Planning-Induced Land Development Opportunities and Rural Household Income Disparities: Evidence from Wuhan’s Urban Development and Wetland Conservation Zones

Abstract

While land development opportunities stemming from planning regulations demonstrably influence rural household income, quantitative evidence quantifying these effects remains limited. Measuring and decomposing these effects can empirically support territorial spatial planning policies aimed at alleviating associated regional development imbalances and advancing sustainable rural development. This study selects Wuhan’s Sino-French Eco-City (urban development zone) and Xiaosi Township (wetland conservation zone) as typical zones. Based on 573 randomly sampled rural households, we explore the effects of land development opportunities on rural household incomes and find that: (1) Land development opportunities for non-agricultural conversion in the urban development zone significantly increase rural households’ total income, wage income, though their corresponding contribution rates are limited. Endogenously accumulated endowments such as human capital and economic status dominate the formation of such income gaps. (2) Planning-induced land development opportunities yield coefficients of 1.0442 for local employment income and −0.4567 for agricultural business income, with both statistically significant at the 1% significance level. Decomposition results show their respective contribution rates of 70.68% and 86.77%, demonstrating that such opportunities primarily account for cross-regional rural household income gaps. (3) Whereas non-agricultural land development opportunities narrow disparities in households’ local employment income, they raise inequality in rural households’ migrant employment, business, property and transfer income. These growth and equality-enhancing effects on local wage income are particularly pronounced for households possessing high-quantity but low-quality human capital. This study recommends supporting protected zones via farmer vocational training, expanded rural public service expenditure, and a benefit-sharing mechanism that channels land development gains to ecological and agricultural regions to strengthen households’ endogenous development capacity.

1. Introduction

Income disparities are a prominent research topic in the global pursuit of sustainable development [1,2,3], as they directly relate to wealth distribution and social inequalities [4,5]. They have become a prevailing research challenge spanning the interdisciplinary fields of economics, sociology and land resource science across the globe. Against the backdrop of global urbanization and the restructuring of spatial governance, planning tools such as ecological regulation and land-use zoning have been widely implemented worldwide. The ensuing rural-urban and intra-rural income gaps substantially affect high-quality economic growth, social equity and ecological conservation outcomes across nations [6,7,8,9]. As the world’s largest developing country, China has achieved remarkable economic growth since its Reform and Opening Up. However, such growth has not been evenly distributed, and the income gap has widened sharply [10,11]. Notably, inequality within rural China is more pronounced than in urban areas [12,13], with the per capita disposable income ratio between the top and bottom 20% of rural households reaching 9.95:1 in 2024 (Data source: China Statistical Yearbook 2025). The axiom that “inequality, not scarcity, is the root of social discontent” underscores rural income disparities as a pivotal socioeconomic challenge during China’s urban-rural transformation, with prolonged large disparities potentially precipitating political instability, economic inefficiency, and social tensions [14].

Extensive literature identifies determinants of rural income disparities, including macro-level factors such as regional development gaps [4,15,16], industrial institutions and policies [17,18], infrastructure construction [1,19,20], and technological change [21], as well as micro-level factors like household economic conditions [16,22,23], labor mobility [24,25], human capital [16,26,27], political capital [28], and land transfer [13,29]. However, quantitative studies on rural income differentiation have largely overlooked the impact of planning controls and the resulting differential land development opportunities. Specifically, few studies have quantitatively analyzed the effects of land use regulations on the various income types of rural households. Within the evolutionary context of global spatial governance systems, the land use control system originated in Germany in the late 19th century, followed by its adoption in San Francisco (1880) and New York City (1916) [9,30,31]. Owing to its strengths in resource allocation, ecological conservation and construction land management, it has been extensively accepted and implemented worldwide. Since the 1990s, China has successively issued spatial zoning policies, including land use master planning, land use regulation, ecological planning, and major function-oriented zoning [31,32,33,34], driving the gradual refinement of its territorial spatial planning system [9,35].

The spatial planning and control system constitutes a key institutional framework in contemporary China [36,37,38]. While leveraging comparative advantages in land development to optimize land resource allocation, it may simultaneously induce imbalanced development patterns that unevenly distribute resources and opportunities [39,40,41], ultimately affecting individual development and welfare outcomes [42,43,44,45]. Existing research has explored the welfare effects and development opportunity disparities arising from planning-induced land development opportunities across agricultural regions with different planning orientations and relevant stakeholder groups, primarily from perspectives of regional economic growth, households’ expected utility, or livelihood capital [43,45,46,47,48,49,50,51]. In recent years, limited studies have noted that planning-induced land development opportunities contribute to rural income differentiation, yet these analyses rely mainly on inductive comparisons or specific planning zones. For instance, Ma et al. identified land development intensity as the channel through which zoning policies influence rural household disposable income, with significant inter-zonal disparities in income growth [7]. Zhu and Cai’s nonparametric analysis revealed higher non-agricultural incomes in key development zones than in other functional zones [50]. Song and Han argued that environmental regulations restricted high-intensity, large-scale industrialization in ecological function zones, widening rural income gaps relative to other zones [52]. Tian and Xia posited that planning-induced development rights disparities exacerbated farmer income polarization [53]. Li et al. quantitatively demonstrated more severe income inequality within protected areas than in non-protected regions [54]. Duan et al. empirically confirmed that protected area establishment worsened local poverty [55].

Collectively, these studies confirm that planning zones with heterogeneous land development opportunities—including key development zones, ecological function zones and prime farmland protection zones—differ in socioeconomic development, rural household livelihood patterns, income differentiation and welfare changes. However, most rely on qualitative methods or descriptive statistics to assess total income impacts, lacking analysis of impacts on disaggregated household income components such as wage, business, property and transfer income. In addition, few studies quantified the contribution of planning-induced land development opportunities to rural household income disparities. As a leading megacity in China, Wuhan has pioneered the development and implementation of its spatial planning system. Under differentiated planning control policies, land development intensity and value realization exhibit significant spatial variation across the city. This study selects the Sino-French Eco-City and Xiaosi Township in Caidian District as representative cases of urban development and wetland conservation based on regional planning orientation, land use, ecological protection, and economic development. Using 2021 randomized survey data from 573 rural households, we apply OLS estimation, Tobit model, propensity score matching, and Shapley decomposition to empirically examine the effects of planning-induced land development opportunities on rural household income and quantify their relative contributions. The findings aim to provide direct evidence for designing territorial spatial planning policies that mitigate planning-induced regional imbalances, thereby contributing to rural revitalization, sustainable regional development and common prosperity goals.

The marginal contributions of this study are reflected in three aspects. First, differing from prior studies confined to total household income, it systematically estimates the impacts of planning-induced land development opportunities on rural households’ disaggregated incomes and corresponding income inequalities. Second, this paper employs contribution decomposition to quantify the contribution share of land development opportunities to household income and confirms their characteristics of low aggregate contribution yet prominent structural effects: household total and wage income growth mainly stems from households’ endogenous endowments, whereas land development opportunities dominate income divergence in other income categories. Third, this research identifies the dual attributes of zoning policies: pro-poor wage convergence alongside polarization of asset-related income, which provides empirical evidence for optimizing supporting policies for territorial spatial zoning and promoting sustainable rural development.

