Agricultural weather insurance undergoing reform

Table of contents

Share this article!

For most of us, insurance is not a very sexy subject to begin with; it makes many people cringe! I don’t know many people who enjoy contacting their insurer – except when something goes wrong – or dealing with their insurance contracts. But we still have a more or less close relationship with the insurance sector because we take out contracts (civil liability, housing, vehicle…) whether it is out of obligation, in anticipation of a risk, or for a whole host of other good or bad reasons.

The agricultural sector also has its share of insurance contracts. For example, there are the classic contracts that you can have as a private individual, but also a whole range of more specific insurance policies such as those for agricultural machinery, buildings, seeds, or crop losses. In this blog post, we will focus exclusively on crop insurance, also known more broadly as climate insurance, for two main reasons. The first is that, in case you haven’t heard, a major climate disruption is underway. And the human race is responsible for it as the IPCC (Intergovernmental Panel on Climate Change) says – without any further ado – in its 2021 AR6 report. This climate deregulation is already starting to hit the agricultural sector in the face. The second reason is that agricultural weather insurance in France is about to be reformed. At least, this is what we should understand from the report by Frédéric Descrozailles, a member of parliament, which was submitted to the Ministry of Agriculture in the summer of 2021, and some of whose amendments were voted on at the beginning of 2022, with a view to implementation in early 2023.

I would like to stress that some elements of this dossier are focused on France – in particular everything that revolves around the reform of agricultural insurance that is taking place there at the moment. Nevertheless, I hope that readers outside our beautiful country will find the background, insights and discussions in this blog post of interest.

As usual, for the readers of the blog, this article is based on video interviews with actors of the sector (whose names you will find at the end of the article) whom I thank for the time they were able to give me. Several articles, reports and seminars have allowed me to complete the feedback from the interviews.

Enjoy reading!

Figure 1. Frost control in Burgundy vineyards in April 2021. The buds had come out very early in 2021 following a particularly mild winter. The April 2021 frost completely reshuffled the cards and put French winegrowers in a catastrophic situation.

Soutenez Agriculture et numérique – Blog Aspexit sur Tipeee

Climate change and the cost of damage

The publication of the first part of the IPCC’s sixth report has once again sounded the alarm. There is no doubt that human activities are causing widespread and rapid global warming (Figure 2). The extent of the warming we will experience will obviously depend on our cumulative CO2 emission scenarios, but experts already agree that the widely publicised threshold of +1.5°C of warming by 2100 compared to the pre-industrial era will almost certainly be exceeded.

Figure 2. Observed and projected temperature changes by 2100 under three GHG emission scenarios. Source: Territorial Resilience Strategy Reports, Shift Project, 2021. Volume 1 – Understanding.

The second part of the IPCC’s sixth report, published at the beginning of 2022, insists on the fact that we have only a very short window of opportunity – of the order of a few years – to act before we enter climate trajectories from which it will no longer be possible to escape (Figure 3).

Figure 3: Illustrative climate or non-climate shock, e.g. COVID-19, drought or floods, that disrupts the development pathway. If we do not act quickly, we can see that the orange and red trajectories never return to the green and/or yellow trajectories. Source: IPCC, 2022.

The damage of ongoing climate deregulation is already widely visible, and it is starting to hurt the wallet. In its 2021 annual report, the insurance broker Aon estimates the damage generated by natural disasters at around 343 billion dollars in 2021 worldwide, after 2017 (519 billion dollars) and 2005 (351 billion dollars) [Aon, 2021]. France Assureurs – the federation of insurers and reinsurers operating in France – has recently highlighted the trend of increasing claims costs paid by insurers since the 1990s (Figure 4). The figure also shows the strong evolution of indemnities for agricultural weather insurance – the “crop” insurance that we will discuss at length later in this blog.

Figure 4: History of claims paid by insurers following natural hazards in France. Source: France Assureurs, 2021

And these trends are not about to stop… The latest IPCC AR6 report of 2021 is very clear on the probability of occurrence and the intensity of the climatic events that could be faced with a few extra degrees on a global scale (Figure 5). It shows, for example, that with only 2°C more warming in 2100 than in the pre-industrial era, an extreme temperature event will occur 5.6 times more often in 2100 than in the period 1850-1900, with this event being on average 2.6°C warmer. We wish our dear farmers around the world well. And to continue on the agricultural sector, the same report predicts that under the same 2°C warming scenario, an agricultural and ecological drought event will occur 2.4 times more often in 2100 than in the 1850-1900 period.

Figure 5: Projected changes in the intensity and frequency of extreme temperatures over land, extreme precipitation over land, and agricultural and ecological droughts in dry regions. Source: IPCC, 2021

And I would like to take this opportunity to add a few keys to the consequences of such warming for the biosphere, in order to break out of our rather too human-centric prism of thought (Figure 6). Whether in terms of the loss of plant species, the loss of insect species (which is widely highlighted in the latest IPBES [Intergovernmental Panel on Biodiversity and Ecosystem Services] report), or the continued decline of coral reefs, we can see that the dynamics are extremely variable and non-linear with a warming planet’s temperature.

Figure 6. Comparative consequences of 1.5° and 2° of global warming since pre-industrial times. Source:, 2021

All this is of course if we limit warming in 2100 to +2°C compared to pre-industrial times. We are not heading in that direction at all… In the “business as usual” scenario we are following, i.e. without any change in our lifestyle, we are at least heading towards +4°C warming scenarios.

Did you think that insurance companies had paid out a lot of money in recent years? But you haven’t seen anything yet, folks! In association with Météo France, the Caisse Centrale de Réassurance (CCR) has produced a map of the evolution of natural disaster damages between 2018 and 2050 (Figure 7). In their 2021 report, France Assureurs (still the same as above), estimate that the cost of natural hazards will continue to grow at the rate of a doubling every 30 years [France Assureurs, 2021]. By 2050, we could expect to have to take twice as much money out of our pockets as we have had to take out in the current years.

Figure 7. Spatial distribution of the evolution of damage from natural disasters between 2018 and 2050. Source: Territorial Resilience Strategy, Shift Project, 2021. Reproduced from JRC and Météo France report (2018).

While the cost of action to limit or adapt to climate disruption may seem high at the moment, it is out of all proportion to the cost of inaction that we may face in the coming years (and are already seeing). In 2006, former World Bank chief economist and vice president Nicholas Stern published his eponymous report in which he argued that the cost of inaction far outweighed the cost of prevention. The Stern Review estimated the cost of inaction, depending on the scenario, to be between 5% and 20% of global GDP, compared to 1% for action. Similarly, if insurance seems expensive to you, non-insurance is just as expensive. As an example, French farmers affected by the exceptional floods of 2016 had to borrow almost 5 billion euros from their bank to get through the bad year. Farmers’ debt capacities were largely saturated by these short-term, non-productive loans, which could have been used for example to finance transitions or simply to invest. And society also bears the losses indirectly through the loss of productivity and competitiveness of the agri-food sectors.

When we talk about agricultural insurance, we should not consider that the risks linked to extreme weather events only affect farmers. It is by considering the entire agricultural sector – input suppliers (seeds, fertilisers, machinery, etc.), cooperatives, farmers, traders and processors – that it becomes clear that all players can be affected if agricultural production is affected by a climatic event. A problem for the farmer and it is the domino effect… Because an uninsured farmer is also a supplier or a client who is not solvent for a cooperative or an input supplier.

Global trade in agricultural products allows financially well-off consumers to be satisfied because a local production deficit due to a climatic event can be compensated for by an import from a non-affected country, regardless of the transport costs. The agricultural risk is then hidden from consumers as long as the climate risk does not become systemic at the global level. The current crisis in Ukraine is a reminder of this. Yield risk due to climatic hazards, on the other hand, permanently affects the individual income of all farmers, anywhere in the world.

Some common insurance knowledge bases

Hazards, risks and vulnerabilities

Many of us take out insurance policies (civil liability, housing, vehicle, etc.) but do we really know what insurance is? Insurance is a transfer of risk. Behind this rather simplistic phrase there are actually many concepts that need to be borne in mind.

The first, and certainly the most important, is that behind insurance there is a risk. If there is no risk, there is no insurance. The insured, who bears a risk (a climatic risk for a farmer, for example), transfers his risk to the insurer by taking out an insurance contract. The insured wants to get rid of his risk, so he sells it. Paradoxically enough, we hear people say that an insurer “buys” a risk, whereas we agree that the insurance contract is paid for by the insured. In reality, it should be understood that the transaction between the insurer and the insured finances the service of transferring the risk from one actor (the insured) to another (the insurer). In France, the insurance supervisory authority (ACPR – Autorité de Contrôle Prudentiel et de Résolution) is very careful to ensure that insurance contracts do indeed involve risk transfers, and not something else, a way of making sure that people do not transfer money of any kind without risk (from one subsidiary to another, for example)

Let’s quickly take the opportunity to make a small aside of vocabulary and clarify the differences between a hazard, a risk, and a vulnerability. A hazard is a phenomenon (natural or technological) that is more or less likely to occur in a given area (Figure 8). A risk, on the other hand, is the possibility of a hazard occurring and affecting a population that is vulnerable to that hazard; vulnerability can be of any kind to infrastructure, health, etc. In other words, if there is no vulnerability, there is no risk. A storm in the middle of the ocean is not a risk for the population. It is a strong climatic hazard, certainly, but since no population is vulnerable to this hazard, we will not consider it a risk. This notion of risk is nevertheless rather subjective because not everyone sees vulnerability in the same way. In the remainder of this document, we will use the terms hazards and risks in a similar manner to make it easier to read, but keep in mind that the concepts are nevertheless different.

Figure 8. Climate hazard – the conjunction of natural variability and climate change. Source: High Council for the Climate, 2021

Insurers and re-insurers

Anyone can insure themselves: individuals, companies, states (for example, states can insure themselves against natural disasters to mutualise the risk of drought, or to protect their budgets to compensate farmers against a future climatic hazard). But perhaps most surprisingly for those unfamiliar with the sector, insurers also insure themselves. They use reinsurers (e.g. Swiss Re, Munich Re, Partner Re; “Re” should be understood as “Reinsurance”) to whom they transfer their risks. One of the reasons for the existence of reinsurers is to cushion natural disasters that affect an entire portfolio of policyholders. The reinsurer pools its risks through a portfolio of policyholders that is as diversified as possible. But isn’t that what insurers already do? Yes, but not on the same scale. When a large area is affected (in the case of a hail storm, for example), it is the entire country that is affected and the losses are enormous at the national level. An insurer whose portfolio of policyholders only included policyholders in this territory would be ruined because it would have to pay an insurance premium to all its policyholders at the same time. This is too big a risk for the insurer to manage alone, so it will tend to reinsure itself. The primary insurers carry the risk. The reinsurers are the second net, they are there in case of a hard blow. In the context of weather insurance, reinsurers diversify risks by insuring several different geographical areas, hoping that there will be compensation between the areas (considering that there is little chance of a storm occurring at the same time in Chile, France and Indonesia). Reinsurers will thus tend to be more international players. Note, however, that some reinsurers may also be insurers on the side for certain classes of business. Note also that insurers may split insurance contracts into several parts (known as proportional insurance) and share them with another insurer or a pool of insurers, in which case it is known as co-insurance (the same thing may happen with reinsurers, in which case it is known as co-reinsurance).

Insurers and reinsurers are not philanthropists. Their job is not to make money for the insured. Insurers are there so that the cost of claims is lower than the insurance premiums (if the insurer makes money, the insured loses money). And the insured, when he or she takes out insurance, is not supposed to consider that he or she will make money. The use of an insurance product is not a means of generating income: it would not make much sense to calculate an “ROI” (return on investment) of an insurance contract for an insured because it is normally supposed to be negative. Over the length of his insurance contract, the policyholder loses money but, in exchange, the policyholder limits the volatility of his income statement. The policyholder buys peace of mind in exchange for a kind of annuity paid to his insurer (unless the insurer decides to lose money on his insurance contract…). It is important to understand that insurance contracts allow to limit the jolts that everyone could have on their profit and loss account (or bank account for individuals). For example, if my house burns down, it is a risk of ruin for me because I cannot assume this risk alone. In the same way, if I kill someone while driving my car, the amounts to be paid will be so huge that I cannot afford not to insure myself. Insuring your car is compulsory for this reason, and it can be said that if this insurance was not compulsory, few people would use their car for fear of having an accident that they could not pay for. On the other hand, if my mobile phone is stolen on the bus, I can consider that this is not a risk of ruin and I can imagine not insuring myself for this type of risk. To draw a parallel with agriculture, if, as a farmer, I can bear the financial ups and downs of several successive bad productions, I might as well keep the risk to myself and not take out an insurance policy. Agricultural weather insurance protects against hard times that cannot be absorbed by the farmer’s savings alone. It is therefore possible to self-insure in order to keep the benefits of the insurance and not to depend on possible variations in insurance rates in the contracts taken out with my insurer.

It is important to understand that insurers and reinsurers will insure random risks. From the moment the risk becomes certain, it is entirely predictable and thus becomes uninsurable. There is therefore a very important distinction between what is in the realm of randomness (and which can be insured) and what is in the realm of tendency (which cannot be insured). Insurance cannot indeed insure a trend (e.g. a global warming trend), but within a given trend, insurance can insure a hazard. Another way of putting it is to distinguish, in the case of climatic hazards, between what comes under the heading of risk volatility, which is typically part of the insurance logic, and what comes under the heading of the level of guarantee allowing for compensation for a loss of production.

Like the banking sector, the insurance sector is also highly regulated. Insurers have equity (i.e. capital available to settle high-cost claims) – as do reinsurers – and are supervised by supervisory authorities (the ACPR in France, for example). There is a professional guarantee fund, but the state is not supposed to intervene. The regulator requires insurers and reinsurers to assess their solvency on an ongoing basis (solvency ratios, risk measures, risk scenarios and impacts on solvency, etc.). The European “Solvency 2” regulation makes it clear that the insurer is obliged to cover claims up to 99.5% within one year (in other words, this also means that the insurer is not supposed to be able to pay claims once every 200 years). The insurer must therefore have sufficient equity to cover the claims. As an insured, we need to make sure that our insurer (or re-insurer) will be able to afford to pay in the event of a claim because if the insurer goes bankrupt, the insured will have to pay the cost of their claim, and the domino effect starts to kick in quite quickly. In the markets, insurance companies are rated. When taking out an insurance contract, policyholders must therefore take an interest in the rating of their insurer. Not all insurers have a rating (in practice, only insurers that are listed on the stock exchange and/or issue bonds on the financial markets have ratings). Every policyholder is entitled to ask about the credit rating of an insurer as defined by the regulations, but you have to know something about it….

How are insurance rates set then? Insurers need to develop risk models (e.g. climate models, earthquake models, etc.) to calculate the risk of a claim falling on them. Insurers anticipate the frequency and cost of each claim in order to price their insurance contract and thus evaluate the insurance premium they will ask to pay. The most difficult thing for insurers to do is to assess the uncertainty associated with climate risks, particularly because these uncertainties vary over time. Insurers constantly assess the so-called loss ratio, i.e. the ratio between the insurance premiums paid by the insured and the claims paid by the insurer. If the ratio is higher than 1, the insurer loses money because it pays out more money than it recovers through its contracts. This is expressed as the ratio between outgoings (management fees, commissions paid, claims provisions and claims reimbursement) and receipts (premiums and contributions received). To give an order of magnitude, the management costs of the insurance contract for the insurer, i.e. excluding the estimated cost of claims, represent about 20% of the total insurance contract price. These include the cost of reinsurance, the cost of tied-up capital, the cost of underwriting contracts (marketing costs), the cost of administering contracts, or even the time spent on managing the expertise and the cost of travelling to the field.

