Transparency in Digital Credit Scoring and Agricultural Investment: Evidence from a Lab-in-the-Field Experiment in India

Digital credit scoring can expand access to finance for smallholder farmers, but its benefits depend on whether borrowers understand how scores are generated and how their actions affect future credit access. Transparency may help farmers respond to digital credit scoring with clear information that push them toward productive practices that are genuinely relevant for creditworthiness. On the other hand, by making the scoring logic more visible, transparency may also encourage farmers to focus strategically on practices they believe are rewarded by the score. Whether the transparency provided to farmers improves investment decisions, induces score-targeting behavior, or has limited behavioral effects is therefore an underexplored area. This paper examines whether making a digital agricultural credit score more transparent improves farmers’ understanding and changes investment behavior. We study Dvara E-Registry (DER)’s KhetScore, a digital credit-scoring system that uses remote sensing and machine learning approaches to assess farmers’ creditworthiness. Using a lab-in-field experiment with 1,210 smallholder farmers, we compare a control group with two information treatments: a basic KhetScore information intervention (Treatment 1) and an enhanced treatment that adds repayment salience (Treatment 2). We hypothesize that KhetScore information will improve farmers’ understanding of the scoring system and may change input choices toward practices they learn as score-relevant by the intervention. We also hypothesize that repayment salience provided in Treatment 2, may affect behavior beyond information alone by making farmers more attentive to repayment risk, future credit access, and the consequences of investment choices under uncertain production conditions. Findings will have immediate relevance for our implementing partner and policymakers.