2. Theoretical Analysis and Hypotheses

Planning regulation is a key policy instrument through which the state leverages its public authority to regulate the direction, intensity and spatial layout of resource utilization based on the public attributes of land [9]. As land serves as both material foundation and spatial carrier for rural households, planning-induced disparities in development opportunities generate divergent land use efficiencies that significantly influence their economic activities and income outcomes [7,31,43,56]. Notably, planning-induced land development opportunities are not the primary source of income disparities among rural households, whereas accumulated endowments of human and physical capital constitute the core determinant of total income divergence. The pronounced income gap across functionally heterogeneous zones stems fundamentally from differential land access: in zones with development opportunities (e.g., urban development zones), low-return agricultural land can be converted into high-yield commercial, industrial, or public-infrastructure uses through land expropriation and redevelopment [43,57,58]. Local stakeholders thereby share economic spillovers from land value appreciation, which promotes household human capital accumulation, property income monetization and non-agricultural employment transition, strengthening their endogenous growth capacity [42,57]. Conversely, zones with restricted development (e.g., wetland conservation zones) bear heavy responsibilities for arable land and ecological conservation, facing prohibitive and restrictive policies on industrial development that severely constrain local resource utilization and regional development, thereby impeding the development of diversified income growth pathways [52,59]. Although national fiscal transfers (e.g., ecological compensation, special funds) have been implemented to offset economic losses in protected areas, their impact remains constrained and fails to fundamentally close resource access and income gaps compared with other agricultural regions [52,60]. Thus, we posit:

Hypothesis 1: 

Non-agricultural land development opportunities significantly increase rural household income, with total household income higher in urban development zones than in wetland conservation zones.

Planning-induced land development opportunities exert heterogeneous impacts on various disaggregated household income components via multiple channels, including improvements in public infrastructure, industrial restructuring, and labor reallocation (Figure 1). Compared to ecological conservation zones or permanent farmland protection zones, urban development zones exhibit superior transportation infrastructure and public services. This enhances village accessibility and locational competitiveness, facilitating regional industrial agglomeration and upgrading [61,62]. Meanwhile, the empowerment effect of improved public service access directly accelerates household human capital accumulation, mitigates farmers’ labor market disadvantages, facilitates labor reallocation to non-agricultural sectors, and boosts household wage income [62,63]. As rational economic actors, rural households prioritize labor allocation to higher-return non-agricultural activities [64]. Increased employment opportunities in development zones accelerate labor shifts to non-agricultural sectors, agricultural land transfer or leasing [65,66], resulting in lower agricultural income but higher land rental income. Non-agricultural land development enhances local transport convenience and service accessibility, driving village industrial upgrading and stimulating market demand for housing and industrial leasing [67]. Rural households in development zones can capitalize idle housing for rentals, thereby increasing property income. Whether households engage in non-agricultural business depends on the net effect of crowding-out and crowding-in effects from urban commercial radiation and market demand on village retailing. Existing literature shows inconclusive evidence regarding whether land development-induced economic growth positively affects transfer income through trickle-down effects. Consequently, comparative non-agricultural business and transfer incomes between development and conservation zones remain indeterminate. Notably, frequent farmland expropriation characterizes urban development zones, where household income is jointly affected by land expropriation compensation and land resource optimization. Integrating these mechanisms, we posit:

Hypothesis 2a: 

Non-agricultural land development opportunities promote optimal labor allocation, significantly increasing wage income; rural households in urban development zones thus earn higher wage income than those in wetland conservation zones.

Hypothesis 2b: 

Non-agricultural land development opportunities reduce agricultural income yet increase housing rental income, while net effects on business, property, and transfer income remain empirically ambiguous.

Heterogeneous land development opportunities under planning regulations also influence intra-rural income disparities. As wage income constitutes the primary livelihood source for most rural households, land development in urban zones particularly benefits low-income families by facilitating full labor reallocation to non-agricultural sectors. Higher-income households—already possessing multiple income channels and having largely exhausted their off-farm employment potential—experience limited marginal gains. Consequently, increased employment opportunities in development zones exhibit pro-poor effects: narrowing wage income gaps between low- and high-income households and mitigating intra-zonal income inequality. Existing studies confirm that the income-enhancing effect of non-farm employment is more pronounced among low-income and vulnerable rural households, as expanded employment opportunities drive substantial growth in their wage income [68,69]. Meanwhile, land asset appreciation derived from land development disproportionately benefits rural households with superior endowments, assets and capabilities, further exacerbating the polarization of property, business and transfer income within regions. This dual heterogeneous effect of narrowing wage income gaps alongside widening disparities in other income categories generates offsetting effects across different income categories and weakens the explanatory power of land development opportunities for overall household income inequality. Thus, we posit:

Hypothesis 3: 

Non-agricultural land development opportunities reduce wage income inequality among local households but exacerbate disparities in households’ property, business and transfer income, which may render the overall effect of land development opportunities on total income inequality insignificant.

3. Study Area, Variables, and Methodology

3.1. Study Area and Data Sources

Wuhan, situated on the eastern Jianghan Plain, is a sub-provincial megacity in Central China. It has consistently aligned its planning with national spatial planning reforms, establishing a functional zoning framework that delineates agricultural, ecological, and urban spaces. Caidian District—one of Wuhan’s six new urban districts—exhibits a clear “east-urban, west-rural” spatial pattern (Figure 2a). The Sino-French Eco-City and Xiaosi Township represent typical planning zones for urban development and ecological conservation within the district: the former lies adjacent to Caidian’s built-up area, while the latter is a distant lakeside rural township (Figure 2b).

In terms of planning orientation, the Sino-French Eco-City, the first Sino-French cooperative eco-city, was proposed for initiation in 2014 and adopts an industry–city integrated development model. In contrast, Xiaosi Township is designated primarily as a wetland conservation and prime farmland protection zone, with land use focused on agriculture and ecological preservation. From the perspective of natural endowments and development baselines, built-up land for urban–rural settlements, industry and mining accounted for 8.43% and 5.08% of total land area in the Sino-French Eco-City and Xiaosi Township respectively in 2005, while cultivated land accounted for 57.16% and 64.45%, respectively (Data sources: 2005 land cover remote sensing monitoring dataset). These statistics indicate that the two regions possessed highly similar land resource endowments before the implementation of land regulation policies. Nevertheless, disparities in land use structure and regional economic development between the two areas widened substantially after policy intervention. Regarding land use and ecological protection status in 2019, Xiaosi Township covers 14,404.17 ha, of which construction land accounts for only 4.89%. Over 70% of its area lies within the Chenhu International Wetland Reserve. Its cultivated land totals 4521.47 ha, representing 14.98% of Caidian’s total farmland. In contrast, Sino-French Eco-City spans 4956.79 ha, with construction land making up 40.34% (Data sources: 2020 Caidian Land Use Change Survey; Caidian Statistical Yearbook). In 2017, its fixed asset investment and fiscal revenue of the Sino-French Eco-City reached ¥6.319 billion and ¥811 million respectively, accounting for 15.37% and 14.54% of Caidian’s totals. Xiaosi Township’s corresponding figures were ¥169 million and ¥24 million, merely 0.41% and 0.43% of the district’s totals. Clearly, the Sino-French Eco-City and Xiaosi Township hold distinct planning orientations for urban development and wetland conservation. The current disparities in land use and socioeconomic development between the two regions stem not from inherent regional gaps but from their divergent long-term planning positioning, which renders the selected research areas highly representative. It should be noted that regions inherently differ in geographical location, resource endowments, and development conditions, and planning regulatory policies further reshape rural household income patterns based on these pre-existing disparities. Instead of identifying the net effect of land regulation, this study focuses on practical institutional contexts to explore the joint impacts of planning regulation and indigenous regional endowments on rural household income and income inequality.

