1. Nature of the scheme

In general, crop insurance models are of two categories: (i) farm specific claim-based insurance products where payouts are based on claimed crop yield losses on individual farms in a region, and (ii) area/region specific (village, district, region) index-based insurance products where the rates of payouts are based on a measured independent index that is strongly correlated to farm level losses in the region. The index can be defined on the basis of output (like average area-yield at the end of the season) or input (weather variables correlated to the area-yield such as average rainfall in the area during the growing season). In index insurance, if an individual farmer suffers a yield loss when the area (on average) does not, the farmer’s loss will not be indemnified. Thus, in contrast to claim-based insurance that protects individual farmers from losses, index based insurance offers protection to farmer communities against large scale shocks.

When numbers of smallholder farmers are large, the claim-based model is not viable because of high cost of crop loss verification, adverse selection (only farmers most at risk may buy insurance), and moral hazard (individual farmers influence loss assessments in their favour). In Index insurance, all farmers in the region receive the same rate of payout when the area-index triggers, irrespective of what actually happens on individual farms. Index insurance therefore has advantages of low monitoring costs (crop losses need not be verified on individual farms), quick uniform payouts, no risk of adverse selection, and also (in principle) limits the scope for moral hazard. However, if the area-index used in the insurance model is not well-designed, that is, it does not offer risk-reduction value to individual farmers in the area, it can undermine their confidence in investing in crop insurance.

PMFBY is an Area-Based Yield Index insurance model for universal comprehensive cover to farming communities against crop failure (yield losses) due to multiple non-preventable risks – climate anomalies, prevented sowing, pests, post harvest losses, and localized natural calamities like hail, floods, drought, fire, inundation, etc. The level of coverage is crop and region specific. The region covered by the area-index is called an Insurance Unit (IU). All insured farmers of a given crop in an IU are covered by the same area-index and receive the same rate of payouts when the index triggers. Production conditions for the insured crop are assumed homogeneous in the IU (despite spatial variations among individual farms in soils, water, pests, resources, and management practices). The insurance covers the period from pre-sowing to post-harvest stages of the crop. It does not cover market risk (price loss) and losses due to inadequate crop management or theft. PMFBY was operationalized in 2016, and covers all food crops, oilseeds, and commercial and horticultural crops grown in both kharif and rabi seasons. The scheme is implemented jointly by Central and State Governments, and private and public insurance companies.

2. Operational Context[1]

For specified crop and season (kharif, rabi) notified by respective State Governments:

• Insurance Unit = Gram Panchayat (GP)  – defined by Govt (~ 240000 GPs in India)
• Insurance Premium = Lowest bid (L1) determined by respective State Governments after an annual public bidding process among insurance companies. The bids are made separately for individual districts or for regional clusters of districts with similar agroclimatic risk profiles for each crop and season. The validity of successful bids has been extended to three years from 2020, to incentivize long term investments in supporting infrastructure by insurance companies, and bring uniformity in premium rates.
• Sum Insured (per ha) = States can choose between the two alternatives: Scale of Finance or Notional Average Value of crop:
  • Scale of Finance or the finance required (cost of cultivation + some profit) per unit cultivated area in a district. It is determined at district level every season by District Level Technical Committees of Banks for giving credit to farmers, and notified by respective State Governments before every season.
  • Notional Average Value of Crop = (Notional Average Yield x MSP).
    (Farm Gate Price can be used in place of MSP, when MSP for the crop is not declared)
• Average yield = Historical average crop yield in the IU/GP. It is determined as the average yield of best 5 years in immediate  past seven years (notified for each IU/GP by respective  State Governments in call for bids by insurance companies)
• Threshold Yield = (Average yield) x (Indemnity level for crop in GP)
• Indemnity level = One of three levels of indemnity based on crop production risk in the IU: (1) low risk, 90%; (2) moderate risk, 80%; and (3) high risk, 70%. This means that farmers will have to bear 10%, 20% and 30% of the loss by themselves in low risk, moderate risk and high risk areas respectively. Indemnity level is notified for each crop and season by respective State Governments at IU level, based on historical yields of the crop in the district/district cluster and their observed variability.
• Premium Subsidy: PMFBY provides for direct subsidy for insurance premiums. Farmers pay only 2% of the premium for kharif crops, 1.5% for rabi crops, and 5% for commercial crops. The balance premium is paid directly to the insurance companies by Central and respective State Governments, 50% by each. (Since 2020 the central subsidy has been reduced to 30 % for irrigated crops and 25% for rainfed crops)
• Actual Yield = average yield of the crop at harvest in current season for entire IU/GP. It is estimated from Crop Cutting Experiments (CCE) at randomly sampled locations in the IU/GP, selected by a statistical sampling process. CCE are carried out under Govt authorities supervision: 4 CCE/IU for major crops; 2 CCE/IU for non-major crops, 8 CCE/IU for commercial crops.
• Yield Loss = [(Threshold Yield) – (Actual Yield)]
• Insurance claim payout/ha= [(Yield Loss) / (Average Yield)] x Sum Insured
• Payment of claims to farmers: Transferred digitally to farmer accounts within three weeks from receipt of CCE crop yield data by insurance company. If Threshold Yield is lower than Actual yield in an IU/GP, no farmer in the IU will receive a payout.

The respective State Governments need to include in the call for bids for insurance premium at the beginning of crop year, the relevant information on: Average Yield, Indemnity Level, Threshold Yield, and Sum Insured for each IU/GP in the districts cluster. This information and other relevant data for the IU (farmer data, sown area, coverage, and Claims) are also updated on the PMFBY National Crop Insurance Portal (NCIP), for the purpose of claims calculation and payment to farmers in a transparent way.

