Any decision-making process on crop insurance involves many stages. These stages and certainly the priorities will differ, depending on which type of body is doing the investigation. This may be a government ministry, a farmers organization, an insurer, a bank or a group of marketing/processing agencies. In any case, some of the more important issues and steps are:
Demand assessment - ensuring that any initiatives are in response to real risk management needs;
identification of the key insured parties; automatic or voluntary cover?
determination of key perils - a key factor in insurance design;
decision on crops to be covered - another key factor in insurance design;
analysis of insurance options, administrative models and loss assessment procedures, together with determination of associated costs;
rating - determining the pure premium required, plus administrative and loss adjustment overheads to derive the initial premium level to be charged;
identifying possible complementary roles for the government and for the private sector.
In any given situation the results of investigating these issues will determine whether or not crop insurance is the most efficient and effective mechanism to manage a particular area of risk. The results will also indicate the type of insurance product which is optimum for a given situation.
The sections below set out some of the arguments, and illustrate, with examples, how some insurance solutions have been developed.
This must come first, but is always difficult, as before a detailed investigation of the incidence and effect on crops of perils, and an assessment of operating costs, it is impossible to give more than a very vague estimate of the likely cost of the insurance. Unless farmers know the details of the product and its price, they are not likely to indicate whether or not they will buy.
Closely linked to this is the need for any crop insurance programme to respond to real needs. As stated in the introduction, crop insurance is a business, and both parties must want to participate. Real needs must be met for this condition to be satisfied.
These needs are changing, as new opportunities arise. An example comes from the increasingly accepted (and, often, trade-mandated) policy to reduce of the use of certain crop pesticides and soil sterilants. In order to overcome growers fears that alternative and safer techniques and products could be used, insurance has a potentially important, if temporary, role to play.
In the face of these needs, the services of an experienced crop insurance team are required when crop insurance is under consideration. Such a specialist team would be able:
to examine the risk structure of certain key crop sectors;
to identify the extent to which the involved parties are vulnerable to these risks;
to draft an outline of an insurance programme, with indicative costs and benefits, and responsibilities; it would also include details of further investigative, publicity and lobbying work required before insurance business could commence.
This team would consult closely with several sectors in the economy, and follow up in detail the issues which are described below.
Farmers are one obvious party to crop insurance. Those who depend on a supply of farm produce for their business are another. The latter group includes processors and crop product buyers. These firms often stand to lose financially if crop product is not available from their local farmer suppliers. In this case they may face increased product acquisition costs; they therefore have an insurable interest in the growing crops. In Canada a baked bean cannery buys insurance where the trigger for an indemnity payment is a shortfall in supply of beans from local growers. The indemnity payment assists the factory in sourcing raw beans from elsewhere - usually at higher cost.
One of the factors identified earlier as leading to an increased demand for crop insurance is the growth of contract farming arrangements. When insurance can economically address some of the production risk involved, risk which affects both growers and contractors, then there may be a case for making crop insurance automatic. This is the same as making it compulsory, but automatic is a better description of the process when insurance becomes just one of a range of services being provided, as a package, to contracted growers.
The Mauritius Sugar Insurance Fund (MSIF) has operated for some 50 years. It provides automatic cover against the main peril to the industry, namely windstorm (cyclone), for all sugar growers on the island. The compulsory nature of the cover has always attracted some criticism from a few growers, but the parastatal MSIF has countered this by appointing strong representation of growers to the Board which overseas the operations of the Fund. Over the years the growers representatives have been responsible for several improvements to the product bought by farmers. This is now generally regarded as being both fair and highly useful. A similar, automatic programme operates for the large numbers of small-scale banana growers in the Windward Islands, in the Caribbean.
In New Zealand crop insurance against frost and hail was automatic for all growers of kiwifruit for several years, from 1988. After a few years with just these two perils included in the policy, some losses were experienced in the industry from geothermal activity. On grower demand an additional peril - damage from volcanic ash - has been added to those included in the policy. The cost of administering kiwifruit insurance is extremely low, as the insurer pays only a very minor amount for the acquisition of the business i.e. advertising expenses and brokers commissions. On the other hand, the nature of the perils being covered mean that individual loss assessment has become the norm, and indeed is necessary. This is costly for the insurer, leading to a higher premium than would be the case if an index type of policy could be designed.
