Towards A Digital Green Revolution in India: Part 3: Agri Insurance

(With Jitendra Singh, IBM Research - India)

This article is a part of a series on the potential for a digital green revolution (esp in India): Full Series: Part 1: Productivity, Part 2: Digital Ag Marketing, Part 3: Agri Insurance, Part 4: Digital Spot / Futures Markets), Part 5: Blockchain meets Agriculture, Part 6: Food vs (Bio)fuels: Evolving the Debate. Other articles include: "A Plant's View to IoT and Digital Agriculture", "Cognitive Drones & Nano-Satellites", "Solar Meets Agriculture: Solar Pumps for Micro-Irrigation".

In this piece, lets take a look at the potential for digitization to drive revolutionary change in agricultural finance and risk management, especially insurance in the Indian context.

Agriculture is a highly risky financial venture on several counts. Input costs (land, labor, water, mechanisation, seeds, fertilizer, pesticide etc) are steadily rising and becoming more volatile and less available. Productivity, i.e. yield per hectare, is subject to a number of risks (including monsoons/rains, disasters of various kinds, pests/diseases). Output value (i.e. revenue / monetization of the yield) is subject to huge market volatility, and power of intermediaries (eg: mandi, traders etc): even in a good year of harvest, if there is aggregate over supply relative to demand in the micro-market, prices can fluctuate wildly despite government notified minimum support prices (MSP) if there is no government procurement. Farmers and their cooperatives have tended not to invest in storage facilities, in part due to capital issues; and some crops like sugarcane need to be milled quickly to capture economic value. On the positive side, with improvement in roads, the costs of transport and availability of more markets for produce become open. The movement of agriculture processing closer to the production is also helpful (see earlier articles).

If Agri is such a risky venture, how is it insured? What are the opportunities to broaden risk management tools and insurance product administration in India?

Agriculture Insurance: Issues and Potential Roles For Digitization

Agriculture in India is subject to variety of risks arising from unexpected rainfall patterns, temperature fluctuations, hailstorms, cyclones and floods. These risks are amplified by price fluctuation, weak rural infrastructure, imperfect markets and limited reach of financial services such as credit and insurance especially to small and marginal farmers.

Agricultural insurance is a risk-management means by which farmers can stabilize farm income and investment and guard against disastrous effect of losses due to natural hazards or low market prices (income insurance). Crop insurance not only stabilizes the farm income but also helps the farmers to initiate production activity after a bad agricultural year. It cushions the shock of crop losses by providing farmers with a minimum amount of protection. It spreads the crop losses over space and time and helps farmers make more investments in agriculture.

The risk bearing capacity of an average farmer in India is very limited. A large farm household or a wealthy farmer is able to spread risk over time and space in several ways; he/she can use stored grains or savings during bad years, and can diversify his crop production across different plots. Insurance, by offering the possibility of shifting risks (across a larger pool, and over time and space), enables individuals to engage in risky activities which they would not undertake otherwise.

Individuals cannot influence the nature and occurrence of the risky event. In the case of agriculture or crop insurance. Unlike most other insurance situations, the incidence of crop risk is not independently or randomly distributed among the insured. Good or bad weather may affect the entire population in the area. A variety of insurance products are prevalent in different parts of the world (see world bank primer):

The two classic problems of insurance: adverse selection and moral hazard play a big role. Adverse selection refers to the situation with voluntary participation, where if premium rates are high (or perceived high) it will discourage majority participation and only high risk clients participate. The risk is not sufficiently diversified relative to the expected payouts. pooled over a Moral hazard refers to the temptation of an insured individual to take less care in preventing the loss than an uninsured counterpart when expected indemnity payments exceed the value of efforts. These problems limit the penetration of insurance products.

Yield-based insurance schemes use the so-called "crop cutting experiments" (CCE) to estimate yield / production in a notified insured unit (IU) such as a district. Current practice includes random selection of crop plots in the IU: the area and size varies across states (eg: 10m side equilateral triangle in UP, circle in WB). A minimum sample size of 24 CCEs is needed for a district; 16 CCEs for a tehsil / taluk, 10 CCEs for a revenue circle; and 4-8 CCEs per village / panchayat. Delays in CCE completion is a key reason for delays in claim settlements. One of the opportunities using digital technology (eg: satellite vegetation indexes) is to optimize and/or reduce the number / location of the CCEs (eg: by clustering crop area according to crop health as determined by vegetation indices).

