	0	0
Er	0.762	0.015	6.884	0.896	0.024	5.009	0.723	0.015	6.884
Hl	1.248	0.000	13.174	1.553	0.000	13.174	1.160	0.000	12.658
Ci	0.001	0.000	1.000	0.001	0.000	0.503	0.001	0.000	1.000
Es	3.805	0.000	9.730	4.519	0.000	9730	3.599	0.000	9.540
Ro	0.080	0.000	1.000	0.080	0.066	1.000	0.080	0.000	0.785
Al	0.000	0.000	1.000	0.000	0.000	0.099	0.000	0.000	1.000
Lc	0.004	0.000	1.000	0.004	0.002	1.000	0.003	0.000	0.036
Tb	2.958	−0.693	12.252	3.236	−0.693	9.944	2.878	−0.693	12.252
Oc	0.009	0.000	1.000	0.009	0.000	0.089	0.009	0.000	1.000
Sc	0.001	0.000	1.000	0.002	0.000	0.467	0.001	0.000	1.000
Mc	0.000	0.000	1.000	0.000	0.000	1.000	0.000	0.000	0.385
Nc	2.840	−40.514	56.883	3.764	−33.442	56.883	2.573	−40.514	37.523
Ta	1.767	−24.866	13.016	3.231	−21.047	12.686	1.345	−24.866	13.016
Sg	8.986	0.405	15.062	9.937	2.197	15.057	8.711	0.405	15.062
Gt	0.735	0.716	0.788	0.735	0.734	0.764	0.734	0.716	0.788

reveals pronounced sample selection characteristics: among 5600 county-level wood processing enterprises, 1067 (19.1%) engage in export activities, confirming heterogeneous export participation. Comparative analysis of core indicators yields three key findings: First, export enterprises exhibit significantly higher average environmental regulation convergence (0.896) than non-exporters (0.723), suggesting that regulatory convergence may differentially incentivize exporting firms. Second, exporters demonstrate systematic advantages in human capital, production scale, and technological innovation, consistent with self-selection of high-productivity firms into export markets. Third, the mean variance inflation factor (VIF) across all variables is 1.42, well below the conventional threshold of 10, indicating no severe multicollinearity concerns. Figure 2 depicts a positive association between urban–rural environmental regulation convergence and both the probability and intensity of firm exports. However, this correlation may reflect endogeneity concerns, including reverse causality (e.g., exporters proactively adapting to stringent standards) or omitted variable bias (e.g., regional enforcement capacity disparities). Thus, the underlying causal relationship requires rigorous empirical verification.

3.2. Positive Effects of Urban–Rural Environmental Regulation Convergence

Table 3. Regression Results of Urban–Rural Environmental Regulation Convergence on County-level Timber Processing Enterprises’ Exports.

	(1)	(2)	(3)	(4)	(5)	(6)
	Export Probability	Export Intensity	Export Probability	Export Intensity	Export Probability	Export Intensity
L.Et	3.604 *** (77.79)		3.336 *** (85.58)		3.487 *** (72.64)
Er	0.058 ** (2.39)	0.024 * (1.89)	0.047 * (1.92)	0.030 *** (4.61)	0.050 ** (1.98)	0.031 *** (2.69)
Hl			−0.008 (−1.15)	0.005 *** (2.81)	0.025 *** (2.79)	−0.000 (−0.00)
Ci			0.000 (1.03)	0.000 ** (2.20)	0.000 (0.93)	0.000 ** (2.38)
Es			0.143 *** (7.97)	0.040 *** (8.01)	0.102 *** (5.14)	0.042 *** (5.93)
Ro			−0.020 (−0.93)	0.016 (0.78)	−0.024 ** (−2.00)	0.020 (0.85)
Al			−0.000 * (−1.80)	0.000 (1.11)	−0.000 (−1.34)	0.000 *** (5.81)
Lc			0.674 *** (2.59)	0.750 *** (3.97)	0.655 *** (2.60)	0.762 *** (2.77)
Tb			−0.100 *** (−9.01)	−0.026 *** (−9.96)	−0.084 *** (−6.99)	−0.027 *** (−7.04)
Oc			−0.002 * (−1.84)	0.000 (0.16)	−0.002 * (−1.83)	0.000 (0.01)
Sc			0.001 *** (3.40)	−0.000 (−0.56)	0.001 ** (2.08)	−0.000 (−1.52)
Mc			−0.000 (−1.27)	−0.000 (−0.49)	−0.000 (−1.28)	−0.000 *** (−4.56)
Nc			0.013 *** (2.76)	−0.005 *** (−3.92)	0.008 (1.48)	−0.004 ** (−2.23)
Gt			11.993 (0.45)	6.402 (0.62)	13.212 * (1.69)	5.862 (1.27)
Sg			0.028 *** (4.57)	0.023 *** (14.30)	0.020 *** (3.26)	0.023 *** (9.47)
Ta			0.076 *** (4.67)	−0.066 *** (−13.71)	0.113 *** (6.78)	−0.069 *** (−9.06)
_cons	−0.951 *** (−16.42)	0.746 *** (19.88)	−11.645 (−0.59)	−3.579 (−0.47)	−11.734 ** (−2.05)	−3.184 (−0.94)
Fixed year	Yes	Yes	No	No	Yes	Yes
Stationary individual	Yes	Yes	No	No	Yes	Yes
F		46.378		59.830		44.618
R2		0.076		0.181		0.186

