Agriculture insurance has a long tradition in mature markets and has been fast expanding in emerging markets such as China, India and Brazil. The total gross written premium for agriculture insurance is estimated at USD 28 billion in 2013 of which USD 5 billion is ceded to reinsurers.
Increasingly, the agriculture supply chain providers (processors, traders, logistic companies, agriculture banks and funds) have started to use crop shortfall products as part of their enterprise risk management to protect against business interruption through low volume throughput or loan defaults from natural disasters. The gross written premium transacted with the agriculture supply chain in 2018 is estimated at USD 80 million.
A variety of different approaches are used for agriculture risk transfer, depending on the risk type, the perils covered and the size of the insured farms or the portfolio of a trader, processor or agriculture bank. Risk transfer products are available in the form of insurance or derivatives.
For crop insurance, four main types of risk transfer products can be distinguished
Indemnity-based risk transfer based on yield
For medium to large-sized farms, the actual yield is determined by an expert loss adjustor at farm level. The loss is determined by comparing the actual yield with the expected yield by multiplying the yield short fall with the area planted and the pre-agreed or the actual commodity price.
As yield reduction is the result of many factors and perils, all perils are insured as it is nearly impossible to reconstruct precisely the impact of one peril from the yield reduction.
The insurance of yield reduction is often called Multi Peril Crop Insurance (MPCI) and if volume risk (yield reduction) and commodity price risk is insured jointly, this is referred to as Revenue Covers. For the commodity sector if volume risk and commodity price risk is covered, this is known as Contingent Price Options.
USA, Canada, Brazil
Indemnity-based risk transfer based on mean damage ratios
Losses are determined by trained loss adjustors based on visual inspections in terms of a mean damage ratio determined over the affected field(s). Loss adjustment needs a large number of trained experts and to overcome this constraint, aerial photography from unmanned aeroplanes is used, as for example in China, to extrapolate loss extents over wider areas with punctual loss adjustment on the ground. Under this product, specific perils such as drought, flood, typhoon, frost, fire, hail or fire are insured with different insurance terms (rate, franchise, deductible) per peril and crop type.
In countries with predominantly large farms, this product is offered at farm level, whereas in countries with smaller farm sizes (e.g., China), the product is sold at village level.
China, South Korea, Australia, New Zealand, Europe, Argentina, Mexico, USA (hail only)
Index-based risk transfer based on yield
For small farms with exposure from the commodity sector in various sub-regions, losses are calculated from actual yields over a certain area (e.g., commune, county or district) from government resources compared to the expected (insured) yield of the same area. The yield reduction is then multiplied by a pre-agreed or the actual commodity price and the area planted. As yield reduction is the result of many factors and perils, all perils are insured as it is nearly impossible to reconstruct precisely the impact of one peril from the yield reduction.
Under this product, all covered risks in the area benefit from a payout relative to the covered area per insured risk.
Index-based insurance introduces basis risk which arises through imperfect correlations between the index and the actual losses at farm level. Consequences of basis risk are that farmers may not receive compensation when they have suffered effective losses or vice versa. In order to design a yield index, yield volatility has to be examined intensively in terms of weather impacts and other factors such as changes in farm management practice and such as seed and fertilizer application.
This product is increasingly used by the agriculture supply chain to manage exposure from adverse weather conditions which impact production levels and margins and government entities can equally benefit in protecting their disaster funds from the impact of natural disasters.
India, Vietnam, parts of Africa (pilots) Beneficiaries:
- Agriculture banks (loan default risk)
- Agriculture funds (investment risk)
- Processors, traders and logistics companies (business interruption through production shortfall)
Index-based risk transfer based on weather parameters
For small farms with exposure from the commodity sector in various regions, weather proxies such as rainfall, temperature, relative humidity and wind measurements from ground weather stations are used to quantify losses within well-defined areas. A trigger in terms of a weather parameter is defined for each crop growth stage and perils such as deficit/excessive rainfall, heat days/cold spells or disease days are determined. A payout is triggered if the actual weather parameters are below/above the triggers after which each unit (e.g., millimeter of rainfall) is compensated by pre-agreed monetary amounts.
Under this product, all covered risks in the area benefit from the payout with the difference in the payout only being the size of the individual planted area. Perils can be selectively covered as long as the weather parameters can explain most of the observed yield volatility. However, perils such as hail, river flooding and landslide cannot be covered as they are not measurable at weather stations.
Similar to yield based index insurance, weather based Index insurance also introduces basis risk which arises through imperfect correlations between the index and the actual losses at farm level.
