Introduction

Over the past few decades, increased fertilizer input has played a vital role in boosting agricultural yields and ensuring food security (Jiang et al., 2018; Ma and Zheng, 2022). Globally, the excessive and inefficient use of chemical fertilizers has become a growing concern due to its environmental and ecological impacts. International organizations, including the FAO and the United Nations, have emphasized improving fertilizer use efficiency as a key priority for achieving sustainable agriculture. Studies have shown that only 30–50% of applied nitrogen is absorbed by crops (Lehnert et al., 2018), with the excess leading to nutrient runoff, soil and water contamination, and greenhouse gas emissions (Reay et al., 2012; Zhang et al., 2016; Yang et al., 2019; Walling and Vaneeckhaute, 2020; Wang et al., 2023). Addressing fertilizer-related pollution is essential not only for sustainable farm development but also for global environmental and food system goals. Improving fertilizer use efficiency has therefore become a priority strategy (Abler, 2015; Huang and Jiang, 2019; Tang et al., 2019; Lin et al., 2022).

In China, fertilizer use efficiency remains considerably low—estimated between 25% and 35%—compared to over 60% in North America and Europe (Ju et al., 2009; Zhang et al., 2015; Lassaletta et al., 2016). To address this, China has implemented a series of reforms combining technological innovation and institutional change, such as the “Action Plan to Achieve Zero Growth of Chemical Pesticides and Fertilizers by 2020” (Ding et al., 2018). Meanwhile, studies have highlighted the role of land management scale expansion and market-oriented agricultural services in reducing fertilizer input while improving efficiency (Zhu et al., 2022). These trends underscore the relevance of structural agricultural policies to fertilizer use efficiency.

High-standard farmland construction has recently emerged as a flagship infrastructure policy aimed at improving land quality, consolidating fragmented plots, and promoting mechanized, environmentally sound farming (Li et al., 2023; Ye et al., 2023). The policy integrates land leveling, water conservancy projects, and soil improvement to support high-yield, resilient agricultural systems (Pu et al., 2019; Song et al., 2019; Xu et al., 2022). As such, it is widely expected to enhance not only land productivity, but also input efficiency—particularly fertilizer utilization. Key components include terrain modification and irrigation infrastructure (Yin et al., 2022; Su et al., 2019), both of which are theorized to improve fertilizer absorption and reduce waste.

The literature related to this article mainly focuses on the following aspects. The first encompasses literature on the methodologies for measuring fertilizer use efficiency, principally employing Stochastic Frontier Analysis (SFA) and Data Envelopment Analysis (DEA). While SFA is a parametric approach requiring the specification of a production function—commonly Cobb-Douglas or Translog—DEA is a non-parametric method less constrained by functional forms but sensitive to data outliers (Wu, 2011, Ma et al., 2014, Abu, 2011). Additionally, a segment of the literature investigates the determinants of fertilizer efficiency, including variables like farm size, farmland costs, urbanization levels, climate change impacts, and educational attainment of the workforce (Zhu et al., 2022, Qi et al., 2022, Pete et al., 2006, Amouzou et al., 2018).

In China, government policy interventions are commonly used to improve fertilizer use efficiency (Wu, 2011). High-standard farmland construction is a typical agricultural infrastructure policy (Ye et al., 2023). Theoretically, the construction of high-standard farmland can improve soil quality and increase the scale of agriculture, thus increasing the efficiency of fertilizer use. However, there are no relevant empirical studies to conduct an in-depth analysis. Existing studies have mainly focused on the output and income effects of the high-standard farmland policy. For example, Li et al. (2023) find that the construction of high-standard farmland significantly increases food production and reduces agricultural carbon emissions by improving resource agglomeration. Similarly, Pu et al. (2019) report that the policy effectively upgraded low- and medium-yield fields in Liaoning Province, thereby improving land use efficiency and farmers’ incomes.

Other studies have highlighted the role of high-standard farmland in promoting disaster resilience and resource conservation. Xu et al. (2025) demonstrated that the policy improved farmers’ recycling behaviors and ecological awareness, while Yin et al. (2022) emphasized how ecological land use improvement facilitated farmers’ participation in environmental protection. Recent studies have further extended the research on high-standard farmland’s effectiveness. Hao et al. (2024) found that high-standard farmland construction has promoted grain production across various regions in China. Liu and Zhang (2024) examined the spatial and temporal variation in the efficiency of high-standard farmland construction, revealing both convergence and divergence trends in its effectiveness. Li et al. (2025) further explored how high-standard farmland construction affects farmland abandonment by farm households, shedding light on the policy’s social implications. Additionally, Sun et al. (2025) highlighted the positive impact of improving element quality on the comprehensive production capacity of grain, reinforcing the importance of soil quality in enhancing agricultural output. Despite these findings, relatively few studies have assessed the environmental impacts of the policy from the perspective of fertilizer use efficiency, leaving an important gap in understanding its ecological effects.

This study attempts to fill this gap by empirically investigating whether and through what mechanisms the construction of high-standard farmland has improved fertilizer use efficiency. This study makes three key contributions to the literature. First, it shifts the focus from traditional fertilizer application volume to fertilizer use efficiency, providing a more comprehensive understanding of agricultural sustainability and the effectiveness of high-standard farmland construction. Second, it evaluates the dynamic and long-term impact of the policy, using a continuous difference-in-difference model to capture the policy’s evolving effects over time. Third, the study introduces land quality and agricultural scale as mediators between the policy and fertilizer use efficiency, offering new insights into the mechanisms through which improvements in land infrastructure and farm scale drive fertilizer efficiency, an area that has been underexplored in existing research.

