How Does the Spatial Structure of the Furniture Industry Shape Urban Residents’ Health? Evidence from China Labor-Force Dynamics Survey and POI Data

Abstract

In the context of advancing sustainable urban development, the spatial organization of industries plays a critical role in shaping environmental quality, economic vitality, and public health. This study examines the health effects of furniture enterprises agglomeration in Chinese cities, using a unique dataset combining point-of-interest (POI) big data and micro-level survey responses from 13,217 individuals. The results show that a one-unit increase in furniture enterprises agglomeration intensity is associated with a 0.656-unit improvement in physical health and a 0.060-unit improvement in mental health. These benefits are driven by three synergistic mechanisms: environmental improvement, income growth, and enhanced public health services. However, the health gains are unevenly distributed, with greater benefits observed in less-developed cities and among vulnerable groups such as low-skilled and middle-aged workers. We further reveal divergent effects between specialized and diversified agglomeration patterns, moderated by environmental regulation. Our findings underscore the need for health-oriented industrial policies that align with sustainable urban planning, emphasizing spatial adaptation, targeted support for vulnerable populations, and innovative regulatory approaches to foster both industrial growth and resident well-being.

1. Introduction

Sustainable urban and regional development is a new international agenda. Increasingly, attention is focused on how the spatial layout of industries, local climate conditions, and the distribution of health resources work together to shape the overall well-being of residents. The United Nations’ Sustainable Development Goal 11 emphasizes the need to build human settlements and cities that are inclusive, safe, resilient, and sustainable. The simultaneous implementation of the Healthy China 2030 program and the Made in China strategy shares a similar vision; they are in line with developing industrial capabilities and improving people’s health, as distinct goals. However, it is the quick industrialization and urbanization that, although generating impressive gains in economics, have created further pressure on nature and increased inequalities in resource distribution in health [1,2,3,4]. These interlinked issues will need a complex method of addressing these problems through a balance in terms of administrative arrangements of the industrial layout, the rehabilitation of the environment, and the fair access and provision of health services as an integral aspect of sustainable development.

The industrial agglomeration in space has also been one of the strongest drivers of urban and regional economic growth over the years, but its extended implications for the sustainability of a city need a closer look. The spatial location of manufacturing operations in industries that are closely related to daily life not only influences the performance of industries but also the environment and availability of health resources among the population [5,6,7,8,9]. The furniture industry is not a typical traditional manufacturing sector. It spans the entire value chain from raw material processing and creative design to end retail services, and it has a distinct “manufacturing plus services” character. As a key livelihood-oriented industry, its sales activities are closely connected to urban living spaces, and its production processes directly affect local environmental quality. These features make the furniture sector a valuable context for studying how industrial activity influences residents’ health and the broader urban living environment. Over the past two decades, China’s furniture industry has grown at an average annual rate of more than 15 percent, forming major industrial clusters in the Pearl River Delta, Yangtze River Delta, Bohai Rim, northeast China, and the central and western regions [10,11,12]. At the same time, the furniture industry, as a provider of essential goods, is widely distributed across cities and towns of all levels nationwide. This spatial pattern, combining broad distribution with regional concentration, not only promotes resource sharing, technology spillovers, and economies of scale, but also significantly affects local environmental quality, public service provision, and the health and well-being of community residents.

Though the literature has studied the furniture industry in terms of time, technology, and environment, including operational efficiency, industrial upgrading, the digital transformation, and the ecological impact, the spatial agglomeration of the furniture enterprises in terms of health has not been given much consideration. There is no clear understanding of the extent to which various agglomeration patterns, including specialization and diversification, impact the physical and mental well-being of residents, and how these health outcomes are distributed between cities and social groups. This study develops a three-step empirical framework using micro-level data from the 2018 China Labor-force Dynamics Survey (CLDS) (https://www.cnsda.org/index.php?r=projects/view&id=75023529 [accessed on 19 October 2025]) and city-industry-level point-of-interest data on furniture enterprises. The approach aims to elucidate the health impacts and underlying mechanisms of furniture industry agglomeration while informing spatial and industrial strategies that balance efficiency and equity in promoting sustainable urban development. A region fixed effects model is utilized to assess the impact of industrial clusters on the physical and mental health of the residents, and also various levels of different urban hierarchies and demographic groups are taken into account. The discussion also investigates the mediating effects of environmental quality, growth of income, and distribution of public health resources in determining these effects and assesses the combination of different agglomeration modes and governmental environmental control to create different health outcomes.

This project adds value to the body of literature in some significant aspects. It introduces the best use of billion-scale point-of-interest information, combined with nationally representative micro-level health information, into a nested city-individual database, entitling it to rigorous empirical study of the effect of industrial spatial arrangements on population health. It is based on this that it suggests an analytical framework encompassing this agglomeration of the furniture enterprises as associated with the physical and mental health of residents, indicating the underlying dynamics through which environmental state, economic satisfaction, and institutional control collectively interact in this relation. The study also offers an evidence-based basis on which place-sensitive, health-oriented industrial and spatial policies should be designed by differentiating between specialized and diversified agglomeration patterns to propel the interest of sustainable urban development.

2. Literature Review and Hypothesis

Enterprise agglomeration or industrial agglomeration is the term denoting the concentration of companies related to the industries in a single geographical location, which helps in sharing resources, reduces costs, and increases competitiveness and innovations [13]. Industrial agglomeration can spur economies of scale and, in imperfect competition, both financial resources and core competitiveness to develop urban and regional agglomeration through the technologies of shared fixed inputs, matching in the labor market, and learning [14,15]. Ideally, the first type of industrial agglomeration is specialized agglomeration, or what we commonly refer to as the Marshallian externalities [16]. Many empirical studies have proved that the geographic concentration of homogenous firms can contribute to a significant decrease in the costs of production and transactions [17] and the rise in the productivity of the region, which is provided through knowledge spillovers and the diffusion of technology [18]. In addition, the special agglomeration would foment full employment, augment the earnings of workers, and hike the disposable incomes of residents, which would allocate the structure of consumption and investment in the most optimal manner. This shift becomes the basis of urban economic development, which is no longer investment-driven but consumption-driven, and has a final result in the ability to develop socioeconomically sustainably [19].

On the basis of a substantial body of literature related to the macroeconomic advantages of specialization, recent research has begun focusing more on its micro-implications regarding the quality of life of residents. According to Zheng [20], people living in areas where industrial agglomeration is established obtain more purchasing power, which is a gain that allows them to afford to receive a better quality of healthcare and nutrition. Also, tax revenues of industries are increasing the fiscal capacity of the localities, motivating governments to invest in local health benefits, which promote both the welfare of residents and long-term economic growth [21]. Simultaneously, the environmental externalities of furniture industry agglomeration are a severe and rather intricate determinant of the health outcomes. As an example, heavy industrial agglomeration can add to the risk of exposure to pollution, but light furnishing industry agglomeration can reduce environmental health risks by means of extensive control of pollution and development of green technologies [22,23]. This highlights that the health effects of industrial agglomeration are closely tied to sector-specific production practices and the adoption of green technologies, emphasizing the importance of conducting independent analyses for each major industry subsector.

Being a symbolic light industry of labor-intensive character, the furniture enterprises present unique health transmission processes in its unique agglomeration. In terms of the attributes of industrial chains, the agglomeration of the furniture enterprises implies a broad range of services and a high value-added potential, and these features increase the efficiency of the intermediary inputs utilization and final goods delivery, thereby increasing the innovation capacity of the city [24]. Simultaneously, the industrial chain displays a multi-tiered structure led by manufacturing and processing, with cultural and design activities as complementary components. This arrangement creates numerous inclusive employment opportunities, broadening income growth benefits for vulnerable populations [25] while also generating culturally oriented, high-end positions, such as contemporary furniture design, that directly advantage high-skilled professionals [26]. Applying the production paradigm, such key polluters as painting and woodworking are extremely standardized, which means that pollution control is more practical on a centralized basis of more stringent environmental policies and intelligent manufacturing [27]. Moreover, it depends on the homogeneity of production processes, which makes it easier to standardize occupational safety, meaning that health threats related to work are reduced to considerable levels [28]. On the whole, these attributes constitute the health-enabling potential of furniture enterprises agglomeration, which is mainly attained through enhancing the ability to invest in health and systematically minimizing environmental-based health risks. The hypothesis of the given study is as follows:

H1. 

