The utilization of APS has the potential to bolster agricultural net income through cost savings and improved agricultural productivity, thereby aligning with the objectives of increasing farmers' total income. The strength of APS may lie in the integration of advanced technologies into every stage of agricultural production, encompassing crop breeding and selection, soil testing for fertilizer formulations, pest control strategies, and grain harvesting techniques (Sahoo et al., 2021). On one hand, the application of these advanced technologies may contribute to a reduction in fertilizer usage (Chen et al., 2023) and labor input, thereby resulting in cost savings for agricultural production. On the other hand, the implementation of these advanced technologies has the potential to enhance agricultural output, consequently improving agricultural productivity. However, the improving-agricultural-net-income effect of APS may be heterogeneous at the farm household level. There are two theoretical considerations.

First, there exists discernible heterogeneity in terms of both timeliness and quality levels in the supply of APS, particularly when comparing large-scale farmers with small-scale farmers. This discrepancy may primarily be attributed to the substantial differences in land scale between these two types of farmer households. Notably, the likelihood of receiving timely and premium APS offerings is significantly higher for expansive and level farmland within the APS market. One fundamental rationale for this observation stems from the fragmentation of farmland, resulting in APS suppliers incurring elevated communication costs and service costs (such as those related to fuel expenditures, notably oil costs) when servicing multiple small-scale farmers at the same land scale level, in contrast to their large-scale counterparts. Consequently, APS suppliers may tend to prioritize large-scale farmers, as they offer greater profitability potential under the same APS pricing structure. Furthermore, it is noteworthy that large-scale farmers typically maintain farmland that exhibits greater uniformity and standardization when contrasted with their small-scale counterparts. This characteristic facilitates the implementation of standardized operations by APS providers for large-scale farmers, thereby enabling the provision of enhanced premium services.

Second, there exists heterogeneity in the technical content and quantity of APS adopted by large-scale farmers and small-scale farmers. In comparison to small-scale farmers, large-scale farmers typically possess a more substantial reservoir of human capital and financial resources at their disposal. This advantageous position equips them with the capability to cultivate a deep understanding of APS and the financial capacity to invest in advanced APS technologies. Moreover, it is worth noting that large-scale farmers frequently exhibit a higher degree of land consolidation and standardization. This characteristic augment their ability to effectively incorporate larger quantities of APS and readily adopt advanced APS technologies, including but not limited to machine transplanting and drone-based spraying. These technologies, in turn, contribute to improved agricultural productivity. Consequently, large-scale farmers are better positioned to achieve higher agricultural net income levels than small-scale farmers. In light of these considerations, this study proposes Hypothesis 1. Hypothesis 1 (H1): APS increase the total income of both large-scale and small-scale farmers primarily by boosting agricultural net income. However, the impact of APS on agricultural net income is expected to be greater for large-scale farmers compared to small-scale farmers.

APS have the potential to enhance wage-business income by exerting influence on the labor market, and subsequently augment total household income. APS exert influence on the labor market through two distinct mechanisms. In the first instance, APS has the potential to augment labor supplementation. This phenomenon arises from the substitution of manual labor with machine-based engineering technology in the framework of APS (Wang et al., 2016). This substitution results in substantial reductions in agricultural labor inputs. The labor thus saved may subsequently be channeled back into the labor market, spanning both the agricultural and non-agricultural sectors. This may involve engagement in hired work at a nearby farm or agricultural enterprise, or the management of small retail establishments, among other possibilities.

Second, APS may have generated supplementary employment prospects within rural areas. It is pertinent to consider that APS play a pivotal role in broadening the array of employment opportunities available to the family labor force. These opportunities encompass various roles, including but not limited to those of agricultural machinery operators, agricultural machinery maintenance personnel, and transportation workers of agricultural products. Consequently, APS has the potential to lead to an increase in wage-business income for farm households. Nevertheless, the impact of APS on wage-business income improvement may vary at the level of farm households. As farm scale increases, factors such as marketing channels and farm management assume heightened significance in influencing agricultural net income. Consequently, the number of tasks for farmers tends to grow. Consequently, the time saved from reduced participation in agricultural production activities due to the adoption of APS is more likely to be allocated to performing these tasks. Essentially, this implies that large-scale farms have limited availability to engage in activities that generate wage-business income, while small-scale farmers possess relatively more time to dedicate toward pursuing wage-business income-generating endeavors. Therefore, this study proposes the Hypothesis 2.

Hypothesis 2 (H2): APS is expected to positively influence the total income of all types of farm households by promoting an increase in wage-business income. This positive effect is anticipated to be more pronounced among small-scale farmers compared to large-scale farmers.