We conducted a field questionnaire survey of rural households in the two study areas using a combined full-coverage and stratified random sampling approach. All villages within each zone were included, with households randomly sampled at approximately 8–10% of each village’s permanent resident households to ensure representativeness and distributional balance. Trained researchers conducted 40 to 60 min face-to-face interviews using standardized questionnaires, primarily targeting household heads or knowledgeable adults after obtaining informed consent. Between March and May 2021, we distributed 298 questionnaires across 5 retained villages within the Sino-French Eco-City and 344 questionnaires across 12 villages in Xiaosi Township (Figure 2c). After excluding invalid questionnaires and those covering only elderly residents, valid response rates reached 93.96% and 88.66% respectively. To reduce the influence of extreme income values, data were winsorized at the 1st and 99th percentiles, yielding a final sample of 573 households. The questionnaire covered household demographics, contracted land status, employment, livelihood capital, income, and attitudes toward household development.

3.2. Variables

3.2.1. Dependent Variables

The study examines rural household income and income inequality. Household income comprises total income and its four subcomponents: wage income, business income, property income, and transfer income. Wage income refers to earnings from local or non-local employment, distinguished based on whether the job permits returning to the village daily. Business income is net of business costs, covering both agricultural and non-agricultural activities. Property income includes land rent and housing rental income, excluding financial investment returns. Transfer income includes government subsidies, allowances, and pensions. Total income is the sum of these components. All income data refer to 2020 values reported during 2021 surveys. Income variables are log-transformed to mitigate heteroscedasticity. Income inequality is measured by the Kakwani index (bounded between 0 and 1), which satisfies scale invariance and normalization properties. Higher values indicate greater relative deprivation and inter-household income inequality. The metric is computed as

$${RD}_k={\displaystyle \frac{1}{n{\mu }_I}}{\displaystyle \sum _{i=k+1}^n}(I_i-I_k)={\gamma }_{I_K}^{+}{\displaystyle \frac{({\mu }_{I_K}^{+}-I_k)}{{\mu }_I}}$$

(1)

In Equation (1), $i$ and $k$ denote the i-th and k-th households in the sample; RD_k represents the income inequality degree of household ($k$) measured by the Kakwani index; n is the total sample size, with the corresponding income vector $\mathit{I}=(I_1,I_2,...,I_n)$ordered ascendingly. ${\mu }_{I_K}^{+}$is the mean income of households with income exceeding $I_k$, and ${\gamma }_{I_K}^{+}$denotes the proportion of such households to total sample size $n$. ${\mu }_I$ is the mean income of the full sample.

3.2.2. Independent Variables

The independent variable captures heterogeneous land development opportunities arising from planning regulations. The Sino-French Eco-City, adjacent to Caidian’s urban center, is a demonstration zone for sustainable urban development featuring multiple concentrated construction areas with active non-agricultural land development opportunities. It is defined as the treatment group (coded 1). Xiaosi Township, bordering Chenhu International Wetland and located within a prime farmland protection zone, faces land development restrictions due to ecological and agricultural preservation mandates, serving as the control group (coded 0).

3.2.3. Control Variables

Following the literature [70,71,72], we control for four types of capital—human, physical, financial, and social—all of which significantly influence rural household income. As a key determinant of rural household income, human capital is measured by household size, labor force ratio, health status, education attainment and allowance support. Given that physical, financial, and social capital are also key drivers of household income, we further control for farmhouse area, dwelling quality, urban housing, credit accessibility, and social networks. Variable definitions and descriptive statistics are reported in

Table 1. Descriptive Statistics of the Control Variables.

Variables	Variable Definitions	Full Sample (N = 573)		Urban Development Zone (N = 274)	Wetland Conservation Zone (N = 299)
		Mean	SD	Mean	Mean
Household size	Number of household members (persons)	4.7504	1.1880	4.8358	4.6722
Labor force ratio	(No. of persons aged 20–69 × 1 + No. aged 70–79 × 0.5 + No. aged ≥80 × 0.2)/Household size	0.7735	0.1397	0.7664	0.7801
Health status	Average health status of household members aged 20–69 (1–5 = very poor to very good)	4.5197	0.6701	4.5774	4.4669
Education attainment	Highest education level in household: Illiterate = 1; Primary = 2; Junior high = 3; Senior/vocational high = 4; Junior/technical college = 5; Bachelor = 6; Postgraduate = 7	4.4066	1.1887	4.5109	4.3110
Allowance support	Receipt of minimum living allowance: 1 = yes, 0 = no	0.0576	0.2332	0.0547	0.0602
Farmhouse area	Building area of rural housing (m2/household)	222.9367	107.9914	254.7153	193.8151
Dwelling quality	Composite index based on the building’s age, renovation recency, and structural type, higher values indicate better quality	0.1129	0.1454	0.1008	0.1239
Urban housing	Building area of urban housing (m2/household)	30.2485	51.1935	28.4453	31.9010
Credit accessibility	Difficulty in obtaining informal loans (1–5 = difficult to easy)	3.1449	0.9507	3.2226	3.0736
Social networks	Having immediate relatives as town cadres or corporate executives: 1 = yes, 0 = no	0.2723	0.4455	0.2956	0.2508

.

3.3. Methodology

3.3.1. Baseline Regression Model

To analyze the impact of planning-induced land development opportunities on rural household total income and its components, we establish the OLS regression model:

$$ln{y}_i={\alpha }_0+{\alpha }_1{X}_i+{\alpha }_2{\mathit{D}}_{\mathit{i}}+{\mu }_i$$

(2)

In Equation (2), $y_i$ is the dependent variable, representing household total income or each income component; $X_i$ is the core explanatory variable measuring land development opportunities; ${\mathit{D}}_{\mathit{i}}$ is a vector of control variables affecting rural household income; ${\alpha }_0$ is the intercept; ${\alpha }_1$ and ${\alpha }_2$ are coefficients to be estimated; ${\mu }_i$ is the random disturbance term.

Since the Kakwani index measuring inter-household income inequality is bounded between 0 and 1, OLS estimates would be biased. We therefore employ a Tobit model estimated via maximum likelihood to analyze land development opportunities’ impact on income inequality. The specification is:

$$y_{ki}={\beta }_0+{\beta }_1{x}_i+{\beta }_2{\mathit{D}}_{\mathit{i}}+{\mu }_i,{\mu }_i~N(0,{\sigma }^2)$$

(3)

In Equation (3), $y_{ki}$ is the dependent variable representing household income inequality; $N(0,{\sigma }^2)$ denotes the normal distribution with mean 0 and variance ${\sigma }^2$; other terms are defined consistently with the OLS model in Equation (2).

3.3.2. Shapley Value Decomposition

The Shapley value decomposition method is further used to analyze the contribution rates of land development opportunities to rural household income and its inequality. This method applies cooperative game theory to measure each regressor’s contribution share to overall variation [73,74]. This is achieved by examining changes in the model’s $R^2$under different subsets of independent variables and calculating each variable’s contribution accordingly. The marginal contribution of independent variable $X_i$ to R², denoted ${MR}_i^2$, is expressed as follows:

$${MR}_i^2=E\left[R^2(y=\alpha +\sum _{j\in S}{b}_j{x}_j+b_i{x}_i+e)-R^2(y={\alpha }^{*}+\sum _{j\in S}{b_j^{*}{x}_j^{*}}_j+e^{*})\right]$$

(4)

In Equation (4), $x_j$ represents the combination of independent variables excluding $x_i$. To measure the marginal contribution of $x_i$, we first compute the model’s $R^2$ when $x_i$ is included, and then evaluate the change in $R^2$ when $x_i$ is omitted from the model. It is important to note that—when the model contains multiple independent variables—the change in $R^2$ from excluding a variable $x_k$ differs across different variable subsets. The marginal contribution ${MR}_k^2$ of that variable is obtained by taking the average of these varying changes.