Agglomeration of the furniture enterprises facilitates the enhancement of physical and mental health of the residents, thus increasing the overall health conditions.

Using relevant literature and theoretical arguments, this paper introduces three main mechanisms by which the agglomeration of furniture enterprises can impact the physical and mental health of residents, such as the effect of the ecological environment, the effect on the growth of income of households, and the effect on the allocation of public health resources.

First, the effect on the ecological environment. The environmental exposure approach assumes that urban people live and work in a certain ecological environment, and environmental exposure, especially pollution, is one of the significant threats to the population in terms of its health [29]. The inflow of production factors and the expansion of output that accompany the clustering of furniture enterprises can, in the short term, accelerate resource consumption, increase congestion, and raise localized pollution emissions. These effects may further worsen the urban ecological environment [30]. However, in the long run, the furniture industry, as a light manufacturing sector with relatively standardized pollution processes and clear governance pathways, is more likely to generate scale effects, coordinated governance benefits, and green technology gains under agglomeration conditions [31]. According to the Porter Hypothesis, agglomeration in industry stimulates green technological innovation through the concentration of rivalry, whereas shared pollution-control facilities may considerably reduce the number of abatement units [32]. Here, the agglomeration of the furniture enterprises can theoretically increase technological innovation, lower the cost of transportation and information, and improve the economic and energy efficiency, which in the end will promote ecological restoration and the management of the environment in the urban environment [33,34]. In addition, industrial agglomeration can also contribute to the fact individuals become more aware of environmental issues and, thus, encourage people to start engaging in the fight against pollution and drive businesses to invest more in green technology [35]. Under an increasingly mature governance system and stricter environmental regulations, the agglomeration of the furniture industry is more likely to reduce per-unit industrial emissions, and thereby improve overall urban environmental quality [36].

Second, the income growth effect. The employment opportunities and wage rates tend to be better in industrial agglomeration places as compared to peripheral areas, and the presence of income disparities is one of the main contributors to the process of labor migration to agglomeration centers [37,38]. The systematic growth of furniture enterprise agglomeration, therefore, may not only increase the level of individual income but also increase the level of labor productivity in the region, which may provide adequate amounts of labor inputs to industrial and service industries. The individual demand as well as society to cater to health investments, including medical insurance and healthcare, is enhanced as the general income increases. Moreover, Grossman’s model of health capital [39] also indicates that the growth in income enhances the elasticity of demand in healthcare, especially for low-income groups, which promotes increased investment in health and reduces health vulnerability.

Third, the impact on resource allocation of public health. Furniture enterprise agglomeration has the ability to enhance the physical and mental health of residents by improving collective provision of public health services, both at a fiscal and market level. Industrial tax revenue on the fiscal side increases the ability of local governments to spend socially, such as to increase the number of hospital beds per capita and raise the percentage of medical insurance reimbursement [38]. On the market side, due to population concentration caused by enterprises agglomeration, a scale effect in the demand is formed because an increase in population density in agglomerated areas increases the demand for healthcare facilities and professionals [40,41]. In turn, the government investment and market-provided supply of resources can establish a synergistic relationship, which facilitates the desirable correlation between health service provision and the demand, which will provoke positive outcomes in terms of the well-being of residents. Nevertheless, because of the characteristic of a public good that is health resources, over-agglomeration can increase pressure on urban governance and crowd out public health expenditure [42]. According to the analysis of the case provided above, the following hypothesis can be developed by this study:

H2. 

Agglomeration in the furniture enterprises impacts not only the physical and mental health of residents based on the ecological environment, but also in terms of income growth and distribution of health resources to the population.

For a long time, regional coordinated development has been regarded as a strategic task and a key national project in China. This emphasis reflects the objective reality that significant disparities persist among regions in terms of their industrial evolution and patterns of economic agglomeration. Correspondingly, variations in the spatial organization of the furniture enterprises may lead to heterogeneous impacts on the health of urban residents. In terms of its conceptual foundations, the agglomeration of the furniture enterprises can be broadly categorized into two distinct forms. The initial industrial specialization agglomeration, also known as Marshallian externalities, goes back to the efforts of the British economist Alfred Marshall. According to this view, the grouping of like companies increases economic change in the urban areas in terms of sharing of intermediate production factors, pooling of labor, and inter-firm knowledge transfer [16]. Industrial diversification, agglomeration, the second one, and as famous as Jacobs’ externalities, is proposed by American scholar Jane Jacobs. Her argument was that, by virtue of co-location of different industries, knowledge sharing across sectors and technological advancement are enhanced, which has the effect of determining the flow of city development and enhancing the competitiveness of regions [43]. Cumulatively, these theoretical strands imply that the various modes of agglomeration and related externalities can have different effects on the physical and mental health of residents.

Industrial diversification agglomeration is a spatial concentration of heterogeneous enterprises in a particular area, where frequent information exchange, technological cooperation, and knowledge sharing among various industrial segments strengthen the efficiency of the industrial chain and various supporting structures [44]. Unlike specialization, where urban industrial systems are based on homogeneity, the concept of diversification influences urban industrial systems in an optimizing way, reducing path dependence, inter-industrial coordination, and stimulation of the effect of demonstration and imitation [45,46]. These processes not only increase productivity but also enhance the well-being of the residents due to the enhancement of the economic structure and quality of jobs. Moreover, diversified agglomeration is also likely to draw a professional and heterogeneous labor supply, enhance the complementary distribution of production factors, and have a culture of innovativeness in urban culture that supports long-term further growth.

However, according to the findings of interdisciplinary studies, the benefits of diversification can be nonlinear and reach a threshold beyond which their advantages are reduced [47]. In space and resource-sensitive situations, too much diversification may lead to unhealthy inter-industry competition, incongruence on production standards, and the rising cost of coordination [48]. In the furniture enterprises, the inequality of pollution levels, including the differences in wood-coating and metal-finishing processes, and the unequal distribution of the resource needs among upstream, midstream, and downstream companies can further erode collective environmental governance and increase the risk to health. These forces suggest that diversification should reach a critical level of economies of scope before being converted into actual health gains. It is based on this theoretical assumption that the following hypothesis can be formulated.

H3. 

The heterogeneous nature of the relationship between agglomeration and physical and mental health of the residents might be a result of different ways of furniture enterprises agglomeration and the related externalities, which will imply that the agglomeration-health association is both structural and intensive.

Massive industrialization and high rates of urbanization have brought about serious environmental issues such as air, water, and soil pollution, which have been extremely pronounced, threatening human health [49]. Due to the nature of environmental quality as a type of public good, the market forces do not suffice to reduce the pollution level, which is why the influence of the government is unavoidable to cope with collective action failures. In that regard, the key actors in environmental governance are the local governments, who should develop and implement regulatory tools that limit as well as encourage the environmental performance of firms [49].

Environmental regulation has, therefore, become a major factor that defines the process of pollution and the organization of areas where industry is located. On the one hand, environmental regulations, such as emission standards and punitive taxes, force high-polluting enterprises to either exit the market or upgrade their technologies [48]. The most striking instance is the nation-wide campaign of the volatile organic compounds (VOCs) in China that helped cut down the emissions produced by the furniture industry by 38 percent. Simultaneously, strict regulations tend to encourage the close physical location of companies in specialized eco-industrial parks and, in that way, promote and allow the treatment of pollution together and implement joint environmental policy through the proximity of uniform geographical locations [50].

The consequences of such dynamics are directly associated with the issue of health in the population. The synergistic impact between the enterprise agglomeration and environmental pollution determines the exposure of the population to health hazards, in particular, those involving air quality degradation. There is extensive empirical data that shows a rise in total suspended particulates leads to more mortality, the prevalence of respiratory illnesses, reduced life expectancy, as well as overall population health deterioration [51,52].

Combined with the other results, it can be concluded that environmental regulation can positively moderate the correlation between the agglomeration of furniture enterprises and the physical and mental well-being of residents. Regulation will also reduce the adverse externalities of agglomeration by either encouraging the use of cleaner technologies and the creation of environmentally friendly industrial clusters or maximizing the potential health benefits. However, on the basis of this theoretical explanation, the hypothesis is the following:

H4. 

