Driving Financial Inclusion for Smallholder Farmers by Leveraging Satellite Data and Machine Learning

Blog Post

Driving Financial Inclusion for Smallholder Farmers by Leveraging Satellite Data and Machine Learning

Globally, there are an estimated 500 million smallholder farmers -- and some 2.5 billion people living in smallholder farmer households -- who rely on agricultural production for their livelihood. These farmers feed 80% of the population in developing countries. By 2050 the world’s population is expected to reach 9 billion people with the global demand for food increasing by 70%.

Currently, less than 10% of smallholder farmers have access to credit which would allow them to buy improved inputs, like fertilizers and certified seeds, as well as invest in technologies like crop renovation and water irrigation systems. With quality inputs and technologies, farmers could improve their yields and income.

The current gap between the demand for, and supply, of credit to smallholder farmers is due primarily to the information asymmetry between lenders and smallholder farmers. Financial institutions traditionally rely on historical financial data such as, loan repayments, savings deposit activity, and other payment activity and behavior, bill payments, salary slips- to assess the creditworthiness of a potential borrower.

More recently, newer lending platforms have introduced alternative data such as digital and social media footprints as a means to assess a potential borrower. Both traditional and more recent methodologies, however, effectively exclude smallholder farmers since these farmers typically lack transaction or payment ‘paper trails’ and do not have robust digital footprints.

Even for financial institutions serving smallholder farmers, there are information challenges and costs to overcome, including (i) having staff trained in agricultural lending and with the skills necessary to assess farm activity as well as ag-related risks (price, market, climate) and (ii) the high costs of serving remote populations.

Given these information asymmetries and costs, Harvesting is creating and providing data and decision-making tools to help lenders better understand the activity of smallholder farmers, more efficiently and transparently, with the goal of providing greater financial access to these underserved communities.

Leveraging Big Data and Machine Learning

Specifically, Harvesting’s AgIntel (Agricultural Intelligence) Engine platform leverages satellite data, ancillary data, and a machine learning algorithm to help lenders efficiently understand the past, present and future activities of an individual farm allowing them to make informed lending decisions and develop services tailored for smallholder farmers.

Below is a sample output of our agriculture intelligence system, from Ekiti State, Nigeria, where our system provided a detailed assessment of the crop activity of an individual farm which can be updated every two weeks. The data included the number of crop cycles, crop dominance and current growth status of the crop (as in fig. 2).

Increasing advancements in satellite imaging data and machine learning are facilitating huge developments in this space and at Harvesting we are excited to leverage thisprogress to enable financial inclusion for smallholder farmers in emerging markets.

Post your comment

Log in or register to post comments