The expected signs of the effect of APS on income inequality within rural areas are indeterminate. It is evident that APS have a more substantial positive influence on the agricultural net income of large-scale farmers compared to small-scale farmers. However, the positive effect on the wage-business income of large-scale farmers is relatively less pronounced when contrasted with that of small-scale farmers. Whether the net effect of APS on income inequality within rural areas is positive or negative is contingent upon empirical investigation. Therefore, this study proposes Hypothesis 3.

The dataset employed in the empirical analysis was procured through a household survey conducted in the year 2017. The main purpose was to investigate the relationship between APS and the livelihood strategies of farmers within the framework of rural land reform. A multistage sampling procedure was applied to select the households. Initially, this study chose two typical counties in teams of both APS development and marketable grain production. They are Sheyang County, located in Yancheng City and Haian County, located in Nantong City. Subsequently, all townships within the aforementioned sample county were divided into high-, medium- and low-level groups based on the per capita agricultural net income level. Three townships were randomly selected from each group, and then 18 townships were included in the sample. Finally, all villages in each sample township were divided into high-, medium- and low-level groups according to the per capita agricultural income, two villages were selected from each group, and then 36 villages were selected. Finally, within each chosen village, a random sampling technique was used to select between 10 and 12 households for inclusion in the investigation. In total, the survey encompassed 368 households in the final dataset, after excluding households not actively involved in farmland cultivation. Of this sample, 355 households reported the utilization of APS practices in the year 2016, while the remaining 13 households confirmed the absence of APS adoption in the same year.

Table 1 displays the descriptive statistics of this sample. Among all the farm households included in this study, the average total income per capita in 2016 was 52,890 CNY. Of this, 52% was derived from agricultural production, while 37% came from off-farm business activities and employment. It indicates that the primary sources of farmers' total income are agricultural net income and wage-business income. On average, farm households' expenditure on APS accounts for 31.4% of the total costs of household agriculture production. The average age of agricultural production decision-makers was 58.85 years, with 20.1% of them being members of the CPC. The households had an average of 3.546 family members, received an average of 8.27 years of education, and possessed agricultural equipment valued at 126,180 CNY on average. In 2016, 91.3% of these households received agricultural subsidies.

Our sample comprises 60 large-scale farmers and 308 small-scale farmers. Significant differences are observed in the mean values of various variables across the two groups. Specifically, the average per capita total income of large-scale farmers exceeds that of small-scale farmers by 201,124 CNY. In terms of specific components, large-scale farmers exhibit an average per capita agricultural net income that is 156,470 CNY higher than that of small-scale farmers, along with an average per capita wage-business income that is 28,250 CNY greater. Furthermore, on average, the proportion of APS costs paid by large-scale farmers is 12.5 percentage points lower than that paid by small-scale farmers. In terms of other characteristics, the agricultural production decision-makers of large-scale farmers are, on average, 10.48 years younger and are 4.1% less likely to be members of the CPC when compared to small-scale farmers. Large-scale farmers, on average, comprise 0.72 more family members and receive 1.53 more years of education than small-scale farmers. Additionally, large-scale farmers are less likely to receive agricultural subsidies than small-scale farmers, and the value of agricultural equipment in large-scale farmers average 764,960 CNY more than that in small-scale farmers.

The estimated results of the income determination function are presented in Table 2. The results of Model 1 represent the baseline regression results. The empirical results indicate that APS has a significantly positive effect on both per capita total income and per capita agricultural net income, with significance levels of 1 and 5%, respectively. It's consistent with the inference of hypothesis H1. This finding suggests that households allocating a greater portion of their agricultural production expenditure to APS observe a corresponding increase in their total income and agricultural net income. Additionally, APS has a significant effect on per capita wage-business income, aligning with the expectations outlined in hypothesis H2. As household heads age, farm households tend to generate lower income across all three income types. If the household head is a member of the CPC, the household may potentially realize enhanced financial returns from agricultural production. Nonetheless, there exists no statistically significant disparity in per capita total income and per capita wage-business income between households led by CPC members and those led by non-members. Furthermore, households with more years of education tend to report increased total income and wage-business income. An expanded family size is linked to heightened total income and wage-business income, concomitant with a decrease in agricultural net income. Additionally, Households possessing higher-value agricultural equipment or receiving agricultural subsidies tend to exhibit higher total income and agricultural net income.