Theoretical framework and research hypotheses

Within the framework of the high-standard farmland construction policy, measures such as land leveling, irrigation facility construction, and soil improvement aim to optimize the agricultural production environment. These initiatives not only enhance farmland productivity by modernizing infrastructure, but also provide a foundation for improving fertilizer use efficiency. From the perspective of agricultural production economics, these improvements function as non-fertilizer input enhancements that increase the marginal productivity of fertilizer (Zhang et al., 2013; Ju et al., 2016), thereby allowing farmers to produce more output with the same or even reduced fertilizer input. As the core of the policy, high-standard farmland construction directly creates conditions for improving land quality and promoting large-scale farming through infrastructure optimization. However, the effectiveness of these interventions is expected to vary across regions due to differences in development levels, infrastructure, and implementation capacity (Liu and Zhang, 2024; Sun et al., 2025). Therefore, we construct the theoretical analysis framework of Fig. 1.

Fig. 1
figure 1

Theoretical analysis framework diagram.

Full size image

A key objective of high-standard farmland construction is to improve production conditions and environmental quality. By enhancing land quality, farmers can achieve higher crop yields with the same amount of fertilizer, directly increasing fertilizer efficiency. Several factors contribute to this improvement (Oyetunji et al., 2022). First, soil structure is enhanced through land leveling, soil improvement, and the upgrading of irrigation systems. These measures improve soil water retention and fertility, reduce nutrient loss, and enable crops to absorb fertilizer more efficiently, thus raising fertilizer utilization (Lal, 2004; Zhang et al., 2013; Sun et al., 2025). Second, soil fertility is improved through scientific land management, which helps maintain organic matter and nutrient supply (Ju et al., 2016). This reduces farmers’ reliance on chemical fertilizers and enables more precise fertilization. Third, the modernization of irrigation systems, such as drip and sprinkler irrigation, promotes the wider use of water-fertilizer integration technology. This technology adjusts fertilizer application based on crop needs, reducing waste and improving fertilizer efficiency (Gu et al., 2017).

In addition to improving land quality, high-standard farmland construction facilitates large-scale farming, which further enhances fertilizer efficiency by optimizing resource allocation and improving crop management. Several mechanisms contribute to this (Wu, 2011; Li et al., 2023;Xu et al., 2025). First, large-scale farming reduces unit costs. Farmers benefit from centralized purchasing and mechanized operations, which lower the cost of fertilizer and enable more accurate application, minimizing over-fertilization and improving efficiency. Second, large-scale operations accelerate the dissemination of agricultural technologies, such as soil testing, formula fertilization, and intelligent management. These technologies are more easily adopted across larger areas, enabling precise fertilization and reducing waste (Ye et al., 2023). Third, intensive management and specialized services are promoted through farmland transfer and farm expansion. Farmers and cooperatives gain access to specialized agronomic services, such as soil testing and scientific irrigation techniques, which further improve fertilizer efficiency and minimize unnecessary use. Accordingly, we propose hypotheses 1–3:

H1: High-standard farmland construction policies can significantly improve fertilizer use efficiency.

H2: High-standard farmland construction can improve the fertilizer use efficiency by improving land quality.

H3: High-standard farmland construction can improve fertilizer use efficiency by promoting large-scale agriculture.

Due to substantial regional disparities in natural resource endowments, agricultural development, and economic infrastructure, the impact of high-standard farmland construction on fertilizer use efficiency is expected to exhibit significant heterogeneity. This heterogeneity manifests along three main dimensions:

First, from a geospatial perspective, the policy effect is more pronounced in eastern provinces, which benefit from better infrastructure, stronger institutional capacity, and a higher degree of agricultural modernization (Xu et al., 2022). In contrast, western regions face natural and infrastructural constraints that limit the effective realization of policy goals (Li et al., 2023), leading to a clear gradient of “strong in the east, weak in the west”. Second, in terms of agricultural functional zoning, the policy impact is stronger in major grain-producing areas, which are the primary focus of high-standard farmland construction aimed at ensuring food security. In these regions, infrastructure improvements are more closely aligned with staple crop production, thereby enhancing fertilizer efficiency more effectively (Hao et al., 2024). Conversely, in non-grain-producing areas with more diversified cropping systems, the policy exhibits weaker effects. Third, heterogeneity is also shaped by the initial baseline level of fertilizer use efficiency. Regions with higher initial efficiency levels—often equipped with stronger absorptive capacity for technology and better infrastructure—tend to benefit more rapidly, reinforcing a “Matthew effect” in policy outcomes (Zheng et al., 2021; Ye et al., 2023). This dynamic risks widening interregional gaps in fertilizer efficiency over time. These heterogeneous characteristics highlight the need for region-specific and precision-oriented policy design, ensuring that the benefits of high-standard farmland construction are both effective and equitably distributed. Accordingly, we propose hypotheses 4–6:

H4: The effect of high-standard farmland construction on fertilizer use efficiency shows East-Central-West heterogeneity.

H5: The fertilizer use efficiency effect of high-standard farmland construction is more significant in main grain producing areas, while the effect is not significant in non-main grain producing areas.

H6: Policy effects increase incrementally with regional levels of fertilizer use efficiency and may exacerbate interregional efficiency gaps.

Materials and methods

Contextualizing high-standard farmland construction policy

Development of high-standard farmland construction

High-standard farmland construction is essential to China’s land improvement program (Dong and Ju, 2023). In 1998, China’s State Council decided to set up a land development and construction fund. It began to explore the transformation of medium- and low-yield fields and the construction of high-standard farmland. In 2005, the No.1 document of the central government of China issued in January put forward the idea of creating an excellent ecological barrier for the construction of high-standard farmland and mentioned the concept of high-standard farmland for the first time. In 2006, China pushed forward the demonstration project of high-standard basic farmland in some counties with large grain growing areas and carried out the land improvement covering eight aspects, including fields, soil, water, roads, forests, electric power, technology, and management, to upgrade and reform the agricultural production conditions of the existing basic farmland and low and medium-yield farmland. These initiatives are all part of the exploratory phase of high-standard farmland construction. In 2012, the Chinese government promulgated the “High-standard Basic Farmland Construction Standards (for Trial Implementation),” fully implemented nationwide in 2013. In the same year, the Chinese government compiled the General Rules for the Construction of High-standard Farmland, which stipulates the basic principles, construction areas, technical requirements, acceptance standards, and other detailed requirements for constructing high-standard farmland. In accordance with the Chinese government’s official documents and the timeline of policy execution, this study categorizes the development of high-standard farmland into two distinct phases: the exploratory stage, occurring prior to 2013, and the standardized implementation stage, commencing in 2013 and onwards. To gain a comprehensive understanding of high-standard farmland construction in China, this study collates data on the area dedicated to such construction over several years, as depicted in Fig. 2. The trends illustrated in Fig. 2 show an initial increase followed by a decrease in the area of high-standard farmland construction across the 2008–2017 period. Notably, the central region exhibits a higher extent of construction compared to the eastern and western regions. Figure 2 also shows the area of high-standard farmland construction and corresponding fertilizer use in China from 2008 to 2017. The indicators represent the total area under high-standard farmland construction in each province, alongside the fertilizer use per hectare. These values help illustrate the correlation between policy implementation and fertilizer efficiency.