Pricing is not easy either because we do not always know exactly what has happened in the past and what will happen in the future. An example is asbestos. When asbestos was considered to be carcinogenic, plaintiffs started to ask for compensation. Between the moment the problem is understood and the trial takes place, many years can pass. The insurer then finds himself pricing in a rather obscure environment where he does not always have a good idea of the risk involved.

Insurance works in cycles and contracts are re-priced every year. Re-pricing every year is not a legal requirement. However, as insurers’ solvency models are 1 year in regulation, insurance contracts tend to be too. In theory, insurers could very well be able to run their risk models over 5 or 10 years. And insurers would be well advised to do so over a number of years in order to mutualise risks, both in space and in time. However, in the case of subsidised insurance contracts (we will see the case of agricultural insurance below), the insurer is dependent on eligibility for the subsidy and this eligibility is fixed by decree every year. If the State decided to modify the specifications for obtaining the subsidy, the insurer would no longer be able to provide the insured with the subsidy. Even if some insurance contracts are offered over many years, there would in fact always be a clause to redefine the rates annually if there were problems. Let’s get back to the point. If there are several years in a row of low claims, prices will fall. If, on the other hand, the number of claims increases, the prices of insurance contracts also increase. When the insurance premiums paid by policyholders are not sufficient in relation to the claims experience of the insurers, the insurers increase their prices and regain their health over time. At the time of the World Trade Center attack, insurance on the airline fleet increased enormously. The few insurers who had the courage to continue offering insurance policies made a lot of money because there were very few claims afterwards.

Insurers and reinsurers use in-house teams and external consultants (data scientists, climatologists, etc.) who offer modelling services and require skills and expertise in fluid mechanics, weather forecasting and data analysis. Agrometeorology is important here to find damage modelling formulas adapted to the crops that are covered. But not all risks are so easily modelled, specific risks are the most complicated to consider. Modelling the exposure of a territory to a risk remains a difficult task.

Insurance in the face of climate deregulation in agriculture

In France, in the agricultural sector, total insurance premiums can be roughly divided into three large blocks of equal size. The first third for building and animal insurance. The second third for agricultural machinery. And the third for weather insurance. With the climate deregulation we are experiencing, the situation will change tomorrow because crops will be the assets most exposed to climatic events. The risks to which French agriculture is exposed are actually going to get worse. There is a trend towards a form of tropicalisation of the climate, which is combined with a historic halt to the increase or stability of production potential. In France, the recent consultations of the “Water Conferences” and the “Varennes de l’eau” have shown that the subject is completely topical.

It is clear that insurers have a short-term vision, except perhaps for civil liability. If previously, it was the impact of a company’s direct activity that was compensated, we are starting to see indirect activities come to the fore. For example, recently the Port of Los Angeles had to compensate residents living near the port for the CO2 and fine particles emitted by the trucks feeding the port. With the ongoing awareness of climate change, any negative impacts that certain types of activity will amplify will be subject to litigation.

We have seen above that weather events are increasingly hurting the wallet (Figure 3). These events are becoming more frequent and more expensive, not least because as we build more and more infrastructure we are undeniably increasing our vulnerability. Insurers do not have enough equity to manage major weather disasters and therefore insure themselves “in excess” with reinsurers (as an insurer, I consider, for example, that beyond a loss of 2 million euros, I can no longer deal with the settlement of claims alone). Re-insurers are stronger with a more diversified portfolio of policyholders, have more equity and capital, and can therefore take this risk in excess.

Nevertheless, insurers and reinsurers are beginning to ask more and more questions. In the area of risk forecasting, they are realising that climate models built with past data to predict what might happen in the future are no longer relevant. There is no longer any stationarity in the climate. Models must become deterministic, including physical, economic and sociological parameters and not just statistical parameters to understand the risk. Crossing issues and hazards requires the development of much more complex models, to anticipate and better predict, and price risks. One of the people I interviewed made the following joke: “the statistics we use to price risks are a lantern to light up the future and we carry it behind our backs”. The fact is that the past no longer explains the future in terms of hazards, which may also explain why a good part of the results are chronically loss-making, with the possibility that reinsurers and insurers are losing interest in certain types of risks.

With climate change, regulations are also changing. Insurers are starting to be rated (ESG ratings – Environment, Social and Governance). The carbon impact of insurers’ assets (particularly the type of bonds purchased) must be published. In short, we can dream that with climate change, certain types of polluting activities will no longer be insured (even if we agree that this is not really the trend we are seeing at the moment…).

Insurance, one of the many aspects of risk management in agriculture

Agriculture is one of the sectors that emits the most greenhouse gases, but also the most sensitive to the consequences of climate change. It is therefore absolutely necessary for the agricultural sector to reduce its footprint but also, and above all, to be able to adapt to the risks incurred. And these risks are numerous: frost, heavy rain, flooding, excess humidity, storms, heavy or excessive snow, hail, excess temperature, drought, or lack of radiation. One could even imagine many others, not directly climatic, such as: no germination, absence of fruit set…

This dossier is devoted to insurance in agriculture, and more particularly to climatic insurance (we will come back to all this very quickly). However, it is important to understand that agricultural insurance is part of an overall risk management strategy and that there are many levers that can be used to adapt to the current climate change. These levers – in many forms – can be used independently or in combination:

  • Agronomic levers: Choice of varieties adapted to the local environment and resistant to climate change, varietal mix, sowing density, direct sowing under cover, staggering sowing dates, preserving the value of grassland, crop diversification, changing the breed of animal, bringing forward the sowing period, reducing the number of livestock
  • Technical levers: Installation of irrigation and/or drainage systems, anti-hail nets, photovoltaic panels, etc.
  • Economic levers: Increase in self-financing capacity, setting up of precautionary savings (DEP) during difficult years, subscription to one or more insurance contracts,
  • Strategic levers: diversification of activities on the farm, reorientation of marketing channels, reorientation towards labelled production systems, autonomisation of one’s farming system, carrying out processing activities on the farm to increase the added value produced, reduction of energy dependence on the farm, anticipation and analysis of risks on the farm through training

One of the first actions, perhaps the most important one, consists in acting in prevention of climatic hazards thanks to different agricultural techniques which allow to be less dependent on climatic conditions.

Agricultural insurance

Indemnity insurance

There are two types of “weather” or “crop” insurance in agriculture:

  • So-called “indemnity” insurance is insurance that compensates farmers on the basis of an assessment of the damage carried out by an expert on the farm. We will come back to hail insurance and multi-risk climate insurance (MRC)
  • So-called “parametric” or “index” insurance, which compensates farmers on the basis of the deviation of an “index” or “parameter” from the expected

Note that the term “crop insurance” does not mean the same thing to everyone. I will use it here in a rather broad sense (hail, storm, index insurance, multi-risk…). Some people will only see or think of multi-risk weather insurance, wrongly.

We will review their main characteristics here.

The hail contract

Insurers offer the agricultural sector contracts to insure against the risk of hail. In France, this type of contract represents about half of the climate insurance business. It is an indemnity contract in the sense that the expertise is human and that it is managed in more or less the same way by everyone.

The world of hail insurance has been known for many years by the major insurers in the market. Hail is a ‘named’ peril, i.e. it is a risk that is explicitly listed. Hail insurance operates without subsidy and is a contract that, overall, is balanced for insurers (the loss ratio is around 70 to 80%). This means that the policyholders’ contributions are sufficient to pay the year’s claims and to make provisions for claims where the loss ratio is around 200%, 300% or even more (roughly speaking, when it starts to hail severely…).

Hail insurance contracts are differentiated by hail zone. Insurers have zoniers (maps of hail zones) because the hail corridors are well known. For example, it is clearly known that, in the commune of La Chapelle-en-Vercors, in France, hail falls one year out of two. The apricot trees there are therefore very expensive to insure. Even without a subsidy for this insurance contract, the farmers subscribe because they are aware of the risk incurred and the damage potentially caused to their production. It’s a very real risk, which is frightening, and farmers insure themselves all the more because they know that the state will not intervene in the event of a problem.

The Multi-Risk Climate Contract (MRC)

Multi-risk climate insurance (MRC) was launched in France in 2005 following the droughts and heat waves of 2003, which most of us still remember (report by the Member of Parliament Christian Ménard). Before 2005, the State alone covered this climatic risk thanks to the Agricultural Disaster scheme (this scheme was set up in 1964). This calamity scheme – sometimes called the agricultural calamity fund for short – is financed by the Fonds National de Gestion des Risques en Agriculture (FNGRA); this solidarity fund is fed by taxes paid by farmers and supplemented by the national state budget (this is why there is no VAT in agriculture, as there are already taxes to feed this agricultural calamity scheme). Following the drought of 2003, the budget for the agricultural disasters scheme literally exploded – by a factor of four. The State then organised the deployment of a private insurance industry dedicated to this sector, alongside the State; the MRC insurance was born.

Unlike hail insurance, MRC insurance is subsidised, currently at 65%. Many countries choose to subsidise crop insurance because agriculture is a strategic sector from a government perspective. As discussed above, insurance is one voice among others in the risk management strategy. At the European level, the choice has been made to move from a direct subsidy mode to agriculture (direct CAP aids) to an indirect mode by favouring investment in protective equipment or by subsidising crop insurance. The idea behind the subsidy is to ensure that insurance contracts are spread within farms so that, once they have become commonplace, this subsidy is gradually reduced and the insurance system runs itself.

To return to the MRC, for an MRC insurance contract costing €100, the farmer currently pays only €35 (the contract is 65% subsidised). For a contract set up for the 2022 harvest, the farmer pays his insurance premium in October 2022 and the subsidy is paid around March 2023. In the decision to subsidise an insurance contract, it was decided by the government that a subsidisable contract was one that addressed not just one risk but several (the MRC contract must contain 17 weather hazards to be eligible for subsidy). The idea behind this is that there is no point in subsidising a contract that only takes into account one hazard if the final objective is to secure the farms. Hail is included in MRC insurance but some insurers have a subsidised basic policy and an additional hail policy (a specific hail policy as discussed in the previous section) with a specific deductible for each plot independently (the basic MRC policy covers hail with a deductible per appellation or crop, which is not necessarily advantageous…) We will come back to the concept of deductibles a little later.

The MRC insurance thus enables farmers to benefit from risk coverage that covers all climatic risks and is tailored to their individual needs. Until 2022, there were a number of requirements for subscribing to the MRC insurance policy. In particular, a winegrower had to have all of his land covered, and for field crops at least 70%. In addition to the primary objective of securing farms, the MRC contract had several associated interests. For the farmer, it is already a gain in visibility and management, in the sense that it is simpler to insure his entire farm at once than a small part here and a small part there. For the insurer, this contract is a way of diversifying his customer portfolio; he sells more insurance than if he only sold hail insurance. For the administration, but also for the insurer, it is also a dimension of simplicity that must be seen. The DDTs (departmental directorate for the territories, french acronym) will be responsible for collecting and verifying that the contracts are eligible, for checking the unit prices of the contracts, for verifying that the areas correspond to the declared CAP areas, and for verifying that all of the vineyard areas of a farm are covered if it takes out an MRC insurance contract.

At the same time as crop insurance was launched, legislative support did not intervene much to develop the agricultural disaster fund and to allow disaster and insurance to coexist. The crop insurance system was initially conceived as a new system added to the agricultural disasters system, with the two systems coexisting and articulating each other as best they could, with practices varying by department. Initially, the State’s intention was to let private insurance develop until insurance was sufficiently widespread, and then to remove the crops concerned by the insurance from the agricultural disaster scheme. The idea was that, in the long run, only private insurance would remain in place and the agricultural disaster scheme would gradually disappear from the scene. Thus, in 2009, field crops were removed from the agricultural disaster scheme. Viticulture followed in 2011. The scope of the agricultural disaster scheme was thus reduced, as the wine-growing and field crop sectors were deemed to suffer from insurable risks. The positioning of the calamity scheme or, more rigorously, the scope of intervention of the part of the FNGRA that compensates agricultural calamities, is centred on what is deemed uninsurable. Other crops (meadows, arboriculture, market gardening) remained eligible.

Since 2005, about 30% of field crops and vineyards are insured on the crop contract. For other crops, the figure is less than 5%. The situation has not changed since 2011. The level of insurance is far from being the same everywhere on the planet. Worldwide, the United States, China and India account for three quarters of agricultural insurance premiums paid by the agricultural sector (CICA, 2020). Although the population of these countries is still very large, it is clear that the entire world’s agricultural population is not insured in the same way. In a 2020 report, the International Confederation of Agricultural Credit stated that the value of agricultural insurance premiums in 2016 was around 7% of the value of agricultural production (CICA, 2020). Worldwide, some of my interviewees gave me the figure of 20% of crops insured, and barely 2% of grassland production.

Currently, to have access to insurance, farms must have a grazing number (and therefore have already made a CAP declaration) – the application for insurance is made during the declaration of the CAP file.

Why aren’t more farmers insured with MRC?

It will be difficult to fully explain why only 30% of wine and arable farmers are currently insured with an MRC policy (and even fewer in fruit, vegetable and grassland farming). The explanations are extremely broad and multi-dimensional. A first explanation is perhaps rather psychological. A sort of denial in which the farmer convinces himself that the risk will never happen to him or his crops. In the Beauce region, in France, some farmers felt they were out of the loop until the catastrophic floods of 2016. It is true that the culture of risk among farmers is not extremely well developed, and the training offered is almost non-existent (we will talk about this in the last part of this blog post). But it is also true that farmers are also gamblers in many ways. This is also understandable in the sense that they evolve in a particularly unstable and risky environment.

There is already a first risk linked to the production itself or to the yield on the farm, each crop being subject in its own way to a very large number of events that influence the level of production. When contracting the sale of their production, farmers are also subject to harvest quality criteria. The risks to prices are particularly numerous. For example, there is so-called “imported” instability, linked to the global financial markets and the variability of raw material prices, “natural” instability with climatic and natural hazards that impact production in terms of yield and quality, and “endogenous” instability, which pushes farmers to anticipate the selling price of a product and which produces fluctuations that are uncorrelated with the market. For example, the explosion of the selling price of a cereal will push many farmers to cultivate this cereal the following year, which will generate an overproduction and consequently a sharp drop in the prices of the following year’s crops. Another source of uncertainty is production costs, i.e. all the operational and structural costs required to cover a full production season. We can even add to this the risks on a slightly wider scale, with operational risks, linked to the farmer himself (death, illness, disability) or to his farm (fire, theft, damage, etc.), financial risks since the farmer may be subject to variations in rates on his various contracts or to problems with the liquidity of his farm account, or institutional risks linked to changes in policies, regulations and standards (e.g. a new regulation on an input, a change in the price of a product, a change in the price of a product, etc.): A new regulation on an input, a regulation on the import or export of certain commodities. .)

Farmers are also not necessarily very used to taking out insurance policies for their crops and will therefore prefer to self-insure. Farmers are more used to insuring their buildings and equipment.

Another reason given is that farmers believe that the state will always be there to intervene if there is a real problem. We saw above that field crops and vines were removed from the agricultural disaster scheme in 2009 and 2011 respectively. However, following the spring frost in 2021, we saw the state intervene by putting money back on the table. In 2011, the agricultural disaster funds were also triggered in anticipation of a drought that did not occur. By intervening ex-post, the State sends the message that it will always be there, and farmers no longer necessarily see the point of protecting themselves with a private insurance policy.