This study uses the interaction term of APS and farm household types to examine the heterogeneous effects of APS on the three types of income between large-scale and small-scale farm households. To this end, Model 2 involves this interaction term based on Model 1. The outcomes of estimation indicate that, as the proportion of APS expenditure rises by one-unit, significant disparities emerge in three types of income between large-scale and small-scale farmers. Specifically, these findings demonstrate that large-scale farmers realize an increase of 224.7% in total income per capita and a 257.3% surge in agricultural net income per capita compared to their small-scale counterparts. This disparity likely arises because larger-scale farmers, endowed with superior human capital, are better able to identify and adopt APS packages that are optimally timed and tailored to their production conditions, thereby securing disproportionately higher returns. In addition, the scale economies of land amplify the marginal net agricultural benefits generated by APS interventions. These results align with the anticipated outcomes posited in hypothesis H1. Conversely, concerning wage-business income per capita, large-scale farmers experience a 297.9% reduction compared to small-scale farmers under the same conditions, consistent with hypothesis H2.

The results of income inequality decomposition are presented in Table 3. Based on the results reported in the first line of Table 4, the income determination function can explain 72.98% of the income inequality generation. This indicates that the variables involved in the income determination function can explain the income inequality well. The focus of this study lies in the effect of APS on rural income inequality. The decomposition results reveal that the contribution of APS to income inequality is −1.35%. This finding suggests that APS has the potential to narrow rural income inequality, albeit with potential limitations on its effectiveness. Considering the results from Model 2, a plausible explanation emerges: the incremental improvement brought about by APS on the agricultural net income of large-scale farmers is less pronounced than its impact on the wage-business income of small-scale farmers.

Farm household types contribute over 50% to rural income inequality, ranking highest among contributing factors. According to the definition of farm household types in this paper, this finding implies a pivotal role for cultivated farmland size in the widening of rural income inequality. However, alternative explanations for this phenomenon may exist. For instance, research suggests that differences in available resources and skill levels among farm households of varying operational scales may result in disparate impacts of income improvement through the adoption of new technologies, thereby potentially exacerbating rural income inequality (Shita et al., 2020). Section 4.1 establishes significant distinctions in characteristics between large-scale and small-scale households. Nevertheless, further research is required to ascertain whether and why these differences in characteristics influence rural income inequality. Among the other factors in the decomposition model, factors such as agricultural devices and the average education level of family members emerge as noteworthy contributors to rural income inequality, with respective shares of 19.45 and 20.50%. These findings suggest that agricultural devices and human capital play a significant role in shaping rural income inequality. The aforementioned finding once again underscores the significant influence of education on income inequality, aligning with the findings of Morduch and Sicular (2002). The remaining factors are beyond the scope of this study; thus, this conclusion is presented without providing detailed explanations. The variables of township dummies controls, age of household head, family size, CPC membership, and agricultural subsidy exhibit a relatively smaller impact on income inequality.

Given the grounding of variable selection in prior literature and local context, the representativeness of farm household types and income inequality variables necessitates meticulous attention. To assess the robustness of regression outcomes, this study employs a new measurement methodology for both farm household types and income inequality in robustness testing. For the classification of farm household types, this study adopts the categorization system established by the Food and Agriculture Organization of the United Nations (FAO). This system utilizes a delineation threshold of 30 mu (2 ha) to demarcate distinct categories among farm households. This methodological adjustment aims to enhance the precision and robustness of the analysis by aligning with internationally recognized standards in the characterization of farm households. Table 5 reports the results of Model 2 subsequent to the alteration in the measurement of farm household types.

These findings align consistently with the empirical results. For the income inequality variable, this study employs two additional widely acknowledged indicators of income inequality, namely the Atkinson index and the Generalized entropy index. These new income inequality indicators are used in Equation 3. Table 4 reports the explanatory capacity of variables involved in the income determination function on the income inequality generation.

Notably, these variables exhibit a robust explanatory capacity for income inequality. Moving forward to Table 6, this study elucidate the contributions of variables to income inequality subsequent to the alteration in the measurement of income inequality. The revised ranking of variables in Table 6 maintains congruence with empirical findings.

Based on the aforesaid discussion, the study findings reinforce the global relevance of APS in advancing equitable agricultural growth. National strategies like China's land trusteeship and international examples such as CHCs in India or service cooperatives in Africa exemplify scalable models (Attipoe et al., 2021; World Bank, 2023). To realize the full potential of APS in promoting common prosperity, targeted policy support, investment in rural infrastructure, and institutional innovations are needed globally.