Fig. 2
figure 2

Area of high-standard farmland construction and fertilizer use in China from 2008 to 2017.

Full size image

Theory of change for high-standard farmland construction

The theory of change behind high-standard farmland construction explains how this policy intervention enhances fertilizer use efficiency—a key objective in sustainable agricultural development. By modernizing farmland infrastructure, including land leveling, soil improvement, irrigation systems, and mechanization support, the policy directly addresses soil degradation and nutrient loss (Lal, 2004; Gu et al., 2017). These improvements create optimal conditions for precision farming, leading to higher fertilizer absorption rates, reduced environmental runoff, and increased crop yields (Zhang et al., 2013; Ju et al., 2016).

However, the policy’s success hinges on coordinated action among stakeholders. Farmers must see tangible benefits, such as lower input costs and higher productivity, to adopt new practices. Agricultural service providers require adequate infrastructure to deliver technical support, including soil testing and tailored recommendations. Meanwhile, local governments play a critical role in funding, monitoring, and maintaining these systems (Sun et al., 2025; Xu et al., 2025). A breakdown in any part of this chain—whether due to poor implementation, lack of training, or weak interagency coordination (Liu and Zhang, 2024)—could undermine efficiency gains.

Ultimately, high-standard farmland construction must translate physical upgrades into lasting behavioral and technical changes at the farm level. Only through integrated efforts—combining infrastructure, knowledge transfer, and farmer engagement—can the policy achieve its dual goals of boosting agricultural productivity and reducing environmental harm (Abler, 2015; Huang and Jiang, 2019). Sustained investment, institutional collaboration, and post-construction support are essential to maximize its impact.

Methodological approaches

Analytical model for estimating fertilizer use efficiency

To estimate fertilizer use efficiency, we adopt the Stochastic Frontier Analysis (SFA) approach, which allows us to separate random noise from technical inefficiency(Battese and Coelli, 1992; Battese and Coelli, 1995). Specifically, we use a Cobb-Douglas production frontier, where grain yield serves as the output, and fertilizer input is one of the core inputs alongside land, labor, and capital. The inefficiency term captures deviations from the optimal fertilizer use given the production frontier. This method is widely used in agricultural efficiency studies and provides a robust foundation for constructing a comparable fertilizer use efficiency indicator across regions (Hu et al., 2019). The analytical model is defined as:

$${y}_{{it}}=f({x}_{{it}},t,\beta )\exp ({v}_{{it}}-{u}_{{it}})$$
(1)

In Eq. (1), x denotes the vector of input variables, encompassing land, labor, machinery, fertilizer, and irrigation inputs. The variable y signifies output, i is for province, t represents time, and β is the parameter to be estimated. The terms v and u represent random error and technical inefficiency, respectively, and are assumed to follow normal distributions \(v \sim N\left(0,{\sigma }_{v}^{2}\right)\) and \(u \sim N(u,{\sigma }_{u}^{2})\). The variance function \(r=\frac{{\sigma }_{u}^{2}}{{\sigma }_{u}^{2}+{\sigma }_{v}^{2}}\) is subsequently tested, and maximum likelihood methods are employed for parameter estimation. Fertilizer use efficiency is then calculated as:

$${{FUE}}_{{it}}=\frac{{F}_{{it}}^{* }}{{F}_{{it}}}=\exp \left(-\frac{{u}_{{it}}}{{\beta }_{F}}\right)$$
(2)

Equation (2) designates FUE as fertilizer use efficiency, F∗ as the optimal fertilizer use, F as the actual fertilizer use, and βF as the elasticity coefficient of fertilizer input in the production function.

Impact assessment of high-standard farmland construction policy

This section outlines the methodology employed to assess the impact of China’s high-standard farmland construction policy on fertilizer use efficiency. Recognizing the significant regional variations in policy implementation across Chinese provinces, our study approaches the policy as a quasi-natural experiment. To quantify the causal effect of this policy on fertilizer use efficiency, we adopt the continuous Difference-in-Difference (DID) model. This approach is particularly well-suited for our analysis for several reasons. Firstly, the staggered implementation of the high-standard farmland construction policy across provinces and over time aligns well with the classic features of DID models, allowing us to exploit this variation to estimate the policy’s causal effect (Nunn and Qian, 2011; Callaway et al., 2024). Secondly, the continuous DID model effectively addresses data heterogeneity, a concern in our study due to the diverse agricultural practices and resource endowments across provinces. This ensures more reliable and robust estimations of the policy’s impact. Finally, and most importantly, the continuous DID model mitigates the endogeneity issue often associated with policy changes. This issue arises when unobserved factors simultaneously influence both the policy adoption and the outcome variable (fertilizer use efficiency), potentially leading to biased estimates. By employing the continuous DID model, we can control for these unobserved factors and obtain more accurate estimations of the policy’s true effect.