There may well be a misunderstanding of the system, with an offer that often appears complex, with subsidy rules that are difficult to understand and a system that is far from obvious. Some contracts require a very early subscription period in relation to the agricultural season (e.g. a subscription at the end of December for a season that starts in March/April), which requires farmers to position themselves in advance and to anticipate risks. Although this blog post is absolutely clear (you have to know how to throw yourself a few flowers), understanding the insurance mechanisms in place was not an easy task… The current reform seeks to harmonise and simplify the processes in place for multi-risk climate insurance, but you still have to roll up your sleeves and spend some time to really understand what is going on.

Some farmers will also tend to complain that the insurance premiums offered are too high, and that the policies are poorly priced for their risk. The concept of the ‘Olympic benchmark’, which is the average level of yield that is insurable on the farm, is often brought up. With several bad years in a row, the Olympic benchmark decreases and with it the level of insurable yield, and the insurance contract becomes much less attractive to the farmer.

Even if this type of contract is subsidised, farmers consider the price too high and do not want to insure themselves. Once again, there is the idea that you are only happy to pay if you are sure that it will do you some good. Can you really blame them? Is the argument about the cost of insurance contracts really justified? We will also come back to this later.

Let’s also add that in some fields, the current offers are not numerous or considered relevant enough to be interested in.

Anti-selection and moral hazard: the main phenomena of indemnity insurance

Four phenomena or types of risk associated with indemnity insurance can be distinguished.

The idiosyncratic or ‘specific‘ risk is a component specific to each farmer. It is an independent risk that does not affect all people at risk at the same time. It is a sort of specific marker of a farmer. For example, every farmer has a risk of having one of his farm buildings burn down. It can be agreed that this risk is independent of the risk that his neighbour has of his building burning down. The idiosyncratic risk is independent between each insured and is therefore a mutualisable risk.

Systematic or systemic risk is risk that does not affect a particular individual but a group of farmers with a common characteristic, such as geographical area or type of crop. Examples of systemic risks are floods, droughts and diseases (when widespread). The disadvantage of this risk is that it is non-diversifiable and therefore cannot be covered by conventional insurance mechanisms. The commonly used method of hedging is to add a risk margin to the contract price of the insurance, which varies according to the probability of the risk occurring.

Indemnity insurance is subject to problems of information asymmetry, in the sense that the insurer and the insured do not have the same level of information about the riskiness of agricultural production and production practices. Indeed, farmers are expected to be better informed about both simply because they are in the field and manage their farms. This asymmetry of information generates two problems well known to insurers: anti-selection and moral hazard.

Anti-selection or adverse selection is the fact that farmers who take out insurance are those who are exposed to above-average risks. When signing an insurance contract, the insurer does not know the exact risk of the insured, and will therefore be tempted to increase the price of the insurance contract in order to cover itself against the individuals presenting the greatest risk. The insurer then finds himself in a situation where only farmers with bad risks – basically, risks that hurt and are likely to happen – insure themselves because only for them is the cost of the insurance premium justified. A farmer who knows that his land is subject to relatively little risk will not necessarily want to insure himself at too high an insurance premium. An insurer’s portfolio of insureds will therefore tend not to be representative of an average risk in an area. The insurer has more bad risks than good risks in its portfolio. It must be said that within a territory, whatever its size, some areas are subject to greater risks than others, some plots of land are more frosty than others simply because the topography is very different.

Moral hazard, on the other hand, is the phenomenon that leads farmers to adopt riskier production practices when they are insured. Two types of moral hazard can be distinguished: ex-ante (prior) and ex-post (after the fact). Ex-ante risk occurs when the farmer voluntarily modifies his practices – for example by reducing his consumption of inputs – in order to reduce or modify the distribution of yields on his plots and thus receive compensation via his insurance contract. This is referred to as ex-ante risk for the insurer because no loss has occurred on the plots. The ex-post risk concerns the risk of fraud, i.e. a farmer may neglect a damaged plot in order to recover or increase compensation.

To limit the problems of asymmetric information, one way out for insurers is to set up deductibles and have experts visit the field. Although the insurer is often seen as a thief in the sense that the insured has an increasing tendency to wait to win back his initial stake (the idea that because one pays for insurance, one should necessarily get something back), it must be noted that if there were no expert appraisal in the field, the insurer would surely pay much more than he should.

Margin and turnover insurance

The main current crop indemnity insurance formats focus on yield or production risk, i.e. a loss of production quantity. We are entering a field that is more of a banking than an insurance nature.

In terms of coverage, certain “turnover” insurance policies are beginning to develop, the objective not being to cover a price risk directly but to integrate a price dimension into the coverage. Price risk is not covered directly by insurers because it cannot be pooled, even on a global scale. At the climatic level, when a drought occurs in one place and not in another, the events are independent (to a certain extent for the most fussy readers). In terms of price risk, if Nebraska and the Corn Belt have a problem with maize, the global price will rise – the events are no longer independent. The current solution, which is not widely used, is to insure the yield via MRC insurance and to insure the price via futures markets. The possibility for the farmer to hedge already exists. But insurers and reinsurers will not really insure a price risk directly. In the United States, for example, turnover insurance does not meet with unanimous approval. Some dispute its principle on the grounds that it would lead to public expenditure even when crop prices are high. It would be better, in their view, to restrict state aid to the provision of a safety net against climatic hazards, limited to covering the main production costs. Others, on the contrary, denounce the fact that premium subsidies and compensation paid are proportional to production and thus favour large farms and the concentration of agricultural land for the benefit of a few. As a result, they are calling for a reduction in the support paid by the programme to the largest producers.

Some insurers are also working on “margin” insurance, which changes the situation in terms of logic, especially for extensive systems. Let’s take the example of two cereal farmers with tight situations and a risk of poor yield. Let’s imagine that these farmers are wondering whether or not they should apply a final nitrogen application to their plots. Taking the last step to try to get your yield (in this case, applying a last nitrogen application) can be considered as a risky operation for agriculture. There could therefore be many cases where the person who does not take the last step comes out ahead if he or she has the insurance net (we have talked about ex-post hazards, we are in that case here). If, on the contrary, an insurer had secured the production of this cereal farmer with a margin insurance, perhaps this nitrogen contribution would have been carried out. Everyone wins, so it changes the logic of the insurance system.

Agricultural insurance reform in France

How MRC contracts have worked since their creation ?

In France, until 2022, the multi-risk climate risk management contract is based on three main pillars:

  • A first, unsubsidised level, which in fact corresponds to a sort of deductible that the farmer will have to pay anyway if he has a claim (this is the farmer’s remaining liability). This deductible is the way in which the insurer or insurance broker can change the cost of the insurance premium paid by the farmer (by increasing the deductible, the insured reduces the insurance premium and vice versa). Some insurers also offer deductible buy-back, which is an additional guarantee where the farmer pays more for the insurance premium but in exchange pays little or no deductible in case of an accident.
  • The second level corresponds directly to private insurance. It is this second level that is subsidised. In France, it was decided to subsidise climate insurance premiums with a share of the European CAP (Common Agricultural Policy) budget, and more particularly that of the second pillar: the European Agricultural Fund for Rural Development (EAFRD). In the example presented above of a €100 MRC insurance contract subsidised at 65%, the farmer would in fact only pay €35 while the CAP would be used to pay the remaining €65. The farmer thus bears a limited part of the cost of the insurance contract he takes out.
  • The third level corresponds to public intervention by the state with the agricultural disaster scheme. This third level should not be considered as a subsidy but as an aid, in the form of a principle of national solidarity.

Too easy, super clear, I understood everything! Three different levels with a deductible, private insurance and public intervention by the state. Yeah, not so fast… Let’s add a little complexity with the levels of coverage and the different types of contracts that can be faced. If you have followed everything, you will have understood that part of the cost of an insurance contract is directly linked to private insurance with part of the cost of the insurance covered by the CAP (this is the second level we have talked about). I gave the example of a 65% subsidy to get an idea of the underlying mechanisms. This 65% actually corresponds to a first level of guarantee (base level) and is linked to a whole bunch of criteria (national references and a scale to set a ceiling on the insured capital, the proportion of the farm’s surface area to be insured, compensation for quantity losses, the triggering threshold [corresponds to the level of crop loss from which the subsidy takes place – for example, there must be at least 30% yield loss], and a minimum deductible) which we are not going to go into so as not to get too lost.

It should therefore be understood here that France mobilises the second pillar of the CAP (EAFRD) to finance up to 65% of the amount of the insurance contribution corresponding to the first level of guarantee (base level). This first level of guarantee aims to give the farmer the means to restart a production cycle. France has also gone further in subsidising this private insurance contract by adding the possibility of subscribing to optional supplementary cover, subsidised at a lower rate than the basic level (45% here compared to 65% for the basic level), allowing the farmer to reduce the excess and adapt the cover to his own risks. A third level of cover, known as optional, also exists to allow the farmer to further adapt the contract taken out to his characteristics. This third level is not subsidised by the EAFRD. Note that crop insurance contracts benefiting from these subsidies cannot receive any other aid financed by State, local authority or European Union credits

Let us also add a point on the fact that two main types of crop insurance contracts are subsidizable:

  • the so-called “crop group” contracts, the principle of which is to pay specific compensation for each type of crop as soon as the production loss observed following a disaster for this crop is higher than the triggering threshold. It is not a contract per plot but per batch of plots on which the same crop is grown. Note that for this type of contract, a minimum area per type of crop must be insured. For example, in order to have an MRC Arable contract on a farm, at least 70% of the farm’s arable area must be covered by the contract
  • the so-called “on-farm” contracts, the principle of which is to pay compensation if the total losses on the types of crops insured following a disaster are higher than the trigger threshold. The main difference with the so-called “crop group” contract is that here there is a mutualisation between the types of crops insured, as gains on one type of crop can compensate for losses on another. To subscribe to this contract, there are, as for the previous contract, coverage obligations. It is required that at least 80% of the area of the farm’s sold crops be insured and that at least two different types of crops be insured.

Why reform?

The main advocates of the current reform point out that the current system of compensation for losses due to climatic hazards is out of date. Several reasons are given:

The historical players in agricultural insurance are losing money. From 2016 onwards, the system of the main insurers has gone into a tailspin (we spoke above about the major floods that took place). Claims escalated to premium loss levels of almost 200% (insurers paid out twice as much as they had received in insurance premiums). Until 2021, the ratios did not fall below 100%. Years like 2016 are no longer exceptional and the consequences of climate change continue to develop (events in frequency and intensity). For the major insurers, the system has become unsustainable. Insurers have had to review their tariff conditions and cover under pressure from their reinsurers and the supervisory authority (ACPR).

The MRC multi-risk weather policy has a penetration rate of only 30% among farmers. There are not enough farms that subscribe to the policy to be able to balance the insurers’ risk portfolio. This is the anti-selection phenomenon we have been talking about, which means that the good risks will leave the portfolio, leaving only the bad risks in the majority. In traditional insurance, rates are smoothed by zone (municipality, department, etc.). If smoothing is only done with bad risks, a farm with a low risk will have a very unbalanced rate compared to one with a much higher risk. However, some will argue that the fact that there is ‘only’ 30% take-up of crop insurance contracts at present for field crops and vines is not necessarily a sign of malfunctioning. It may also be the fact that some farms prefer to self-insure and manage the risk on their own because they are strong enough.

The rather awkward cohabitation between private insurance and public intervention by the State via the Agricultural Disaster regime is said to be the cause of an excessive eviction effect on farmers in the sense that they rely too much on State support in the event of a hard blow (we have spoken, for example, of State intervention for the 2021 frost on vines when viticulture was no longer covered by the Agricultural Disaster regime) The link between the DDTs (departmental directorate for the territories, french acronym), which are in charge of managing the agricultural disaster scheme on the ground, and private insurance is not always very effective, and has sometimes made them more competitive than anything else. The loss rates measured in the two cases are not always the same because local assessments and trigger criteria can be very different (insurers used Olympic yield averages but the agricultural disaster scheme did not take them into account). Farmers can then turn against their insurer or the State, calling one or the other a thief.

The boundary between what is insurable (private insurance) and what is not (public intervention) is absolutely decisive, not least because it is not entirely clear and, moreover, it changes over time (some risks are – or will no longer be – insurable). It should be noted that public intervention by the state in the event of a climatic hazard does not have to take an economic form. Several actions are conceivable (and some have already taken place). Fallow land declared to be of ecological interest can be used in the event of drought. Neonicotinoids have for example also recently been authorised on certain crops. The “Varennes de l’eau” exercise proposed to regulate the supply of methanisers when there were tensions on fodder resources during climatic hazards.

The current complexity lies in the fact that all of these thresholds (deductible threshold, trigger threshold, etc.) and subsidy and aid criteria vary from one crop to another. Up to now, it is indeed by sector that it was decided to change the nature of compensation from the Fonds National de Gestion des Risques en Agriculture (FNGRA) and not by type of risk, even if, strictly speaking, it is a hazard/sector cross-reference that specifies the risks classified as insurable. Some crops, for example, have left the calamities regime because it was considered that the risk associated with these crops was insurable, whereas for others the calamities regime is still available. And yet, as we have discussed, although vines were removed from the scheme in 2011, the state has stepped in for weather hazards in the year 2021

Developments in the reform

The report by Frédéric Descrozailles, Member of Parliament, submitted to the Ministry of Agriculture in the summer of 2021, suggests a number of changes to the current agricultural insurance system. This report already points out three fundamental principles of the upcoming reform.

  • First of all, it is a principle of national solidarity, or in other words, the fact that it is agreed that the State will officially reinvest in the current insurance system. I recall here once again that certain crops were taken out of the agricultural disaster regime and were therefore not supposed to be covered by the State in case of strong climatic hazards. By this principle of solidarity, it is considered that agricultural workers cannot finance the insurance schemes they might need by taking part of the wealth they create. However, we could come back to this game of vocabulary in the sense that the agricultural disaster scheme is still financed in part by taxes on farmers. Nevertheless, it is explained that the majority of the budget of this fund will be financed by national solidarity.
  • The second founding principle of the report is that of the legitimacy of the State to intervene in agricultural insurance in the sense that it is important to make a clear distinction between what is due to the hazard (and which is insurable by private insurance) and what is due to the trend (not insurable by the private sector and requiring public intervention). The State is therefore present to get involved in the management of exceptional or systemic risks, i.e. those that cannot be insured or that require public reinsurance in the case of phenomena of exceptional scope.
  • Finally, the reform is based on the principle of universality, insofar as any farmer wishing to be insured will be able to subscribe to an insurance contract, which was not the case until now.

We agree that the proposed reform seeks to harmonise and simplify an initial, relatively complex system, which was not really easy to navigate. Nevertheless, from there to say that the expected result at the end of the reform is biblically simple, I would tend not to get my hopes up… The reform can be summarised in Figure 8!

Figure 8. Crop insurance reform. Source: Pacifica, internal.