These findings indicate that it could generate larger distributional dividends by tailoring the APS framework to overcome the distinctive constraints of smallholders in order to compress intra-rural inequality and rural-urban income divide. However, constrained by remote location (Megerssa et al., 2020; Lusike et al., 2023), underdeveloped agricultural infrastructure (Ai et al., 2023), digital divide and limited agronomic expertise (Yang et al., 2023; Zhou et al., 2023), and weak market leverage (Agholor et al., 2023), small-scale farmers confront multiple barriers to APS utilization, ranging from constrained access channels and difficulties in locating or bargaining with service providers to highly variable service quality. Consequently, the anticipated role of APS in advancing common prosperity remains largely unrealized.

This study offers critical policy insights on the need to widely promote APS to advance global common prosperity. According to the 2020 report by the International Land Coalition and Oxfam, 2.5 billion agricultural smallholders operate under stark land disparities. Moreover, 1% of the largest farms control over 70% of global agricultural land, while 80% of smallholder operate < 2 hectares. Given the limited purchasing power of smallholder farmers, price signals alone cannot be relied upon to induce the market to supply them with high-quality APS. Consequently, constructing an APS system that is genuinely smallholder-friendly becomes a policy imperative, particularly in most of Asia and Africa, where smallholders constitute the majority of agricultural producers and are simultaneously exposed to entrenched poverty and heightened risks of re-impoverishment, such as China, Ethiopia, Kenya, South Africa. Therefore, targeted policies are essential to ensure that smallholders in these regions can access APS effectively. On the whole, smallholders are primarily constrained by a triad of deficits: (i) reliable information channels, (ii) advanced domain-specific expertise coupled with digital literacy, and (iii) conventional endowments of capital and natural assets—namely agricultural infrastructure, irrigation water, arable land, and financial reserves. Consequently, governmental integration strategies must proceed along two mutually reinforcing tracks. First, the architecture of the APS framework must be deliberately calibrated to the heterogeneous characteristics of smallholders. Second, this framework must be co-deployed with a policy portfolio that encompasses fiscal support instruments, infrastructure upgrading programs, soil and water resource governance schemes, and targeted capacity-building initiatives in professional training. Concretely, governments can construct a smallholder-friendly APS framework along four interlocking axes.

First is informational framework. APS will turn smart, precise, and green. Big-data and AI will map smallholders' information gaps-water, soil, weather, markets (Kumar et al., 2017). These information technologies algorithmically translate the resulting insights into bespoke agronomic prescriptions, context-sensitive production services, and evidence-based managerial guidance, thereby strengthening smallholders' adaptive capacity to climatic variability and market volatility.

Second is organizational framework. The emerging APS is best conceived as a “platform-plus-network” configuration that aggregates and reallocates agricultural resources and knowledge. Governments should (a) deploy 5G, cloud computing, and the internet of things to build a digital service platform that seamlessly links smallholders with technology, finance, and markets, enabling resource pooling and optimal allocation (Zhang et al., 2022), and (b) institute a collaborative organizational matrix in which agricultural service providers, intermediary organizations, and farm households form a triangular relationship centered on service delivery, cost minimization, and benefit sharing (Yang et al., 2024). Exemplary of this model are Poland's Agricultural Chambers, which function as bidirectional conduits for transmitting farmers' opportunities and constraints to policymakers while simultaneously disseminating policy information downward (Kasprzyk et al., 2024).

Third is precision-oriented training services. Smallholders juggle subsistence. They often overlook regional crises—climate shocks (Ohenhen et al., 2024), aquifer depletion (Gupta et al., 2022a), chemical runoff (Gupta et al., 2022b). These messages travel digitally. Low digital literacy locks smallholders out. Such unawareness dulls their interest in eco-friendly APS. Therefore, governments should deliver grounded, demand-driven training programs—most notably, capacity-building in digital literacy for smallholders and widespread diffusion of the modern agricultural technologies embedded within APS.

Fourth is aligned fiscal and financial instruments. Fiscal and financial policies must be synchronized with agricultural development and land-use policies to generate coherent policy synergy (Zou et al., 2023). Governments can deploy targeted subsidies and institutional incentives to induce financial institutions to extend concessional loans to both smallholders purchasing APS and the service organizations supplying them (Qiu et al., 2023). Additionally, the introduction of index-based insurance mechanisms can help service providers hedge against natural-disaster risks, thereby stabilizing their operations (Lydiah et al., 2023).

This study is limited by its regional scope and dataset size. It does not fully capture the broader dimensions of common prosperity such as urban-rural income divide or the differential impacts of APS on various income sources across farm types. Future research should leverage large-scale macroeconomic and cross-national data to deepen understanding of APS impacts across diverse socio-economic and agro-ecological contexts.