There is a positive moderate effect of government environmental regulation on furniture enterprises agglomeration and the physical and mental health of residents.

3. Materials and Methods

3.1. Data

This study draws on two primary data sources: Gaode Map, China’s largest online mapping service platform, and the micro-level dataset from the China Labor-force Dynamics Survey (CLDS). Using the official open API of Gaode Map, we collected extensive point-of-interest (POI) data on the furniture enterprises across Chinese cities. The dataset includes enterprise names, geographic coordinates, and administrative locations. After filtering, deduplication, and coordinate correction, a total of 1,150,776 furniture-related POIs were identified as of November 2017. Based on industry value-chain classification criteria, the POI data were categorized into upstream, midstream, and downstream segments. Specifically, upstream enterprises include suppliers of raw materials such as wood, panels, paint, and hardware; midstream enterprises encompass furniture design and manufacturing firms; and downstream enterprises involve furniture sales, warehousing, logistics, and after-sales service providers. Following data preprocessing procedures, including coordinate calibration, a spatial database of city-level furniture enterprise distributions in China was constructed. Based on this, considering the cities covered by the CLDS microdata, the sample cities and the distribution of their furniture enterprises are shown in Figure 1.

Data on residents’ physical and mental health, along with other individual and household control variables, were obtained from the 2018 wave of the CLDS. The survey consists of three modules—personal, household, and community questionnaires. The personal and household modules collect information on demographics, income and consumption, employment and well-being, and occupational and wage characteristics. The community module covers aspects of local governance, land use, culture, and environmental development. The CLDS adopts a multi-stage, multi-level probability sampling strategy proportional to the labor force size, with 16,537 valid samples collected across 28 provinces (including municipalities and autonomous regions) and 249 prefecture-level cities in 2018. Following established research practices [53,54], we processed the CLDS data by (a) matching individual and household information using unique identification codes, (b) removing cases with missing values for key variables such as health, income, and education, and (c) winsorizing annual income at the 1% level to mitigate the influence of outliers. After preprocessing, the final analytical sample comprised 13,217 respondents. City-level control and moderating variables were drawn from the China Urban Statistical Yearbook, prefecture-level statistical reports, and the EPS database. To account for lagged macroeconomic effects, city-level data were matched to the CLDS microdata with a one-year temporal lag.

3.2. Variable Description

3.2.1. Identification of Industrial Agglomeration

Measurement approaches to industrial agglomeration can generally be categorized into three types: aggregate indicator-based methods, geographic distance-based methods, and spatial pattern-based methods. Because industry output data are primarily obtained from provincial or municipal statistical yearbooks and presented in aggregate form, it is difficult to disaggregate them across geographic or spatial dimensions. Consequently, most existing studies rely on methods based on aggregate indicators. Within this category, indicators such as the industry concentration ratio and the Herfindahl–Hirschman Index (HHI) can measure the degree of agglomeration but require detailed firm-level economic data—such as employment or output figures—that are often unavailable, thereby limiting their applicability in specific industries. In addition, the spatial Gini coefficient and the EG index method consider both the spatial distribution and economic characteristics of enterprises, capturing industrial agglomeration in terms of economic scale and spatial pattern to some extent. However, these methods face practical challenges due to data availability.

In contrast, the location quotient method requires only data on enterprise type and quantity, making it effective in measuring the relative geographic concentration of an industry in a specific region compared to the national level. This method aligns well with the strengths of high-precision POI data in identifying enterprise numbers and classifying industrial chains. By combining with POI data, it can accurately reflect the degree of enterprise agglomeration. Therefore, this study ultimately uses the location quotient method to identify the intensity of furniture enterprise clustering. The calculation formula is as follows:

$$L{Q}_{ij}={\displaystyle \frac{X_{ij}/{\displaystyle {\sum }_{i=1}^m{X}_{ij}}}{{\displaystyle {\sum }_{j=1}^n{X}_{ij}}/{\displaystyle \sum _{i=1}^m{\displaystyle \sum _{j=1}^n{X}_{ij}}}}}$$

(1)

The location quotient $L{Q}_{ij}$ represents the value for city $j$ and furniture enterprise type $i$, where $X_{ij}$ is the number of type $i$ furniture enterprises in city $j$.The parameters $n$ and $m$ denote the number of cities and enterprise types, respectively. Specifically, $m=1,2,3$, where m = 1 corresponds to upstream enterprises in the furniture industry chain, m = 2 to midstream enterprises, and m = 3 to downstream enterprises. ${\displaystyle {\sum }_{j=1}^n{X}_{ij}}$ and ${\displaystyle {\sum }_{i=1}^m{X}_{ij}}$ represent the total number of type $i$ furniture enterprises nationwide and the total number of all furniture enterprises in city $j$, respectively. ${\displaystyle {\sum }_{i=1}^m{\displaystyle {\sum }_{j=1}^n{X}_{ij}}}$ denotes the total number of furniture enterprises across the country. Furthermore, this study applies a weighted method to combine the agglomeration intensity of upstream, midstream, and downstream furniture enterprises, resulting in a comprehensive measure of furniture enterprise clustering intensity.

Secondly, the measurement of agglomeration structure adopts the Simpson diversity index, where a higher value indicates greater diversity in industrial agglomeration and a higher degree of mixing among upstream, midstream, and downstream enterprises. The specific formula is as follows:

$$D=1-{\displaystyle \sum _{i=1}^n{P}_i^2}$$

(2)

In the formula, D represents the agglomeration structure of furniture enterprises; P_i denotes the agglomeration intensity level of each type of furniture enterprise; and n is the total number of furniture enterprise types. The closer the value of D is to 1, the more balanced the distribution of agglomeration intensity across the three segments, indicating a higher level of industrial diversification. Conversely, the closer D is to 0, the more the industrial structure is specialized in a particular segment.

3.2.2. Identification of Residents’ Health Status

The main dependent variable in this study is residents’ physical and mental health. We measure these outcomes using self-rated health data from the CLDS survey. Although such measures are subjective and reflect individuals’ perceptions of their own health, they are widely used in health economics and epidemiology because they reliably predict objective health outcomes and offer strong feasibility and comparability in large-scale social surveys. Specifically, health is captured along two dimensions: physical health and psychological well-being. Physical health is evaluated through the questionnaire item, “How would you rate your current health status?” Respondents’ self-evaluations are coded on a 1 to 5 scale corresponding to “very unhealthy”, “unhealthy”, “average”, “healthy”, and “very healthy”, respectively. Mental health measurement follows the approach of Zhang (2021) [54], employing the Kessler Psychological Distress Scale (K-6) standardized to a 0–1 range. This scale inquires about respondents’ mental states over the past month across six aspects: feeling “depressed or so sad that nothing could cheer you up”, “nervous”, “restless or fidgety”, “hopeless about the future”, “everything was an effort”, and “worthless”. Response options include “all of the time”, “most of the time”, “some of the time”, and “none or almost none of the time”, assigned values from 0 to 4 in ascending order. Higher scores indicate better mental health.

3.2.3. Other Variables

Based on the theoretical relationships among the core variables and relevant literature, this study controls for a series of individual, household, and city-level characteristics in the empirical analysis. These include gender (female), age (age), marital status (marriage), local household registration (hukou), medical insurance coverage (health_ins), pension insurance coverage (pension_ins), smoking status (smoking), drinking behavior (drinking), household size (family_members), and per capita household income (household_inc), as well as city-level medical resources (med_resources) and educational resources (edu_resources). Individual characteristics may influence respondents’ labor decisions and living environments [55,56,57]. For instance, males and married individuals often assume greater family economic responsibilities, leading them to choose higher-income occupations with weaker health protection, thereby affecting their physical and mental health. Similarly, larger household size and lower per capita household income may reflect higher economic pressures, prompting individuals to prioritize economic gains over health concerns. At the city level, medical resources directly affect residents’ accessibility to healthcare, while educational resources enhance human capital, improve employment and income levels, and indirectly promote residents’ physical and mental well-being by optimizing their living and working conditions.