Tadaaaa. Got it? The graph on the left is for the farmer who is not insured. The graph on the right is for the farmer who is insured. So far so good:

  • For the uninsured (left graph): Below the trigger level (“seuil” in french), i.e. the percentage of crop loss, the farmer receives no aid and has to pay out of pocket. The trigger levels are set at 50% for all crops but it is not impossible that these levels diverge as shown in Figure 8 (60% for vines, 50% for arable crops and 30% for trees and grassland). It should therefore be understood that for a field crop plot where the yield loss is less than 50%, an uninsured farmer will have to pay everything himself. If the crop loss is above the trigger level, the state, via the national solidarity fund, will pay 45% but only for the loss above the trigger level. For example, if a field crop plot suffers a loss of 70%, the state will pay 45% of the amount of the loss, but only for the amount between the 50% and 70% loss. The rest is paid by the farmer. For the uninsured, it is therefore a two-tier contract (farmer and state).
  • For the insured (right-hand chart): Below the trigger level, i.e. the percentage of crop loss, the farmer receives no aid and has to pay out of pocket. Unlike the uninsured farmer, the insured farmer has taken out an insurance policy and is therefore compensated for crop losses ranging from 25% (deductible threshold) to 50% loss rate. However, this compensation is not total, the insurer pays between 5 and 50% of the losses – again only for crop losses ranging from 25% to 50% loss rate (these are increments, like what happens to our taxes). I remind you that this insurance contract is subsidised at 65% with a basic level of cover. The farmer, together with his insurer, was able to request deductibles of less than 25% – either by negotiating with his insurer or by having his insurer buy back deductibles by increasing the total price of his insurance contract. If the crop loss is above the trigger level, the state, via the national solidarity fund, will pay 90% (i.e. twice as much as for an uninsured farmer) but only for the loss above the trigger level. The last 10 percent is paid by the insurer. For the insured, it is therefore a three-tier contract (farmer, insurer and state)

In Figure 8, we see here that insurance contracts will depend on the trigger levels. Some advocate identical trigger levels across all crops (at 50%) but it is not impossible that these thresholds diverge as shown in Figure 8 (60% for vines, 50% for arable crops and 30% for trees and grassland). The share of state support (shown as 45% for the uninsured and 90% for the insured) may still vary.

Some thresholds and rates are also likely to change as a result of the application of the agricultural part of the EU Omnibus regulation (there are parts of this regulation for sectors other than agriculture). If fully implemented, this regulation could also allow for changes to the levels of excesses and in particular for the minimum excess to be lowered from 25% to 20%. The Omnibus Regulation also allows for insurance contracts to be subsidised up to 70% (instead of the current 65%). The Omnibus Regulation would therefore reduce the price of the insurance contract for the farmer and also reduce his out-of-pocket expenses in the event of a claim. It must be understood that, by construction, this reform and the potential use of the Omnibus regulation represent a substantial increase in public investment if the penetration rate of crop insurance increases among farmers, since the state would intervene more than before. It was not considered reasonable to take all or even most of this increase from the second pillar of the CAP, even if it were decided to transfer funds from the first to the second pillar of the CAP. Several solutions can be envisaged at present:

  • A return to a contribution rate of 11% on agricultural insurance contracts (which had been lowered to 5.5%),
  • An increase of two points in the surcharge on car and home insurance contracts that finances the natural disaster scheme (the acronym of this scheme is “Cat Nat” in French),
  • An increase in some of the contributions that make up the general tax on polluting activities, in accordance with the principle that activities that have an impact on agriculture (pollution or soil artificialisation, for example, with regard to nitrogen and sulphur oxide emissions or extractive industries) should contribute to financing its adaptation to global warming

It should be noted here that the thresholds of the Omnibus Regulation presented here (20% for the franchise and 70% for the subsidy) cannot be extended, because of WTO trade rules. These are therefore thresholds that cannot be exceeded.

Appraisals between private insurers and the state will be harmonised. The 5-year historical reference of the Olympic average (the average level of yield that is insurable on the farm) will be used for both private and public schemes. Following the reform, the Olympic average combined with the high deductible should exclude the possible over-compensation of farmers.

Two tools will also accompany this reform:

  • A pool (or grouping) of insurers that will be set up to mutualise the risks of insurers by crop, but also by territory (provided that the productions are not all sensitive to the same climatic hazards). The pool will mutualise risks and allow actuarial calculations to be made to cover farmers’ risks. With total mutualisation, it is considered that the calculation of insurance premium levels will be as fair as possible. This pooling would allow the insurers, members of the pool, to improve their knowledge of agricultural weather risks, lead to more appropriate pricing and improved loss prevention measures. Insurers could then offer insurance policies that are better adapted to the diverse situations of insured farmers. This pool is also intended to federate all insurers who will distribute crop insurance contracts.
  • A committee for the orientation and development of crop insurance (CODAR), a sort of governance body where insurers, reinsurers, farmers and the State will discuss these premiums. The CODAR can be seen as a thermometer that allows us to watch the evolution of the insurance system’s loss ratio. This system is intended to be totally transparent so that all the players can see the level of finance in the insurance system and identify risks that are no longer insurable. This CODAR is therefore a real tool for questioning agricultural trajectories in relation to observed claims. If the loss ratios drift too far, it will be necessary to jointly implement adaptation measures in the territories: cease production, introduce tropical varieties, move production basins, introduce crop diversification, or relocate certain types of production. It will therefore be fundamental to take into account the IPCC scenarios and data, and ideally to study climate disruption scenarios by production basin in order to think about adaptations and adjust a time step consistent with the commercial reactivity of insurers.

Agricultural insurance is not made compulsory, but many incentives are being put in place to encourage farmers to take out insurance:

  • The insured and the uninsured will no longer be on the same footing. An insured person will benefit from CAP subsidies for his private insurance contract (second stage of Figure 8) and from state intervention based on the principle of national solidarity (third stage of Figure 8). An uninsured person will not receive any subsidies from the CAP for his insurance contract since he will not have taken out any, and will only receive half the amount of national solidarity that he would have received if he had been insured (the famous 45% and 90% in Figure 8).
  • A tax incentive is also in place through the deduction for precautionary savings (DEP). As a reminder, these savings make it possible to deduct a certain amount from one’s income – in particular so as not to pay too much tax in a bad year. These savings are then returned to the farm account in a better year. A farmer who takes out an MRC insurance policy would thus have access to a 100% deduction for precautionary savings up to €50,000 of agricultural profit (compared with €27,000 for a non-insured person), and 30% of profit above €50,000 (compared with 30% of profit above €27,000 for a non-insured person). Be aware, however, that this tax aid may have to be added to the MRC subsidy in order to be included in the total state aid, which is limited by EU regulations.

Insurance & Digital : Index or Parametric Insurance

Principles and functioning of index insurance

Parametric” or “index” insurance is insurance that is triggered, not as a result of the loss itself (as in the case of indemnity contracts), but as a result of the deviation of an index or parameter (weather index, vegetation index, regionalized yield index, disease pressure index, etc.) from the expected. To simplify a little, we could say that indemnity insurance correlates an expert’s estimate of losses with actual losses, whereas parametric insurance correlates the deviation of an index from the expected with actual losses. For example, the insured’s compensation may be triggered by a level of rainfall measured at a weather station near the farm. If this weather station has a sufficiently long data history, both parties, the farmer and the insurer, have identical information on the insured value.

Since parametric insurance contracts are based on the definition of an index, one can be very imaginative – the range of possible solutions is extremely wide. One could insure against a lack of wind for a wind turbine, a lack of radiation for photovoltaic panels, an increase in the price of gas, a lack of water on the Danube. In the agricultural context, one could insure against frost or excess heat (rather with the help of a weather station), drought or excess humidity (rather with the help of satellite data), excess or lack of rain (with a weather station or radar data), hail with hail sensors. We can also think of certain guarantees for actors around the farmer, such as a seed company, with for example guarantees directly integrated into the seeds. The seed company then uses the insurance as a commercial guarantee for the farmer against a failure of the seedling to emerge or a drought. The amount of insurance paid out is often proportional to the percentage change in the actual index value compared to the indicator set in the range of values, or the total amount of the sum insured. If the index used is good in the sense that it correlates well with what is happening on the plot, then parametric insurance works well. If, on the contrary, the relationship is not good, parametric insurance is immediately less interesting… Having an annual review and regular monitoring of the index seems relevant to check that there is no drift.

For the advocates of parametric insurance, this format has many advantages:

  • Tensions and awkward moments between the insured and the insurer are already avoided when the insured does not agree with the insurer’s claims assessment results. Claims handling can often be a source of conflict (which is why insurers sometimes use insurance mediators). Here the index is measurable and objective, it depends neither on the insurer nor on the insured. The claim becomes objectifiable. In addition, the adjuster is not necessarily always educational in his report.
  • The time taken to process the file is greatly reduced in that the adjuster does not need to go to the field and payment can be triggered directly following the achievement or non-achievement of an index level. The farmer, for his part, can therefore receive his compensation more quickly because he does not have to wait for the end of the campaign and/or the visit of an expert. From a logistical point of view, it is also simpler for the insurer who does not need to develop a network of experts, nor does it need to set up a call centre to receive customer requests for assistance. From an economic point of view, insurers can thus lower their insurance premium rates since the costs of expertise and contract management are reduced. However, this aspect must be put into perspective because it is not this part of the insurance that ultimately costs the most in the total insurance premium (management costs would in fact only represent 10 to 20% of the total cost of the insurance premium). Since it is no longer necessary to make damage assessments, insurers can transfer their risk to reinsurers or to the financial markets by using indices. Even if many producers are affected simultaneously, the insurance company does not go bankrupt because it has covered itself by transferring its risk to a reinsurer.
  • Anti-selection and moral hazard phenomena (see the corresponding section) disappear, again because the index is based on objective elements that do not depend on either the insurer or the insured. The risk will be modelled, for example, as purely weather risk. Everything will be defined in the contract and the insurer will not take any risk related to the terrain or the exact type of crops. The insurer will refer to a weather data, a calculation formula and will compensate according to the index.
  • Unlike traditional insurance, which calculates an insurance premium according to the frequency of a claim on an average cost, index insurance removes the average cost part and transforms it instead into a fixed cost: compensation of X euros if a threshold is exceeded (one can also have progressive index insurance after triggering). The index and the pricing can be more or less complicated to produce and implement. In viticulture or arboriculture, certain damage thresholds are now relatively well known in the scientific literature or easily estimated to trigger insurance. Index insurance based on heat stress or precipitation can be constructed quite easily on the basis of connected weather stations.
  • Parametric insurance also meets the constraints set by indemnity insurance because it is much more flexible. For example, in France, in relation to the minimum area covered, a traditional indemnity policy for hail and storm will require that a vineyard covers all of its vineyard area (100%), whether it wants to insure all of it or not. Some farmers will have capital per hectare to cover for certain plots, whereas an insurer will only want to insure a certain maximum amount per hectare, for example. The procedure can be very simplified with index insurance. For example, an insurer can quote for 20 hectares of cover – with such and such a value covered – without specifying the parcels covered, something that traditional insurance might not cover because of anti-selection. The solutions offered by index insurance are much more tailor-made with deductibles and thresholds that the client can easily refine. Index insurance can insure not only a quantity of product but also the value of a product. It is understandable that a great wine producer, given the selling price of its bottle, would prefer to insure the value of a bottle rather than a quantity of grape juice.

However, one cannot have it both ways:

  • Parametric insurance is so different from traditional insurance that simplifying assumptions must also be accepted. There is already an effort to understand and familiarise oneself with the data, and new reflexes to adopt in the sense that a farmer is not obliged to insure 80% of the land on set-aside, nor to consider a past yield over the last five years. The farmer must understand that if the measured data is validated (from a weather station or a satellite), there may not be any compensation despite a loss. However, there is a parallel with traditional insurance in that human expertise is not necessarily perfect either. The index(es) must be understandable to farmers. Simple – even if imperfect – calculation formulas are sometimes more relevant so that the farmer knows how he is covered and can make the link with his agro-meteorological expertise (the risk for index contracts is that they are gas factories because they want to limit a whole bunch of parameters). The issue of the complexity of the index is a reality. For example, let’s imagine a case of freezing temperatures with a 10% loss of yield observed at a temperature of 0°. We can have 0° with 10% humidity, with 90% humidity, or 0° for 1 hour or much more. With parametric, farms accept the notion of reducing risk at a well-priced price. The risk is reduced but not cancelled. It is not the same comfort as with traditional insurance where, in this case, the farmer can protect between 10% and 90% of the crop loss, and it is the actual crop loss that will be compensated. There is therefore a risk that the farmer will interpret index insurance as an ambiguous lottery.
  • The farmer will potentially have to equip himself and invest in equipment (for a parametric insurance on frost, for example, a weather station is needed). To continue with the weather stations, if we wanted to be very precise, we would need a huge mesh on the ground. Between the top and the bottom of a valley, even without many kilometres separating them, you would need several weather stations. And since insurance is triggered at a certain point, it might be necessary to multiply the sensors, and to put at least one station at the entrance and at the end of the valley.
  • One of the big debates going on at the moment is the ability of farmers to challenge the parametric expertise. The tools must be sufficiently credible and not too contested. The reliability rate of the index is going to have a big impact. At 80% reliability, for example – which may seem acceptable – the insurer will face some farmers that should be compensated but will be told that they will not. The insurer may then fear an extremely high rate of recourse and it is easy to imagine that, in the same village, there are two farmers, one who is compensated when he should not be and the other who is not compensated when he should be. The loss of credibility would be total. Note that the error in the index can affect both the compensation threshold (do I compensate or not?) and the level at which the loss is estimated (depending on the value of the index that is calculated). However, it could be argued here that trust in field expertise is not necessarily always safer. It is necessary that all stakeholders understand the risks when signing the contract. Everything must be made clear and all stakeholders must be aware of what they are accepting. It should be noted, however, that provided there is an absolute and 100% accurate reference, which is never the case in agriculture, assessing the reliability of an index is generally complicated and shortcuts are dangerous.
  • Some insurers imagine that they could do without experts for major claims by defining the damage as much as possible and by characterising it with a hazard. An expert could indeed suggest to an arborist that he only insures against frost from the beginning of March to the beginning of April, and sets temperature thresholds for each level of damage. Pragmatically, however, working with living things requires humility. Not everything is as predictable as we would like. Living things are rarely linear – many influencing factors can come into play, some places can freeze, others not, while the temperature is the same. A yield loss can be explained by a combination of several factors, not just one climatic factor in the year (a little dry in autumn, a little cold at some points in the season…). Framing a climatic event with only a temperature threshold can be relatively limiting. A frost stage on 15 March appears on a flowering in a given year while in other years, on the same date, flowering may not have taken place yet. Targeting a factor to build index insurance is not necessarily always relevant. There may also be significant threshold phenomena, differences between old and young vines. In 2021, some plots of late vines had not yet really started to recover on 8 April (right during the frost observed this year). In some cases, winegrowers were able to receive 50% of their insurance compensation even though they finally made a classic yield at the end of the season. In some vineyard plots with 90% estimated frost on 8 April, with eyes out and freezing, some growers were still able to achieve yields of 40 hectolitres per hectare. Once again, agriculture is not linear (although perhaps the yields of subsequent years will be strongly impacted). By making index-based insurance contracts too simple, someone will always get a slap in the face, be it the farmer or the insurer. Insurance is not currently capable of taking into account all the complexity of living things – for example, in vines, the rootstock, the variety, the type of soil, etc. Will it ever really be able to do so?
  • With index insurance, one can unfortunately also engage in speculation because one disconnects the payment from the cost of the real loss. By accepting index insurance, the farmer takes the risk that his loss will be different from what the index tells him. Depending on which index parameter is set, the insured may be compensated with a certain amount of money, but it is up to the insured to assess whether the amount of money is sufficient for him. As a farmer, I could say to myself that if I think the insurer is wrong, I can buy his policy. If the insurer underestimates, for example, the risk of a wind exceeding a certain speed, I, not being a farmer, could ask him to insure me. I could then play the insurer.