Major Insurance Schemes in India: PMFBY and WBCIS

In India, currently two types of crop insurance schemes are prominent: (a) Pradhan Mantri Fasal Bhima Yojana (PMFBY) and (b) Weather Index based Crop Insurance Scheme (WBCIS).

PMFBY replaces the existing two schemes National Agricultural Insurance Scheme (NAIS) as well as the Modified NAIS. It has been designed and marketed to raise the level of crop insurance adoption in India (which reaches only a minority of farmers).

Crop Insurance coverage in India is low. Only two (2) crore of an estimated 12 crore farmers in the country had crop insurance cover in 2014-15, even as the insurance indemnity was just against the cost of cultivation and barely provided any income protection. From a land area perspective < 20 per cent (40 million hectares) of the total agricultural land is insured.

PMFBY is a comprehensive risk insurance scheme covering most non -preventable risks - natural fire, lightening, flood, droughts, landslide, pests, diseases. Improvements over previous insurance schemes include (a) No cap on premium subsidy, (b) Risk cover for prevented sowing, mid season, post-harvest periods (c) Sum insured upto cost of production/TY (d) Higher indemnity levels (e) Insurance units can be defined at a village level (f) Increased focus on use of technology. Farmer's premium rate is capped to 2% of sum insured (SI) (Kharif food and oilseed crops), 1.5% SI (Rabi food, oilseed crops), and 5% SI (commercial crops, horticulture).

The government determines insured units (eg: village/panchayat level), and invites bids from insurance companies. The lowest bids are selected (one company for a insured unit); and the government bridges the gap between farmer premium and the lowest bid from companies). Implementation agency (IA) selection (from an empanelled set of insurance companies) is made through a clustering approach balancing good and bad districts. Banks are requiring (and bundling in) insurance to be taken as part of crop loans / credit. The amounts are withdrawn electronically/cashless automatically as part of the loan administration. Banks also get a fee for originating insurance (independent of loans). If liabilities of the IA increases by 350% of total premium collected at national level or liabilities increase by 35% of total sum insured, the extra amount is borne in equal share by the center and the state.

A range of life-cycle risks are insured. For example, "Prevented Sowing/Planting Risk" is determined through proxy data such as rainfall, weather satellite, and a lump sum payout of 25% of SI is done and insurance cover terminated subsequently for the season. Other risks covered include mid season adversity (if expected yield below 50% of total yield in the insured unit), wide spread claims (based upon season end yield), post-harvest losses and other localized risks.

According to government sources, 24 States and 4 UT's have notified the PMFBY scheme, and it covers 23% of crop area, 370L farmers (Kharif 2016). Premium collection for Kharif 2016 was Rs. 16130 Cr, and for Rabi 2016 Rs. 25000 Cr. Government outlay has increased from Rs. 5500 crores to Rs. 9000 Cr in FY 2017-18 (used to bridge the actual premium and farmer paid premium). This is a promising start, but there is a long way to go especially to cover marginal farmers.

Weather Index Based Crop Insurance Scheme (WBCIS)

The Indian private sector has been a pioneer in Weather-index based crop insurance. This scheme provides protection against adverse weather conditions, set up as weather indices. Example weather indices include rainfall (eg: deficit rainfall, excess rainfall, un-seasonal rainfall, rainy days, dry-spells), temperature (eg: high temperature (heat), low temperature), relative humidity, wind speed, hailstorm, cloud burst, or a combination (Multiple parameter index). No loss assessment (eg: crop cutting experiments) is necessary, and settlements can be fast. It has been a profitable insurance offering (see below), and low moral hazard due to the use of transparent data; but requires historical data for index design. Indexes could also be targeted at pest / disease "conducive weather events": eg: blight conducive events for potato blight insurance. Government’s financial liabilities could be budgeted up-front and close ended, as it supports the premium subsidy (and not claims subsidy). Premium rates are capped for the cultivator; and the premium (rates) beyond the cap are shared by the Central and concerned State government on 50:50 basis.