According to the theoretical analysis and the three proposed mechanisms, this study incorporates three mediating variables: industrial dust and smoke emissions per unit of output (M1), annual total income (M2), and accessibility of public health services (M3). Drawing on existing research and national policy documents such as the 14th Five-Year Plan for National Economic and Social Development and the 2035 Long-Term Vision Goals of the People’s Republic of China (https://www.gov.cn/xinwen/2021-03/13/content%5F5592681.htm [accessed on 21 October 2025]), the accessibility of public health services is identified from two aspects: “availability of medical care” and “availability of elderly care” [58]. Specifically, the composite index is derived by weighting and summing two indicators: the proportion of residents participating in local social medical insurance and the proportion covered by pension insurance. Furthermore, the moderating variable involved is the level of government environmental regulation (Env_regulation). Following Chen and Chen’s [59] (2018) methodology for constructing environmental regulation indicators, this study applies text analysis techniques on municipal government work report texts, including word segmentation and counting of environment-related terms, to calculate the proportion of such terms relative to the total word count in the reports. This proportion is used to characterize the level of government environmental regulation at the city level. Environment-related terms specifically include: environmental protection, eco-protection, pollution, energy consumption, emission reduction, pollutant discharge, ecology, green, low-carbon, air quality, chemical oxygen demand (COD), sulfur dioxide (SO₂), carbon dioxide (CO₂), particulate matter 10 (PM10), and particulate matter 2.5 (PM2.5), among others.

Table 1. Summary statistics.

Variables	Observations	Mean	Std. Dev.	Min	Max
Physical_health	13,217	3.5964	1.0178	1	5
Mental_health	13,217	0.8733	0.1688	0	1
Aggregation1	13,217	0.3789	0.0671	0.2612	0.6714
Aggregation2	13,217	0.0527	0.0100	0.0303	0.0833
M1	13,217	14.0334	31.5850	0.0397	202.8660
M2	9739	31,940.3333	35,105.0100	500	200,000
M3	13,217	0.2617	0.2386	0.0307	1
female	13,217	0.4723	0.4992	0	1
age	13,217	46.8272	14.3729	15	91
marriage	13,217	0.8174	0.3863	0	1
hukou	13,217	0.9433	0.2312	0	1
health_ins	13,217	0.8929	0.3092	0	1
pension_ins	13,217	0.5614	0.4962	0	1
smoking	13,217	0.2635	0.4406	0	1
drinking	13,217	0.1913	0.3934	0	1
family members	13,217	4.5265	2.0892	1	19
household_inc	13,217	1.6335	3.5412	−3.3333	150
med_resources	13,217	0.2643	0.1353	0.0493	0.7625
edu_resources	13,217	165.4727	115.2712	62.4757	736.3219
Env_ regulation	12,834	0.0065	0.0022	0.0016	0.0148

presents descriptive statistics for all major variables involved in this study.

3.3. Model Specification

The CLDS data used in this study is cross-sectional; therefore, a region fixed-effects model is employed to identify the causal relationship between furniture enterprises agglomeration and residents’ physical and mental health levels. The model is specified as follows:

$$Healt{h}_{ij}={\beta }_0+{\beta }_{1}Aggr{e}_j+{\beta }_{2}Control{X}_i+{\beta }_{3}Control{Y}_f+{\beta }_{4}Control{Z}_j+{\mu }_s+{\epsilon }_{ifj}$$

(3)

where ${\beta }_1$ is one of the key parameters of interest in this study; $i$, $f$, and $j$ represent individual residents, households, and cities, respectively. $He{a}_{ij}$ is the dependent variable representing residents’ health status, measured in two dimensions: physical health and mental health. $Ag{g}_j$ is the core explanatory variable, indicating the level of furniture enterprises agglomeration, which encompasses two dimensions: agglomeration intensity and agglomeration structure.$C{X}_i$, $C{Y}_f$, and $C{Z}_j$ are control variables at the individual, household, and city levels, respectively. ${\mu }_s$ denotes province fixed effects.

The baseline model is used to identify the overall average effect of furniture enterprise agglomeration on residents’ physical and mental health. However, this effect may vary across spatial and individual dimensions. At the city level, differences in urban scale, economic development, and industrial governance can lead to variations in the health effects of industrial agglomeration. At the individual level, the impact of industrial agglomeration and related economic activities may differ according to gender, age, household registration, and educational background, indicating significant group-level heterogeneity in health outcomes [60]. For example, residents of different age groups may adapt differently to furniture enterprise agglomeration, resulting in varying degrees of impact on their physical and mental health in response to industrial shocks. These asymmetric effects not only reveal disparities in the distribution of agglomeration benefits but also highlight potential health inequalities during industrial agglomeration and transformation. Understanding these patterns requires considering individual life cycle, human capital attributes, and industry technology characteristics. Therefore, based on the baseline regression, this study conducts grouped regressions across city attributes and individual characteristics to systematically examine the heterogeneous effects of furniture enterprise agglomeration on residents’ physical and mental health.

It should be noted that existing studies generally suggest that the incidence of major diseases, such as cancer, increases significantly after age 35, leading to a notable decline in labor productivity [61]. Accordingly, this study uses age 35 as a cutoff for heterogeneity analysis, classifying samples aged 35 or younger as the younger group and those older than 35 as the older group.

4. Results

4.1. Benchmark Regression Results

This study first investigates the overall average effect of furniture enterprises agglomeration intensity on residents’ physical and mental health.

Table 2. Baseline regression estimation results.

	Physical_Health			Mental_Health
	(1)	(2)	(3)	(4)	(5)	(6)
Agg1	0.687 ***	0.622 ***	0.656 ***	0.065 ***	0.062 ***	0.060 ***
	(0.132)	(0.128)	(0.130)	(0.022)	(0.022)	(0.022)
female		0.152 ***	0.152 ***		0.023 ***	0.023 ***
		(0.021)	(0.021)		(0.004)	(0.004)
age		−0.022 ***	−0.022 ***		−0.000 ***	−0.000 ***
		(0.001)	(0.001)		(0.000)	(0.000)
marriage		0.105 ***	0.105 ***		0.011 **	0.011 **
		(0.026)	(0.026)		(0.005)	(0.005)
hukou		−0.029	−0.033		0.002	0.002
		(0.033)	(0.033)		(0.006)	(0.006)
health_ins		−0.049 *	−0.049 *		0.004	0.004
		(0.029)	(0.029)		(0.005)	(0.005)
pension_ins		0.010	0.009		0.005 *	0.005 *
		(0.018)	(0.018)		(0.003)	(0.003)
smoking		−0.071 ***	−0.071 ***		−0.005	−0.005
		(0.025)	(0.025)		(0.004)	(0.004)
drinking		0.017	0.016		−0.003	−0.003
		(0.024)	(0.024)		(0.004)	(0.004)
family members		−0.030 ***	−0.030 ***		−0.002 ***	−0.002 ***
		(0.004)	(0.004)		(0.001)	(0.001)
household_inc		0.022 ***	0.022 ***		0.003 ***	0.003 ***
		(0.004)	(0.004)		(0.000)	(0.000)
med_resources		0.094	0.081		−0.010	−0.010
		(0.067)	(0.068)		(0.012)	(0.012)
edu_resources		0.000 ***	0.000 ***		0.000 ***	0.000 ***
		(0.000)	(0.000)		(0.000)	(0.000)
_cons	3.336 ***	4.348 ***	4.342 ***	0.849 ***	0.840 ***	0.840 ***
	(0.051)	(0.075)	(0.075)	(0.008)	(0.013)	(0.013)
Province FE	NO	NO	YES	NO	NO	YES
F Stats.	27.140	130.656	130.571	8.703	14.020	14.004
R2	0.002	0.107	0.108	0.001	0.014	0.014
Obs.	13,217	13,217	13,217	13,217	13,217	13,217

Note: *** p < 0.01, ** p < 0.05, * p < 0.1, standard errors in parentheses.

presents the estimation results of the baseline regression models under different explanatory variables and model specifications. Columns (1) and (4) report parsimonious models containing only the core explanatory variables; columns (2) and (5) introduce all individual-, household-, and city-level control variables but exclude regional fixed effects; and columns (3) and (6) further incorporate province fixed effects.