Basis risk

When we discussed indemnity insurance, we pointed out a whole host of risks or phenomena inherent in this contract format, notably the phenomena of anti-selection and moral hazard. Parametric insurance also has its own: it is called “basis risk”. The basic risk can be defined as the fact that there are differences between the compensation received by farmers and the losses suffered by the farm. Basically, the insurer has not triggered the insurance while the client has suffered damage or, on the contrary, the insurance has been triggered while the insured has not suffered any damage. The insurance index may therefore be weakly correlated with individual loss outcomes. This basic risk can be divided into three categories:

  • The ‘spatial’ basis risk represents differences between the conditions measured at a given location and those actually observed at the farm level, for example due to the spatial distance between the farm and the location where weather data was collected. The index in this case does not take into account the fact that weather conditions in the study area are potentially very variable.
  • The “temporal” basis risk is that the time window that was chosen to determine the index or insurance parameter may not be the most relevant, for example between rainfall over the whole year rather than just during the growing season. Temporal basis risk arises from the way a contract selects particular calendar intervals for assessment, while excluding others, or how the contract aggregates observations over time periods such as decades (10 days) or wet or dry seasons
  • The basic “design” risk roughly captures all the remaining sources of error, e.g. missing important variables (weather or otherwise) or implementing the index calculation with poorly adapted and/or biased methods or techniques.

The question of the spatial scale of work is therefore fundamental. The larger the area covered, the more heterogeneous the individual risks faced by the insured. The variance of farmers’ situations (or the idiosyncratic variance in reference to the idiosyncratic risk we defined above) will therefore be all the greater, and the index will be less able to correctly predict individual results. For the index to be of good quality, it is important to be able to correctly estimate the probability distribution of the index and to have good quality data on the variables that determine the index (e.g. meteorological or vegetation data). But data on expected losses are also needed. This data is often difficult to obtain because it is expensive to collect, especially in developing countries, and its availability also depends on the different types of index insurance contracts. Not having the appropriate data leads to a misrepresentation of the true probability distribution of the index and may result in poor correlation between the index and individual loss outcomes. Proxy data is needed to fill in the years where no data is available, which can be problematic in some countries where historical data is not available.

To improve data quality, it is suggested not to rely on proxies or single variables to determine individual losses for weather-based contracts and to complement country-level aggregate data with farm-level observations for area-based yield index insurance. Farmers can be the ultimate source of information for limiting basis risk. Participatory approaches that involve community groups and farmers in the index design phase can result in products with low basis risk.

As the geographical scale narrows, for example from country to region to department, the variability of farmers’ situations can be expected to reduce. Pushing the envelope, when a satellite measurement is perfectly calibrated to capture variation in production and the area is reduced to the size of an individual plot and a single crop type, idiosyncratic variation disappears by definition. Accurate satellite estimation of crop yields is important not only because it directly reduces design risk, but also because the ever-increasing resolution of satellite measurements makes it possible to narrow down insurance zones to more homogeneous areas, each with its own yield index, without a significant increase in operating costs. However, it should be borne in mind that reducing the insurance area to the scale of a single plot or farmer would reintroduce problems of moral hazard, although double-trigger auditing methods can be adapted to remote sensing and used to discipline moral hazard by denying payments to farms whose yields differ significantly from those of randomly selected neighbours. The challenge for index designers is therefore to identify a spatial unit that is small enough for abstraction to be meaningful, but large enough to facilitate the implementation of index insurance contracts.

Some critical literature, however, has criticised index insurance projects for paying ex gratia payments to farmers for basic risk problems to avoid the risk of losing the confidence of their clients, their lenders or the local government (Johnson, 2021). A kind of script of the promise of insurance (which will be there when things go wrong) to stabilise a company’s reputation while avoiding legal liability, regulatory censure or disenchantment with the insurance industry.
In general, the extent of basis risk for different index products is difficult to quantify. It is therefore difficult to conclude whether index insurance products actually reduce risk.

The example of the index contract in grassland

Index insurance contracts have developed mainly on grasslands. But why? It is because grassland is not a crop like any other. It is a crop that is harvested several times a year (there are several cuts) and is grazed. Measuring production is therefore not really possible, except for farmers who count their harvest bales. And in any case, as 95% of the production is self-consumed on the farm, the grassland is not really considered as a sales crop – so it does not really appear in the farm accounts – and the farmer is not necessarily interested in a drop in production. If the level of production is not measured, then there is no history to go by, and so traditional indemnity insurance is almost impossible to price.

Many countries have moved into index insurance contracts on grassland, particularly using satellite remote sensing technologies, because this is what, from experience, seems to work best. Grassland biomass can indeed be measured by vegetation indices. Most of the players involved in grassland insurance think in terms of fairly large spatial scales, for example at the level of municipalities, agronomic zones or small forage regions.

The principle of the grassland index is to follow the growth profile of the grassland – globally between February and October in France, with triggers at the start of growth until the end. The idea is to compare the production profile of a given year with those of the same area over previous years – so we are comparing biomass production integrals over different years. It is important to understand that you still need a history of satellite data to be able to go back in time. Several satellite vectors can be used depending on the spatial and temporal resolutions used. Some readers may immediately think of the Sentinel 2 satellites, even though they are not actually the ones that are currently used too much, firstly because the data history is not yet extremely large (Sentinel 2A was launched in 2015, and Sentinel 2B in 2017), but also and above all because the fairly fine spatial resolution of 10 metres can lead to anti-selection and moral hazard phenomena. The index also needs to be corrected or harmonised for cloud, topography or soil type effects. Weather indicators complement the model and are used to artificially penalise biomass production during the season.

The experience of French actors on the subject would tend to show that the low use of index contracts in grassland (there is no indemnity contract for the reasons mentioned at the beginning of this section) would not be due to the quality of the index. The indexes would be rather well correlated to biomass and some index services would even go through a commission every year with institutions from several different sectors (Idele, Meteo France…) to validate their relevance. Farmers would actually have confidence in a satellite measurement but would have a poor knowledge of their production profiles, this being due to the fact that farmers make economic trade-offs – admittedly quite legitimate – but for which they do not necessarily make the link with their grassland production profile. Grassland indices measure what has grown, not what has been harvested – and that is the big difference. In the alpine pastures, if the grassland has grown 2cm, I can ask myself as a farmer to go and harvest. I might only harvest 1 tonne of dry matter, but if I have 15 km of driving to do, I might decide not to go and harvest as it’s not worth it. But if you think about the grassland, 1 tonne is still a production that cannot be completely neglected, and whose condition does not depend on not growing (because the grassland has really grown by 2 cm) but rather on the farmer’s decision or practice not to harvest. With forage stocks built up in the spring, autumn grassland regrowth is not always harvested, sometimes because it is not considered to be of the same quality as the original grassland. There may also be cases where harvesting is not possible – for example during a flood. In short, all these cases will cause the index, which measures a level of growth, to diverge from what has actually been harvested.

Other digital technologies for insurance

Data providers can supply insurers in a variety of formats. During the season, this can be providing weather indicators, validating that a plot has been planted, assessing the percentage of the plot that has emerged, estimating the date of emergence, comparing the development of biomass on several plots, or supporting the experts with indicators and maps to respond to farmers’ complaints.

Digital tools make it possible to play on management economies. Some of them propose the concept of picture-based indices, in the sense that a farmer could send a photo of his plot of land, time-stamped and geolocated, to highlight the impact of a loss. Experts can also use digital tools to measure hail and game damage directly on a map, and agree with the farmer on the delimited areas. High-resolution images (drone, aeroplane) can also be used to assess emergence percentages or seedling counts to provide reseeding advice to a farmer. Note that these drones can be used for much broader insurance purposes for the farmer (legal or administrative problems, land issues, etc.) but in this blog post we will focus on the climatic part.

What is expected of data from digital technologies? First of all, that they are reliable: the measuring tool must not break down and the data must be certified by the supplier. The fact that the data must be certified implies somewhat indirectly that the insurer should not develop its own measurement tools in the sense that it could be accused of playing both sides. For example, it might be preferable for an insurer not to deploy its own network of weather stations, but rather to offer an insurance package that complements a weather station from a third party provider. The insurer must go through a trusted third party. In the example of weather data, this may be data from a third party operator (such as météo France) – this is known as an oracle, i.e. a third party source of information whose sincerity and accuracy is recognised by all. It can also be the data of an insured person (the weather station on his field) that the insurer would pass through a trusted third party to validate the quality, this trusted third party could be a manager of weather stations or data, or experts in data analysis who can confirm that the data is good (maintenance correctly carried out, statistical validations, validation with spatialized climatic data…). For the insurer, it does not matter what method the trusted third party uses if the data is certified. If the trusted third party takes the risk of certifying the data, the contract can be fulfilled. For weather stations, however, it should be noted that global models (Météo France or others) are fairly reliable – and locally too, which also opens up ways of checking that the data sent by the station does not present unexplained anomalies. Another point not to be overlooked is that these data must be representative of the entity being measured, in particular to avoid the basic risk. It is up to the insurer to check that the data is representative.

By looking at the insurance field, I thought I would be able to glean feedback on decentralised insurance formats and crypto-currency payments to compensate farmers. I thought I would hear about the use of blockchain and smart contract technologies. I soon realised that this was not going to be the case because these tools are still only used at the margin.

The smart contract is basically a computer program in which the terms of the insurance contract have been specified (for example: “pay an indemnity of so much to Mr X as soon as the plot of land has frozen”) and which will execute a transaction (the indemnification of the insured) in a blockchain to secure it as soon as all the indemnification conditions of the predefined contract are met (in our example as soon as the plot of land has frozen). The smart contract is not the contract in the legal sense, but a tool for automating and securing it through the use of blockchain. Smart contracts can also be automatically supplemented by data from a third party organisation – the oracle, which we mentioned earlier, and which is a third party source of information whose sincerity and accuracy everyone recognises.

In the insurance field, blockchains and smart contracts could thus allow :

  • Reduce insurance premium rates by reducing the costs of managing and structuring insurance contracts because the system is automated (processing of the declaration, verification of the reality of the claim, management of any disputes via an oracle)
  • Accelerate insurance procedures thanks to the use of third-party data from oracles or objects connected to the plots (weather stations). In the event of a large-scale event, such as a natural disaster, one could imagine filtering out cases that present no ambiguity in order to concentrate on the most delicate situations
  • Imagine agricultural insurance in a collaborative way, with policyholders grouping together in a community, declaring their assets and paying a contribution to insure them. Most of the money could be placed in an escrow account and the rest invested in an insurance contract (with a reinsurer for example). In the event of a claim, the insured would be paid out of the money in the escrow account and then out of the insurance contract when the escrow account is empty. The following year, community members would only pay the amount that would replenish the escrow account and pay the reinsurance policy. The aim of collaborative insurance is to make insurance meaningful again by making policyholders responsible and avoiding the feeling of paying without ever getting anything back (although we did discuss the fact that the purpose of insurance is not to make money for a policyholder).

For some, these blockchain and smart contract technologies are too oversold. If blockchain-based insurance contracts use oracles, the link with a third party would still be too present, and there would not be too many differences with a classic insurance contract. Others wonder what blockchain could really add to current contracts in the sense that the issues of the moment are not about the tangibility or robustness of the contract. Legally, contracts stand up, and insurers and policyholders would not have a problem with trusted third parties or insolvency. Trust between an insured and an insurer would be based on the insured’s perception of the promise made by the insurer. Several interviewees also pointed out that the human dimension is preponderant in the insurance field (climate risk has a very emotional and traumatic dimension). Compared to other risks, climate risk is much more expensive and more frequent. Some insurers therefore consider that not everything can be settled with purely automatic aspects. Operators will necessarily have to be accompanied in difficult situations.

From a logistical point of view, total automation can be made difficult by the fact that it is necessary to wait for the result of the field expertise, particularly in the context of indemnity contracts. Some insurers, in order to compensate for certain delays in triggering payment, would make advance payments during the season when the loss rates detected (in the framework of an index contract, for example) are already too high.

It should also be noted that farmers, particularly in developing countries, may not have access to the infrastructure needed to participate in a decentralised blockchain-based insurance system.

The legal issues surrounding blockchain are also not obvious and are not necessarily clear-cut (supervision of relations between participants, governance of blockchains, management of personal data, sharing of information between insured parties, etc.). One could also question the motivation of the parties involved in the transaction (insurer and insured) to provide authentic and accurate information, especially for small farms (I refer you to one of my last big blog entries on digital technology as seen by the humanities and social sciences).

Index-based microinsurance in developing countries

Insurance mechanisms in agriculture are also developing in developing countries. Here, we tend to speak of micro-insurance rather than insurance, in the sense that the insurance premiums and compensation correspond to relatively small amounts. However, the modalities and operating principles are the same for both insurance and micro-insurance. The insurer insures smaller amounts because the financial levels are also lower. The difference is made in terms of the level of return insured and the yield level of the farmer. Some of the interviews allowed me to play with vocabulary, notably by questioning the real difference between what is considered insurance in developed countries – which is often subsidised and allows lower insurance premiums to be paid by the farmer – and micro-insurance in developing countries – which is sometimes unsubsidised and insures lower amounts. Criticising insurance in developed countries but not micro-insurance in developing countries might seem paradoxical.

The principle of index or parametric insurance is of interest in developing countries because it provides relief in places where administrative capacity is weak, credit is limited and past disaster relief may have been delayed. Here, the law of large numbers prevails, i.e. the insurer is remunerated on a small margin but which, multiplied by a large number of insured farmers, allows the insurer to find his interest. It is a mass insurance with a low risk for the insurers because the policyholders insure small amounts. But is this format profitable for the insurer? One would think that it has to be, but on the other hand, the insurer is also there to provide a security service and farmers in developing countries also face serious risks and may need to be insured. So there is also a notion of responsibility of insurers and reinsurers here. Index insurance also allows the insurer to have fewer control costs. The insurer would not be able to assess all the areas of the insured farmers anyway, firstly because the number of farmers is much greater since they cultivate small areas, and secondly because the networks of experts in the field are not always developed anyway. In situations where little historical data and few statistics are available, index insurance products are more relevant than traditional insurance products.

Nevertheless, the same limitations as for classical index insurance are present, especially with the basic risk. The more geographically dispersed the policyholders, the greater the variability of agricultural conditions, and the greater the basis risk. It is therefore important for insurers and reinsurers to have a diversified portfolio of policyholders in several different regions because, even if the amounts of microinsurance are small, a flood or storm can affect a lot of small farmers at once.

However, it is important to remember that microinsurance is not the solution to all ills either. West Africa, for example, presents a contrasting landscape where nearly 60% of the population is rural, with an agricultural sector that accounts for 40% of GDP, even though agriculture accounts for barely 5% of credit to the economy (CICA, 2020). Microinsurance can hardly miraculously remedy all these disorders, but it can, on the other hand, if it is supported, help to improve cooperation between the actors in the area. In developing countries, this index insurance can be interesting in that it can guarantee credits taken out by producers, particularly seasonal credits (such as inputs), for example when environmental conditions – particularly rainfall – are insufficient. Since lack of finance and access to credit can be one of the reasons for subsistence farming and very low yields, credit guarantees can be very interesting. It works as follows: the micro-insurance is used to pre-finance inputs (seeds) and, if rainfall is insufficient, producers do not have to repay the input credit. However, micro-insurance premiums should be affordable enough for small-scale producers, and premiums should not increase too much when a major climatic event is announced.