For the purpose of this analysis, provinces are categorized into two distinct groups: the experimental group, consisting of provinces with a proportion of high-standard farmland above the median of all provinces, and the control group, comprised of provinces with a proportion below the median. This classification ensures a clear distinction between provinces with higher and lower levels of policy implementation. The mathematical representation of the model is:

$$FU{E}_{it}={\alpha }_{1}+{\alpha }_{2}DI{D}_{it}+\beta {X}_{it}+{\mu }_{i}+{\theta }_{t}+{\varepsilon }_{it}$$
(3)

In Eq. (3), FUE represents fertilizer use efficiency, DID stands for the policy variable related to the high-standard farmland construction, and X encompasses a set of control variables. The terms μi and θt serve as fixed effects for time and province, respectively. ϵit is the random error term. Both α and β are parameters to be estimated, with α2 being the key parameter representing the policy’s impact on fertilizer use efficiency. Estimations are performed using robust standard errors.

Mediation analysis of policy impact on fertilizer use efficiency

To further explore how the high-standard farmland construction policy influences fertilizer use efficiency, this study undertakes a mediation analysis. Land quality and agricultural scale were selected as mediators, as both are directly impacted by the policy and have been shown to significantly influence fertilizer use efficiency in prior studies (Wu, 2011; Ju et al., 2016; Oyetunji et al., 2022). The high-standard farmland construction policy is designed to improve land quality by addressing issues like land fragmentation and soil degradation, while also promoting the expansion of agricultural operations through land consolidation and infrastructure development. Both improved land quality and larger operational scales have been shown to contribute to increased fertilizer use efficiency. Therefore, it is crucial to examine whether these two factors act as mediating pathways through which the policy exerts its influence on fertilizer utilization.

We employ a series of regression equations to quantify the mediating effects. The analysis is structured as follows:

Step 1: Policy’s Impact on Mediating Variables:

$$L{Q}_{{it}}={\gamma }_{1}+{\gamma }_{2}{DI}{D}_{{it}}+\delta {X}_{{it}}+{\mu }_{i}+{\theta }_{t}+{\epsilon }_{{it}}$$
(4)
$$A{S}_{{it}}={\eta }_{1}+{\eta }_{2}{DI}{D}_{{it}}+\lambda {X}_{{it}}+{\mu }_{i}+{\theta }_{t}+{\epsilon }_{{it}}$$
(5)

In these equations, LQ denotes land quality, defined as the ratio of effective irrigated area to total sown area in each province. AS indicates the scale of agricultural operations, measured by the total area allocated for crop production in each province. The coefficients γ2 and η2 capture the direct effect of the policy (DID) on land quality and agricultural operation scale, respectively.

Step 2: Combined Effect of Policy and Mediating Variables on Fertilizer Use Efficiency:

$${FU}{E}_{{it}}={\rho }_{1}+{\rho }_{2}{DI}{D}_{{it}}+{\rho }_{3}L{Q}_{{it}}+{\rho }_{4}A{S}_{{it}}+{X}_{{it}}\tau +{\mu }_{i}+{\theta }_{t}+{\epsilon }_{{it}}$$
(6)

In this equation, ρ2 represents the direct effect of the policy on fertilizer use efficiency, while ρ3 and ρ4 capture the indirect effects operating through land quality and agricultural operation scale, respectively.

Variable definitions

Dependent variables

The dependent variable for this study is Fertilizer Use Efficiency, computed via the Stochastic Frontier Analysis method. Selection of appropriate input and output variables is essential for the accuracy of this metric. Building upon existing literature (Huang and Jiang, 2019; Fan et al., 2023), the output variable is defined as the Total Agricultural Output Value (Output). The input variables comprise: (1) Machinery Input (Input1), measured as the total power of agricultural machinery in 10,000 kilowatts; (2) Fertilizer Input (Input2), denoted by the amount of chemical fertilizer used in 10,000 tons; (3) Labor Input (Input3), represented by year-end labor force in 10,000 persons; (4) Land Input (Input4), defined by agricultural sown area in 1000 km2; (5) Pesticide Input (Input5), gauged by pesticide usage in 10,000 tons; (6) Agricultural Film Input (Input6), measured in agricultural plastic film usage in 10,000 tons; (7) Diesel Fuel Input, quantified by agricultural diesel fuel usage in 10,000 tons. Price indices are employed to deflate variables involving prices.

Core independent variables

The primary independent variable is the High-Standard Farmland Construction Policy Variable (DID), formulated through the interaction between a time dummy variable for policy implementation and the area of regional high-standard farmland. The coefficient of this variable will signify the policy’s effect on fertilizer use efficiency.

Control variables

Informed by previous research (Pete et al., 2006), this study integrates a range of control variables to ensure a comprehensive analysis: (1) Regional Human Capital (EDU), quantified by the average years of education within the labor force. Higher educational attainment is associated with greater awareness of input efficiency and better adoption of sustainable practices (Amouzou et al., 2018); (2) Urbanization Rate (UR), the urban population’s proportion to the total population. Urbanization will affect fertilizer use efficiency through labor transfer and technological progress (Qi et al., 2022); (3) Financial Support for Agriculture (AF), represented by the total amount of regional agricultural financial aid. Financial support will affect fertilizer use decisions, thereby affecting fertilizer use efficiency(Fan et al., 2023); (4) Economic Level (EL), gauged through the per capita income of rural inhabitants. Higher income levels may encourage farmers to adopt more advanced technologies and efficient input combinations(Zhu et al., 2022); (5) Fertilizer Price (FP), based on the fertilizer production price index. The input price directly affects farmers’ marginal input decisions and cost-effective strategies, which will directly affect fertilizer utilization efficiency (Fan et al., 2023); (6) Agricultural Planting Structure (AS), the ratio of grain-sown area to the overall crop-sown area.Different crop types require different fertilizer levels, and grain-dominated systems tend to be more fertilizer-intensive(Zhang et al., 2013). Data reliability was enhanced by removing outliers and utilizing interpolation for missing data. Table 1 presents a descriptive statistical analysis of this data. Quantitative analysis was conducted using Stata16 (StataCorp LLC, College Station, Texas) software.