Notes: Export probability refers to the extensive margin and is estimated using a Probit model in the first stage of the Heckman procedure. Export intensity refers to the intensive margin and is estimated using the second-stage Heckman regression. L.Et denotes the one-period lagged export participation indicator. Columns (1)–(2), (3)–(4), and (5)–(6) report different model specifications. Year and firm fixed effects are included where indicated. t-statistics are reported in parentheses. *, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively.

reports a positive and statistically significant coefficient on the one-period lagged export indicator for county-level wood processing enterprises. This finding demonstrates that prior export probability and export intensity exerts a significant positive influence on current export decisions. The result aligns with Roberts’ [43], who argues that past export experience helps firms overcome sunk costs and sustain export activities. This pattern confirms the path dependence of enterprise export probability and export intensity, consistent with dynamic capability theory. Models (1)–(2) in Table 3 present baseline estimates without additional firm-level controls, focusing on the unconditional relationship between urban–rural environmental regulation convergence and export outcomes. Models (3)–(4) incorporate firm-level control variables, while Models (5)–(6) further include lagged export participation and fixed effects to assess the robustness of the estimated effects.

The results across specifications demonstrate that urban–rural environmental regulation convergence significantly enhances both the export probability (extensive margin) and export intensity (intensive margin) of county-level wood processing enterprises, supporting Hypothesis H1. Specifically, controlling for time, individual, and other relevant effects, the regulation convergence coefficients equal 0.050 and 0.031 for export probability and export intensity, respectively, and are significant at the 95% and 99% confidence levels. This positive relationship contrasts with findings by Zhang et al. and Cagatay et al. [26,44], who report that environmental regulations inhibit firm exports. The discrepancy likely reflects that regulation convergence mitigates urban–rural dualism, reduces institutional barriers, and optimizes regional resource allocation, thereby facilitating export expansion for county-level enterprises. The results align, however, with subsequent empirical literature emphasizing innovation compensation effects and first-mover advantages [45,46,47]. This perspective suggests that regulation convergence stimulates technological innovation, optimizes resource allocation, fosters organizational change, and enhances firm capabilities, ultimately transforming external environmental pressure into green competitiveness and creating a win-win outcome for export.

For control variables, government support, firm size, lending capacity, and green total factor productivity all significantly enhance export probability and intensity, suggesting that firms with strong resource integration capabilities, robust innovation foundations, or economies of scale are better positioned to reduce marginal export costs and strengthen competitiveness. Conversely, tax burden exhibits a significantly negative coefficient, indicating that heightened financial pressure constrains resources available for technological upgrading and market expansion, thereby dampening export willingness and performance.

3.2.1. Sustainable Effects of Urban–Rural Environmental Regulation Convergence

Table 4. Test of the Persistent Effect of Urban–Rural Environmental Regulation Convergence.