Beware, however, of the possible deviations of these insurance formats in these countries. When this insurance is a way for agribusiness firms to penetrate new markets and sell hybrid seeds, the issues of poverty alleviation, inequality and natural resource conservation must be put back at the centre of the debate.

Let’s try to take a step back

What can we say about the reform in place?

A first quick assessment

The first challenge of the proposed reform in France is now that of numbers and mobilisation. In order to double the number of insured people, and in particular to reach 60% of insured areas in field crops and viticulture, and 30% of grasslands and arboriculture, it will be necessary to convince 70,000 livestock farmers, farmers and winegrowers who are currently not insured under the MRC to take the plunge. According to my interviewees, the objective seems achievable in arable farming by extending the hail contracts in place. In viticulture, the 2021 frost will degrade the Olympic average, reinforcing the temptation to stay with hail contracts based on yield to potential. In arboriculture, it remains to be seen whether the offers proposed will be sufficiently attractive to increase the low rate of insureds. As for livestock farming, farmers seem to be attached to the agricultural disaster scheme despite its shortcomings and have not yet been convinced by the parametric insurance for grasslands.

The devil is in the details. The reform could indeed be played out on the criteria and modalities chosen. And in terms of thresholds, we have seen that things could go well (trigger thresholds per crop, level of state intervention, levels of deductibles, taking into account the Omnibus regulation, etc.). Those who are sure to see their situation improve in the short term are perhaps the insurers, in the sense that their risk will be capped because their commitment will be more limited. At the level of farmers, we will probably have to wait for the full set of modalities before drawing any real conclusions. However, experience since 2005 shows that pricing alone will not be enough. At this stage, there does not seem to be a clearly defined means of promoting the reform to the uninsured, and especially to farmers.

Can we ask a farmer to pay for a risk to which he is subject, knowing that all citizens depend on agricultural production? The reform proposed by the government recognises the principle of solidarity by restoring the national solidarity fund. The insured, like the uninsured, will have access to it in the event of a major crop loss (even if not in the same proportion). It cannot therefore be said that the entire burden of insurance is borne solely by the farmer. Some actors, notably the Confédération Paysanne, are defending even more solidarity-based models, based for example on the mechanism of the compulsory voluntary contribution (CVO), with the idea of making the actors in the food chain who make a lot of money contribute. The compulsory voluntary contribution is not the responsibility of the State but is the result of inter-professional agreements. It is therefore the professionals concerned who must ask the State to extend the CVO to all players in the sector. In other words, more than 70% of the players in each downstream agricultural production chain would have to agree to ask the government to extend this contribution scheme to all. With the number of sectors in place, is this solution really plausible? One may also wonder whether the players in the agri-food sector, even if they paid additional contributions in the framework of the CVO, would be ready to assume a price increase for the consumer. There is such an imbalance in the sector that it is possible that this cost could be passed on to the farmer…

Setting up the insurance pool

The proposed pool of insurers should respond to a lack of visibility and a demand for greater transparency in the insurance system. And this transparency is multiform. First of all, it is transparency on tariffs so that the players are aware of the results of the MRC insurance, the evolution of the insurance tariffs and its options, its state of equilibrium or imbalance, or even what is the insurable risk (by the private sector) and the non-insurable risk (paid by national solidarity). Comparable systems are needed so that policyholders do not feel that they are being cheated and to change attitudes.

Nevertheless, the insurance pool, under the pretext of a positive desire to make the market more dynamic (with strong data sharing between insurers), should not result in a standardisation of offers; the risk being that all the small insurers will withdraw from the market. It is through a greater diversity of offers and competition between offers that each farmer will be able to find his or her account. It must be said that at present, contracts are already fairly homogeneous because the insurance system is regulated. The contracts differ slightly in the tariff models and the mathematical models used. In addition, the experts in the field are independent, not employees of the insurance companies. These adjusters can therefore work for several companies, and each time use the adjusting method of the company for which they are carrying out a claim assessment. With the reform, contracts should be even more homogeneous, not only because the specifications will remain identical but also because the public authorities will require that the claims management methods be the same. Methodologies will also be harmonised and institutionalised. Tariffs may then be very similar but not exactly the same in terms of competition law. Sharing data could also make it possible to specify loss rates and potentially to exempt oneself from certain expert assessments. It is conceivable that, for a homogeneous pedoclimatic zone with loss rates actually measured by an expert, if a farmer located in this zone calls for a climatic hazard, the measurements already taken by the expert could be used as a reference value. Of course, care will have to be taken as to what can be transposed from one plot to another (an error at the margin accepted by policyholders and insurers) and what cannot, with plots that are too different and measurements that are not representative.

However, some would argue that the pool of insurers is also a way for the major insurers in the market to share their bad results with the rest of the pool (the major insurers are currently losing money on crop insurance). It is important to understand that the risks will be pooled within this insurance pool. This means that the insurer who earns 5% of the insurance premiums on the pool will also have to manage 5% of the claims of the whole pool. But if, as a small insurer, I have made sure that I have a diversified portfolio of clients, I will also have to share on average the claims of the big insurers who will have a portfolio that is perhaps riskier than mine (because the major insurers are currently suffering from the anti-selection and moral hazard phenomena we have been talking about). The main insurers in the market are multi-branch insurers and do not only offer weather insurance to farmers, they also insure their cars, tractors and buildings. Crop insurance is therefore also counterbalanced by other insurance formats.

For others, the insurance pool will allow all insurers to come in with their claims, premiums, and loss experience. New insurers in the market could benefit from the data, history and experience of insurers who have been in the market for a long time. New entrants would not need to develop tariff models and could benefit from the data of the old ones.
The ongoing insurance reform could potentially lead to an increase in insurance tariffs. As the technical balances of the main insurers are not reached (loss ratios), farmers will certainly be insured not with 25-30% deductible but with 20% deductible (if the European Omnibus regulation is implemented). The insurance premium will therefore be increased for the insured. It is to be feared that the first reaction of farmers will be to refuse to accept this price increase. The responsibility for the poor technical balances of the main insurers may thus appear to be more diluted through the insurance pool.

However, it should be noted that the details of this pool are not yet fully established. The Spanish, within their AgroSeguro system, have pooled data, data management methods, premiums, claims, and the information system, actuarial work, or research and development. The pool planned in France does not seem to go that far. It should be remembered that in Spain, for example, an insurance pool does exist, but it was created at the time because there was no private reinsurance capacity. A pool had to be created to create competence. In France, this capacity already exists.

Is the budget allocated sufficient?

The State recently agreed that the national solidarity fund would be endowed with almost 600 million euros, i.e. a doubling of the budget allocated until now. One third of this budget will be taken from the CAP, with the rest coming directly from the state. Is this a large amount? Is it a sufficient amount? To give an order of magnitude, 600 million euros is the second or third budget line after the CAP. It’s about the same level as the tax exemption for biofuels. So it’s not a completely negligible amount. This amount, which is indicative, is in any case quite theoretical. With the simulations of the DGPE (Directorate General for Economic and Environmental Performance of Enterprises, French Acronym), if all things were equal (with today’s loss data and the same distribution of insured crops), several interviewees agreed that the budget could cover, on average, the loss data, for a doubling of the covered surfaces, i.e. for twice as many crop insurance contracts taken out by farmers.

However, several unknowns remain in the equation. What will be the impact on the budget if the penetration rate of insurance in fruit growing and grassland increases (it is currently less than 5%)? If the insured area doubles but the risk portfolio also changes, the cost of insurance is likely to be higher than expected. Not to mention the fact that the ongoing climate change will most likely increase the loss ratio in the years to come. From a more down-to-earth point of view, the choice to follow the European Omnibus regulation by reducing the deductibles of insurance contracts should increase the insurance budget quite significantly (the application of the Omnibus regulation is a real improvement of the risk but it is expensive). Omnibus allows for a reduction in insurance deductibles from 25-30% (current threshold) to 20%. At present, a farmer could quite easily ask to reduce the deductible of his contract from 30% to 20%. The price of the insurance contract (the insurance premium) would rise quite sharply with this deductible reduction because the distribution of claims roughly follows a normal distribution (a Gaussian) around a 25% deductible. However, it is important to understand that at present the state subsidy does not cover the 20-30% excess, it only applies to the part above the 30% excess. If the Omnibus Regulation were applied, the state would have to subsidise the excess from 20% – the subsidy would increase sharply by construction as the price of the insurance contract would also increase substantially. With a lot of subscriptions, and low deductibles, the state budget could therefore prove insufficient.

The subject of insurance is so important that some people are even calling for it to become an international issue, and why not create a third pillar of the CAP. Agricultural production is a very important issue for geopolitical stability – the current situation in Ukraine is a painful reminder of this (some institutions are already predicting shortages in certain parts of the world in the coming months). With the risks that climate deregulation poses to global agricultural production (food security, malnutrition, etc.), this subject cannot be taken lightly.

The issue of yield references

The levels of insurable crops in private insurance contracts are calculated from a so-called ‘Olympic’ average yield, an average based on the yields observed over the last five years, taking out the best and worst years (this is also known as a three-year average since the average is taken over three of these five years). As we have seen in this dossier, the use of this Olympic average is rather criticised in the field because with the current climatic deregulation, the observed yields are likely to decrease sharply, and therefore the level of insurable yield as well. Some people propose instead to use the ten-year average, which would mean calculating the average over the last 10 years, giving a slightly more homogeneous horizon to the past yield calculation. Technically, for an insurer using a 10-year average (instead of 5) is not very complicated. The difficulty lies in finding old historical data (e.g. in a farmer’s records).

It is nevertheless reasonable to think that this gap between the Olympic average and the ten-year average will gradually be reduced over time. We are indeed witnessing an acceleration of the degradation of production potential. Optimistically, one might think that this degradation is not destined to continue indefinitely if new practices, itineraries and rotations stabilise new farming systems. Unfortunately, in reality, we can rather fear an exponential acceleration of the damage linked to climate change with strong positive feedbacks between climate, yield, diseases, pests, or even invasive species. These feedbacks could lead to a domino effect making it very difficult to evolve and adapt. From a legal point of view, compensation for the acceleration of the deterioration in production potential, as measured by the difference between the Olympic average and the ten-year average, could be considered to fall within the category of aid intended to remedy the damage caused by natural disasters in the agricultural sector. In order to link this aid to the natural disaster regime in French (the “Cat Nat” regime), one could play on the relative imprecision of the notion of natural disaster, which provides for leaving it to the government to interpret and qualify the facts.

Note that parametric insurance does not rely on the insured’s Olympic averages. It is the historical losses over the longest possible history (30 years if possible) that are taken into account. The average trend of these historical series is removed (known as “detrending”) to take account of past developments (trend towards improved production, climate change already felt, etc.)

The concept of the Olympic average is imposed by Community regulations applicable to compensation from the Fonds National de Gestion des Risques en Agriculture (FNGRA) by virtue of the 1995 Marrakech agreements within the framework of the World Trade Organisation (WTO). For some, moving away from the Olympic average would change the nature of the compensation paid; this compensation would then become sectoral aid and no longer exceptional aid. Stepping back from the ongoing climate deregulation, it can also be considered that shifting the timeframe of the Olympic average (e.g. to 10 years) would have a momentarý interest in the sense that with a long trend of global warming, it would actually only postpone the problem by a few years. In a context where crop yield performance is deteriorating and reference yields will evolve downwards, one may nevertheless wonder whether limits have been determined or foreseen on the decrease of reference yields.

In traditional crop insurance contracts, the object of the guarantee is the losses of year ‘n’. These are only crop losses (often quantity losses, although there are provisions for quality loss, reseeding costs, or additional harvest costs). On the subsidised multi-risk climatic insurance, there is nothing on substantive losses. If vines are destroyed, this is not taken into account. In the case of parametric insurance, it is a little different because the object of the guarantee takes into account any pecuniary loss relating to the events. So it’s not just a drop in yield but it can be additional costs of fighting global warming (that would be related to a pecuniary loss here).

In general, the question of reference is always a taboo subject. The reference reassures us because we feel we are comparing ourselves to something that is authoritative, something precise that cannot be questioned. In the field of insurance, this notion of reference is actually quite subjective. A benchmark of performance level based on cuts or field sampling will depend on the expertise of the operator. According to several of the people I interviewed, some expert tests carried out under controlled conditions during insurance days and/or events (e.g. by the International Hail Insurance Association) have revealed differences of up to 30% between the same claims assessed by different experts. The idea here is not to question the expertise of field operators, quite the contrary, but to question the notion of reference which is perhaps too often considered as the absolute truth. Moreover, it should be borne in mind that when there is no reference for a given area for a crop of interest (and if the farmer has no reference), it is to public or neighbouring sources that one must turn (the yield level on a neighbouring plot, for example), which – you will agree – is not necessarily the most accurate. This notion of reference remains important in any case because insurers must have opposable values to put in their insurance contracts.

Towards a demutualisation of risk and an individualisation of society?

Are we not witnessing a trend towards the demutualisation of insurance risks and an individualisation of society? One might get the impression that the insured refuses to pay for insurance without a return on his or her investment, which, as we have already discussed, goes against the very principle of insurance. Demutualisation seems to creep in quietly through selection by making insurance premiums tailored to each insured. And it is the reinsurers who reconstitute the mutualisation through a diversified portfolio of insureds. The digital economy and associated technologies offer the possibility of creating individual offers without significantly increasing the value of a product. The individualisation of the insurance offer implies both the individual assessment of the risk by increasing the amount of information collected on the insured and the object of the insurance, but also the preparation of an individual insurance offer at the request of the insured. Could we end up with hyper-individualised insurance, such as insurance based on the genetic profile of the insured, on smokers, or even on road contracts to distinguish between heavy and light drivers? Wouldn’t we then have reached a point of no return? The climate change problem we are facing is so systemic and requires such profound changes in our behaviour that the trend towards individualisation in society is frightening.

Further discussion

Today, insurers regulate by zone, but perhaps insurance should already be regulated at national level. Climate deregulation requires a certain amount of distance. Climatic risks are very heterogeneous between regions. Even if certain climatic hazards can cause catastrophic damage on a yearly scale in certain regions – we saw this in 2021 with the frost episode – the whole of France was not affected. A risk portfolio at national level would make it possible to continue pooling risks.

The agricultural insurance market in France is very unbalanced, with two insurers sharing the majority of the market share. A more balanced market, say five players with more than 10% of the market share, could perhaps make the sector more dynamic. The influence of the players in the market is very important, with the possibility to put pressure on their current insurance capacity. One could fear that the main insurers in the market only offer solutions within the framework of the CAP (the hail contract, for example, is not subsidised by the CAP, unlike the multi-risk climate contract).

Some crops – forage and vegetable seedlings – are currently not covered by crop insurance (or with prohibitively high premiums). One can also think of small-scale diversified systems and complex agrosystems (which are known to be resilient factors on a farm) for which it still seems complicated at present to offer systemic insurance. Crop insurance is often accused of being an obstacle to better ecological sustainability of agricultural practices, because it would encourage the turning over of grasslands and discourage crop diversification. How will these systems and/or crops be covered under the reform? We obviously do not know much about this.

Indemnity insurance or index insurance?

The insurance market seems to be rather organised between the proponents of the all indemnity or the all parametric, most insurance, brokerage or insuretech firms are specialised in one type of insurance. But do we really need to choose definitively between these two insurance formats? Can we not rather imagine that these two formats are complementary and that, in the end, neither replaces the other?