The results show that, after adequately controlling for potential confounding factors and eliminating unobserved regional heterogeneity, furniture enterprises agglomeration intensity exerts a significant positive effect on both residents’ physical and mental health. A one-unit increase in the intensity of furniture enterprises agglomeration corresponds to an estimated 0.656-unit improvement in residents’ physical health and a 0.060-unit improvement in mental health, both statistically significant. These results indicate that, on average, industrial agglomeration in the furniture sector not only stimulates economic growth but also generates positive social welfare spillovers, contributing meaningfully to residents’ physical and psychological well-being. The stronger association with physical health—nearly twice that observed for mental health—suggests that the material and environmental improvements accompanying agglomeration exert a more immediate and tangible influence on physical conditions than on psychological outcomes.

The robust and repeatedly significant total impact has important practical importance as it probably represents the sum of several socioeconomic processes that are triggered by the agglomeration of the furniture enterprises. Being one of the most labor-intensive industries, where manufacturing and services are interconnected, the concentration of furniture businesses offers a significant number of jobs locally, and, owing to the establishment of full industrial chains, provides differentiated and comparatively stable working positions. The above-mentioned developments directly boost household levels of income, improve the financial capacity of the residents to spend on health, and increase the capacity to limit health risks. Simultaneously, the high level of industrialization tends to bring about infrastructure, and the delivery of services to the population because more investment flows with industrial agglomeration. Concurrently, the localization of industrial systems leads to the development of specialized communities and more stable social economic networks. With shorter commuting distances and time stress, this agglomeration will ease the physical and mental strains of elevated travel expenses or work insecurity. It further fosters a sense of belonging among the residents and allows the amassing of social capital that leads to psychological well-being, as the social interaction increases and the community becomes more integrated [62,63]. In general, the positive effect of furniture enterprises agglomeration on the physical and mental health of the residents is a multidimensional phenomenon that is formed due to the combination of economic, infrastructural, and social factors. The following discussion further explores these mechanisms to explain the detailed channels in which the effect of industrial agglomeration impacts the outcome of public health.

4.2. Robustness and Endogeneity Tests

To ensure the reliability of the baseline regression results, this study conducts a series of robustness tests using multiple approaches. First, the explained variables are replaced. To avoid potential statistical bias from using questionnaire-based health scores, this study replaces the health indicator with a binary variable. Respondents who rate their health as “very healthy”, “healthy”, or “average” are assigned a value of 1, and all others are assigned 0. Although self-rated health indicators are widely used proxies for physical health in the existing literature, they are inherently subjective. To address this limitation, following [64], this study remeasures residents’ physical health using the Body Mass Index (BMI) as an objective indicator. In addition, the unstandardized K–6 scale scores are employed to re-identify residents’ mental health levels. Second, alternative estimation models are applied. Given that the self-rated health indicator is transformed into a binary variable, both ordered Logit and ordered Probit models are used for re-estimation. Third, multidimensional fixed effects are controlled for. Beyond incorporating a series of individual, household, and city-level control variables, the analysis further includes industry fixed effects to mitigate potential estimation bias arising from unobserved industry characteristics.

Table 3. Results of robustness and endogeneity tests.

	Replace the Dependent Variable			Replace Estimation Model		High-Dimensional Fixed Effects		IV1		IV2
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)
Agg1	0.121 **	0.165 *	1.081 ***	1.279 ***	0.771 ***	0.752 ***	0.094 ***	4.460 **	0.674 ***	6.533 ***	0.439 **
	(0.049)	(0.086)	(0.404)	(0.251)	(0.144)	(0.157)	(0.028)	(2.061)	(0.274)	(1.051)	(0.155)
_cons	1.071 ***	1.687 ***	21.123 ***	-	-	4.035 ***	0.783 ***	-	-	-	-
	(0.027)	(0.052)	(0.227)	-	-	(0.096)	(0.017)	-	-	-	-
Control	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES
Province FE	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES
Industry FE	NO	NO	NO	NO	NO	YES	YES	NO	NO	NO	NO
F Stats.	81.020	40.968	14.004	-	-	30.489	7.279	35.890	6.130	109.60	11.50
K-P F-stat	-	-	-	-	-	-	-	8.597	8.597	228.354	274.577
R2	0.068	0.038	0.014	-	-	0.075	0.023	-	-	-	-
Obs.	13,217	13,217	13,217	13,217	13,217	13,217	13,217	13,217	13,217	13,217	13,217

Note: *** p < 0.01, ** p < 0.05, * p < 0.1, standard errors in parentheses.

reports the estimation results of the above robustness tests. The results in columns (1)–(7) indicate that the robustness test findings are largely consistent with those of the baseline regression, thereby confirming the stability of the main conclusions.

Inequality in socioeconomic development across a region might itself be a contributor to differences in the agglomeration of the furniture enterprises, which will result in the possibility of endogeneity, and subsequently it may also provide misguided estimates of regression and lead to the policy being misinterpreted. In consideration of this issue, the current study uses terrain undulation as a measuring rod that is used in determining the spatial concentration of the furniture enterprises. Terrain undulation, which is the objective geographic measure, exogenously determined is what is defined by the difference in the maximum elevation and the lowest elevation in an area, the percentage of flatland, and overall land area. As the major factor that predetermines the spatial distribution of the population and industrial processes in China [65], terrain undulation influences the distribution of factors of production, including labor and capital, and therefore meets the relevance criterion. Simultaneously, it does not depend on any time-varying economic factors and individual health features, and thus the employment of terrain undulation as a furniture enterprises agglomeration instrument is theoretically reasonable and empirically credible.

In addition, following existing studies [66], this paper introduces a historical variable indicating whether the city had a railway in 1933 as an additional instrumental variable. At that time, given the poor road infrastructure in China, the opening of a railway had a significant impact on industrial agglomeration in cities, making this historical variable relevant to the current clustering of furniture firms. Moreover, since this instrumental variable predates the study period by approximately 70 years, it does not directly affect current residents’ health, thus satisfying the exogeneity requirement. The results of the endogeneity tests are presented in Table 3, with columns (8) to (11) reporting the estimates obtained using the instrumental variable under the two-stage least squares method. These findings indicate that, the higher the level of the furniture enterprises agglomeration, the more positive change it has on the physical and mental health of the residents, although the influence on the latter has been more evident. These results support the strength and internal validity of the results of the baseline regression.

4.3. Heterogeneous Analysis

The health exposure of agglomeration of the furniture enterprises’ exhibits a significant degree of heterogeneity regarding urban hierarchies. Residents of non-urban agglomeration land and small cities are the individuals who get the greatest health returns, as physical health coefficients stand at 1.008 and 0.853, respectively (see

Table 4. Results of urban heterogeneous analysis.

	Panel A: Physical_Health
	Urban Agglomeration		Big City		High Aggregation Level
	(1) Yes	(2) No	(3) Yes	(4) No	(5) Yes	(6) No
Agg1	0.419 **	1.008 ***	0.371 *	0.853 ***	−0.329	0.954 **
	(0.186)	(0.190)	(0.207)	(0.183)	(0.223)	(0.420)
Control	YES	YES	YES	YES	YES	YES
Province FE	YES	YES	YES	YES	YES	YES
Obs.	6278	6939	4801	8416	1564	11,653
	Panel B: Mental_Health
	Urban Agglomeration		Big City		High Aggregation Level
	(7) Yes	(8) No	(9) Yes	(10) No	(11) Yes	(12) No
Agg1	0.139 ***	−0.014	0.100 ***	0.016	−0.100 **	0.144 *
	(0.031)	(0.034)	(0.034)	(0.032)	(0.040)	(0.074)
Control	YES	YES	YES	YES	YES	YES
Province FE	YES	YES	YES	YES	YES	YES
Obs.	10,189	3028	3812	9405	1564	11,653

Note: *** p < 0.01, ** p < 0.05, * p < 0.1, standard errors in parentheses (same below).