India and parts of Africa (pilots) Beneficiaries: Same us under Index-based risk transfer based on yield Risk Transfer:
Modelling risks for biological systems (crops, forestry and livestock), is very different from traditional natural catastrophe modelling of property risks. Biological systems have abilities to adapt to different types of hazards, as well as recover from perils depending on the time of impact in the growth phase. This is imperative, as the dynamics define the elasticity of the exposure to the perils. For example, a wheat crop will behave vastly differently to a rice crop when exposed to a natural peril and the same wheat crop reacts differently depending on the soil type and irrigation level. Similarly, the same amount of rainfall will have different impacts on crop growth dynamics (single risk type) depending on the characteristics of the dry spells. Livestock and forestry systems and their respective perils are different from each other and need to be modelled as separate systems but need to reflect the correlations of common perils with crops. For example forestry damage from fire will have completely different underlying principles and dynamics when compared with the impact of an epidemic disease on a cattle or pig population.
Since original policy conditions change frequently and more risk types are insured in markets like China and India, an agriculture model has to be permanently updated to reflect these changes. The challenges in developing agriculture risk models are multiple and complex. It is important such models capture the main hazards and drivers of the losses, as observed by insurers, and the uncertainty associated with inherent limitations of the models. It is most important to calibrate the output of risk models with industry losses, when available, and other loss proxies when industry losses are limited, as in the case in emerging countries where insurance schemes are recent.
Crop growth and related indemnity paid under risk transfer products depend on:
- Crop type
- Agro-climate zone (primarily characterized by rainfall, temperature and radiation)
- Soil type and topography
- Cultivar (Seed varieties that determine various adaptation traits to perils)
- Management practices (fertilizer, crop protection, sowing)
- Access to irrigation and presence of flood mitigation
Perils commonly covered under risk transfer products include drought, forest, typhoon/strong wind, hail, frost, fire and in some instances plant pests and diseases (for selected risk types).
While some perils impact the quantity (yield), others additionally impact the quality (protein and moisture content) of the production. Given the exposure to multiple perils, risk transfer for crops are always based on a growing season, hence an “event” is a growing season (4-10 months).
Crop insurance is either based on
- Yield reduction (for either indemnity- and index-based products) from multi perils,
- Mean damage ratios for named perils or
- Weather indices and other proxies of the loss.
Given the high vulnerability of crops, a total loss at a given surface unit (e.g., a hectare) is quickly achieved and a catastrophic event is therefore driven by the geographical extent of an event and the simultaneous impact on various crop types and other risks such as livestock and forestry. Due to the high correlation of agricultural risks, mitigating risk exposures across several different policies and geographies is increasingly hard, and hence exposure to catastrophic loss is high.
A robust crop model has to be based on the hazards that are relevant for the risk transfer product, i.e., multi-peril based if the indemnity is yield reduction or named-peril based if the loss payout is defined in terms of mean damage ratios. Using a model based on yield outputs for named peril is challenging as risk transfer conditions under named-peril policies are peril specific and cannot be distracted with high confidence from yield reduction, which is the result of impacts of various perils as well as farming management practices over the crop growth cycle.
The loss extent on a forest, whether natural wood lands or timber plantations, depend on:
- Tree species and age which determines root depth
- Composition of the stand
- Topography and
- Fire season and fuel load (undergrowth)
- Proximity to fire sources (urban and agriculture areas) besides lighting sources from severe thunderstorms
- Management standards such as fire watching towers, fire breaks, pest and disease protection measures and government forest fire contingency plans
Losses in forestry are often distinguished to be partial or total losses and often a salvage value has to be deducted from the loss amount. For fruit-bearing or industrial tree types (such as rubber), the tree itself can be insured together with the loss of harvest (fruits or rubber production). Following disaster impact, there is often a residual value (salvage) which has to be deducted from the loss if the risk transfer product is defined on full value.
A robust forestry model has to be based on the hazards that affect forests. Wildfire (either from natural causes such as lightning and man-made) is the main hazard in many geographies. However, extreme wind and, in some cases, pests and diseases can also be important hazards.
Livestock populations are exposed to a variety of hazards including:
- Epidemic (contagious) diseases
- Natural hazards
- Management practices and government contingency plans including emergency vaccinations
The livestock density and the general mobility of livestock is a key driver in the spread of epidemic diseases. In emerging markets, epidemic diseases are often endemic, i.e., permanently present in the livestock population and outbreaks are addressed by emergency vaccinations of non-affected livestock – while in OECD countries, affected livestock populations are rigorously culled under strict contingency plans.
Small livestock farms have a different risk profile than industrial livestock operations and imported animals often have a different vulnerability compared with local animals.
The last 20 years have seen a significant improvement in sanitary livestock conditions in emerging markets but the increased international mobility and globalization of livestock trades have increased the risk profile in general.
A robust livestock model addresses base mortality (mainly diseases), epidemic diseases including government contingency plans and natural disasters, which in turn accumulate with drought and flood impact on other agriculture risk types.