India’s dominant crop-insurance programme presents insurers and reinsurers with unique challenges of risk assessment. Planned changes to the programme may not help, say AIR Worldwide’s Jeffrey Amthor and Kazi Farzan Ahmed.
Pradhan Mantri Fasal Bima Yojana (PMFBY), the prime minister’s crop insurance programme, accounts for more than 80% of the country’s crop insurance liability. This scheme largely covers end-of-season crop yield loss relative to a target (threshold) yield established by state governments.
Initiated in 2016, the multiple-peril PMFBY covers more farmers and collects far more premium than its predecessors. In addition, PMFBY greatly reduced loss ratios through significantly increased premium rates, with premium payments highly subsidised by both central and states’ governments.
Despite these improvements, PMFBY is often criticised. The country’s third largest food grain-producing state, Punjab, declined participation from the start, and the states Andhra Pradesh, Bihar, and West Bengal withdrew from the programme. In addition to current uncertainties about which states will participate in future, several operational aspects of the PMFBY hamper robust and stable quantification of risk by insurers and even more so by reinsurers.
The low bid wins
The state-run bidding process to select implementing agencies (IA) allowed to sell insurance in particular districts (only one IA per district is permitted) is an important PMFBY feature. Insurers wishing to do business in a district bid premium rates for each crop insured in that district.
Each state groups several districts into a ‘cluster’, with a state divided into several clusters and with actual bidding taking place for whole clusters. The low bidder is selected as IA, which can favour understatement of risk and present a challenge for reinsurers who may have limited basis for evaluating whether the winning bid adequately reflects risk in this young scheme.
What makes the reinsurer’s job more difficult is the fact that most reinsurance agreements are finalised before the bidding process is complete. As such, reinsurers are required to sign up for treaties based on stated desires of IAs to do business in specific clusters rather than knowledge of which clusters the IA will actually do business in.
Indeed, ceding companies themselves may have an unclear picture of their own underwriting risks when planning the composition of their portfolios, as they do not know in most cases which clusters they will win.
The premium rates to be bid are decided by each insurer based on their own calculus. In most cases, the rates are based on crop-yield histories within each notified area (NFA), which is a sub-district geographic unit for which insurance claims are calculated; PMFBY uses an ‘area approach’ in which all farmers growing a specific crop in an NFA are insured together.
The yield histories, provided by the states, are based on field measurements called crop cutting experiments (CCEs) conducted during the past 10 years. This relatively short yield history may cause premium rate volatility associated with reactionary pricing in response to short-term swings in yield, which is a common feature of crop insurance underwriting in many markets. The short CCE history also excludes the all-India droughts in 1987 and 2002, which certainly could be repeated in the near future and typify significant risk.
Change is to come
PMFBY’s operational guidelines were revised in 2018. One change was to calculate threshold yield in each NFA based on the five largest yields of a crop during the past seven years. Prior to that, states were allowed to use the best five, six, or seven yields during the past seven years to establish threshold yields. The 2018 standardisation introduced some stability to risk assessment.
More operational changes are now planned, beginning with the 2020 kharif season. One planned change would affect CCEs, which in future might only be conducted in areas suspected of having low yields based on satellite imagery of vegetation and weather observations during the crop growing season.
Importance of streamlining
As described by the Union Cabinet in a press release issued in mid-February of this year, areas with favourable weather and good-looking satellite imagery might go without CCEs. The intention is to streamline operations, which on the surface seems reasonable, but unintended negative consequences might accrue.
The use of CCE data to quantify risk is already hampered by persistent questions about CCE quality, uniformity, and completeness as well as the changes to CCE geographic resolution over time that makes comparisons of yields in the first half of most CCE records incompatible with the second half.
If high-yielding areas are systematically excluded from future CCEs, the use of CCE data to both assess risk of yield loss and to establish the threshold yields triggering claims could be compromised. The magnitude of these potential effects presently cannot be assessed due to lack of details about how, when, and where future CCEs are to be conducted.
Nonetheless, changes to CCE methodology could introduce additional uncertainty in risk assessments that are based on CCE data because of additional inconsistencies in characteristics of the data over time.
A new source of uncertainty could also be introduced during revamping of the PMFBY. The new plan may allow coverage for single, named perils such as hail. Because there is very limited historical cause-of-loss data for most crops in most districts there is little basis for assessing risk for any future named peril policies. Again, the intent is worthy, but the implementation could add greatly to uncertainty in the assessment of risk of loss. A March memorandum from Ministry of Agriculture clarified that ‘basic yield based cover…is mandatory’, with prevented planting, post-harvest losses, and localised calamities now being at the option of states. The new plan, however, may alleviate one of the big uncertainties faced by reinsurers. Starting from this year’s kharif season, the allocation of clusters to IAs will be for three consecutive years, although some states are already doing that.
This amendment may help stabilise the market by reducing uncertainties related to annual biddings. Reinsurance underwriters could have a clearer picture of risks associated with ceding companies because they will know which clusters will be associated with each IA, and have better visibility on likely premium rates, while writing treaties for the second and third years. Here too, however, this amendment may have a downside; the terms of a March memorandum allow states to choose from multiple methods of updating threshold yields and premium rates during the second and third years of the three-year arrangements. Risk of claims will vary across those methods, potentially reducing programme uniformity between states.
Final harvest valuations
Of potential significant benefit to farmers, states may now choose from two methods of establishing sum insured per hectare of crop. Previously the ‘scale of finance’ was the insured value, which accounted for the cost of growing the crop including purchases of seeds, fertilizers, pesticides and such. States may opt to set sum insured equal to the value of the final harvest as established by multiplication of average yield and the government-established minimum support price (MSP). For major crops, MSP typically exceeds scale of finance. A
Dr Jeffrey Amthor is an assistant vice president and Dr Kazi Ahmed Farzan is a senior scientist at AIR Worldwide.