Weather based Crop Insurance Scheme (WBCIS) operates on the concept of “Area Approach” i.e., for the purposes of compensation, a ‘Reference Unit Area (RUA)’ shall be deemed to be a homogeneous unit of Insurance. This RUA is notified before the commencement of the season by the State Government and all the insured cultivators of a particular insured crop in that Area will be deemed to be on par in the assessment of claims. Each RUA is linked to a Reference Weather Station (RWS), on the basis of which current weather data and the claims would be processed. Adverse Weather Incidences, if any during the current season would entitle the insured a payout, subject to the weather triggers defined in the ‘Payout Structure’ and the terms & conditions of the Scheme.

WBCIS has a high "basis risk", i.e. locations farther away from the reference weather station may have different micro-weather conditions, but are indexed to the station's observations. Another view of basis risk is that the risk that the loss for which a farmer is compensated is very different from the loss that he/she actually suffers. Note again that weather-insurance is not a yield guarantee insurance (unlike the multi-peril PMFBY scheme). This is a source of grievances (see below). Basis risk with regard to weather could be high for rainfall and moderate for others like frost, heat, humidity etc.

Opportunities with digitization: Finer-Grained Insurance, Bundling with Finance

Historically, schemes like NAIS, use a larger insurance unit (eg: district or taluk) for administrative convenience and conduct of crop cutting experiments. For example, in a district-based multi-peril (yield insurance), the indemnity is based on the realized average yield (AY) of an area such as a district, not the actual yield of the insured party. The insured yield is established as a percentage of the average yield for the area. An indemnity is paid if the realized yield for the area is less than the insured yield regardless of the actual yield on a policyholder’s farm.

Farmers would prefer the insured unit to be smaller (eg: a village / panchayat) or even at the level of each individual farmer. The lack of data for actuarial analysis, land records, productivity and idiosyncratic factors have been a bottleneck. With increased digitization (eg: of land records) and remote sensing based indexes, it is possible to get more granular indexes at the level of tens of meters (eg: LANDSAT vegetation index data is available at 30 m resolution). With the growth of the nano-satellite and multi-national satellite industry, more data is becoming available to form the basis of indexes or claim settlement. These will allow the implementation of "micro-insurance" schemes for farmers, especially weather-based where data is increasingly available at a granular level. Drones are not allowed in India, but a drone-based local survey or fine-grained satellite data analysis (pre- and post an event) can be used to pinpoint losses for claim settlement. Participatory sensing by farmers with smart phones in the future offers a new additional modality for risk monitoring.

Awareness amongst small holder farmers is low, and this precisely is the segment who are also less served by the formal credit of banks and in the jaws of money lenders. With increasing rural credit, and Jan Dhan bank accounts, there is a huge opportunity to bundle insurance with credit. Insurance is sold because the farmers have to advance money to the insurer; but in credit the farmer is advanced money by the bank who needs risk protection. In fact, most of the successful PMFBY rollouts involve loanee farmers, who have to purchase insurance mandatorily, and the premium is deducted digitally from their accounts (in a cashless manner). Bundling of insurance with the price of other crop inputs (eg: pesticides, fertilizers, irrigation) allows economies of scope in distribution.

In summary, digital revolution via satellite data, weather forecasts, local sensing (especially soft sensing and participation using mobile smart phones), drone / imaging platforms, cashless finance, jan dhan accounts, digitization of land records, monitoring of crop growth stages, disease and pest risk monitoring opens up new vistas in insurance offerings and their marketing via bundling with credit and other farm products.

Twitter: @shivkuma_k

ps: Related articles: "Towards a Digital Green Revolution for India Part 1: Productivity Drivers", "Towards a Digital Green Revolution for India, Part 2: Digital Ag Markets and Marketing", "A Plant's View to IoT and Digital Agriculture", "Cognitive Drones & Nano-Satellites", "Solar Roofs & Greenhouses...", "Cashless Finance with Universal Payments Interface", "The Second Green Revolution (Jaspreet Bindra)".