). The magnitude of the effects, however, is significantly smaller in urban agglomerations and large cities, which implies that in underdeveloped areas with an unstable industrialization base and scarce public services, the furniture enterprises agglomeration would break the initial equilibrium at low levels through a variety of channels. These are the economies of scale which cause a decrease in the costs of pollution control per unit, increased employment which has a direct positive effect on the household income, and fiscal capacity which makes it possible to spend more on healthcare and infrastructure. In less developed localities, by contrast, the marginal health benefits of additional agglomeration are limited by mature public service institutions and rising factor costs. Additionally, the agglomeration active pressure may contribute to increased urban ailments, and there is a substantial threshold effect on the agglomeration-related enhancement of health (agglomeration-specific heterogeneity). Physical and mental health in the city is significantly enhanced in the cities with low furniture enterprises agglomeration, with the estimated values of 0.954 and 0.144, respectively. Nonetheless, in cities that are highly agglomerated, this is reversed, and hence moderately agglomerated cities are best placed in the creation of health dividends. All-in-all, these results reinforce the value of regionally adaptive governance. In the case of the underdeveloped non-urban and small-city areas, the development of industrial land incentives, investment in infrastructure, and environmental management can be the main focus of the policy work to create synergistic enhancement of the economy and population health. On the other hand, in densely urbanized and agglomerated cities, it is necessary to establish a suitable minimum threshold of density regulation and encourage a spatial relocation of businesses in order to attain inclusive and sustainable health results due to spatial reorganization.

In terms of heterogeneity analysis by individual characteristics, columns (1), (2), (7), and (8) of

Table 5. Results of individual heterogeneous analysis.

	Panel A: Physical_Health
	Age		Skill Level		Industrial Sector
	(1) Old	(2) Young	(3) High	(4) Low	(5) Yes	(6) No
Agg1	0.762 ***	0.351	0.462 **	0.702 ***	0.876 **	0.622 ***
	(0.155)	(0.229)	(0.224)	(0.158)	(0.349)	(0.141)
Control	YES	YES	YES	YES	YES	YES
Province FE	YES	YES	YES	YES	YES	YES
Obs.	10,189	3028	3812	9405	1564	11,653
	Panel B: Mental_Health
	Age		Skill Level		Industrial Sector
	(7) Old	(8) Young	(9) High	(10) Low	(11) Yes	(12) No
Agg1	0.089 ***	−0.036	0.021	0.076 ***	0.018	0.066 ***
	(0.026)	(0.043)	(0.039)	(0.027)	(0.063)	(0.024)
Control	YES	YES	YES	YES	YES	YES
Province FE	YES	YES	YES	YES	YES	YES
Obs.	10,189	3028	3812	9405	1564	11,653

Note: *** p < 0.01, ** p < 0.05, standard errors in parentheses.

show that furniture firm agglomeration has no significant effect on the health of younger residents but significantly improves the physical and mental health of individuals aged 35 and above. This division probably indicates the two-fold susceptibility of physical deterioration and skills outmoding in the latter group. Physiological degradation decreases intensive or hazardous employment endurance, giving the older employees an incentive to change the supply of labor or improve their health through agglomeration, which reduces health decline. At the same time, slower renewal in skills as compared to an industrial upgrading may, in addition to causing indirect structural unemployment, prompt behavioral adaptations enhancing health outcomes. In comparison, younger people, with greater physical resilience and more income-motivated, receive shorter health benefits, and thus the latent health risks to younger workers should be taken into greater consideration through industrial policy.

The second dimension is associated with educational heterogeneity. Table 5 columns 3, 4, 9, and 10 demonstrate that agglomeration associated with the furniture enterprises is highly supportive to low-skilled workers (with high school education or less) in terms of physical and mental health, but not for the high-skilled groups (with associate degree or higher). This is an inversion of differentiation due to the inclusive empowerment of low skill labor as well as the implicit restrictions of skill ceiling. Being a labor-intensive industry involving manufacturing, assembling, and logistics, the furniture enterprises have been providing many jobs at the entry level, which has enhanced the wage bargaining power and economic stability of low-skilled laborers, reducing the feeling of financial insecurity and the psychological pressures related to it. Conversely, less demand for highly skilled jobs in conventional manufacturing will result in under-exploitation of human capital and reduced health benefits at an environmentally better performance. Therefore, industrial agglomeration is a health ladder to low skilled workers and ineffective to provide equal benefits to groups of high skilled ones, a structural contradiction between basic labor positions and lack of motivation for higher talents to work in manufacturing transformation.

The third dimension is with regard to occupational differentiation. The agglomeration of furniture enterprises has a statistically significant impact on the health of non-industrial workers, but not on the health of industrial workers, as shown in column (5), column (6), column (11), and column (12). This break reflects the faster pace of development of service-oriented segments against the slower pace of the transformation of production-oriented segments. On the one hand, the growth of the production consumption ecosystem of agglomerated areas has increased the number and an even greater number of jobs in service sectors, including retail, logistics, and installation, that have healthcare benefits because of cleaner working conditions and more flexibility in schedules. Conversely, in industrial industries, industry rhythms will increase production, job insecurity and imposition of environmental compliance costs on a company; as an example, self-funded pollution control can negate the health benefits of automation and cleaner technologies. This has led to the service sector capturing a proportional number of the agglomeration produced health dividend, indicating the beginning of inter-sectoral imbalance redistribution of welfare.

4.4. Effectiveness Channel Analysis

How does furniture enterprises agglomeration influence residents’ health? In addition to environmental improvement, income growth, and public health resource allocation, three transmission pathways with different operating mechanisms, respectively. Columns (1)–(3) in

Table 6. Effectiveness channel verification.

	(1) M1	(2) Phy_Health	(3) Men_Health	(4) M2	(5) Phy_Health	(6) Men_Health	(7) M3	(8) Phy_Health	(9) Men_Health
Agg1	−5.628 ***	0.481 ***	0.024	0.354 **	0.721 ***	0.091 ***	0.043 **	0.502 ***	0.039 *
	(0.171)	(0.136)	(0.024)	(0.174)	(0.148)	(0.025)	(0.019)	(0.132)	(0.023)
LnM1		−0.031 ***	−0.006 ***
		(0.008)	(0.001)
LnM2					0.147 ***	0.018 ***
					(0.009)	(0.002)
M3								0.680 ***	0.075 ***
								(0.107)	(0.018)
BS_1		0.169 ***	0.040 ***		0.045 *	0.006 *		0.038 **	0.004 **
		[0.512]	[0.009]		[0.025]	[0.003]		[0.015]	[0.002]
BS_2		0.453 ***	0.022		0.631 ***	0.086 ***		0.450 ***	0.036 *
		[0.131]	[0.024]		[0.144]	[0.024]		[0.124]	[0.021]
_cons	4.313 ***	4.475 ***	0.868 ***	11.239 ***	2.622 ***	0.613 ***	0.029 ***	4.412 ***	0.850 ***
	(0.091)	(0.082)	(0.014)	(0.118)	(0.134)	(0.024)	(0.011)	(0.073)	(0.012)
Control	YES	YES	YES	YES	YES	YES	YES	YES	YES
Province FE	YES	YES	YES	YES	YES	YES	YES	YES	YES
F Stats.	655.647	122.519	14.699	254.044	99.199	19.596	3562.941	125.522	12.450
R2	0.354	0.109	0.016	0.257	0.108	0.025	0.867	0.113	0.015
Obs.	13,028	13,028	13,028	8854	8854	8854	12,051	12,051	12,051

Note: *** p < 0.01, ** p < 0.05, * p < 0.1, standard errors in parentheses. BS_1 denotes indirect effect, BS_2 denotes direct effect; values within square brackets represent Bootstrap standard errors.

report the estimation results of environmental effect. Estimation results show that furniture enterprises agglomeration can reduce smoke and dust emissions from industry, and improve both physical and mental health. In industrial clusters, the pollution-control facilities are used in common, which not only drives technological upgrading in environmental management, but also directly reduces health risks at the source. Meanwhile, the spatial concentration of enterprises makes pollution more visible and governmental supervision more centralized and efficient. Mechanically, this process could be regarded as a micro-level manifestation of the Environmental Kuznets Curve: once agglomeration overcomes the stage of extensive expansion, the economies of scale would transfer ecological constraints and development pressure into impetus for health improvement. Columns (4)–(6) in Table 6 report the estimation results of income growth effect. Estimation results show that furniture enterprises agglomeration would significantly increase residents’ annual income, which in turn promotes the accumulation of health capital and validates the “economic empowerment–health investment” pathway. Higher wage levels and stronger employment stability in agglomerated areas make households’ financial situation more resilient, and they would be able to spend more on preventive healthcare, diet, and insurance, which would help the poor escape the “health poverty trap”. Columns (7)–(9) in Table 6 report the estimation results of public health resource allocation effect. Estimation results show that furniture enterprises agglomeration would stimulate local governments to increase investment in medical facilities and other public health services. The improvement in access to these resources would mitigate the “urban disease” problems brought about by population inflows into agglomeration areas, and transform spatial concentration into momentum for high-quality socioeconomic development and health.