A listing of key perils and risks for agriculture across the world would be long. For the present purposes it is useful to focus on those which are of major concern to developing countries. Further, they can be clustered into a number of groups. One such clustering would produce a list as follows:
natural resource risks;
marketing and price risks.
Production and natural resource risks are relevant to this discussion of crop insurance, and are discussed in greater detail below. Financial and marketing/price risks fall outside the scope of the present publication, except in the case of crop/revenue insurance products, as discussed above.
This is the main category of insurable risks. Both quantity and quality losses can result. Perils included are:
Adverse climate conditions: drought, excessive rain, flood, windstorm, frost, hail, sunburn, snow;
pest and disease attack;
These warrant separate discussion, under the headings below.
Drought is both a major concern of many developing countries, and the natural weather event which causes most problems for insurers. The reasons for this are many. Firstly, insurers feel most confidence when an adverse event has a clearly defined time of impact, coupled with a clearly defined geographical area. The classic example is hail, which may do its damage in a matter of a few minutes, or even seconds, and will typically impact an area confined to a few hundred square metres up to a few square kilometres. Hail damage is clearly attributable to the adverse weather event, and is readily verified as such provided that a field inspection is undertaken.
By contrast drought has a vague beginning, its effects linger for a very long time, and can extend over more than one growing season. Moreover it typically impacts a very wide land area. Production loss caused by drought can be aggravated by the incidence of other problems, e.g. diseases attacking plants weakened by water stress.
From a purely underwriting point of view drought poses great difficulties for a standard crop insurer offering what is in effect a yield guarantee. Firstly, because drought affects a large number of growers in the same season - perhaps the whole of a country - the production losses are very large. This systemic or catastrophe exposure means there are problems in mobilising sufficient insurance capacity to cover the sum at risk, even with recourse to substantial reinsurance. Secondly, droughts in recent years, at least in many parts of Africa, have tended to extend over more than one year. This experience means that it is extremely hard for insurance companies to obtain reinsurance for crop insurance portfolios which carry drought risk. Thirdly, the magnitude of the risk in most developing countries means that actuarially calculated premiums would be very high - too high perhaps to attract all but the most at-risk growers. No insurer wants to build a portfolio based entirely on such a clientele.
For these reasons insurers are very wary of covering drought as an inclusion in standard crop insurance policies. This is particularly the case in those parts of the developing world where drought is the major weather constraint to crop production: Southern and Eastern Africa, Sahelian Africa, Horn of Africa, North Africa/Near East, Eastern Europe, Central and East Asia, South Asia, Central and South America. The list illustrates the key role which drought plays in the lives of much of the developing worlds rural population.
Given the almost insurmountable problems involved in including drought in standard crop insurance policies for developing areas, attention in recent years has turned to examining whether index (coupon) policies could provide a useful measure of security. Initial developmental work in this field is promising.
As described in Section 2.5.2 above, index insurance involves using a meteorological measurement as the trigger for indemnity payments. In the case of an index policy covering drought, the most likely form would be a series of indemnity steps, each step corresponding to a given level of rainfall deficit. The assumption is that growers could select a level of indemnity suited to individual circumstances. Thus the indemnity payable would increase as the rainfall shortfall increased from a defined drought trigger amount. At the time of writing, index policies covering drought or other climate risks cannot be described as being a standard product for developing countries. Rather they are in the nature of a promising new insurance technique, attracting much interest among risk management professionals.
Crops need water, and much of the developing worlds arable and horticultural production relies on rainfall. Too much rain at any time can damage a crop, but there are periods of special vulnerability, described below.
The first danger point is excessive rain just after germination and emergence. Entire crops can be washed out of the ground, necessitating resowing. This is an insurable risk, where the indemnity which would be written into the policy would be the costs of re-sowing, plus a possible additional amount in those cropping situations (common in tropical, rainfed agriculture) where a delay in sowing means that the eventual harvested crop is smaller than would have been the case had the crop been able to take advantage of the whole of the normal growing season.
The next common point of vulnerability is at or near to harvest. Maize and other grains can sprout prematurely while still growing in the field. Various fruits (e.g. cherries) can be damaged by excessive rain or even any rain just prior to harvest. Other crops can be lost when excessive rain prevents harvest. An example is a crop such as tomatoes grown for processing. The processing factory schedule of crops for harvesting means that the date of harvest is fixed.