Table 1 Descriptive statistics of the main variables.
Full size table

Data sources

This study conducts an empirical analysis across 30 provinces in mainland China, excluding Hong Kong, Macao, Taiwan, and Tibet, due to differences in agricultural practices and policy frameworks. We choose statistics from 2008 to 2017 for our empirical analysis. Due to limitations in the availability of statistical data, the most recent data we were able to collect for this study extends only up to 2017. This limitation arises from the availability of publicly accessible datasets within China, and it is also consistent with the data used in most related research studies, which similarly focus on pre-2017 data. To evaluate fertilizer use efficiency, the research utilizes historical data extracted from the China Statistical Yearbook and the China Rural Statistical Yearbook. In addition, data pertaining to control variables are derived from the China Agricultural Statistical Yearbook and the EPS database, ensuring a thorough and accurate assessment of the policy’s impact within the specified time frame and regions.

Empirical results

Elasticity coefficients and regional fertilizer use efficiency trends in agriculture

To analyze the elasticity of various inputs in agricultural production, this study integrates a Cobb-Douglas production function with a Stochastic Frontier Analysis (SFA) model. The inputs considered include machinery, fertilizer, labor, land, pesticide, agricultural film, and diesel fuel. Table 2 presents the results of the elasticity coefficient estimations. The significance of these coefficients for most inputs validates the model’s robustness. A notable finding is the coefficient of 0.349 for fertilizer input, highlighting its critical role in agricultural output. Excluding land input, all other factors contribute positively to agricultural production, aligning with China’s current agricultural practices. In addition, the study employs the Data Envelopment Analysis (DEA) method for robustness checks, yielding results consistent with the SFA method.

Table 2 Results of the SFA estimation.
Full size table

The study further calculates the fertilizer use efficiency for each province over the years using Eq. (2). The data indicate that, although there is a slight upward trend from 2008 to 2017, the overall fertilizer use efficiency in China remains low, increasing from 0.529 in 2008 to 0.733 in 2017, with an average annual growth rate of 2.26%. This slow growth rate suggests a continuing inefficiency in fertilizer application, leading to substantial resource wastage.

Sub-regional disparities in fertilizer efficiency are also evident. The eastern region exhibits the highest efficiency, followed by the central and western regions. To provide a detailed temporal analysis, the study includes a visual examination of data from 2013 to 2017 (Fig. 3). A closer examination of the data reveals significant changes in the efficiency rankings of various provinces. In 2013, the provinces leading in fertilizer use efficiency were Sichuan, Liaoning, and Beijing. However, by 2017, this landscape had altered, with Sichuan and Jiangsu rising to the forefront as the most efficient provinces.

Fig. 3
figure 3

Changes of fertilizer use efficiency in China.

Full size image

Impact analysis of high-standard farmland construction policy on fertilizer use efficiency

Baseline regression results: examining the policy’s overall impact

This section presents the results of the continuous DID regression analysis, which estimates the overall impact of the high-standard farmland construction policy on fertilizer use efficiency. We employ a stepwise regression approach, starting with a baseline model without control variables (Model 1) and then introducing control variables in Model 2.

The results, presented in Table 3, reveal that the DID variable in Model 2 yields an estimated coefficient of 1.055, which is statistically significant at the 10% level. This finding suggests a positive and statistically significant influence of the high-standard farmland construction policy on fertilizer use efficiency. From an economic perspective, this implies that provinces with a higher proportion of high-standard farmland tend to exhibit higher levels of fertilizer use efficiency compared to provinces with less high-standard farmland, after controlling for other relevant factors. The improvement in fertilizer efficiency is economically beneficial as it leads to reduced fertilizer costs for farmers, allowing for more sustainable farming practices and potentially higher profits due to more efficient use of inputs. In summary, we test hypothesis 1.

Table 3 Baseline regression results.
Full size table

Further analysis of control variables reveals that increased urbanization rates positively impact fertilizer use efficiency, aligning with Ju et al. (2016). Economically, this suggests that urbanization may facilitate labor mobility and the spread of modern fertilization technology, reducing inefficiencies in fertilizer application. Conversely, higher financial support for agriculture seems to decrease efficiency, in line with Jayne et al. (2013), potentially due to over-application of fertilizer, which may lead to both cost savings and higher agricultural output. Moreover, an increase in rural income, as discussed in Ji (2023), appears to improve efficiency, likely through the adoption of better fertilization practices. Fertilizer prices and changes in planting structures also contribute to efficiency, echoing findings from Kormawa et al. (2003) and Li et al. (2018). These economic factors indicate that policies aimed at improving efficiency are not only environmentally beneficial but can also provide direct financial benefits to farmers. However, human capital variables do not show significant statistical impact, indicating areas for further research.

Temporal effects of the high-standard farmland construction policy

To gain a deeper understanding of how the policy’s impact on fertilizer use efficiency evolves over time, this section examines the year-on-year effects of the high-standard farmland construction policy. We incorporate a series of policy dummy variables for the years 2013 to 2017, interacting them with the DID variable to capture the policy’s effect in each year following its initiation (Ye et al., 2023; Han et al., 2021).

The results, presented in Table 4, reveal two key findings. Firstly, only the estimated parameters of DID × 2016 and DID × 2017 in Model 2 (with control variables) are statistically significant. This suggests a lagged effect in the policy’s influence on fertilizer use efficiency, with significant improvements becoming evident from the third year onwards. The delay can be attributed to the time needed for infrastructure improvements, such as land leveling, irrigation system upgrades, and changes in farmers’ behavior, to fully manifest their impact on fertilizer use efficiency. Secondly, the estimated parameters in Model 2 show a year-over-year increase, signifying a stable and enduring improvement in fertilizer use efficiency attributable to the high-standard farmland construction policy.

Table 4 Temporal effects of the high-standard farmland construction policy.
Full size table

These findings are consistent with those of other studies of similar land improvement policies. For example, Ye et al. (2023) found that the full effects of infrastructure-based policies, such as the construction of high-standard farmland, exhibit a time lag, usually appearing several years after implementation, and that the effects show an upward trend. The year-on-year increase in fertilizer use efficiency observed in this study further supports the conclusion that HSSC policies can increase agricultural productivity and sustainability in the long term and in a sustained manner.