	(1)	(2)
	Export Probability	Export Intensity
L.Er	0.050 ** (2.01)	0.030 *** (2.67)
L2.Er	0.035 (1.26)	0.037 *** (2.94)
L3.Er	0.018 (0.46)	0.041 *** (2.90)
Control variable	YES	YES
Fixed year	YES	YES
Stationary individual	YES	YES

** and *** denote significance at the 5% and 1% levels, respectively.

reports that in the export probability model, the lagged regulation convergence coefficient is 0.050, significant at the 95% confidence level. Higher-order lag coefficients decline and become insignificant, suggesting a one-year lag in the effect on export probability. In the export intensity model, the first-, second-, and third-order lag coefficients are 0.030, 0.037, and 0.041, respectively, all significant at the 99% confidence level. In the export intensity model (Table 4), the coefficients on the first-, second-, and third-order lagged regulation convergence variables (L1.Er, L2.Er, and L3.Er) are 0.030, 0.037, and 0.041, respectively, all statistically significant at the 1% level. The increasing coefficients suggest a cumulative effect on export intensity. This sustained effect stems from long-term returns to green investment and steady improvements in market competitiveness. Firms gradually realize marginal gains through technological upgrading, workforce training, and institutional improvements, creating a virtuous cycle of environmental empowerment and export efficiency enhancement. Meanwhile, policy lags and path dependence in international markets jointly promote firms’ transition to high-quality export strategies. Overall, urban–rural environmental regulation convergence significantly contributes to the sustainable export development of county-level wood processing enterprises.

3.2.2. Robustness Tests Validate the Stability of Baseline Results

This study conducts robustness tests via four approaches—double machine learning, instrumental variables, alternative model specifications, and 1% Winsorization—to mitigate potential confounding factors, reduce model bias, and enhance estimation reliability, as shown in

Table 5. Robustness test results.

	Dual Machine Learning Tests		Instrumental Variables Test (IVT)	Replacement Model		1% Indentation
	Export Probability	Export Intensity	Export Intensity	Export Probability	Export Intensity	Export Probability	Export Intensity
L.Et						3.466 *** (71.46)
Er	0.035 *** (8.64)	0.039 *** (7.28)	1.141 *** (6.68)	0.124 ** (2.55)	0.029 ** (2.30)	0.076 ** (2.92)	0.035 ** (3.04)
Control variable	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Control year	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Controlling individuals	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Phase I (F-value)			15.779
Over-identification (p-value)			0.794 (0.07)
Phase II (p-value)			12.907 *** (122.05)

** and *** denote significance at the 5% and 1% levels, respectively.

.

First, double machine learning (DML) addresses potential estimation bias from high-dimensional controls and nonlinearities. DML effectively handles multicollinearity and feature selection in high-dimensional settings, thereby improving causal identification [48]. The results show that the coefficients on regulation convergence for export probability and intensity are 0.039 and 0.035, respectively—both significant and consistent with baseline estimates—suggesting that the effects are not artifacts of model specification or high-dimensional confounders.

Second, to address potential endogeneity arising from unobserved factors, instrumental variable (IV) estimation is adopted. Following Du et al. [49], instrumental variables are constructed from regional coal consumption and regulation convergence indicators. IV estimates show that the regulation convergence coefficient is 1.141, significant at the 1% level, indicating that after addressing reverse causality and omitted variable bias, regulation convergence remains a significant driver of enterprise exports, enhancing causal inference.

Third, we further validate the results through alternative model specifications. A Tobit model addresses truncation in export intensity, while a high-dimensional fixed-effects model controls for individual heterogeneity and time trends. Both specifications yield significant positive results: the convergence coefficients for export probability and intensity are 0.124 and 0.029, respectively, and are significant at the 5% level. These results confirm that regulation convergence enhances exports and underscore the robustness of our findings.

4. Discussion

4.1. Mechanism Test

To identify the transmission mechanisms through which urban–rural environmental regulation convergence affects county-level wood processing enterprise exports, its effects via green credit and peer quality channels using Equations (3)–(5). Theoretically, regulation convergence should enhance export probability through two channels: green credit alleviating financing constraints, and peer quality generating knowledge spillovers.

Table 6. Results of the mechanism test.

	Green Credit Equation (3) Fitting Result	Cohort Mass Equation (4) Fitting Results
Three-step method
Indirect effect	0.001 * (1.355)	/
Direct effect	0.030 *** (2.73)	/
Sobel test
Indirect effect	/	0.006 *** (10.73)
Direct effect	/	0.031 *** (11.85)
Bootstrap test
Indirect effect	/	0.006 *** (9.81)
Direct effect	/	0.031 *** (10.80)

* and *** denote significance at the 10% and 1% levels, respectively.

validates these theoretical predictions. First, regulation convergence significantly enhances green credit availability (coefficient δ₂), reflecting environmental policies’ signaling effect in guiding financial resource reallocation. Furthermore, green credit’s partial effects on export probability and intensity are 0.001 (p < 0.10) and 0.030 (p < 0.01), respectively, indicating that improved financing channels help marginal firms overcome fixed export costs while strengthening incumbent exporters’ intensive margin expansion. This aligns with Zhou et al. (2023), who find that green finance alleviates liquidity constraints [50].