3.3.3. Propensity Score Matching

Propensity Score Matching (PSM) is employed to estimate the Average Treatment Effect on the Treated (ATT) of planning-induced land development opportunities on rural household income and its inequality. The ATT is specified as follows:

$$ATT=E\left[Y_1|D=1\right]-E\left[Y_0|D=1\right]$$

(5)

In Equation (5), D is a binary variable where D = 1 denotes urban development zone and D = 0 indicates wetland conservation zone, capturing heterogeneous land development opportunities. $Y_1$ and $Y_0$ correspond to household income and income inequality levels in the development zone and the conservation zone, respectively. $E\left[Y_0|D=1\right]$ represents the counterfactual income of households in the development zone had they been located in the conservation zone, which is unobservable in reality. It is estimated by matching observations from the control group with similar characteristics based on propensity scores.

4. Results

4.1. Descriptive Analysis

As shown in

Table 2. Comparison of Income and Income Inequality between Rural Households in Urban Development Zone and Wetland Conservation Zone.

Income Types	Mean Income (104 yuan)		T-Value	Income Share (%)		T-Value	Kakwani Index		T-Value
	UDZ	WCZ		UDZ	WCZ		UDZ	WCZ
Total income	15.8744	13.8026	2.0718 **	/	/	/	0.2419	0.2772	0.0354 ***
Wage income	13.8364	11.3921	2.4443 **	87.02	80.31	6.71 **	0.2652	0.3259	0.0607 ***
Local employment	7.0953	2.0951	5.0002 ***	50.02	15.99	34.03 ***	0.4868	0.7744	0.2877 ***
Migrant employment	6.7411	9.2970	−2.5559 ***	37.00	64.32	−27.31 ***	0.6316	0.4030	−0.2286 ***
Business income	0.6042	1.4996	−0.8954 **	3.14	11.22	−8.08 **	0.9384	0.6902	−0.2482 ***
Non-agricultural	0.4418	0.5324	−0.0906	2.17	3.17	−1.00	0.9761	0.9614	−0.0147 *
Agricultural	0.1625	0.9672	−0.8047 ***	0.97	8.05	−7.08 ***	0.9229	0.6290	−0.2939 ***
Property income	0.1093	0.2259	−0.1166 **	0.57	2.09	−1.52 **	0.9510	0.8364	−0.1146 ***
Farmland rent	0.0292	0.2259	−0.1967 ***	0.21	2.09	−1.88 ***	0.9460	0.8364	−0.1096 ***
Housing rent	0.0801	0.0000	0.0801 ***	0.36	0.00	0.36 **	0.9757	0.0000	−0.9757 ***
Transfer income	1.3245	0.6850	0.6395 **	9.28	6.39	2.89 **	0.5589	0.5054	−0.0536 **
Government subsidies	0.0161	0.0619	−0.0458 ***	0.12	0.58	−0.46 ***	0.6594	0.2954	−0.3640 ***
Government allowances	0.0343	0.0859	−0.0516 **	0.36	1.51	−1.15 **	0.9559	0.9682	0.0123
Pensions	1.2741	0.5372	0.7369 ***	8.80	4.31	4.50 ***	0.5804	0.5370	−0.0434 *

Note: *, **, and *** denote statistical significance at the 10%, 5%, and 1% levels, respectively. UDZ and WCZ represent the Urban Development Zone and Wetland Conservation Zone, respectively.

, rural households in urban development zone and wetland conservation zone exhibit significant t-test differences in the means and shares of all income types except non-agricultural business income. The mean total household income in development zone is 158,744 yuan, 20,718 yuan higher than that in conservation zone, with the t-test significant at the 5% level. Wage income is the dominant income source for households, accounting for 87.02% and 80.31% of total income in the development and conservation zones, respectively. Households in the development zone rely on both local (50.02%) and migrant (37.00%) employment income, whereas those in the conservation zone depend mainly on migrant employment (64.32%), with local employment income accounting for only 15.99%. The second-largest income source is transfer income in the development zone and business income in the conservation zone. Transfer income in the development zone is dominated by pensions, with a mean of 12,741 yuan, while business income in the conservation zone mainly comes from agricultural activities, with a mean of 9672 yuan. Property income accounts for just 0.57% and 2.09% of total income in the two zones respectively, both at a low level.

Regarding income inequality, t-tests reveal significant differences in the Kakwani index for all income components except government subsidies. The Kakwani index of total income is 0.2419 in the development zone and 0.2772 in the conservation zone, significant at the 1% level, indicating higher total income inequality in the conservation zone. Regarding income components, the development zone shows significantly higher Kakwani indices for business, property and transfer income, but its Kakwani index of wage income is 0.0607 lower than that of the conservation zone. These results imply greater income inequality in business, property, and transfer income among households in development zone, but lower wage-income inequality compared with households in the conservation zone.

4.2. Baseline Results

4.2.1. Effects and Contributions on Rural Household Income

Table 3. OLS Estimates and Shapley Decomposition of Land Development Opportunities’ impact on Rural Household Income.

Variables	Total Income	Wage Income	Local Employment	Migrant Employment	Business Income	Agricultural Income	Property Income	Transfer Income	Pensions
Land Development Opportunities	0.0952 ***	0.2115 ***	1.0442 ***	−0.7790 ***	−0.5096 ***	−0.4567 ***	−0.0807 ***	0.2316 ***	0.2792 ***
	(3.69%)	(8.60%)	(70.68%)	(35.32%)	(77.19%)	(86.77%)	(28.84%)	(21.03%)	(31.55%)
Household Size	0.1476 ***	0.1935 ***	0.1589 ***	0.0523	−0.0100	0.0105	−0.0023	0.0555 ***	0.0571 ***
	(24.61%)	(26.15%)	(4.15%)	(1.72%)	(0.60%)	(0.35%)	(0.15%)	(14.10%)	(17.48%)
Labor force ratio	0.6464 ***	1.1536 ***	1.6985 ***	0.2354	0.2202	0.3419 **	0.0031	−0.7795 ***	−0.8128 ***
	(3.30%)	(7.98%)	(7.73%)	(0.31%)	(1.70%)	(4.11%)	(0.08%)	(23.80%)	(27.65%)
Health status	0.1397 ***	0.1689 ***	0.0660	0.1644 **	0.0948 ***	0.0473 *	−0.0115	−0.1000 ***	−0.0722 ***
	(12.76%)	(13.62%)	(0.77%)	(4.89%)	(4.93%)	(1.30%)	(1.07%)	(7.46%)	(2.15%)
Education attainment	0.0296 **	0.0412 **	−0.0228	0.0840 **	−0.0280	−0.0157	−0.0186 *	−0.0036	−0.0043
	(6.71%)	(7.32%)	(0.22%)	(5.37%)	(0.90%)	(0.86%)	(4.96%)	(0.35%)	(0.30%)
Allowance support	−0.0953	−0.3439 ***	−0.1409	−0.3225	0.0195	0.0046	−0.0083	0.4440 ***	−0.0259
	(3.95%)	(10.07%)	(0.34%)	(4.11%)	(0.31%)	(0.12%)	(0.17%)	(16.13%)	(0.19%)
Farmhouse area	0.0005 ***	−0.0001	−0.0003	0.0008 *	0.0009 **	0.0006 ***	0.0000	−0.0001	−0.0001
	(8.15%)	(2.62%)	(2.91%)	(1.31%)	(7.79%)	(3.97%)	(1.28%)	(0.49%)	(1.02%)
Dwelling quality	0.0455	0.0848	0.9578 ***	−0.4681	−0.0829	0.0097	−0.0281	−0.3232 ***	−0.3191 ***
	(0.35%)	(0.24%)	(3.76%)	(0.60%)	(0.40%)	(0.48%)	(0.18%)	(3.69%)	(3.77%)
Urban housing	0.0026 ***	0.0024 ***	−0.0036 ***	0.0080 ***	−0.0005	−0.0006 *	0.0013 ***	0.0016 ***	0.0016 ***
	(24.90%)	(15.99%)	(8.18%)	(43.83%)	(0.35%)	(1.39%)	(59.77%)	(9.66%)	(10.93%)
Credit accessibility	0.0427 ***	0.0544 **	0.0292	0.0479	0.0109	0.0157	−0.0107	0.0166	0.0214
	(5.62%)	(5.58%)	(1.01%)	(1.41%)	(0.32%)	(0.21%)	(1.19%)	(0.47%)	(1.39%)
Social networks	0.1001 ***	0.0180	0.0600	−0.1786	0.1704 **	−0.0106	0.0120	0.0826 **	0.0790 *
	(5.98%)	(1.82%)	(0.24%)	(1.14%)	(5.52%)	(0.44%)	(2.30%)	(2.82%)	(3.56%)
_cons	0.3059 *	−0.6222 **	−1.6592 ***	0.0354	−0.0326	−0.0914	0.2810 **	1.1543 ***	0.9892 ***
N	573	573	573	573	573	573	573	573	573
R2	0.4859	0.3599	0.2988	0.2670	0.1820	0.2510	0.0829	0.2907	0.2745
Prob > F	0.0000	0.0000	0.0000	0.0000	0.0000	0.0000	0.0000	0.0000	0.0000

