Bridging data gaps in the Indian agri-lending value chain

During our travels to the hinterlands of Madhya Pradesh, we met with Shaktiman, a marginal farmer with less than one hectare of land, growing wheat and soya. He struggled for years to secure formal working capital credit for his input purchases and suffered from paying high interest to local moneylenders. Despite his persistent efforts to secure formal credit, he has been denied a bank loan for years, as he has little collateral and bank officials made decisions based on their subjective judgement rather than tangible data.

Shaktiman’s story is neither new nor isolated; 85% of farmers in India are small and marginal. Less than 40% of these farmers have access to institutional credit channels, and fewer than 30% possess Kisan Credit Cards (KCCs). KCCs enable farmers to obtain timely and cost-effective short-term credit for crop production and related expenses. Even amongst the smaller subset of ‘banked’ farmers, issues such as prolonged loan processing times and inadequate ticket size persist.

It’s not that banks and non-banking financial companies don’t want to serve this segment. They too encounter challenges across the lending value chain. For example, limited or poor credit histories and rules requiring physical documentation, especially for new-to-credit and tenant farmers, impede effective underwriting and consequently impact their ability to extend credit.

However, it is not all gloom in the fields. The surge in digital data and AgTech ecosystem, coupled with systemic efforts like land record digitisation, promises a new dawn for India’s small farmer. Amongst many emerging solutions across the lending value chain, leveraging alternative data to enhance credit risk assessment holds significant potential.

Alternative data encompasses a diverse range of non-traditional data sources that have become increasingly accessible due to the widespread democratisation of data. With numerous small farmers often overlooked by traditional business rule engines (BREs), alternative data offers hope for broadening financial inclusion. These datasets offer insights into the farmers’ payment potential and can include:

· Environmental and remote sensing data such as satellite imagery and weather data.

· Market and transactional data such as input purchase and produce offtake.

· Financial behaviour and payment history data such as telecom data, utility payments, e-commerce transactions and insurance data.

· Social data such as social media interactions, household behavioral surveys.

One prominent source gaining traction is the use of satellite imagery. Advanced machine learning algorithms are leveraged to analyse this data and provide insights on crop growth and health, yield estimates, soil moisture levels, weather anomalies, harvest readiness, feeding directly into lending decisions. The aim is to minimise the unpredictability and information asymmetry in agricultural lending and aid the underwriting process. AgTechs are pioneering this solution and forging partnerships with lenders to embed these insights in their traditional underwriting models.

Such solutions can significantly elevate the agri-lending ecosystem by adding value on various fronts:

· Optimise credit risk evaluation of farmers with no/limited credit histories.

· Enhance lead generation capabilities by identifying farm areas with high productivity potential.

· Tailor financial products that are more aligned with the farmers' seasonal cashflows and risk profiles.

A recent announcement by a state government asking banks to remove the CIBIL requirement for crop loans only enhances the utility of such alternative datasets for banks.

However, few challenges persist in scaling the use of alternative data. Our recent discussions with leading banks highlighted:

· Reluctance to leverage satellite data for small and marginal farmers due to the lower ticket size of loans.

· Concerns about the granularity and reliability of satellite data with accuracy limited at a village-level rather than at an individual farm-level.

· Considerable time and effort required to integrate solutions with the banks’ systems.

· Increasing need for tools that can effectively analyse data and distill insights from diverse alternative data sources.

Separately, our conversations with AgTech founders highlighted the constrained ability to provide precise information due to limited availability of data such as standardised land records across states, crop-cutting experiment (CCE) datasets, and prohibitive costs of high-resolution satellite imagery.

Let us imagine a scenario--what if there was a middleware platform facilitating end-to-end digitisation of the loan origination and underwriting journey, especially for new-to-credit farmers. The front-end could function in a simple manner. All the relevant farmer data required for loan origination, such as land records and Aadhaar, could easily be retrieved digitally through application programming interfaces (APIs). At the back-end, banks could also retrieve multiple alternative datasets via APIs, conduct a credit evaluation, and sanction the loan within minutes.

The good news: Such a digital public infra (DPI) already exists, developed by the Reserve Bank Innovation Hub (RBIH), and initial pilots have yielded positive results. In our field visits, we witnessed how Shaktiman was sanctioned a KCC loan in less than 10 minutes. These platforms are continuously enriching themselves through various datasets including payments data and warehousing data.

The foundations are in place, and proof of concepts (PoCs) have been promising; now it’s time to scale. The government can institute a framework for the standardization of critical data such as digital land records while ensuring access to crop-cutting experiment (CCE) data and subsidizing costs for high-resolution satellite imagery. State governments can further their progress on digitizing land records and cadastral maps. Development Finance Institutions can offer First Loss Default Guarantees (FLDG) for loans disbursed to small farmers that are backed by alternative data. Lenders can adopt a saturation approach for PoCs in select states to disburse agri-loans. Successful pilots can serve as a blueprint for a broader adoption. The startup ecosystem can develop an alternative data analyser or digestor that can synthesise multiple alternative databases and facilitate automated insights for the banks BRE.

If we can get these building blocks for scale in place, the woes of farmers like Shaktiman could become history. We are at the cusp of rewriting the future of the Indian farmer by democratising access to institutional credit, and we must seize the moment.

This article is authored by Sushma Vasudevan and Aparna Bijapurkar, managing directors and partners and Varad Pande, partner & director, sustainable finance & investment, BCG India.