Moreover, it is now common practice with commercial tomato crops to spray with ethrel (ethephon) in order to accelerate the ripening (reddening) of fruit which are still green, in order to allow once-over harvesting. If excessive rainfall is experienced just when the critical readiness for harvest is achieved, then harvest may be prevented, and the crop lost.
Flood damage may be due to on-site excessive rainfall, but it can also be caused by excessive precipitation elsewhere, and the subsequent rise of river and lake levels, to cause flooding of crop land. The risk is usually insurable. Exceptions would be crop land which is insufficiently drained or where existing drains are not maintained, and also flood plains exposed to a very high risk of flooding.
Flood is sometimes one of the results of severe storms. Examples are the frequent tropical cyclones experienced in the Bay of Bengal. These usually cause flooding of low-lying farmland along the affected coastal zone. Records indicate that although the fundamental peril is windstorm, the actual losses on farms - to livestock as well as to crops, have been due to flood damage resulting in turn from wind-induced high sea levels, which are known as storm surges.
Crop insurance programmes in the Windward Islands (bananas) and in Mauritius (sugar cane) have already been mentioned. Both were set up to assist in managing the losses from excessive wind - cyclones in Mauritius and hurricanes in the Caribbean. High wind speeds affects nearly all crops - and can cause serious damage in forests.
As with other weather perils, the first move in risk management lies in appropriate farm management - correct attention to plant density (for mutual support), to the provision of shelter belts for those crops highly sensitive to wind (e.g. kiwifruit), and care with harvesting in the case of forests. It is not uncommon for problems to arise when partial harvesting takes place in forests. Those trees that are too immature to harvest are suddenly exposed, and may be blown over by high winds.
In writing wind-storm insurance, insurers take these sorts of management practices into account. They make certain that it is only exceptional events that will trigger the insurance, when normal practices are insufficient to prevent damage.
Windstorm is associated with catastrophic losses to life and property, as well as to crops. Hurricane Andrew, one of the most destructive storms ever recorded, hit Florida and Louisiana on 25 August, 1992. Storms of this magnitude, and lesser but still serious weather events of this nature are believed to be increasing in frequency.
Although not at all common in developing countries generally, there are some regions where this is an occasional risk, especially to vegetable and fruit crops. This applies especially to Eastern Europe and the Middle East.
Frost causes damage by the freezing of the water content of plant cells, and their subsequent rupture. It will be evident that it is not only the temperature which matters; it is also the time when the temperature is below a certain minimum level which causes a damaging event. Crop insurers write policies accordingly, sometimes constructing a damage point (i.e. insurance trigger) curve which plots temperature against time.
Frost conditions can impact a wide area, causing extensive damage. However, the micro-climate in a given site can increase the likelihood of frost damage. For example, fruit and vegetable production often takes place in valleys because of the presence of deep topsoil, washed down from surrounding hills, together with the availability of water from surface or groundwater sources. These same valleys can also be frost-pockets because freezing, still air accumulates readily in this type of topography.
Again, an insurer may expect growers to take normal precautions against frost damage, through the use of devices to move the air (burning frost pots as commonly used in Eastern Europe, propellers mounted on towers, as being introduced in some of the fertile fruit growing valleys in Syria). Perhaps the most effective preventative of frost damage for horticultural crops especially is the use of sprinkler irrigation.
It will be clear to the reader that all of these measures involve a cost. Design of an insurance policy to respond to frost damage will take into account the inevitable trade-off between the costs of physical and financial measures of managing the risk. Usually the most cost-effective approach is a blend of the two, with insurance acting as a final safety net, to be triggered if the physical devices fail to prevent damage.
Hail holds a special place in the history and also the current practice of crop insurance. It was the first crop peril to be insured by a modern insurance company - the first policies being issued, in Germany, in 1791. It is also the simplest of weather perils to handle from an insurance point of view. Its incidence is readily confirmed by observation of damage, and compensatory growth factors are reasonably well understood for most major insured crops (see also under Loss Assessment below).
Moreover, over time, the likelihood of hail events in any given agricultural area can be estimated in a manner that permits actuaries to confidently set premium levels at values which both sides, insured and insurer, find reasonable. This is due also to its long history, and the manner by which records of damage have been prepared and retained over the years. This means that there is a wealth of data on the incidence of the peril, and of the crop damage which has been caused as a result.