Second, peer quality’s mediating effect exhibits nonlinear spillover characteristics: although regulation convergence’s direct effect on peer quality (Equation (4)) is statistically insignificant, Sobel and bootstrap tests show that the indirect spillover effect is 0.006 (p < 0.01), representing 16.2% of the total effect. This suggests that regulation convergence drives continuous improvement in green technology, product quality, and management standards. Technology sharing, market experience exchange, and cooperative quality improvement among firms create a virtuous cycle, providing strong impetus for international market expansion and export growth.

4.2. Heterogeneity Test

Considering the potential heterogeneity in the impact of urban–rural environmental regulation convergence on exports of different types of county-level timber processing enterprises, the benchmark model structure was refitted, and subsample tests were conducted based on four dimensions: enterprise growth cycle, mechanization level, government intervention, and geographical location. The results are shown in

Table 7. Results of heterogeneity test.

Categorical Variable	Sample Classification	Impact Factor		Control Variable	Control Year	Controlling Individuals
		Export Probability	Export Intensity
Business Growth Cycle	Founding period	0.188 ** (2.71)	0.035 * (1.68)	Yes	Yes	Yes
	Growth period	0.067 (0.95)	0.007 (0.25)	Yes	Yes	Yes
	Maturity period	−0.170 (−1.15)	0.079 ** (2.10)	Yes	Yes	Yes
	Recession period	0.243 ** (2.47)	0.072 ** (2.07)	Yes	Yes	Yes
Level of government intervention	Low level of government intervention	0.092 ** (2.47)	0.039 ** (2.31)	Yes	Yes	Yes
	Higher level of government intervention	0.009 (0.26)	0.019 (1.20)	Yes	Yes	Yes
Geographic location	Eastern part	0.072 ** (1.98)	0.026 ** (2.01)	Yes	Yes	Yes
	Central Region	0.031 (0.64)	0.037 (1.38)	Yes	Yes	Yes
	Western Region	0.102 (1.44)	0.05 (0.99)	Yes	Yes	Yes
Economic environment	Yangtze River Economic Belt	−0.015 (−0.33)	0.033 (1.35)	Yes	Yes	Yes
	Non-Yangtze Economic Belt	0.080 ** (2.70)	0.031 ** (2.40)	Yes	Yes	Yes

* and ** denote significance at the 10% and 5% levels, respectively.

.

First, from the enterprise growth cycle perspective, regulation convergence significantly promotes exports of start-up and declining firms but not growth or mature firms. Coefficients are 0.092 for start-ups and 0.087 for declining firms, consistent with Hu and Chen [51]. Start-ups face dual constraints from sunk export costs and environmental compliance costs; regulation convergence reduces institutional frictions, alleviating resource constraints. For declining firms at critical transformation junctures, green credit support from regulation convergence funds technological upgrading and promotes exports. In contrast, growth and mature firms have achieved environmental compliance economies of scale, where marginal institutional benefits of regulation convergence diminish, confirming diminishing institutional returns.

Second, from the government intervention perspective, regulation convergence significantly promotes exports in low-intervention regions. Where fiscal expenditure/GDP is below average, regulation convergence coefficients are 0.092 for export probability and 0.039 for intensity, but effects are insignificant in high-intervention regions, consistent with Zhang and Li [52]. This supports the policy complementarity hypothesis: when governments reduce direct resource allocation, environmental regulations effectively guide firm export probability and export intensity through price signals; conversely, excessive intervention distorts regulatory incentives, creating regulatory capture and resource misallocation that weaken export promotion.

Geographic heterogeneity reveals development gradient constraints. Regulation convergence significantly enhances exports in the eastern region (β = 0.072) but not in central or western regions, consistent with Chen et al. [46]. Eastern enterprises leverage advanced infrastructure and integrated supply chains to reduce costs, enhance green technology adoption, and rapidly convert environmental compliance into product quality advantages. Conversely, central and western enterprises face technological transformation gaps, lagging economic development, and insufficient technological reserves, limiting their effective response to regulation convergence.