Note: *, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively. Numbers in parentheses are contributions (%) from the Shapley decomposition.

presents the OLS and Shapley decomposition results. Given the relatively small shares of property and transfer income, sub-components other than pensions are excluded from regression analyses. All income equations show statistically significant F-statistics at the 1% level with satisfactory R² values. Coefficients sign align with theoretical expectations, confirming the model’s validity.

The coefficient for land development opportunities on total household income is 0.0952 and significant at the 1% level, indicating that non-agricultural development opportunities in urban development zone significantly increase total household income, thus supporting Hypothesis 1. For wage income, the coefficient is 0.2115 (p < 0.001), confirming a positive effect of land development opportunities on wage earnings, thereby verifying Hypothesis 2a. However, the contributions of land development opportunities to total and wage income are relatively low, only 3.69% and 8.60%, respectively. Growth in these income categories is driven mainly by labor force endowments (household size, health status) and economic status (urban housing), indicating that household endogenous endowments, instead of spatial planning, serve as the dominant determinants of gaps in total income and wage income. Disaggregating wage income into local and migrant employment incomes reveals that land development opportunities significantly increase local employment income (coefficient = 1.0442, p < 0.001) but reduce migrant employment income (coefficient = −0.7790, p < 0.001), with corresponding contribution rates of 70.68% and 35.32%, respectively. Local employment income is primarily determined by non-agricultural land development, whereas migrant employment income is also shaped by household economic conditions, which contribute up to 43.83% to migrant employment income. For other income types, land development opportunities negatively affect business income (specifically agricultural business income) and property income, with coefficients of −0.5096 (−0.4567) and −0.0807, respectively, while showing positive effects on transfer income (i.e., pensions). These findings verify Hypothesis 2b regarding the effect on agricultural business income and further clarify the impact directions of land development opportunities on business, property and transfer income. Contribution decomposition results further show that land development opportunities account for 77.19% of business income variation (86.77% of agricultural business), 28.84% of property income variation, and 21.03% of transfer income variation (31.55% of pension). The large contribution to agricultural business income likely reflects farmland expropriation resulting from non-agricultural land development, which substantially reduces agriculture-based earnings.

4.2.2. Effects and Contributions on Rural Household Income Inequality

Table 4. Tobit Estimates and Shapley Decomposition of Land Development Opportunities’ impact on Rural Household Income Inequality.

Variables	Total Income	Wage Income	Local Employment	Migrant Employment	Business Income	Agricultural Income	Property Income	Transfer Income	Pensions
Land Development Opportunities	−0.0106	−0.0452 ***	−0.4709 ***	0.3762 ***	0.5696 ***	0.7487 ***	0.4050 ***	0.0673 ***	0.0278
	(0.82%)	(3.22%)	(64.12%)	(37.24%)	(89.09%)	(87.80%)	(41.87%)	(2.91%)	(2.10%)
Household Size	−0.0631 ***	−0.0753 ***	−0.0677 ***	−0.0235	−0.0120	−0.0393 *	0.0409	−0.0488 ***	−0.0818 ***
	(29.47%)	(29.79%)	(4.65%)	(3.13%)	(0.30%)	(0.50%)	(0.20%)	(21.70%)	(31.46%)
Labor force ratio	−0.2869 ***	−0.4486 ***	−0.9143 ***	−0.1289	−0.2855 *	−0.6258 ***	0.4707	0.6166 ***	0.9505 ***
	(4.39%)	(8.84%)	(9.25%)	(0.50%)	(1.76%)	(4.95%)	(0.81%)	(34.15%)	(40.26%)
Health status	−0.0601 ***	−0.0675 ***	−0.0306	−0.0774 **	−0.0816 ***	−0.0847 **	0.0236	0.0680 ***	0.0644 **
	(15.43%)	(15.34%)	(1.35%)	(5.02%)	(2.44%)	(1.12%)	(1.32%)	(9.27%)	(2.48%)
Education attainment	−0.0085 *	−0.0119 *	0.0112	−0.0294 *	0.0146	0.0132	0.0778 **	0.0104	0.0251
	(5.37%)	(6.02%)	(0.23%)	(4.31%)	(0.52%)	(0.50%)	(5.67%)	(2.27%)	(2.00%)
Allowance support	0.0603 **	0.1373 ***	0.0516	0.1743 *	−0.0777	−0.1021	−0.0478	−0.2494 ***	0.1303 *
	(6.00%)	(11.35%)	(0.40%)	(5.36%)	(0.18%)	(0.08%)	(0.37%)	(12.73%)	(1.68%)
Farmhouse area	−0.0001 **	0.0000	0.0001	−0.0004 *	−0.0006 ***	−0.0007 ***	−0.0005	0.0001	0.0003
	(6.14%)	(2.92%)	(2.39%)	(1.26%)	(2.72%)	(2.65%)	(2.88%)	(1.82%)	(1.39%)
Dwelling quality	−0.0031	−0.0182	−0.5243 ***	0.2131	0.0582	0.0073	0.0099	0.2593 ***	0.3948 ***
	(0.25%)	(0.21%)	(5.58%)	(0.44%)	(0.29%)	(0.23%)	(0.22%)	(4.52%)	(5.77%)
Urban housing	−0.0009 ***	−0.0009 ***	0.0020 ***	−0.0031 ***	0.0008 **	0.0013 **	−0.0039 ***	−0.0010 ***	−0.0014 ***
	(20.69%)	(15.03%)	(10.71%)	(40.21%)	(0.63%)	(1.47%)	(44.86%)	(9.58%)	(11.97%)
Credit accessibility	−0.0198 ***	−0.0225 ***	−0.0185	−0.0240	−0.0126	−0.0272	0.0576	−0.0065	0.0027
	(6.39%)	(5.90%)	(1.07%)	(1.51%)	(0.11%)	(0.22%)	(1.05%)	(0.23%)	(0.12%)
Social networks	−0.0370 ***	−0.0027	−0.0236	0.0766	−0.0901 **	0.0372	0.0010	−0.0351	−0.0247
	(5.07%)	(1.38%)	(0.25%)	(1.02%)	(1.96%)	(0.48%)	(0.76%)	(0.82%)	(0.77%)
_cons	1.2231	1.4722 ***	2.1807 ***	1.3058 ***	1.4716 ***	1.8850 ***	0.4077	−0.0749	−0.1981
N	573	573	573	573	573	573	573	573	573
Prob > chi2	0.0000	0.0000	0.0000	0.0000	0.0000	0.0000	0.0000	0.0000	0.0000

