How geodata-based insurance can help smallholder farmers cope with weather risks

Introduction

Weather-related perils such as drought disproportionately affect smallholder farmers. Apart from the impacts on production in the current season, these farmers may be unable to reinvest in agriculture in the subsequent season as a drought can wipe out their savings.

Another major consequences of experiencing drought is that farmers try to avoid risk; they become hesitant to invest in inputs and other technologies that can increase their productivity.

Not investing in better inputs and technologies keeps many smallholder farmers in a cycle of low agricultural productivity and its effects on socioeconomic outcomes such as poverty and health.

Geodata-based insurance developed through the Geodata for Agriculture and Water (G4AW) program helps smallholder farmers deal with this challenge. The G4AW is a program of the Netherlands Space Office (NSO) and the Dutch Ministry of Foreign Affairs (MoFa).

Designing agricultural insurance

A well-designed crop insurance scheme can help deal with both the impacts of drought on production and the mentality of smallholder farmers. Apart from the risk reduction mentioned above, such a scheme does three things:

(i) protects farmers against their potential inability to repay credit for inputs,

(ii) increases smallholder access to finance to invest in better agricultural technologies, and

(iii) remedies the fear of farmers to use more and better inputs for higher productivity.

However, traditional actuarial-based insurance products require long time series data (typically >30 years) on production and other indicators. This data is not typically available in many areas with smallholder farmers where droughts occur.

Geodata to the rescue

Working with business, corporate, government, and support partners in Ethiopia and Europe, a team at ITC developed the VICI product.

GIACIS is a single-peril drought insurance product based on quantifying the deviation of a season of remotely-sensed vegetation condition from the expected long-term average as a trigger of premium.

The design is based on a toolbox which calculates a greenness indicator called the Normalised Difference Vegetation Index (NDVI). It uses NDVI data from the Copernicus earth observation programme, which is shared through the EUMETCast delivery mechanism. This data is based on the Sentinel-3 satellite's Ocean and Land Colour Instrument (OLCI) and is updated every 10 days.

NDVI is a common reference measurement for “greenness” as influenced by the amount of active chlorophyll in plants seen through sensors. This greenness responds to drought and is, therefore, a reliable indicator of plant conditions across space and time. With the same sensor data and processing routines, past NDVI measurements can be used to interpret new NDVI measurements as an expected NDVI is set, with any deviation from that being an anomaly.

The GIACIS insurance product is calibrated using past NDVI data to determine the thresholds that are indicative of a drought and its severity. These thresholds can then be applied to new NDVI data to determine if, when and where a drought has occurred and if the severity is sufficient to trigger a payout for insured farmers in that area.

With this scheme, insurance pay-out starts once the NDVI-value drops below the preset trigger value (15%-value) for a 1km2 pixel and is 100% if it drops below the preset exit value (5%-value).

Droughts are extreme events requiring information at the tail end of the distribution. Therefore, the NDVI thresholds need to be based on population statistics instead of sample statistics, which is done in GIACIS.

Figure X: Example output from GIACIS showing areas exposed to agronomic drought and where output was required in (a) 2015 and (b) 2016.

What is unique about the GIACIS approach to crop insurance?

Insurance, particularly crop insurance products, is complex. The following are key unique features of GIACIS as a crop insurance product that make it suited for smallholder agriculture:

• VICI  is a self-sustaining, efficient and cheap insurance product. It is suited for areas where usual data metrics used for insurance are not available or, in cases where they are available, they are not of sufficient quality or do not cover required periods or areas. This is true for many developing countries in the global south, where drought and other weather perils are common frustrations in efforts to improve agricultural productivity.
• VICI effectively uses agro-ecological zones for index calibration, reducing spatial variability and enhancing the relevance of its indices. This tailored approach has improved product performance and user satisfaction, demonstrating the potential of localized data integration 
• It is cheaper/more cost-efficient because it is not dependent on field inspections as other crop insurance products. These field inspections are very expensive and raise the possibility of disagreements with farmers and other stakeholders as they are subjective.
• As a geodata-driven insurance product, GIACIS fosters the financial inclusion of thousands of smallholder farmers who do not fit the ‘normal’ financial profile. This is because it covers whole countries, is implemented with mobile technology for registration and claim of payouts, and is accessible to both crop and livestock farmers. Such a model aligns very well with many sustainable development goals.
• This approach does not insure the farmer’s business model: Many of the current insurance products are "indemnity" schemes where a policyholder reports a loss, and the insurer will send their expert to assess damages and use that to determine the payouts. These are only applicable when incidences are small and infrequent. Such an approach falters for perils such as drought because, when it happens, it affects large areas and potentially thousands of farmers. Field assessments at the individual level in this case is impractical. 
• It allows for a typical pay-back time of just 10 years: This means that a farmer will statistically recover paid premiums once every 10 years compared with once in 100 years with other designs. GIACIS, therefore, becomes a revolving fund mechanism.
• GIACIS can be integrated into farmer credit and advisory schemes: Often, farmers are looking for integrated solutions to deal with multiple issues that affect their productivity and livelihoods. With GIACIS being implemented through mobile technology with wide area coverage, it can be easily integrated with farm extension services, credit schemes, training programs, agricultural information services, early warning services, and even farmer health information services.

To learn more, check out our other article on Impact and lessons learned from the GIACIS project