In addition, this study employs the Bootstrap method to conduct robustness tests on the mediating effects of the three mechanism variables in the relationship between furniture firm agglomeration and residents’ physical and mental health, as well as to analyze the direct and indirect effects. The results show that, first, the significance of the mediating effects obtained from the Bootstrap test aligns with the mediating effect model, supporting the reliability of the mechanism analysis. Second, in terms of the magnitude of direct and indirect effects, all three mechanisms generally exhibit larger direct effects than indirect effects, indicating that furniture firm agglomeration influences residents’ health through multiple pathways.

In total, the health benefits of furniture enterprises agglomeration would come from the synergistic interplay of environmental regulation, economic empowerment, and institutional resource allocation. These three transmission pathways could be regarded as forming an integrated “environment–economy–institution” system of health improvement: environment effect would reduce exposure risks, income growth would strengthen health capital, and optimized resource allocation would improve access to services. The integrated effect would then contribute to a cumulative improvement in residents’ health capital in industrial clusters. Our findings break through the traditional linear assumption that “agglomeration definitely causes pollution” or “economic growth damages health”. Our empirical results provide strong support for the coordinated design of industrial and public health governance policies.

4.5. Further Analysis: Agglomeration Structure and Policy Adjustment Effects

Table 7. Analysis of agglomeration structure effect.

	(1) Phy_Health	(2) Men_Health	(3) Phy_Health	(4) Men_Health
Agg2	−3.472 ***	−0.257 *	−30.429 ***	−4.532 ***
	(0.859)	(0.150)	(6.997)	(1.254)
Agg2^2			253.733 ***	40.243 ***
			(65.567)	(11.789)
_cons	4.781 ***	0.877 ***	5.482 ***	0.988 ***
	(0.070)	(0.012)	(0.192)	(0.035)
Control	YES	YES	YES	YES
Province FE	YES	YES	YES	YES
F Stats.	130.303	13.798	121.759	13.385
R2	0.107	0.014	0.108	0.015
Obs.	13,217	13,217	13,217	13,217

Note: *** p < 0.01, * p < 0.1, standard errors in parentheses.

shows the physical and mental health effects of different agglomeration types in the furniture enterprises. Columns (1) and (2) represent the agglomeration structure index, reflecting the degree of industrial diversification; columns (3) and (4) represent the agglomeration intensity. The agglomeration structure has a negative impact on physical and mental health. The marginal effect of physical health is −3.472, and the marginal effect of mental health is −0.257. Compared with agglomeration intensity, the effect of the agglomeration structure on physical and mental health is not ideal. According to the theory of related variety, the externalities generated by industrial diversification depend not only on the types of firms but also on the relatedness of their knowledge bases [67,68]. In other words, industrial diversification within a cluster may result in “unrelated variety” if different types of firms lack technological or knowledge connections. For instance, in the furniture industry, differences in production standards across product categories can increase the cost of coordinated pollution control. At the same time, the need to adapt skills across categories adds cognitive load for workers and raises the complexity and cost of local environmental management, which can negatively affect residents’ physical and mental health.

Columns (3) and (4) of Table 7 extend the model to include a quadratic term to examine whether there exists any nonlinear relationship between the agglomeration structure of the furniture enterprises and residents’ health. It can be seen that the coefficient of the linear term is significantly negative while the coefficient of the quadratic term is significantly positive, revealing a U-shaped relationship. The turning points of physical health and mental health are 0.042 and 0.056 when the agglomeration structure index equals to, respectively. We assume that this turning point represents a qualitative shift in the local industrial ecosystem from unrelated to related variety. Beyond this point, furniture manufacturing may act as the core around which closely connected supporting industries, such as eco-friendly paint R&D, smart home design, and flexible logistics, cluster. This symbiotic industrial environment, built on a shared knowledge base, can significantly facilitate cross-industry knowledge recombination and green technology integration. Related industries are also more likely to reach consensus on unified environmental standards and jointly invest in shared pollution control facilities. At this stage, the “Jacobs externalities” generated by diversification, particularly cross-industry knowledge spillovers and collaborative innovation, begin to outweigh coordination costs, turning industrial diversity into coordinated advantages for environmental management and sustainable healthy productivity. This U-shaped pattern offers clear guidance for furniture industry planning. In regions with low diversification, priority should be given to developing specialized industrial parks to reduce coordination costs, while in highly diversified regions, efforts should focus on accelerating technological integration and supply chain coordination to transform diversity into sustainable healthy productivity.

Moreover, the different inflection points for physical and mental health underscore the asymmetric physiological and psychological responses to agglomeration. Physical health benefits emerge earlier through technological and infrastructural synergies, while improvements in mental health require a more advanced stage of ecological and industrial integration. Notably, the identified thresholds (0.042 and 0.056) closely approximate the sample mean (0.0527), suggesting that many Chinese cities currently stand near a critical transition point in the evolution of furniture enterprises agglomeration.

The results in

Table 8. Analysis of policy adjustment effect.

	(1) Phy_Health	(2) Men_Health	(3) Phy_Health	(4) Men_Health
Agg1	0.877 ***	0.060 ***
	(0.136)	(0.023)
Agg2			−5.035 ***	−0.281 *
			(0.892)	(0.155)
Env_regulation	−23.623 ***	−0.290	−23.129 ***	−0.259
	(4.214)	(0.714)	(4.208)	(0.719)
Interact1	−99.244	21.561 *
	(68.462)	(11.805)
Interact2			282.160	−146.530 **
			(376.222)	(66.316)
_cons	4.578 ***	0.865 ***	4.592 ***	0.865 ***
	(0.055)	(0.009)	(0.055)	(0.009)
Control	YES	YES	YES	YES
Province FE	YES	YES	YES	YES
F Stats.	116.324	12.096	116.094	11.945
R2	0.113	0.015	0.112	0.015
Obs.	13,217	13,217	13,217	13,217

Note: *** p < 0.01, ** p < 0.05, * p < 0.1, standard errors in parentheses.

reveal the heterogeneous moderating role of environmental regulation in the relationship between industrial agglomeration and residents’ health. Specifically, its effect on mental health varies: environmental regulation strengthens the positive impact of specialized agglomeration on mental health but intensifies the negative impact of low-level diversified agglomeration. This seemingly contradictory finding highlights that the effectiveness of environmental regulation as an institutional signal strongly depends on the industrial structure. In specialized clusters, high internal homogeneity allows strict and consistent environmental regulations to convey clear and credible signals of green industrial transformation and community environmental improvement to residents. This, in turn, enhances local identification and development confidence, producing positive effects on psychological well-being. Conversely, in regions with low diversification and looser industrial structures, unrelated variety results in substantial differences in technological standards and diverse pollution types across firms. In this context, uniform environmental regulations are perceived primarily as increasing coordination costs and compliance uncertainty. These regulatory pressures translate into higher operational costs and governance challenges for firms, which then create economic risks and governance instability for local residents, ultimately triggering anxiety and increasing mental health burdens.

Taken together, the interaction between furniture enterprises agglomeration and environmental regulation highlights the need for differentiated policy frameworks that reconcile industrial development with public health and well-being. For specialized industrial clusters, environmental regulation should prioritize its psychological empowerment function by converting regulatory compliance costs into identity-based benefits through visible environmental achievements and green branding initiatives. In contrast, for diversified clusters, policy efforts should focus on differentiated regulatory design—such as establishing emission coefficients by product category—and on promoting supply chain decoupling and structural reconfiguration to mitigate systemic complexity at its source. This “targeted regulation–structural optimization” paradigm conceptually situates industrial specialization as the micro-foundation of health welfare and positions environmental regulation as an adaptive institutional mechanism. Together, these elements form a complementary governance architecture that not only prevents the physical and psychological costs arising from excessive industrial diversification but also facilitates the sustainable alignment of industrial transformation with residents’ overall well-being.