Note: *, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively. Numbers in parentheses are contributions (%) from the Shapley decomposition.

presents Tobit estimates and Shapley decomposition results. All equations show satisfactory overall significance, supporting the validity of the empirical analysis. The Tobit results indicate that land development opportunities contribute to lower total income inequality (coefficient = −0.0106) and significantly lower wage income inequality (coefficient = −0.0452, p < 0.01), implying that non-agricultural development opportunities in urban development zone help reduce wage income disparities. However, their contributions to total and wage income inequality are modest, at 0.82% and 3.22%, respectively. Disaggregating wage inequality into local and migrant employment reveals that land development opportunities reduce local employment income inequality but increase migrant employment income inequality, with contribution rates of 64.12% and 37.24%, respectively. Local employment income inequality appears primarily driven by land development opportunities, whereas migrant employment income inequality is substantially explained by household economic status, with contribution rates of urban housing reaching 40.21%. For other income types, land development opportunities exacerbate inequality in migrant employment income, business income, property income, and transfer income for the positive coefficients on the Kakwani index. Land development opportunities account for 89.09% (87.80%) of inequality in (specifically agricultural) business income, suggesting that business income disparities are largely attributable to unequal planning-induced non-agricultural development opportunities. The contribution to property income inequality is 41.87%, while the contribution to transfer income inequality is relatively smaller. Collectively, these results indicate that reduced total income inequality in development zone is mainly driven by declining local employment inequality, since land-development-induced local employment growth mitigates overall income disparity. Nevertheless, land development simultaneously widens disparities in households’ migrant employment income, business income, property income and transfer income. Such offsetting forces render the aggregate effect of land development opportunities on total income inequality statistically insignificant, thereby validating Hypothesis 3.

4.3. Robustness Checks

To verify the robustness of the impacts of land development opportunities on rural household income disparities, we implement three robustness checks. First, we examine the effect of non-agricultural land development opportunities on non-expropriated households’ income to eliminate potential income distortions arising from prevalent land acquisition in development zone. Second, we re-estimate effects after trimming the top and bottom 1% of the total household income distribution. Third, we adopt an alternative estimation method, namely propensity score matching (PSM), to match treatment and control groups, and then estimate the Average Treatment Effect on the Treated (ATT) on income and income inequality. Results from four matching methods are highly consistent, and their arithmetic mean is used to reflect the treatment effect. As shown in

Table 5. Robustness Checks for Income and Income Inequality.

Dependent Variable	Approach	Variable/Matching Method	Total Income	Wage Income	Local Employment	Migrant Employment	Business Income	Agricultural Income	Property Income	Transfer Income	Pensions
Income	Excluding the Impact of Land Expropriation	Land Development Opportunities	0.1114 ***	0.2704 ***	1.0115 ***	−0.6217 ***	−0.4960 ***	−0.4066 ***	−0.0180	0.0589	0.0904 **
			(0.0347)	(0.0498)	(0.1076)	(0.1238)	(0.0588)	(0.0443)	(0.0296)	(0.0388)	(0.0396)
		Controls	YSE	YSE	YSE	YSE	YSE	YSE	YSE	YSE	YSE
		_cons	0.1567	−0.8644 ***	−1.4335 **	−0.4889	−0.0695	−0.1972	0.3482 **	0.6989 ***	0.4724 *
			(0.2123)	(0.3312)	(0.5946)	(0.6244)	(0.3492)	(0.2802)	(0.1636)	(0.2370)	(0.2575)
		N	434	434	434	434	434	434	434	434	434
		R2	0.5061	0.3944	0.2534	0.2134	0.1587	0.1610	0.0800	0.3197	0.2173
		Prob > F	0.0000	0.0000	0.0000	0.0000	0.0000	0.0000	0.0000	0.0000	0.0000
	Trimming 1% from Both Tails	Land Development Opportunities	0.0956 ***	0.1905 ***	1.0373 ***	−0.8065 ***	−0.5013 ***	−0.4594 ***	−0.0846 ***	0.2375 ***	0.2840 ***
			(0.0277)	(0.0400)	(0.0839)	(0.0925)	(0.0476)	(0.0348)	(0.0215)	(0.0348)	(0.0355)
		Controls	YSE	YSE	YSE	YSE	YSE	YSE	YSE	YSE	YSE
		_cons	0.4019 **	−0.3612	−1.5575 ***	0.2251	−0.0872	−0.1249	0.2580 **	1.1016 ***	0.9355
			(0.1672)	(0.2461)	(0.5319)	(0.5802)	(0.2909)	(0.2253)	(0.1306)	(0.2215)	(0.2308)
		N	561	561	561	561	561	561	561	561	561
		R2	0.4572	0.3453	0.2885	0.2644	0.1880	0.2564	0.0796	0.3002	0.2807
		Prob > F	0.0000	0.0000	0.0000	0.0000	0.0000	0.0000	0.0000	0.0000	0.0000
	Propensity Score Matching Estimation	Unmatched	0.1612 *** (0.0366)	0.2489 *** (0.0497)	1.0279 *** (0.0808)	−0.7042 *** (0.0957)	−0.4396 *** (0.0483)	−0.4169 *** (0.0339)	−0.0899 *** (0.0223)	0.2405 *** (0.0377)	0.2937 *** (0.0374)
		Kernel Matching	0.1106 *** (0.0423)	0.2311 *** (0.0573)	1.0034 *** (0.0915)	−0.6845 *** (0.1070)	−0.5083 *** (0.0549)	−0.4446 *** (0.0405)	−0.0582 ** (0.0261)	0.2504 *** (0.0414)	0.2886 *** (0.0410)
		k-Nearest Neighbor Matching	0.1229 *** (0.0472)	0.2104 *** (0.0632)	1.0645 *** (0.0991)	−0.7414 *** (0.1150)	−0.4999 *** (0.0590)	−0.4537 *** (0.0442)	−0.0378 (0.0282)	0.2584 *** (0.0433)	0.2881 *** (0.0435)
		Caliper Matching	0.1112 *** (0.0423)	0.2323 *** (0.0573)	1.0002 *** (0.0915)	−0.6824 *** (0.1071)	−0.5055 *** (0.0549)	−0.4431 *** (0.0405)	−0.0586 ** (0.0261)	0.2507 *** (0.0414)	0.2885 *** (0.0410)
		Local Linear Regression Matching	0.1042 ** (0.0519)	0.2174 *** (0.0711)	0.9884 *** (0.1172)	−0.6830 *** (0.1312)	−0.5015 *** (0.0688)	−0.4403 *** (0.0482)	−0.0628 * (0.0340)	0.2551 *** (0.0478)	0.2934 *** (0.0481)
		Mean	0.1122	0.2228	1.0141	−0.6978	−0.5038	−0.4454	−0.0544	0.2536	0.2897
Income Inequality	Excluding the Impact of Land Expropriation	Land Development Opportunities	−0.0146	−0.0670 ***	−0.4778	0.2881	0.4588 ***	0.5136	0.0680	0.1570 ***	0.1559
			(0.0141)	(0.0192)	0.0654	0.0452	(0.0435)	0.0593	(0.0973)	(0.0267)	0.0445
		Controls	YES	YES	YES	YES	YES	YES	YES	YES	YES
		_cons	1.2689 ***	1.5343 ***	2.2814	1.4722	1.3868 ***	1.7276	0.2258	0.1264	0.0183
			(0.0776)	(0.1063)	0.3827	0.2477	(0.2176)	0.2981	(0.5241)	(0.1471)	0.2425
		N	434	434	434	434	434	434	434	434	434
		Prob > F	0.0000	0.0000	0.0000	0.0000	0.0000	0.0000	0.0000	0.0000	0.0000
	Trimming 1% from Both Tails	Land Development Opportunities	−0.0064	−0.0363 **	−0.4636	0.3885	0.5698 ***	0.7537	0.4190 ***	0.0638 ***	0.0290
			(0.0112)	(0.0149)	0.0482	0.0403	(0.0423)	0.0591	(0.0918)	(0.0237)	0.0348
		Controls	YES	YES	YES	YES	YES	YES	YES	YES	YES
		_cons	1.1743 ***	1.3912 ***	2.0928	1.2085	1.5584 ***	2.0078	0.4189	−0.0370	−0.1664
			(0.0671)	(0.0893)	0.2910	0.2403	(0.2280)	0.3109	(0.5193)	(0.1417)	0.2075
		N	561	561	561	561	561	561	561	561	561
		Prob > F	0.0000	0.0000	0.0000	0.0000	0.0000	0.0000	0.0000	0.0000	0.0000
	Propensity Score Matching Estimation	Unmatched	−0.0354 **	−0.0607 ***	−0.2877 ***	0.2286 ***	0.2482 ***	0.2939 ***	0.1146 ***	0.0536 **	0.0434
			(0.0145)	(0.0179)	(0.0282)	(0.0290)	(0.0179)	(0.0217)	(0.0207)	(0.0248)	(0.0272)
		Kernel Matching	−0.0174	−0.0489 **	−0.2744 ***	0.2242 ***	0.2706 ***	0.3078 ***	0.0803 ***	0.0449	0.0431
			(0.0167)	(0.0206)	(0.0321)	(0.0327)	(0.0209)	(0.0256)	(0.0248)	(0.0285)	(0.0316)
		k-Nearest Neighbor Matching	−0.0231	−0.0454 **	−0.2964 ***	0.2407 ***	0.2772 ***	0.3278 ***	0.0586 **	0.0462	0.0523
			(0.0184)	(0.0225)	(0.0350)	(0.0353)	(0.0227)	(0.0277)	(0.0268)	(0.0307)	(0.0345)
		Caliper Matching	−0.0174	−0.0490 **	−0.2732 ***	0.2238 ***	0.2694 ***	0.3065 ***	0.0809 ***	0.0447	0.0434
			(0.0167)	(0.0206)	(0.0321)	(0.0328)	(0.0209)	(0.0256)	(0.0248)	(0.0285)	(0.0316)
		Local Linear Regression Matching	−0.0153	−0.0449 *	−0.2686 ***	0.2243 ***	0.2659 ***	0.3027 ***	0.0857 **	0.0389	0.0362
			(0.0203)	(0.0258)	(0.0421)	(0.0408)	(0.0283)	(0.0337)	(0.0334)	(0.0381)	(0.0430)
		Mean	−0.0183	−0.0470	−0.2782	0.2282	0.2708	0.3112	0.0764	0.0437	0.0438