Heterogeneity analysis

Recognizing the diverse geographical, agricultural, and economic characteristics of Chinese provinces, this section investigates potential heterogeneities in the effectiveness of the high-standard farmland construction policy on fertilizer use efficiency. We explore these variations through three distinct dimensions.

Geospatial analysis: exploring variations in policy effectiveness

First, we examine whether the policy’s impact on fertilizer use efficiency differs across China’s eastern, central, and western regions. These regions exhibit distinct economic development levels, agricultural practices, and resource endowments, which can result in variations in the effectiveness of policy implementation. Specifically, the eastern region has more advanced infrastructure, higher access to technology, and a more developed agricultural sector, which makes it more likely to benefit from the policy. In contrast, the western region, with relatively lower levels of development and more challenging geographic conditions, faces greater barriers to fully realizing the benefits of high-standard farmland construction.

The results, presented in Table 5, reveal a significant improvement in fertilizer use efficiency attributable to the policy in the eastern region. This is consistent with the view that economic development and modern infrastructure play a crucial role in the successful implementation of agricultural policies (Xu et al., 2022). However, the policy’s impact on enhancing efficiency is notably less pronounced in the western region, suggesting regional disparities in policy effectiveness. These findings highlight the need for targeted policy interventions in the western regions, where additional support may be required to overcome barriers related to infrastructure and resource limitations. In summary, we test hypothesis 4.

Table 5 Results of heterogeneity by natural geographic location.
Full size table

Economically, this disparity could be attributed to differences in agricultural practices and the level of technological adoption across regions. Provinces in the eastern region are more likely to adopt modern farming technologies, such as precision farming techniques and optimized fertilizer application methods, which lead to higher fertilizer use efficiency. In contrast, the western region may still rely more on traditional farming practices, which are less efficient in their use of fertilizers (Li et al., 2023).

Differential policy impacts in grain and non-grain producing regions

Next, we investigate whether the policy’s impact differs between grain-producing and non-grain-producing areas. This analysis is important because the high-standard farmland construction policy is primarily designed to enhance grain production, and its effectiveness might differ in areas focused on cash crops or other non-grain crops. Grain-producing regions are more dependent on fertilizers for achieving higher yields, and thus, the policy’s effect on fertilizer use efficiency is expected to be more significant in these areas.

The results, presented in Table 6, indicate a positive and significant effect of the policy on fertilizer use efficiency in grain-producing areas, while the impact in non-grain-producing areas is not statistically significant. This suggests that the policy is particularly effective in achieving its intended goal of improving fertilizer use efficiency in regions focused on grain cultivation, where fertilizer inputs play a crucial role in crop productivity. Economically, this finding indicates that areas with higher fertilizer dependency, such as those engaged in grain production, will benefit more from policy interventions aimed at improving fertilizer use efficiency (Hao et al., 2024). In summary, we test hypothesis 5.

Table 6 Heterogeneity in policy impact across agricultural functional areas.
Full size table

In contrast, non-grain-producing areas may not experience the same level of improvement because their agricultural systems are more diversified, and fertilizer use may not be as critical for production. Cash crops or horticultural products often have different fertilization needs and practices, which could explain the lack of significant effects in these regions. These findings underscore the importance of tailoring agricultural policies to regional agricultural structures and specific crop requirements.

Regional variations in baseline fertilizer use efficiency levels

Finally, we explore whether the policy’s impact varies depending on the pre-existing levels of fertilizer use efficiency in different provinces. This analysis helps us understand whether the policy is equally effective across regions with varying baseline efficiency levels or if it has a differential impact. We employ panel unconditional quantile regression, which allows us to examine the policy’s effect at different points in the distribution of fertilizer use efficiency (Zheng et al., 2021; Ye et al., 2023). The results, presented in Table 7, indicate that the policy’s influence on fertilizer use efficiency is more pronounced at higher quantiles, suggesting that the policy may potentially accentuate inter-regional disparities in fertilizer use efficiency. Economically, this suggests that regions already utilizing fertilizers more efficiently have better conditions, such as advanced farming technologies, better soil quality, and efficient irrigation systems, which enable them to optimize fertilizer use even further. This finding emphasizes the need for targeted policy interventions to ensure that the benefits of improved fertilizer use are distributed more equitably across regions. In summary, we test hypothesis 6.

Table 7 Results of unconditional quantile regression.
Full size table

Underlying mechanisms of the policy’s impact on fertilizer efficiency

This section further explores how the high-standard farmland construction policy affects fertilizer use efficiency. We focus on two key mediating factors: land quality and the scale of agricultural operations. By examining these factors, we aim to uncover the causal pathways through which the policy exerts its influence on fertilizer utilization.

To empirically assess these mediating pathways, we employ a series of regression equations. First, we estimate the policy’s impact on land quality and agricultural operation scale separately. Then, we illustrate the effect of land quality and scale of agricultural operations on fertilizer use efficiency through existing studies. In this study, land quality is proxied by the ratio of effective irrigated area to total cultivated land, following prior literature (Zhang et al., 2015; Xu et al., 2022). This indicator captures key improvements in agricultural infrastructure and soil-water management capacity, which are core targets of the high-standard farmland construction policy. Greater access to irrigation not only stabilizes yields but also enables more efficient fertilizer absorption, thereby enhancing fertilizer use efficiency.

The results, presented in Table 8, provide strong evidence for the proposed mediating mechanisms. Model 1 shows a significant positive effect of the policy on land quality, while Model 2 demonstrates a significant positive effect on the scale of agricultural operations. It has been shown that improved land quality and increased scale can directly enhance fertilizer use efficiency (Wu, 2011; Ju et al., 2016; Oyetunji et al., 2022). These findings suggest that the high-standard farmland construction policy improves fertilizer use efficiency by enhancing land quality and expanding the scale of agricultural operations. In summary, we test hypothesis 2 and 3.

Table 8 Empirical analysis of impact mechanisms.
Full size table

This analysis sheds light on the causal pathways through which the high-standard farmland construction policy achieves its intended effects. By understanding these mechanisms, policymakers can design more targeted interventions to further improve land quality, promote agricultural scale operation, and ultimately maximize the policy’s impact on fertilizer use efficiency and environmental sustainability.

