Abstract

This study examines the causal impact of digital technology adoption and internet services on farmers’ income in Indonesia using the Propensity Score Matching (PSM) method with microdata from the 2023 National Labor Force Survey (SAKERNAS). Agriculture remains an important sector in Indonesia, contributing 12.53% to GDP and employing about 28% of the national workforce in 2023. However, many farmers face serious challenges because more than 62% of them manage very small plots of less than half a hectare, which limits productivity and income. At the same time, digital technology and internet services offer opportunities to improve efficiency and help farmers reach wider markets. But the adoption rate is still low, with only about 12.8% of farmers using digital tools and internet services, showing a significant digital divide.
Earlier studies found that farmers using the internet tend to earn higher income, but most of these studies only showed correlation, not causation. To address this gap, this study applies PSM, because farmers who adopt digital tools and internet services are often different in characteristics such as age, education, household size, and farm scale compared to those who do not. The dataset covers 105,771 farmers, with income as the dependent variable and an interaction of digital technology and internet use as the main treatment. Covariates include age, gender, years of schooling, experience, labor input, and training participation. Propensity scores were estimated through logistic regression and matched using nearest neighbor with a caliper of 0.1.
The results indicate that farmers who use digital technology and internet services earn on average 64.4% more than farmers who do not. Descriptive statistics already suggested that adopters gain approximately IDR 1.3 million more per month, but the PSM analysis confirms that the difference reflects a causal effect. The heterogeneity analysis by age provides additional insights. Younger farmers are more likely to adopt technology, with nearly 20% adoption, while adoption among older farmers is only about 11%. However, income gains are larger among older farmers, with an increase of about 71%, compared to 44.2% for younger farmers. These findings suggest that digital adoption not only empowers younger farmers who are more familiar with technology but also delivers significant welfare improvements for older farmers who may face greater productivity constraints.
The findings add to the existing literature by providing causal evidence from Indonesia. Policy implications are clear. First, rural internet and digital infrastructure must be expanded to ensure equal access across regions. Second, training programs are needed to help farmers, especially older ones, learn how to use digital tools effectively. Third, financial support such as credit schemes or subsidies could encourage adoption among small farmers with limited resources. Without these interventions, digital divides may increase inequality rather than reduce it.
Although the study offers strong evidence, it also has limitations. Some important unobserved factors may still influence the results, and the cross-sectional nature of the data does not allow for measuring long-term effects. Future research could use panel data or other econometric methods to provide deeper insights. Even with these limitations, the study clearly shows that digital technology and internet adoption significantly increase farmers’ income and can support a more inclusive digital transformation in Indonesia’s agricultural sector.