Note: *, **, and *** denote statistical significance at the 10%, 5%, and 1% levels, respectively. Values in parentheses are standard errors.

, after excluding disturbances arising from land expropriation, the significant positive effects of land development opportunities on rural households’ total and wage income remain robust, and the core empirical results remain highly consistent with the benchmark regression. This evidence implies that income growth of rural households in development zones does not merely rely on short-term income gains and passive livelihood substitution driven by land acquisition compensation. Instead, such income growth originates from the endogenous impetus for sustainable rural development generated by spatial planning via optimizing regional industrial layouts, improving public infrastructure and expanding non-farm employment channels. Results from the trimmed sample, and the propensity score matched sample are highly consistent with our main findings. Collectively, these results confirm the robustness of the empirical results.

4.4. Heterogeneity Analysis

To assess heterogeneity in the effects of land development opportunities on rural household income and inequality from two dimensions of human capital endowment: quantity (the household labor force size is calculated with age-based adjustments, and the specific weighting methodology is detailed in Table 1, measured by household labor force size) and quality (proxied by average years of schooling). For each dimension, households are split into low and high groups based on whether their value exceed the sample mean, with regression estimates presented in

Table 6. Human Capital Heterogeneity Tests for Income and Income Inequality.

Dependent Variables	Group	Total Income	Wage Income	Local Employment	Migrant Employment	Business Income	Agricultural Income	Property Income	Transfer Income	Pensions
Income	Low Quantity	0.0736 *	0.2082 ***	0.7875 ***	−0.5960 ***	−0.4865 ***	−0.4118 ***	−0.0903 ***	0.2638 ***	0.3187 ***
		(0.0409)	(0.0673)	(0.1237)	(0.1388)	(0.0694)	(0.0466)	(0.0319)	(0.0471)	(0.0482)
	High Quantity	0.0976 **	0.1901 ***	1.2923 ***	−0.9920 ***	−0.5256 ***	−0.4897 ***	−0.0746 ***	0.1989 ***	0.2374 ***
		(0.0414)	(0.0551)	(0.1108)	(0.1255)	(0.0645)	(0.0498)	(0.0285)	(0.0494)	(0.0502)
	Low Quality	0.0688 **	0.2328 ***	1.1175 ***	−0.8153 ***	−0.5255 ***	−0.5012 ***	−0.1087 ***	0.1962 ***	0.2520 ***
		(0.0391)	(0.0582)	(0.1081)	(0.1234)	(0.0636)	(0.0500)	(0.0270)	(0.0413)	(0.0429)
	High Quality	0.1191 **	0.1637 **	0.9269 ***	−0.7183 ***	−0.4634 ***	−0.3904 ***	−0.0499	0.2529 ***	0.2895 ***
		(0.0463)	(0.0649)	(0.1316)	(0.1426)	(0.0761)	(0.0476)	(0.0339)	(0.0559)	(0.0564)
Income Inequality	Low Quantity	0.0052	−0.0369	−0.3936 ***	0.3011 ***	0.5882 ***	0.7645 ***	0.3956 ***	0.0266	−0.0461
		(0.0169)	(0.0244)	(0.0766)	(0.0583)	(0.0643)	(0.0885)	(0.1150)	(0.0300)	(0.0456)
	High Quantity	−0.0159	−0.0421 **	−0.5376 ***	0.4566 ***	0.5460 ***	0.7196 ***	0.4228 ***	0.1050 ***	0.1005 **
		(0.0142)	(0.0186)	(0.0593)	(0.0542)	(0.0540)	(0.0760)	(0.1413)	(0.0357)	(0.0510)
	Low Quality	0.0018	−0.0475 **	−0.5331 ***	0.3943 ***	0.5778 ***	0.7525 ***	0.3722 ***	0.0943 ***	0.0834 **
		(0.0153)	(0.0207)	(0.0647)	(0.0541)	(0.0553)	(0.0755)	(0.1130)	(0.0272)	(0.0366)
	High Quality	−0.0252	−0.0381 *	−0.3788 ***	0.3454 ***	0.5398 ***	0.7203 ***	0.4730 ***	0.0542	−0.0197
		(0.0169)	(0.0228)	(0.0706)	(0.0598)	(0.0637)	(0.0898)	(0.1560)	(0.0398)	(0.0692)

Note: *, **, and *** denote statistical significance at the 10%, 5%, and 1% levels, respectively. Values in parentheses are standard errors.