Indemnity insurance and parametric insurance do not play on the same spatial scales. At the level of a French farm (50 to 100 hectares), in view of the indices available and the diversity of soil and climate on the farms, the granularity is limited enough to hope for a perfect parametric correlation between an index and the reality of the loss. On the other hand, at the level of a production basin, a producers’ union or a cooperative, the parametric seems to be much more relevant. From a technical point of view, parametric insurance does indeed make it possible to make substantial savings in terms of field expertise and management costs. This insurance format nevertheless obliges the parametric insurer to do a lot of work upstream of the contract to link the indices to the operator’s losses and reduce the basic risk as much as possible. In France, the index-based approach could also replace the current indemnity-based approach for the intervention of the Caisse Centrale de Réassurance (CCR) in activating the new solidarity fund for agriculture. If the agricultural insurance reform works as planned, a substantial increase in the number of insured persons can be expected – especially since insurers will now be obliged to insure a farmer who requests it.

At present, insurers will not have the capacity to meet such a demand for insurance contracts. The population of experts is sized to appraise 30% of the surface area (this is the current rate of subscription to insurance contracts for field crops and vineyards). On 1 January 2023, by increasing from 30 to 100% of potentially insurable areas, whether the client chooses additional guarantees or whether he chooses to be under the agricultural disaster fund, the experts will not be able to triple their numbers. The development of indexing could be one way to respond. It may already be necessary to make market players aware that parametric insurance is not a break with the existing insurance system but a technical approach like any other. This will nevertheless require training – particularly for advisers in the field – so that they can feel comfortable with this type of contract. Between an indemnity contract and an index contract that would protect a crop during only the most risky 3-week period in terms of drought, it is to be expected that the advisor may not feel comfortable proposing an index contract to the farmer and would rather push a more classical multi-risk contract. However, the insurance reform does not mention much about the deployment and promotion of parametric insurance.

The major difference between the two insurances is that indemnity insurance is subsidised, whereas parametric insurance is not. It is therefore necessary that the basic risk of parametric insurance is very low for it to be preferred, or that there is no indemnity offer. For example, one can think of uninsurable risks such as disease risks. As this risk is not insurable under an indemnity scheme, it could be insured under a parametric scheme with a disease pressure index. It does not matter what the loss of production or yield is. It is the value of the disease pressure index that is insured. Once again, moral hazard can modify the behaviour of the insured in the sense that a farmer insured against fungicide risk may not monitor his plot as carefully as if he were not insured. The same example applies to the quality of cereals. For a farmer under contract with his cooperative or trader and committed to a grain quality level (specific weight, protein, etc.), one can imagine retroactively reconstructing a correlation between indices and the quality of the cooperative’s collection with a weather database (it is in the farmer’s interest to have a good correlation so that the contract is renewed). In other cases, there is no compensation offer simply because there is no field expertise. This is the case, for example, with the grassland mentioned above, because the expert can redo a forage assessment for the current year, but not for previous years.

Parametric insurance is also particularly interesting in cases where historical data is missing and the risk cannot be correctly modelled. Parametric insurance makes it possible to answer very concrete questions from policyholders about measurable things: insurance on rainfall intensity, flood heights, frost. In Africa, where there is not necessarily a network of experts and historical series, it seems interesting to combine pure index data (vegetation index, soil moisture, temperature) with specific field measurements carried out by the clients themselves. In the framework of certain contracts, it is the policyholder who takes a photo of his crop and then sends it to qualify the types of damage. It is clear that, in any case, there will always be very complex parametric cases with a basic risk that will require an expert.

For the fervent defenders of parametric insurance, it seems difficult to make them accept the use of a human-sized counter-expertise, in the sense that this counter-expertise would tarnish the image of purity of index insurance. On the scale of a structure such as a cooperative or a trade, it would be acceptable, for example, for 5% of claims to be counter-examined. However, it is possible to imagine in the future, for simple claims or with a hazard that would affect an entire basin (in the case of a drought, for example, as opposed to a hail hazard that is more likely to materialise in the form of hailstorms), that we switch to parametric insurance systems but that we keep an indemnity part for claims that are more complicated to model and to counter-estimate possible results. In the end, therefore, it is perhaps a hybrid approach that should be facilitated, where the index will help or augment human expertise, but never really replace it. Digital technologies could make it possible to generate situation maps that cross-reference weather data, vegetation observations and the impact of an event, and which, when crossed with a given crop, would make it possible to give the expert a range of areas that are really affected or not.

Insurance and climate change

Is it necessary to take out insurance?

Perhaps the first question is whether one can insure oneself. A farmer can only have access to an insurance product, just as he can only build up savings – whether precautionary or not – if he generates enough income to do so. A farmer with a gross operating surplus (GOS) of 10,000 euros, and a volatile one at that, cannot save or be insured. It is clear that the percentage of farms that do not generate such an income or that do not generate it on a stable basis is quite terrifying.

At the other end of the scale, those who do not need insurance are those who are strong enough to withstand several successive bad productions – one can think, for example, of the great wine producers. The principle of insurance is to pay the value of the product in order to replace it. If a grand cru loses its entire production following a hail storm and is insured, it will certainly receive substantial compensation. But will it be able to buy back the same grand cru? Nothing is less certain, and one can then ask oneself if such insurance makes sense. If, on the other hand, I am a cattle producer and grow silage maize, and it hails, I need capital to buy back silage maize (or straw or whatever). In this case, hail insurance plays its primary role.

For a more traditional farm with a reasonable gross operating surplus, it still makes sense to protect a sales contract. There are situations where it is not necessary to insure and farmers are right not to do so. There are others, on the contrary, where farmers should or should have insured themselves. Indeed, one may wonder how some farms can survive without insurance in the face of regular climatic hazards on the farm. Some situations are more complicated, where farmers are insured but where insurance alone is not sufficient.

In addition to farmers in dire economic situations, the argument of the cost of insurance – in the sense that it is too expensive – is questionable. For well-off farmers, given the selling price of certain products, the cost of insurance actually seems quite derisory. In the case of great wines, the insurer will of course not insure at the price of the bottle, but that does not seem to be the real debate. When you see the cheques from some insurers for large claims, you do wonder how uninsured farmers manage to survive. The cost of insurance must be assessed in the context of a much broader reflection of the farm. The farmer must ask himself what he expects from the insurance, must take into account his expenses and loans, and consider where he is in his professional career (it is not the same for a newly installed farmer with a lot of loans and a farmer who has already been established for a long time). There are many different types of insurance. On this cost argument, however, we must be careful not to shift too much blame onto a climatic risk that would be too expensive and would necessarily involve state intervention. Insurers may also have mispriced their risk or not proposed sufficiently relevant offers for the sector.

The multi-risk weather policy is subsidised at 65% (at least for the basic level of cover). With a tariff of this order, a farm can expect to receive money. When the farmer pays €100 for crop insurance, the insurer receives the €100, which it will then use to pay the insurance broker (say €10), the management fee (say €20) and to pay claims (the remaining €70). The farmer pays €100, receives €65 in state subsidies (65% subsidised contract) and can expect to receive €70 in compensation in the event of a claim. The farmer can therefore expect to make money in the event of a claim; he will have paid €35 for the insurance policy (100-65) and will have received the insurance compensation following the claim. For small farms, or even conventional-sized farms, the insurance premium contributions may not be negligible in relation to the farm’s turnover. However, it can be assumed that by pooling risks – with more insurance policies being taken out – the costs of insurance policies would decrease as a result of very low deductibles.

Achieving a high coverage rate raises the recurring question of the insurance obligation. Simple in principle, compulsory insurance is debated in many countries. In France, this solution was not adopted because it would have been politically burdensome. This proposal can only become secondary if insurance becomes very widespread and we can no longer expect much from the mutualisation of insurance risks. Let us also add that today, young farmers are not granted loans if they have not taken out crop insurance.

How are risk and insurance changing with climate deregulation?

Existing reports on insurance results and insurers’ expenses related to climate deregulation are quite clear, the budget and costs are exploding. Risks that used to be exceptional are now becoming recurrent or even completely normal. Insurance works on the law of large numbers, where hazards are distributed equally and independently and the number of insureds is large. If these new costs become the norm, there are basically two solutions. Either insurers continue to pay as they did before and as insurance premiums no longer cover claims over a long period, the insurer will go under. Or the insured continue to pay and for the system to survive, insurance premiums will have to be increased significantly. So how can we adapt to climate change and how can we use insurance as an incentive? With insurance becoming more and more expensive, the temptation will be to subsidise it more and more (as is the case, for example, with the subsidy of the price at the petrol pump). The trend will be inflationary and may mask the muted evolution of increasingly present climatic risks which in fact call for an adaptation of production systems. We cannot simply compensate. Compensating is fine for a while, but it will be increasingly expensive and it is clearly not always the best solution.

We have already discussed this, but insurers cannot insure something that is becoming commonplace. A fire that occurs every 20 to 30 years is insurable. When the frequency is every two years, insurability is no longer really possible, either for moral hazard or practical reasons because the risk has disappeared. While certain situations were still insurable at a good price, climate deregulation will push insurers to consider that such and such a risk is realised and that it is therefore no longer really a risk (this also poses a problem for a certain number of farmers for whom farming conditions are changing and who can no longer produce the same crops, for example). There can be insurance on a climatic trend, but the insurance must be on something that is not foreseen. If, for example, it is predicted that we will reach +2°C of warming, we could imagine insuring the fact that we will reach 3°C instead of 2°C, i.e. we would insure the fact that the warming is faster than expected (you will admit that this is not very satisfactory for everyone because between +2°C and +3°C, it will sting a bit). No insurance company would pay for an increase in a global benchmark if that benchmark was expected to increase. In other words, there has to be a deviation from the expected. Once the external parameters become too strong and the system drifts, it becomes necessary to make the system independent of the external parameters in order for the insurance to continue to work. For example, it is necessary to stop building in flood-prone areas, to take precautions with regard to agricultural practices, or to stop producing certain crops in risk areas.

Climate change will simultaneously increase farmers’ need for protection and the cost of insurance. Insurers, too, will have to adapt. Firstly, just in the way they take into account historical risks (still too much modelled by historical time series when the climate is no longer stationary), risk assessment and pricing models will have to evolve. Farmers’ demand for new insurance solutions will require a broadening of the range of cover in terms of production and the nature of the risks (for example, we have mentioned turnover insurance and margin insurance). The challenge for insurers is to respond to these needs and expand their business into new territories, while keeping premium increases under control. While parametric insurance is one of the avenues we have discussed extensively for reducing insurer management costs, there are many others (and it should be remembered that the management costs of an insurance contract are not that high in relation to the full cost of the insurance premium). One can think in particular of the improvement of insurance distribution channels, which will be increasingly coupled with the provision of credit and inputs, probably within the framework of contract farming, and which will mobilise the various links in the agri-food chains. Insurers will be able to offer new services to insured producers, based on meteorological information, advice on agricultural practices, or even the analysis of climate models to help the insured to understand their future risk and advise them on how to protect themselves or adapt to it. The insurer will also be able to strongly encourage risk prevention behaviour – through awareness-raising and training – in the sense that certain risks will soon no longer be insurable. For an insurer, supporting its policyholders through training seems to make sense. For example, insurers train drivers to reduce the risk of accidents. Whether in agriculture, housing or motor, it is in the insurer’s interest to reduce the (biggest) risks to ensure mutualisation. Investing in prevention is a way for insurers to invest in risk management.

For an insurer, it is the volatility of risks around the average that is important, not the expectation (unless there is an upward trend in risks and the series of claims is not stationary, as is the case with the current climate deregulation). In the face of climate risk, we could see insurers disengage (by means of packages, by management savings with index insurance, etc.). This is what is happening with life insurance contracts, for example. In savings products, life insurance contracts have reduced the financial risks for insurers because the contracts have moved from majority contracts with guaranteed minimum rates (the policyholder took out a life insurance contract and the insurer offered him an annual rate of 2% for 8 years) to contracts where the policyholder is the one who bears the risk (the policyholder defines the type of investment desired and the insurer gives him the rate of return obtained by the fund). Majority contracts with guaranteed minimum rates worked well for insurers when interest rates on bonds bought by insurers (bought with insurance premium money) were very high. When interest rates fell, the guaranteed minimum rates became proportionately more important to insurers than what could be achieved. From the 2000s onwards, as soon as bond interest rates fell below 4%, insurers became aware of the risk because they had to guarantee policyholders annual rates of 2% on their life insurance policies (before the 2000s, interest rates on these bonds could be over 10%). When bond interest rates fall, it should be understood that insurers invest in bonds with lower and lower interest rates. When the rates go up, the insurer finds himself with capital losses on the bonds he has because he bought them with lower interest rates than the current ones. This is a problem for the insurer because if the policyholders decide to break their insurance contract, the insurer will have to sell its bonds to reimburse the policyholders, but it will have to sell them at a loss because of the low interest rates on its bonds. The insurer, faced with rising risks, reduces its share of risk by transferring part of it to the policyholders, and hides behind management activities.

Reinsurers are wary of the idea that the State could take market share from them, which means that the State is compensating for risks that are insurable. For reinsurers, the State should remain in the trend part of climate deregulation, by absorbing losses that are certain to occur due to climate change. Reinsurers, on the other hand, must position themselves on the random part. The risk must remain insurable and therefore random, i.e. care must be taken to ensure that the farmer takes responsibility and that the insurance does not create the risk. In Morocco, for example, insurance has developed and some farmers have started to plant cereals on hillside areas. Although the example presented here is a caricature, similar cases exist in France with grain maize sown in flood-prone areas. In order to earn a return on their investment, reinsurers need appropriate pricing, risk selection (and thus to make farmers responsible), and long-term visibility.

Can we insure anything other than a climate risk?

Climate change is happening; it is a fact of life, and the human race is responsible for it. Climate risks are increasing. Isn’t it time to look at insurance in a different light? In other words, instead of insuring a climate risk which, as we have widely discussed, will soon be uninsurable, would it not be better to insure something else? Wouldn’t it be better to insure something else? For example, to insure a farmer’s transition to more resilient practices in the face of climate deregulation? On this open question, the interviewees are clearly not all on the same wavelength.

One answer is that we will always need to eat. As a result of climate change, there will be (and already are) many areas where it may no longer be appropriate to farm, for economic or other reasons. Climate change may also open up new areas that are currently not suitable for farming (the IPCC is very clear that adaptation to climate deregulation will also take place by taking advantage of the benefits of climate change). In some other places, the insurance premium will certainly increase, but it is likely that climate risks will still be insurable.

For some insurers, pricing a transition risk is still a big deal. To assess a risk, you need to find a baseline yield, have access to a history (the crop will be differently sensitive to hazards so you need long term statistics to know how often the insurance would have intervened), and understand the impact that these practices may have on the yield or on the resilience of the farmer. If there is no history, the insurer has the right to say that the property is not being managed in a ‘prudent manner’ and will therefore be entitled to refuse to insure. There are cases where insurers insure voluntarily, others where they do so by default and still others where they will not take the risk of doing so. If production is more resilient to climate deregulation, this should be reflected in the premium, but it doesn’t seem that pricing systems are that advanced.