Robustness tests

This section presents a series of robustness tests conducted to ensure the reliability and validity of our findings regarding the impact of the high-standard farmland construction policy on fertilizer use efficiency. These tests help us rule out alternative explanations for the observed results and strengthen our confidence in the causal claims made in the study.

Parallel trend examination

A fundamental assumption of the Difference-in-Differences (DID) model is the presence of parallel trends between the treatment and control groups in the absence of the policy intervention. This means that the trends in fertilizer use efficiency would have been similar for both groups had the policy not been implemented. To test this assumption, we examine the pre-policy trends in fertilizer use efficiency for both groups.

The results, presented in Table 9, show that the interaction terms representing the pre-policy years (DID × 2009, DID × 2010, DID × 2011, and DID × 2012) are not statistically significant. This lack of significance suggests that there were no pre-existing differences in trends between the treatment and control groups before the policy implementation, thus supporting the parallel trend assumption and reinforcing the reliability of our DID regression findings.

Table 9 Results of the parallel trend test.
Full size table

Variable substitution and timing sensitivity tests

We conduct additional robustness tests to further validate our findings. Table 10 presents the results of these tests. First, we employ the Translog production function as an alternative method to calculate fertilizer use efficiency and re-estimate the DID model with this new measure. The results remain consistent with our initial findings, indicating that the policy’s positive impact on fertilizer use efficiency is robust to different efficiency measurement methods

Table 10 Regression results for the replacement variables and samples.
Full size table

Second, we perform a timing sensitivity test by re-estimating the model with 2009 as the hypothetical policy implementation year for DID regression. The absence of statistical significance in these estimates further affirms the robustness of our initial regression results and suggests that the observed effects are indeed attributable to the high-standard farmland construction policy implemented in 2013.

Placebo test

In order to eliminate the interference of unobservable factors on the benchmark regression results, this study adopted a spatial placebo test method: randomly generate high-standard farmland construction area in each province (maintaining the proportion of treatment groups consistent with the actual sample) and re-estimate the policy effect. After repeating this process 200 times, the distribution of virtual processing effects is obtained as shown in Fig. 4. The results show that the estimated coefficients of most virtual policy variables are concentrated near the zero value, and none of them have passed the significance test to confirm that the role of benchmark regression in promoting fertilizer utilization efficiency by promoting medium and high-standard farmland construction is not driven by unobservable spatial heterogeneity or random factors, and the research conclusions are robust.

Fig. 4
figure 4

Placebo test.

Full size image

Discussion

This research significantly enhances the understanding of fertilizer use efficiency in China’s agriculture, which has important implications for environmental sustainability. The study reveals a mean fertilizer use efficiency of only 0.608 in Chinese agriculture, suggesting that fertilizer application could potentially be reduced by about 39.2% without adversely impacting other variables or outputs. This estimate is based on a stochastic frontier analysis (SFA) using provincial-level panel data and reflects the technical efficiency of fertilizer input from an economic output perspective. It is important to note that this differs from the nutrient uptake efficiency figures (typically around 30–40%) commonly reported by government sources and agronomic studies, which are derived from field experiments and refer specifically to nitrogen absorption rates.

This inefficiency is a major contributor to environmental issues such as soil degradation and water pollution resulting from agricultural runoff, as highlighted by Smith et al. (2020). Historically, China’s strategy for food security has involved increasing chemical fertilizer usage (Zhu et al., 2022), but such practices have led to inefficiencies. Unlike Zhang et al. (2013), who reported low nitrogen fertilizer use efficiency based on localized field trials, this study leverages national agricultural input-output data to provide a more comprehensive assessment of China’s fertilizer utilization efficiency. This approach not only broadens the scope of existing research but also offers critical insights for environmental policies focused on reducing agricultural pollution and fostering sustainable practices.

Our analysis of the 2008 to 2017 period reveals a positive trend in fertilizer efficiency in China, aligning with the findings of Huang and Jiang (2019) and Yan et al. (2022). Distinctly, while many studies have concentrated on the efficiency of specific nutrients like nitrogen, our research takes a broader approach by evaluating the collective efficiency of various fertilizer types. This comprehensive perspective is essential for achieving both environmental preservation and agricultural sustainability, as emphasized by Erisman et al. (2011) and Xia et al. (2016).

With environmental degradation becoming an increasingly pressing concern in China, the government has been actively implementing strategies to address the issue. This includes promoting the use of organic fertilizers as a means to reduce reliance on chemical variants, a change backed by research from Godfray et al. (2010), Gu et al. (2015), and Oyetunji et al. (2022). Such government-led initiatives, including subsidies for organic fertilizers, signify a pivotal shift towards more sustainable agricultural practices.

This study goes beyond analyzing trends in fertilizer use efficiency and investigates the impact of a specific policy intervention: the high-standard farmland construction policy. The policy, aimed at improving land quality and expanding agricultural operations, offers a strategic approach to optimize fertilizer use (Zhu et al., 2022; Hu et al., 2019). Our research, employing a robust quasi-natural experiment methodology, demonstrates that the policy has a significant and positive impact on fertilizer use efficiency. This finding aligns with our first research objective, which aimed to determine whether the policy enhances fertilizer use efficiency across various provinces.

In examining the policy’s temporal effects, our empirical analysis identifies a lag in the observable effects, inked to regional variations in implementation and standardization. His finding aligns with Ye et al. (2023) and underscores the policy’s capacity for long-term enhancement of fertilizer use efficiency. The continuous improvement over time indicates the policy’s enduring and cumulative benefits, potentially leading to more sustainable agricultural practices and a reduction in environmental harm due to agricultural runoff.

The effectiveness of the high-standard farmland construction policy, as our results show, is primarily due to its ability to improve the quality of arable land and expand the scale of agricultural operations. Sun et al. (2025) emphasize that improving the quality of land elements such as soil and irrigation can directly enhance comprehensive grain productivity and indirectly contribute to more efficient fertilizer use. In the context of Chinese agriculture, often marked by small and fragmented farming operations leading to reduced efficiency (Xin, 2022), this policy emerges as a valuable solution.