.

Overall, the effects of non-agricultural land development opportunities in urban development zone on income types and inequality in the high- and low-quantity or quality subgroups are largely consistent with those in the full sample, confirming the robustness of the empirical results. In terms of human capital quantity, land development opportunities exhibit a stronger income-enhancing effect on total income in high-quantity households (0.0976) than in low-quantity households (0.0736), manifested primarily through increased local employment income. Moreover, land development opportunities reduce wage income inequality more markedly among high-quantity households. These findings indicate that households endowed with greater human capital quantity are more likely to secure local employment in urban development zones, thereby boosting overall household income and narrowing within-group income disparities. Compared with low-quantity households, land development opportunities demonstrate a more pronounced widening effect on migrant employment income inequality among high-quantity households. This phenomenon may relate to expanded local job opportunities in development zones offering lower wages than migrant employment, coupled with some laborers from high-quantity households shifting from migrant employment to local employment due to living cost pressures and family responsibilities.

In terms of human capital quality, the estimated coefficients of land development opportunities on wage income are 0.1637 and 0.2328 for high- and low-quality groups, respectively, with the latter group demonstrating significantly stronger local employment income gains. This indicates that low-quality households are more likely to benefit from local employment opportunities in development zones. The disparity may be explained by the lower entry barriers of local jobs in development zones and the fact that high-quality labor is already fully employed. Consequently, land development opportunities manifest more pronounced income-enhancing effects for households with abundant labor or lower education. Non-agricultural land development within urban development zones generates local job opportunities and reduces rural households’ reliance on migrant employment. Empirically, laborers with superior human capital endowments exhibit stronger competitiveness and a higher willingness to engage in off-site employment, tending to secure high-paying jobs outside their hometowns. Against this backdrop, this study holds that land development opportunities deliver larger local income gains for low-quality households than for high-quality households, thereby significantly narrowing wage and local employment income inequality for low-quality households.

5. Discussion

From a global territorial governance perspective, the restructuring of rural income patterns driven by spatial regulation constitutes a prevalent worldwide phenomenon [75,76,77,78,79,80,81,82,83,84,85], though regulatory frameworks, implementation pathways and governance outcomes vary markedly across economies. Existing literature has not reached a consensus regarding the impacts of planning regulation on rural household income and income inequality. Most studies document that stringent ecological protection and land-use control policies can limit land development and reduce non-agricultural employment opportunities, leading to persistently lower incomes for farm households in regulated areas compared to those in adjacent unregulated zones, thereby creating a spatial income gap [75,76,77,78,79]. However, several studies argue that well-designed ecological protection regulations, together with targeted supportive policies, can optimize rural households’ income structure and stabilize local livelihoods via ecological compensation and green industry development, thereby mitigating inter-regional and intra-household income inequality [79,80,81,82,83,84,85]. The divergence in the aforementioned findings stems from substantial cross-country and cross-regional differences in planning regulation intensity, governance models, and supporting policy systems. Unlike market-oriented and flexible spatial governance instruments prevalently adopted in many foreign countries, China’s territorial spatial regulation rests on distinctive institutional arrangements including administrative zoning management, government-led land expropriation, and fiscal-based ecological compensation, evolving into a governance logic featured as intensive intervention, differentiated zoning and targeted compensation. Nevertheless, China’s unique institutional context only affects the magnitude and formative pathways of income disparities, it does not change the core mechanism linking spatial zoning to income distribution, thereby granting these empirical findings moderate external validity and generalizability.

This study reveals that households in urban development zones earn significantly higher total and wage incomes than those residing in wetland conservation zones. However, the contribution attributable to land development opportunities remains relatively modest, and no statistically significant difference is observed in within-group total income inequality between the two regions. The above findings do not indicate that land development opportunities exert weak influences on rural household income and income inequality. Instead, these results stem from the superposition and countervailing effects of diverse income components. Specifically, increased non-farm employment in urban development zones narrows household wage income disparities, while dividends from land development, industrial agglomeration and asset appreciation simultaneously aggravate property and business income polarization, ultimately rendering overall income inequality statistically insignificant. Empirical results demonstrate that planning-induced land development opportunities have a significant restructuring effect on the income composition of rural households. Spatial governance should focus on achieving employment equity and balanced asset income growth, so as to mitigate structural income polarization induced by land development.

Metropolitan suburbs represent the most pronounced spatial context for such planning-induced disparities in land development opportunities. Our selection of the Sino-French Eco-City and Xiaosi Township serves as a paradigmatic case to reveal the impacts of planning-induced land development opportunities on rural household incomes, providing evidence for coordinated policy design. It should be noted that several limitations remain in the present study for further improvement. First, the empirical verification of theoretical mechanisms is inadequate. Although the theoretical framework proposes several pathways including optimized labor allocation and improved public infrastructure services, this study only examines the impacts of land development opportunities on rural household income and income inequality, without conducting empirical tests for the aforementioned intermediate pathways. Follow-up research can expand survey indicators and adopt mediation effect models to systematically quantify the magnitude of each channel, thereby refining the theoretical mechanism through which planning regulation influences rural household income. Second, this study is constrained by unavoidable locational heterogeneity and the unobserved black-box problem concerning core explanatory variables. We acknowledge that planning intensity is potentially confounded by locational attributes, resource endowments, and macroeconomic policies. Although efforts were made to control for these variables as much as possible during the selection of survey sample sites, it is acknowledged that this sampling study does not constitute a quasi-natural experiment. Furthermore, the lack of pre-regulation income data for rural households makes it difficult to precisely disentangle the effects of planning regulations from those of other confounding factors. Future research will adopt multi-wave survey data and quasi-natural experimental models such as the difference-in-differences approach to further disentangle inherent regional endowment disparities from the net effects of planning policies, thereby unpacking the black box of core influencing variables.

6. Conclusions

Based on a random sample of 573 rural households in Caidian District, Wuhan, this study quantitatively evaluates the impacts and contributions of planning-induced land development opportunities on rural household income and income inequality. The main findings are as follows: (1) Compared with the wetland conservation zone (Xiaosi Township), land development opportunities in the urban development zone (Sino-French Eco-City) significantly increase rural households’ total income, wage income (specifically local employment income), and transfer income, and these findings are robust to a series of robustness tests. Although household labor endowments and economic status dominate the growth of total and wage income, planning-induced land development opportunities explain 70.68% of the local employment income gap and 86.77% of the agricultural business income gap between the two regions. (2) The Sino-French Eco-City exhibits lower overall income inequality, with a Kakwani index of 0.2419, compared with 0.2772 in Xiaosi Township. Notably, land development opportunities reduce local employment income inequality with a contribution rate of 64.12%, but exacerbate inequality in business, property, and transfer incomes. (3) Land development opportunities exert dual effects on rural households with low-quality but high-quantity human capital, promoting their local employment income growth and alleviating corresponding income inequality.

These findings yield important policy implications. The inequalities in development rights induced by spatial regulation call for corresponding economic coordination mechanisms. First, a co-construction and benefit-sharing mechanism should be established to redistribute land development gains towards ecological conservation zones and agricultural cultivation areas via policy tools such as land taxes and ecological compensation payments. Second, protected areas should shift from fiscal dependency to ecological industry empowerment. Local governments can foster industries such as eco-tourism and leisure agriculture to create stable local employment opportunities, while scaling up fiscal support for vocational skill training, infrastructure construction, and public services. Third, urban development zones should fully utilize their inclusive employment advantages, improve the local non-agricultural labor market, and strengthen regulatory oversight of land transfer and housing rental markets to protect disadvantaged farmers’ asset and property rights.