Impact of the degree of agricultural green production technology adoption on income: evidence from Sichuan citrus growers

The study addresses how the degree of agricultural green production technology (AGPT) adoption affects farmers’ income and through what mechanisms. Against the backdrop of agriculture’s greenhouse gas emissions and the push for green transformation, AGPTs can reduce resource use and pollution but remain under-adopted in developing countries such as China. Smallholders' adoption decisions are driven primarily by expected economic benefits, yet risk aversion and attention to higher upfront costs can deter adoption. This research posits that multiple AGPTs may have complementary effects on product quality, yield, and prices, thereby increasing income. The study focuses on citrus growers in Sichuan, using citrus sales income to precisely capture technology-induced income changes. It further examines mediating pathways (yield, price, costs) and moderating roles of marketing channels, with an emphasis on adoption intensity rather than binary adoption.

Prior work shows AGPTs improve the environment and can increase farm incomes and reduce poverty. Economic mechanisms identified include higher government transfers, increased prices for higher-quality products, and greater production. However, much of the literature examines single technologies (e.g., improved varieties, IPM, conservation agriculture, tissue culture, chemical fertilizer) rather than combined adoption. Because AGPTs can be complementary, focusing on adoption intensity may better capture economic benefits. Studies also highlight determinants of adoption, including regulation, resource endowment, cognition, and training. This study extends the literature by examining multi-technology adoption intensity, crop-specific income effects (citrus sales income), and the mediating roles of yield, price, and costs, as well as moderating effects of marketing channels.

Design: Observational cross-sectional study using a stratified equal-probability random sample of 805 citrus farmers from 73 villages in 52 townships across the top ten citrus-producing counties in Sichuan Province, China (survey year 2020). Data collected via face-to-face interviews covered personal and household characteristics, production, AGPT adoption, marketing, and income. Treatment (multivalued): Degree of AGPT adoption, defined by use of any of three practices: commercial organic fertilizers, biopesticides, and physical pest control. Degrees: 0 (none), 1, 2, or 3 technologies adopted. Outcome: Citrus sales income per mu (log-transformed). Controls: Gender, age, education (years), health status (1–5), household size, planting experience (years), land transfer (1/0), planting scale (mu), cooperative membership (1/0). Causal estimator: Inverse Probability Weighted Regression Adjustment (IPWRA) for multivalued treatments to address selection on observables. Treatment model: multinomial logit to estimate P(T=n|X); outcome model: regression of income on treatment indicators and covariates, weighted by inverse probabilities; doubly robust. Average treatment effects (ATEs) computed for shifts from degree 0 to 1, 2, and 3. Mediation: Improved stepwise regression (Hayes framework) estimated via IPWRA to test mediating effects of materialized cost per mu (log), yield per mu (log), and sales price per catty (log). Moderation: Interaction terms between degree of adoption and marketing channels (e-commerce, farmers’ markets, agricultural professional managers) included in the IPWRA outcome model. Heterogeneity: Disaggregated IPWRA by region (Chengdu Plain Economic Zone; Northeast Sichuan; Southern Sichuan) and by operating scale (smaller vs larger cultivated area). Robustness: Re-estimation via regression adjustment (RA) and nonparametric bootstrap (1,000 replications). Assumptions: Conditional independence (selection on observables), common support; correct specification of either treatment or outcome model (double robustness).

- Main income effects (IPWRA, relative to no adoption): adopting 1, 2, and 3 AGPTs increases citrus sales income by 7.1%, 10.5%, and 19.8%, respectively. Log ATEs: 0.677, 1.000, 1.907. - Mediation—costs: Materialized cost per mu rises with greater adoption by 2.4%, 5.7%, and 6.6% for 1, 2, and 3 technologies, respectively (all significant). - Mediation—yield: Yield per mu increases significantly only at adoption degree 3, by 3.7% (log ATE 0.294); effects for degrees 1 and 2 are not significant. - Mediation—price: Average sales price per catty increases by 1.7%, 2.3%, and 3.0% for 1, 2, and 3 technologies, respectively (all significant). - Net mechanism: Increases in yield and price more than offset higher materialized costs, producing net income gains; effects strengthen with adoption degree. - Moderation by marketing channel: • E-commerce negatively moderates the income-increasing effect of AGPT adoption (significant negative interaction), likely due to information asymmetry and price competition online. • Farmers’ markets show no significant moderating effect. • Agricultural professional managers positively moderate price and income effects (significant positive interactions), aiding certification and quality signaling. - Regional heterogeneity: Adopting two or more AGPTs significantly raises income across regions; in Southern Sichuan (Region 3), adopting only one AGPT does not significantly increase income; increasing returns to adoption degree evident in Regions 2 and 3 more than Region 1. - Scale heterogeneity: Income gains rise with adoption degree for both smaller- and larger-scale operators; smaller-scale operators exhibit larger income gains at comparable adoption degrees. - Determinants of adoption (multinomial logit relative to non-adopters): Education (strong positive for degrees 2 and 3), health status (positive), planting scale (positive), cooperative membership (positive), age (negative for degree 2), land transfer (negative for degree 2); planting experience modestly positive for degree 2; gender and household size not significant.

The findings confirm that AGPT adoption increases citrus farmers’ income and that higher adoption intensity amplifies benefits. Mechanistically, AGPTs improve product quality and agronomic performance, raising prices and yields. Although AGPTs increase materialized costs, the combined gains in yield and price more than compensate, addressing smallholders’ concerns over cost burdens. The negative moderation by e-commerce suggests that undifferentiated online markets and information asymmetries suppress quality premiums for green products, whereas professional agricultural managers can facilitate certification and better market signaling to unlock price premia. Heterogeneity across regions and scales indicates that local economic conditions, resource endowments, and farm size shape the realized benefits; notably, smaller-scale operators capture relatively larger gains, implying strong potential for poverty-alleviation and inclusive green growth. Overall, the results substantiate the research hypothesis that adoption degree positively affects income and illuminate the channels—price, yield, and cost—through which these effects occur.

This study demonstrates that greater intensity of AGPT adoption—combining organic fertilizers, biopesticides, and physical pest control—significantly increases citrus sales income in Sichuan, primarily via higher sales prices and yields that offset increased production costs. The benefits are particularly strong for smallholders and vary by region and marketing channel; professional agricultural managers enhance, while e-commerce can dampen, the income effects. Contributions include shifting focus from binary adoption to adoption intensity, employing citrus sales income for precise effect measurement, and unpacking mediating and moderating mechanisms using IPWRA. Future research should broaden the portfolio of AGPTs examined, test across diverse crops and production contexts, and assess external validity in other developing countries to inform targeted promotion strategies.

- Technology scope: Only three AGPTs (commercial organic fertilizers, biopesticides, physical pest control) were considered; other green technologies were not included. - Crop specificity: Evidence is from citrus growers; effects may differ for other crops. - External validity: Results are based on Sichuan, China; applicability to other developing countries remains to be tested. - Observational design: Despite doubly robust estimation, unobserved confounding cannot be entirely ruled out.