In addition, our study uncovers significant regional variation in the policy’s impact on fertilizer use efficiency. It has a beneficial effect in the economically advanced eastern region, likely due to better resource allocation for high-standard farmland construction, leading to more effective policy implementation. Conversely, in the western region, the policy appears to be less effective, potentially due to suboptimal quality in farmland construction, which inversely affects fertilizer use efficiency(liu and Zhang, 2024). Additionally, the positive influence of the policy is predominantly seen in major grain-producing areas, aligning with its objective to enhance grain production. Such specificity in the policy’s impact, as noted by Ye et al. (2023) and Hao et al. (2024), emphasizes its targeted effectiveness in areas primarily focused on grain cultivation.

In provinces where fertilizer use efficiency is already high, the high-standard farmland construction policy demonstrates a more pronounced impact. However, this increased efficiency in specific regions may inadvertently widen regional disparities, potentially conflicting with China’s goal of achieving common prosperity. Such a development calls for careful policy recalibration to address these growing disparities and ensure that the benefits of improved fertilizer use efficiency are distributed more equitably across regions.

The regional differences in the effectiveness of the policy also underscore the importance of incorporating environmental concerns into sustainable farming strategies (Ye et al., 2023). By improving fertilizer efficiency, it is possible to reduce chemical runoff, thereby alleviating water and soil pollution, as noted by Xiang et al. (2008) and Yu et al. (2021). Therefore, embedding environmental considerations into the high-standard farmland policy could serve dual purposes: achieving agricultural productivity while also protecting the environment. To promote balanced and sustainable development, policy interventions are required to address the growing regional differences in fertilizer use efficiency and ensure that environmental protection remains a central focus in agricultural development strategies.

Compared with the existing literature, this study offers several important advances. Prior studies on high-standard farmland policy have primarily focused on yield effects, income improvements, or land use change (Pu et al., 2019; Li et al., 2023; Xu et al., 2025; Li et al., 2025), while environmental efficiency outcomes—particularly fertilizer use efficiency—remain underexplored. This study shifts the focus to this critical sustainability dimension, providing new evidence that complements and extends the existing research. Moreover, while many studies use binary policy indicators and static models, our use of a continuous difference-in-differences framework captures the temporal dynamics of policy impact. Finally, the identification of land quality and agricultural scale as mechanism channels adds explanatory depth, revealing how farmland infrastructure contributes to environmental efficiency. These contributions help position this research as a meaningful step forward in the policy evaluation literature on sustainable agriculture in China.

Conclusions and policy recommendations

Conclusion

This study investigates the impact of China’s high-standard farmland construction policy on fertilizer use efficiency by leveraging a continuous difference-in-difference model and treating the policy as a quasi-natural experiment. Our findings demonstrate that the policy significantly improves fertilizer use efficiency, with the effects becoming statistically significant from the third year of implementation onward. These results remain robust across multiple checks, including placebo tests and alternative specifications.

Importantly, we observe substantial regional heterogeneity: the policy’s effects are more evident in the eastern region, in key grain-producing provinces, and in areas with already high baseline efficiency. This reflects the unequal readiness and capacity of different regions to absorb policy impacts. Additionally, the analysis reveals that the policy operates primarily through two mechanisms: enhancing land quality and expanding agricultural scale, both of which are critical structural factors in modern agriculture.

By demonstrating both the effectiveness and the conditional nature of the policy’s impact, this study provides empirical support for differentiated, long-term strategies in farmland infrastructure planning. It also underscores the broader environmental value of improving fertilizer efficiency, not only in raising productivity but also in reducing pollution and preserving soil and water health.

Policy recommendations

The findings of this study underscore the effectiveness of China’s high-standard farmland construction policy in enhancing fertilizer use efficiency, thereby contributing to environmental sustainability. Given these outcomes, we suggest a multipronged policy approach to expedite the benefits:

Firstly, the government should continue to promote the high-standard farmland construction policy with a focus on long-term continuity and stable policy implementation, given that its effects become more prominent in the third year and beyond. Strengthening the institutional foundation and maintaining sustained investment efforts are crucial to fully realizing the policy’s long-term benefits.

Secondly, a region-specific approach to high-standard farmland construction is advised. The government should tailor the policy to the agricultural economic landscape and topographical conditions of different regions. For instance, in the economically advanced eastern region, the focus could be on the adoption of high-technology agricultural practices, while in the western hilly areas, the emphasis should be on promoting locally practical technology.

Thirdly, improving scale operation in agriculture is pivotal for maximizing fertilizer use efficiency. During the high-standard farmland construction process, the government should encourage agricultural scale operations through mechanisms like mergers and centralized, continuous operations.

Limitations of the study and future research directions

While this study offers robust evidence in favor of enhancing agro-environmental sustainability through the high-standard farmland construction policy, it is not without limitations. First, although the analysis controls for key confounding factors and uses a difference-in-differences framework, the observed improvement in fertilizer use efficiency may also be partially driven by other concurrent developments—such as advances in fertilization techniques, improved crop varieties, or changes in planting structures. Future studies could incorporate more granular, farm-level data on technology adoption to better isolate the policy’s net contribution.

Second, due to data availability constraints at the provincial level, we were unable to include direct measures of land transfer or land consolidation in the empirical model. Yet, land transfer has been shown to affect production scale and input allocation efficiency, and may interact with infrastructure policies such as high-standard farmland construction. We therefore acknowledge that omitting this factor may bias the estimated effect. Future research should incorporate micro-level or household data to explore the potential interaction effects between land consolidation and farmland infrastructure on fertilizer use efficiency.

Third, the scope of this study is confined to fertilizer use efficiency, leaving out other important agro-environmental factors such as pesticide usage and agricultural film application, which are also significant contributors to non-point source pollution. Therefore, a comprehensive assessment of agricultural sustainability would benefit from extending the empirical framework to examine the effects of high-standard farmland construction on these additional inputs. Such extensions would help provide a more complete picture of the policy’s environmental implications.