Current insurance does not make investments (does not take investment risks) or risks of non-performance of an activity. Insurance insures a loss. And damage linked to external causes, whether climatic or not. In the case of an agricultural transition, if the risk of non-performance of the transition is directly linked to the farmer, the insurers consider that it is not insurable in the sense that there would be too many non-external hazards (technical skills, will or health of the farmer). For insurance to work, the risk must have a random dimension. The risk should not be systematic or too dependent on the farmer’s behaviour. For these internal causes, the farmer should turn to investment rather than insurance. More specifically on changes in practices, some interviewees consider that the farmer will find a way to remunerate himself, for example via the CAP (with future eco-schemes) or with voluntary systems (the low carbon label – I invite you to review another of my blog articles on carbon storage). Existing schemes should therefore allow for remuneration of farmers who have changed their practices. This remuneration is not seen here as a way to bear the risk of yield loss associated with a transition, but rather to support the investment in equipment and practices. For a transition to organic farming, for example, insurers will tend to consider that the risk of seeing production decline is almost certain in the first few years. Hence the existence of aid for the conversion and maintenance of organic farming (maintenance aid was abolished once it was considered that the market was sufficiently mature). And it can be added that the farmer who implements another practice may also be subject to an additional risk. By applying less treatment, an organic farmer may be more sensitive to certain external aggressions during a drought or a frost. In general, insurers seem to have difficulties in establishing what such transitional insurance would actually cover.

On the contrary, some insurers consider that it is possible to establish risk profiles before, during and after transition phases (organic, conservation agriculture, dose reduction, etc.). The main problem is not the modelling of risk (and therefore insurability) but rather the fact that enough farmers will have to be involved for the sector to mobilise (suppliers, distributors, etc.).

In the long term, it is possible that insurance will no longer bring in much revenue for insurers and that it will be necessary to change the concepts. It would be necessary to provide security or adaptation to climate change rather than to subsidise insurance. By insuring a risk of transition to a new crop or a new agricultural practice (conservation agriculture, transition to organic farming, etc.), the insurer loses out at the moment “t” if he only thinks about the crop he has just insured because the farmer necessarily lacks skills and does not always have the right references. But is it really necessary to think in the short term? Today, property insurance is annual in most cases. As an agricultural transition is considered over several years, the situation is changing. By creating a partnership with its insured over a long period of time, the insurer can assume that the sector will change its practices, return to a very acceptable level of yield, and be much more resilient in the face of climatic deregulation thanks to its changing practices. The principle is not to see the crop alone but to see it at the heart of a much more systemic system. In general, the question is whether the evolution of practices can counterbalance the effects of climate change and make an insurance activity in the agricultural sector viable.

To engage in a transition, it is also fundamental to ensure risk parity between new practices – agroecological or otherwise – and conventional practices. Because they lack yield and historical references, the so-called “non-conventional” practices are not yet sufficiently interesting from an insurance point of view for farmers, either because the level of insured yield is too low (the insurance does not really cover much or you have to pay a lot to be covered), or because there are too few offers. Traditional insurance therefore penalises these practices in some way, directly or indirectly. Perhaps part of the solidarity fund or the future scheme should be dedicated to supporting transitions and new practices. This scheme could also be given a mission to support risk-taking in agro-ecological transition investments: in market start-up situations where insurers have no reference allowing them to price products, for example when investing in new crops on land where they have not been produced for too many years. The Caisse Centrale de Réassurance (CCR) could guarantee reinsurance above a threshold of deterioration in the loss ratio that would allow insurers to offer products that make it possible to take risks.

Where insurers could play an interesting role is in giving a more interesting premium to farmers who have virtuous practices, by setting up an incentive system, for example for farmers who follow precise specifications (HVE, organic, uncoated seeds, setting up infrastructures or prevention means, etc.) – although these specifications must be sufficiently explicit and available. Could we set up glyphosate insurance for farmers who do without it? Some investors, for example, ask insurers to ensure that insured plots of land are not deforested. So the whole logic of insurance should be changed to make it a real incentive tool – by covering innovative approaches better than traditional ones. This is obviously not the logic of current insurers, who prefer to secure their own interests first. If, in the future missions of insurance, it was both the promotion of the system and the consideration of new practices that were put forward, the implementation of new practices could be secured. Otherwise, it is quite possible that everyone will pass the quid, with insurers blaming the public authorities for not taking their responsibility and vice versa. This may also be the type of transition that needs to be looked at. For a farmer who is going to make the transition to a more resilient agriculture in the sense of producing less greenhouse gases, the community may for example find it in its interest, but this does not necessarily mean that this transition will change the farmer’s climate risk.

There is even talk of insuring carbon in the sense that changes in practices lead to the generation of carbon credits (I refer you here to another carbon post on my blog). Let’s imagine a forest that I harvest every year. If I stop cutting it down, I potentially generate carbon credits; the bet being that I will generate income from the carbon credits of growing trees instead of exploiting the timber industry. The important question is what is actually insured if my forest burns down. Is it the cost of restoration to regenerate the biotope? Or is it the cost of the carbon credit?

Innovation and risk culture

A lack of risk culture in the agricultural sector

The world has changed. Those who thought they were spared because of their geographical location are now being hit. The resilience of farms requires control of the risk to farmers’ income, not only in terms of volatility but also of loss thresholds relative to consumption levels. Risk management gives farmers the confidence they need (less mental burden and stress) to develop their capacity to invest in traditional assets (equipment, buildings, livestock) or in assets that create value for the consumer (organic conversion, short circuits, etc.) because bank loans are guaranteed, in addition to the traditional and often partial guarantees, by the multi-year stability of the farm’s results. Climate insurance thus provides a guarantee of annual financial stability for the farm in that it allows the farm’s results to be smoothed over several years. It avoids blocking the farmer’s capital – which he would have kept for self-insurance – and allows him to use this capital to invest (in resilient practices for example). Insurance can also be seen as not just an approach to analysing overall risks but also as part of a wider proposition, seeing insurance as part of a wider package, for example in relation to a financing offer. For example, we could take the case of a grape harvesting machine manufacturer who would like to integrate into the price of his machine a piece of index contract so that, in the event of a bad harvest, the farmer receives a small capital sum because the machine is less depreciated. There would be a link here between the use of the asset and the climatic event. As we have discussed, insurance is only one of many risk management strategies. Between technical, agronomic, financial and strategic levers, the farmer has many ways to increase his resilience to climate deregulation.

The concept of risk management is still not sufficiently considered in agricultural training. In view of the current climate change, this may come as a surprise. However, it is essential that the content of initial and continuing training courses evolve in the direction of transmitting knowledge and raising awareness of the challenges of risks in agriculture. Insurance is still too much considered in a transactional approach where the insured want their insurance to be profitable, whereas this is not what insurance was set up for. Farmers should see insurance as a financial tool rather than as a tool to protect their yield and income for the year. Advisory networks also need to be developed to support farm business and risk management strategies for farm managers. For example, most multi-risk insurance contracts are currently at the crop level, whereas contracts at the farm level exist. In the context of the search for resilience, doesn’t it make more sense to consider all crops at the same time and to insure only hard times? Farmers must have the means to project themselves into 20-year agro-pedo-climatic scenarios in their locality, in an educational and practical way, in order to assess the risks to which they are (or will be) subjected and to imagine ways of adapting to them. Farmers must learn to map their risks, analyse them and control them, or even transfer part of them (to insurance companies, for example). This is where foresight exercises come into their own.

It is a real risk culture that farmers must develop. To be slightly provocative, one might also ask whether subsidising insurance does not lead farmers to lock themselves into their subsidy and not take risks. Isn’t it the nature of a farmer to take risks, even if we know that the beginning of a transition phase will be difficult anyway? Some farmers are of course in catastrophic economic situations, and it is necessary that there are measures to support them, but we must also bear in mind that, when it comes to adapting to risks, it is also the responsibility of each farmer to act.

A shared role for the players in the agricultural sector in insurance

Insurance is not just the responsibility of the farmer. The whole ecosystem around him can have a role to play. A global reflection on insurance between all stakeholders can be beneficial for all actors:

  • For the insurer, it is an opportunity to leverage existing non-insurance services to increase awareness and penetration of contracts, to compensate for the lack of staff and own distribution channels in rural markets, to use the positive image of a partner and get customers to try the insurance offer, to reduce distribution costs, or to reduce anti-selection phenomena.
  • For cooperatives, equipment manufacturers and other distributors, it is an opportunity to use insurance to promote sales of agricultural inputs and to differentiate themselves from their competitors, to increase product loyalty, or to generate an additional revenue stream in terms of incentives provided by the insurer. The cooperative may not see the insurance business as a profit-making activity and use it as a means to maintain or increase its market share, an additional reason to offer more attractive insurance premiums to farmers. Seed companies and cooperatives can sell seed with insurance against a risk of drought that would cause the seed not to germinate. The service would be similar to a “money back guarantee” with, if the seed does not grow, a replacement of the seed for the farmer. One can also imagine offers in the event of a supply shortage to guarantee a user (a cooperative or a more specific sector) a level of volume and/or quality for its sourcing.
  • For the farmer, it is an opportunity to reduce the transaction costs of insurance by negotiating with only one entity rather than two if he goes through his cooperative, for example, to have easier access to credit and better agricultural inputs, to have a reduction in the repayment of the loan and access to another loan for the following season, or even an easier payment of the premium if the grouped service provider pre-finances or subsidises the premium.

I would like to take this opportunity to remind you once again that the players in the agricultural sector have other means than insurance to help farmers manage their risks.

Un manque de compétences et d’innovation des assureurs

If farm managers lack a real risk culture, insurers could be criticised for being relatively uninnovative in terms of proposing insurance offers. It is understandable that, with loss ratios below the break-even point, the market leaders are taking a defensive stance and not developing their offers too much. However, standard offers could evolve towards more refined offers, just like the customer paths. A maize farmer in the Landes region with a hundred hectares or so of land does not in fact grow just one type of maize, but perhaps seed maize or popcorn, all different crops that call for different risk profiles. Similarly, as we have discussed, traditional MRC insurance focuses on yield insurance. Other offers, already used in some countries – margin insurance or turnover insurance – could be tested.

There is also an important step to be taken by agricultural insurers to digitise their processes. Some insurers are not yet ready, from an information system point of view, to manage geometric entities such as parcels, which will in any case be an obstacle to the deployment of index insurance services. With information systems based on postal codes, it is impossible to manage the spatial characteristics of parcels or even farms in detail. The deployment of index-based offers, through the construction of multi-source indices (vegetation, weather, etc.) generally requires mathematical modelling, data manipulation and expertise in crop simulation, all of which are different skills that insurers have not necessarily internalised. These skills have sometimes been taken over by re-insurers who work on portfolios of high-risk and highly diversified clients, and then bring the offers back down to the networks of traditional insurance operators.

The provision of data to experts in the field is also made difficult by the fact that several insurers are not yet at the stage of equipping their experts with digital tools. An insurer can certainly subcontract the expertise to specialised companies, but it will still be necessary to have interconnected information systems.

In conclusion

Man-made climate change is causing insurers and reinsurers to shudder. Climatic events are becoming more frequent and uncertain, and the cost of claims is increasing. In the agricultural sector, the consequences of this deregulation are already being felt and could prove dramatic in the years to come, not only for the farmers themselves, but also for the entire population, which will always need to eat. Add to this geopolitical tensions that are increasing the pressure on food – some actors are predicting future food shortages in North Africa as a result of the Russian-Ukrainian conflict, and food security will no longer be guaranteed.

Nevertheless, there seems to be a collective awareness that agriculture needs to make the transition and that the issue of resilience is fundamental. Nevertheless, in a period of transition, a farmer is fragile. In order to be able to commit to transition paths and have the will to invest, they must have confidence and feel protected. In France, the reform in place intends to tackle the limits of the current agricultural insurance and has the will to increase the number of insured farmers, all sectors included. There are a huge number of insurance formats – between the diversity of risks covered, the choice of deductibles and guarantees, or the indemnity or index insurance formats, there seems to be no shortage of choices.

Insurance was historically created to enable trade in the Mediterranean because there was a guarantee that in the event of the ship sinking, the trader would not have to bear the loss of the goods alone. Transition requires risk-taking and insurance is one of the levers to protect against it. But perhaps it is time to think about insurance differently. Does it still make sense to insure climate risk directly? Wouldn’t it be better, on the contrary, to ensure a more resilient agricultural transition in the face of climatic events in order to avoid putting off the problem for the umpteenth time?

Soutenez Agriculture et numérique – Blog Aspexit sur Tipeee

Bibliography complementary to the interviews

Aon (2021). Weather, Climate and Catastrophe Insight :

Benami, E., Carter, M.R. (2020). Can digital technologies reshape rural microfinance ? Implications for savings, credit, & insurance. Agricultural & Applied Economics Association

Bouton (2017). Solutions indicielles de protection du revenu agricole – Modélisation et Tarification. Mémoire d’actuariat, ISFA.

CCR & Météo France (2018). Conséquences du changement climatique du le cout des catastrophes naturelles en France à horizon 2050.

Clement, K.Y., Botzen, W.J.W, Brouwer, R., Aerts, J.C.J.H (2018). A global review of the impact of basis risk on the functioning of and demand for index insurance

Confédération Internationale du Crédit Agricole (2020). Une contribution aux objectifs de développement durable. Pour une agriculture plus résilience et mieux protégée face aux aléas climatiques. Livre blanc sur l’assurance agricole.

Descrozailles, F. (2021). Rapport sur la gestion des risque en agriculture.

ENESA (2018). 40 ans de Système d’Assurances Agricoles. Entité Gouvernementale d’Assurances Agricoles. O. A. (ENESA)

France Assureurs (2016). Quel impact sur l’assurance contre les aléas naturels à l’horizon 2040 ? Synthèse de l’étude changement climatique et assurance.

France Assureurs (2021). Quel impact sur l’assurance contre les aléas naturels à l’horizon 2050 ? Synthèse de l’étude changement climatique et assurance.

GIEC (2021). Climate Change 2021. The Physical Science Basis. Summary for Policymakers. Working Group I Contribution to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change.

GIEC (2022). Climate Change 2022. Impact, Adaptation et Vulnerability. Summary for Policymakers. Working Group II Contribution to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change.

Haut Conseil pour le Climat (2021). Renforcer l’atténuation, Engager l’adaptation. La version grand public. Un résumé du troisième rapport annuel du Haut conseil pour le climat.

Lepoivre, B. (2020). L’assurance des prairies. Fourrages, 244, 87-92

Ministère de l’agriculture et de l’alimentation (2021). L’assurance multirisque des récoltes

Mortensen, E. & Block, P. (2018). ENSO Index-Based Insurance for Agricultural Protection in Southern Peru. Geosciences, vol 8 – 64.

Johnson, L. (2021). Paying ex gratia : Parametric insurance after calculative devices fail. Geoforum, 125

Schuster, C.E. (2020). Weedy Finance : Weather insurance and Parametric Life on Unstable Grounds. Cultural Anthropology, 36:4.

Shift Project (2021). Rapport sur la stratégie de résilience des territoires. Volume 1, Comprendre. Xiong, H., Dalhaus, T., Wang, P., & Huang, J. (2020). Blockchain Technology for Agriculture : Applications and Rationale. Frontiers in Blockchain, vol 3.


Luc BOUCHERDiagoRisk
Nicolas CHATELAINPartner Re et APREF
Timothée CRAIGExo Expert
Frédéric DESCROZAILLESAssemblée Nationale
Jean-Marie DETERREAxa
Bernard FINASJola
Olivier LELIEVREExpert Indépendant
Jérôme LEROYWeenat
Bruno LEPOIVRECrédit Agricole
Charles MAURYClimate Insurance
Jean-Baptiste ORNONAxa Climate
Antoine POUPARTInvivo
Aymeric ROUCHERDescartes Underwriting
Antoine ROUMIGUIEAirbus
Cécile TARTARINGeosys
Pascal VINEGroupama
Ewan WHEELERAcre Africa

Share this article!

Leave a Reply

Your email address will not be published. Required fields are marked *