5. Discussion

5.1. Key Findings

This study uses 13,217 micro-level samples from the 2018 CLDS dataset along with POI big data on furniture enterprises to examine how the spatial clustering of the furniture industry affects the physical and mental health of urban residents. For the first time, it reveals the “double-edged sword” effect of furniture enterprise clustering on residents’ health by looking at both specialization and diversification clustering. The discoveries offer a health-focused assessment model towards the spatial planning of the furniture enterprises and offer empirical data on overcoming the disparities within the health welfare because of the clustering and development of industries, and contribute to overcoming the paradox that exists between the growth of the economy and health inequality, and promote the combined achievement of the Healthy China 2030 and the transformation of manufacturing. The general conclusions are as follows.

To start with, the twin impacts on the health of residents posed by the agglomeration of the furniture enterprises indicate a dialectical connection between the benefits of specialized clustering and the dispersion impacts of diversified clustering. Particularly, physical and mental health are both considerably improved in the furniture enterprises agglomeration, and the coefficients of 0.656 and 0.060 are significant. This finding strongly supports the theoretical expectations of Marshallian externalities. The geographic concentration of similar firms can enhance economic benefits through shared labor markets, specialized supply chains, and knowledge spillovers, while also generating health gains via three channels: environmental regulation, economic empowerment, and resource allocation. In comparison, health outcomes are significantly suppressed in diversified furniture enterprises agglomeration with coefficients of −3.472 and −0.257, which indicates the inability to control health risks over a complicated industrial ecology, in which the structural redundancy will be converted into health burdens. Moreover, the U-shaped results of the diversified clustering and health relationship show that diversification through clustering can only yield sustainable health benefits when it reaches a certain level, showing that it is the resource intensity and coordination that are critical to the impacts of industrial spatial organization on health. This finding complements Jacobs externalities by showing that the health benefits of diversification are not obtained linearly. The further revealed U-shaped relationship indicates that the positive effects only emerge when diversification surpasses a critical threshold and fosters an effective industrial ecosystem and innovation network [69,70]. Overall, the contrasting health effects across different furniture enterprise agglomeration patterns highlight that the health outcomes of industrial spatial organization depend on both the intensity of resource allocation and the degree of coordination.

Heterogeneity analysis shows two imbalances when it comes to the distribution of health benefits in space and micro level. On the urban hierarchy, the findings show a marginal diminishing returns effect in that in non-urban clusters and small urban areas with weak industrial foundation and lack of full access to urban services, furniture enterprises agglomeration visibly destabilizes the low-level equilibrium by scale economies, and the inhabitants of the areas receive the highest health benefits. On the other hand, the benefits are lowered in the metropolitan and large cities where the saturation of the public services and overcapacity of the environment is observed. City health benefits can go against expectations with high levels of concentration and the implication of this is that the density of industry increases to an unsustainable level that undermines the health of citizens. Health gains are more evident in socially vulnerable groups at the individual level. The gain is significant for the residents who are above the age of 35 because they are physiologically vulnerable, and there is no increase in the young residents. Low-skilled workers achieve health improvements via the relative abundance of non-skill-intensive jobs, while high-skilled workers experience no significant effect. Additionally, non-industrial employees enjoy more substantial health benefits, indicating that a portion of the health gains generated by furniture enterprises agglomeration is captured by the service sector, reflecting inter-industry inequities in welfare distribution.

Government environmental regulation shows distinct moderating effects across the dimensions of furniture enterprises agglomeration. Environmental regulation positively catalyzes the mental health benefits of specialized clustering, likely due to improvements in environmental quality and the sense of “green brand” identity, which enhance residents’ psychological well-being and social capital [71]. However, it exacerbates the health harms of low-level diversified clustering. This suggests that as a public policy, environmental regulation involves a tension between its “social-psychological effects” and “economic cost effects”, with the ultimate impact depending on the fit between regulatory policies and the local industrial structure. These results indicate that policy design should follow a differentiated strategy: strengthening the psychological empowerment role of environmental regulation in specialized clusters while reducing its adverse effects and systemic complexity in diversified clusters. Overall, the health impacts of furniture enterprises agglomeration depend on the interaction between intensity and coordination, and between specialization and diversification. Specialized clustering generates widespread benefits through three reinforcing mechanisms, whereas diversified clustering must surpass innovation thresholds to yield health improvements. Therefore, public health–oriented industrial policies should balance efficiency and equity to ensure that industrial growth translates into inclusive health welfare.

5.2. Study Limitations

This study has several limitations. The health assessment data are based on respondents’ self-reported evaluations, including the five-point physical health scale and the K-6 psychological distress scale from the CLDS, which may introduce social desirability and recall biases. Individuals often avoid extremely negative ratings such as “very unhealthy”, and retrospective evaluations of chronic conditions may be distorted by memory errors. These systematic measurement biases may weaken the observed association between industrial agglomeration and health outcomes. Nonetheless, the validated survey instruments used have shown strong reliability in previous studies, and the robustness tests in this research incorporate more objective health indicators that produce consistent results. Second, the cross-sectional data structure constrains the ability to fully resolve endogeneity. Although terrain ruggedness is used as an instrumental variable to mitigate selection bias and the K-P F-statistics indicate the absence of weak instruments, unobserved time-varying confounders may still affect estimate consistency. Cross-sectional data cannot capture the long-term dynamic effects of industrial agglomeration policies or environmental regulations. Future research should use panel data to establish a more robust causal chain linking industrial agglomeration and residents’ health. Third, although the model controls for a range of variables, omitted variable bias may still be present. Some macro-level factors influencing both furniture industry agglomeration and residents’ health may not be fully accounted for.

5.3. Policy Implications

Based on the above findings, this study proposes several policy implications. First of all, for highly specialized furniture enterprises’ clusters such as Shunde in Guangdong, a comprehensive “health enhancement” initiative should be implemented. This initiative should mandate the construction of shared pollution control facilities, incorporate occupational health standards into tax policy frameworks, and deploy real-time monitoring systems for high-emission processes such as painting and coating. At the same time, a threshold-based management mechanism should be established to guide cities seeking to develop diversified furniture enterprises’ clusters. Cities with limited diversification should adopt stringent entry qualifications and approval procedures for new product categories, whereas those with a well-developed industrial base should receive targeted financial and institutional support—through industrial funds or similar instruments—to strengthen the entire furniture value chain. This differentiated approach can enhance resource efficiency, mitigate early-stage health risks, and facilitate the formation of regional growth poles within the sector. Secondly, a targeted protection framework should be created for vulnerable workers through the establishment of specialized occupational health accounts. High-risk groups, including painters and polishers, should have dedicated health savings funds that cover occupational health examinations, preventive checkups, and rehabilitation services, thereby reinforcing long-term health security in high-exposure occupations. Thirdly, it is essential to align environmental regulation tools with the specific patterns of industrial agglomeration. Since the same regulatory intensity can produce opposite social-psychological effects depending on the industrial structure, policies must move beyond a “one-size-fits-all” approach. For specialized clusters, the focus should be on using the signaling value of regulation, such as establishing green furniture demonstration zones, certifying compliance, and publishing environmental performance rankings, to transform compliant behavior into market reputation and public recognition. For clusters at a low level of diversification, the priority should be reducing coordination costs by providing technical support, building shared pollution-control facilities, and enhancing policy communication to stabilize expectations and prevent mass anxiety caused by sudden closures or production restrictions. This requires stronger collaborative governance between environmental authorities and industrial planning departments.

In summary, this study recommends a triadic policy framework of “spatial adaptation, targeted protection, and regulatory innovation.” Specialized furniture enterprises agglomeration should strengthen health-related infrastructure, while diversified regions should emphasize innovation incentives. Industrial workers’ health rights should be prioritized, particularly for low-skilled groups, and regulatory systems should integrate health certification with flexible standards. By deeply integrating industrial policy and health governance, furniture clusters can transform from traditional high-pollution manufacturing bases into engines of healthy productivity, contributing to the coordinated advancement of the “Healthy China 2030” initiative and the national strategy for manufacturing transformation.