From an economic environment perspective, regulation convergence significantly promotes exports outside the Yangtze River Economic Belt (β = 0.080) but not within it, consistent with Wen and Lin [53]. Enterprises outside the Belt capitalize on policy support and market expansion opportunities to drive green technology transformation and enhance competitiveness. In contrast, enterprises within the Belt experience industrial transfer effects that create crowding-out effects through factor outflows and innovation capacity erosion, preventing them from capturing regulation convergence benefits.

5. Conclusions and Implications

5.1. Conclusions of the Study

This study is the first to construct a prefecture-level index of urban (municipal districts) and rural (non-municipal counties) environmental regulation convergence, drawing on an unbalanced panel dataset of 5600 timber-processing enterprises from 2007 to 2016. Using a Heckman two-stage model, the analysis examines the current status, evolutionary trends, and impacts of urban–rural environmental regulation convergence on the export probability and export intensity of county-level timber-processing enterprises. The findings indicate that China has exhibited a clear trend toward convergence—and even intensified competition in urban and rural environmental regulation over the study period. Moreover, urban and rural environmental regulation convergence significantly increases both the export probability and export intensity of county-level enterprises, and this promoting effect demonstrates a persistently divergent pattern. A series of robustness checks—including instrumental variable estimation, double machine learning, model substitution, and 1% trimming—further confirms the reliability of the baseline results. Mechanism analysis reveals that environmental regulation convergence promotes enterprise exports through two channels: green credit and peer quality. Heterogeneity analysis shows that convergence significantly stimulates exports only for start-up and declining enterprises, with no significant effects observed for growing or mature firms. Geographically, the promotion effect is evident for enterprises located in the eastern region but not for those in the central or western regions. Overall, this study enriches the theoretical understanding of how urban–rural environmental regulation convergence affects the export of county-level timber-processing enterprises and provides micro-level evidence on its mechanisms—particularly green credit, industry agglomeration, and peer quality—thereby offering policy-relevant insights for enhancing enterprise export competitiveness. Although this study focuses on China, its pronounced urban–rural dual structure provides a useful setting to identify the mechanisms of regulatory convergence. The analytical framework is not country-specific, but the applicability of the findings to other contexts should be interpreted with caution.

5.2. Policy Implications

First, environmental regulation in non-municipal districts should be strengthened and aligned with urban standards to accelerate regulatory convergence. As ecological-civilization policies deepen, this convergence has become a key driver of export growth for county-level wood-processing firms. Governments should set location-specific emission standards that mirror national objectives, tighten monitoring of non-municipal plants, and block the relocation of polluting facilities, thereby ensuring a level competitive field. Concurrently, China should establish an urban–rural regulatory convergence protocol that is compatible with international norms; this would raise firms’ environmental capabilities, increase foreign-market recognition, and allow county-level processors to move into higher value-chain segments. Such measures will ease the foreign-trade bottlenecks that currently suppress export scale and export propensity, delivering a “dual leap” in both margins.

Second, calibrate green-credit policies to the export-specific needs of wood-processing firms. Panel evidence shows that targeted green credit eases financing constraints, stimulates R&D spending, and raises export intensity. Regulators should therefore (i) build a credit-rating framework that weights sector-specific technological and environmental performance, (ii) dynamically reallocate credit limits as borrowers’ environmental and social-risk scores evolve, and (iii) channel funds to the most productive uses. Parallel expansion of green bonds, export-credit insurance, and specialized funds will broaden and deepen the green-financial system. A well-designed incentive-and-risk-sharing regime can generate a virtuous cycle that sustains export growth by county-level processors.

Third, raise average peer quality and create an environmental information-sharing platform. Governments should spatially reallocate capacity to foster clusters of high-productivity, low-emission wood-processing plants, thereby intensifying localized rivalry and collaboration. Peer effects within such clusters lift green-innovation efficiency and overall competitiveness. Concurrently, an information-sharing platform should disclose verified environmental metrics, integrate fragmented resources, and coordinate R&D among county-level processors. The platform will upgrade environmental management, align green production systems with international standards, and mitigate green-trade-barrier risk. By leveraging these network advantages, firms can strengthen brand equity and market reputation, facilitating their breakthrough into international markets and delivering a joint dividend of green growth and global engagement.