In addition to this base insurance cover, PMFBY also provides for four add-on covers:

1. Prevented sowing/failed germination due to prolonged adverse early season weather ( up to 25 per cent of sum insured)
2. Mid-season adversity (flood or prolonged drought) resulting in yield losses of > 50% of normal yield (up to 25% of sum insured is covered)
3. Crop loss due to localised calamity (hailstorm, landslide, inundation, cloudburst, and natural fire due to lightning affecting part of a notified plot),and
4. Post-harvest loss cover up to two weeks after harvest against unseasonal rain

Note that for 3 and 4, the IU is the affected field of the insured farmer.

3. Issues related to Area-Index Design and Measurement

Note that three of the four factors that trigger the area-index for insurance payouts in an IU/GP, namely Average Yield, Threshold Yield and Sum Insured, are based on historical data. They are explicitly specified apriori in the call for bids for premiums from insurance companies. Only the fourth factor, Actual Yield, is determined for the current season at crop harvest, by taking the average of 4 CCE at randomly sampled locations within the IU/GP. Actual Yield is considered to represent the crop yield value for all the insured farmers in the IU, for payout purposes. It is thus the single most important component of the PMFBY insurance model for fair and rapid settlement of crop insurance claims.

In practice, the CCE sampling process in PMFBY is prone to several delays, biases, errors, and conflicts in claims settlements, and moral hazard, leading to questions about its credibility and reliability. These arise from the following:

1. The large number of diverse IUs/GPs in India (~250000) make it physically impossible and prohibitively expensive to carry out the requisite number of CCEs (4 CCE/IU) per insured crop in the limited time widow available before harvest. This has led to delays, and even non-settlement of claims. (Note that conventional CCE used for crop production estimation annually are carried out at Block level, and crop information of previous years is used for sampling CCE locations).
2. Like all index based insurance models, PMFBY also suffers from two types of intrinsic risks, basis risk and systemic risk. Basis risk affects individual insured farmers (idiosyncratic risk), while systemic risk affects insurance companies and markets.
3. Basis risk for farmers refers to the potential mismatch between insurance payouts based on the area-yield index, and actual losses experienced on individual farms. It arises from the underlying assumption of agro-climatic homogeneity over the IU/GP, based on which a uniform Actual Yield of a crop in current season is estimated from CCE. But in reality, such homogeneity of cropping conditions is rare, as natural resources, management practices, pests, costs, and other factors vary spatially, leading to wide range of yield variability among individual farms even within the IU/GP. Basis risk therefore affects individual farmers differentially. So, farmers in the IU/GP can experience a yield losses on their individual farm, but not receive a payout (false negatives) as the area-yield index is not triggered by the CCE measured Actual Yield value (Actual Yield in IU from limited CCE >= Threshold Yield for the IU, Type 1 error) [. Basis risk can therefore make some farmers worse off than without crop insurance. Alternately, other farmers in the same IU/GP who have not suffered any loss may receive a payout (false positives; Type II error). Such errors can lead to conflicts in claims settlement and also encourage some farmers to influence the choice of CCE locations and measurements (moral hazard) to lower the estimated yield to become eligible for higher insurance payouts. Insurance companies would have similar vested interest in raising the estimated Actual Yield to withhold payouts. Thus, if the positive correlation between individual farm yield and Actual Yield in the IU, is high, the lower is the basis risk.
4. Systemic risk for insurance companies arises when insurance companies cannot diversify their risk over large number of farms. Such situations arise after widespread disaster events like droughts, floods or cyclones, when yield losses among farms over wide regions can be highly correlated (the converse of the conditions for basis risk). Large numbers of insurance payouts will need to be made by the companies as large numbers of farms suffer significant yield losses. Given the high start-up costs of insurance companies, such large payouts can lead to a crash of crop insurance markets. This is why private insurers have been generally reluctant to engage in crop insurance without Government support. PMFBY encourages insurance companies to cover this risk through reinsurance, and supports their costs by paying insurance premium subsidy (>80% of premium on average) to them directly. Further, in case of catastrophic losses (claims to premium ratio > 1:3.5 or percentage of claims to Sum Insured > 35%, whichever is higher at the national level in a crop season), the Government (Central + State) steps in as the reinsurer, to provide insurance cover to the companies (MOAFW, 2020).

Thus at present, in PMFBY, systemic risk is largely covered by the State through direct payment of premium subsidy to companies and providing for their reinsurance during catastrophic losses to stabilize insurance markets.  On the other hand, PMFBY has no specific provisions to protect farmers from basis risk. It is just assumed that, by prescribing the IU at GP/Village level, the area of the IU would be sufficiently small, for homogeneity to hold among individual farms so that basis risk is minimized. However, even at the village level, agroclimatic conditions and management practices can vary greatly among individual farms. As a result, individual farmers can suffer negative impacts on incomes and livelihoods despite insuring their crops. Such losses can lead to loss of confidence and acceptance among farmers in crop insurance. This can be disastrous not only for farmers, but also for the insurance industry and the government which invested heavily in the scheme. For successful implementation of PMFBY, the challenge therefore is to keep basis risk at non-significant levels while protecting against large scale community wide losses, and retaining the operational advantages of the area-index.

[1] PMFBY Revamped Operational Guidelines (2020) Ministry of Agriculture and Farmer Welfare