Again, when hail strikes it is usually very confined in terms of the damage zone. This can be just a few square metres, a few hundred square metres, or, more rarely a few square kilometres. It is seldom larger than this.
There is little that a grower can do against hail damage. Lengthy research has proven that injecting hail clouds with silver iodide via rockets or planes is not very effective. Areas with very high hail exposure and expensive crops can resort to hail nets.
Sunscald, under exceptionally adverse conditions, causes damage to fruits such as pip and stone fruit, grapes and nuts. It is associated with the premature loss of foliage from the plant. The risk is insurable, often as an extra-cost option under multi-risk policies.
Snow can damage all types of crops, including fruit trees and it also a peril of note in forests, where excessive weight loading can cause breakage of parts of trees, or even toppling of the whole tree. Developing countries vulnerable include those in Central Asia, Eastern Europe and the Middle East regions. Snow is an insurable peril in many circumstances. In forests damaged by breakage through snow loading, the presence of broken tree parts can facilitate the build up of pest and disease organisms.
Insurance cannot substitute for sound management of the risk of pests, parasites and diseases. Indeed, this is a significant area of modern farm and forest management, with very substantial losses resulting from failures in this area.
Moreover the growing importance of international trade in agricultural commodities impacts on the pest and disease issue in developing country farming in several ways:
Phytosanitary regulations mean that any evidence of pest or disease in a consignment may disqualify produce from entry to the country of destination;
similarly, pesticide residues are subject to very tight limits under the standards for international trade;
competition in the market is fierce, and even if produce is allowed to enter, blemishes on fruit etc. mean the produce is unlikely to find a buyer.
Insurance implications can similarly be summarised in a brief list:
It is sometimes possible for growers to obtain cover against pests and diseases where there is no generally accepted management control;
in an attempt to reduce the adverse environmental impact of some well-established chemical spray routines for pest and disease control (e.g. certain chlorinated hydrocarbons) alternative, benign regimes have been developed. Insurance may be utilised in the future in order to provide temporary risk assurance to growers using the new routines;
frequently damage to fruit and other crop products provides an entry point for disease organisms. Perforation of the skin due to hail damage is a common example. In this case any hail policy needs to be clear as to whether the consequential loss from disease is also covered.
One of the oldest perils to be covered in property insurance, fire is a major peril for many crops (especially broad field crops such as grains) and for virtually all forests. It is commonly included in multi-peril crop insurance, and is frequently the key peril under forestry covers (which may also include wind and snow damage).
Fires are caused by human action (and carelessness) and also by lightning strikes during electrical storms. Whatever the cause, there are control measures to reduce any losses. These may be through early detection and the subsequent means to take action and/or through the use of cleared firebreaks.
Insurance policies will normally state the expectations under the policy of the means to control fire losses. Again this is an example of insurance being just a part of a cluster of measures used to control risk.
Adverse soil conditions, e.g. salinity, erosion of topsoil and loss of soil nutrients;
deterioration in water quality e.g. due to pollution of the water table or natural water courses;
lack of water from the irrigation source.
In the main these risks are best addressed by farm management practices. However, some of the underlying causes of these problems may themselves be insurable. For example, soil erosion may follow excessive rainfall and/or wind. Pollution of water may be beyond the control of the farmer drawing from wells or rivers.
Related to this is the risk that a water source used for irrigation may fail. Prolonged drought means that water tables fall, necessitating the boring of deeper wells. Similarly rivers and streams can dry up, due again to drought, or to an increase in uptake of water upstream. Where this involves another country then this falls into the political risk zone, something that many insurance policies specifically exclude.
As with most assets or production processes, virtually any crop can be insured, against virtually any peril, but only at a price. At the time of writing, with squeezed profit margins on the production of many crop commodities, a paradoxical situation arises. The tight margins highlight the need for risk management, including insurance, but also reduce the ability of growers to buy the desired level of protection.
The discussion below will focus on four main groups of crops, annual field crops, perennial crops (including horticultural tree crops), glasshouse crops and finally, forests. The focus will be on identifying those areas of risk which the nature of the crop, and of its common perils, could predispose it for insurance as part of a risk management strategy. In this discussion, crop insurance relates to the various types of contract which make up the more traditional type of cover, as opposed to index policies. With the latter, the nature of the crop is not an issue, since the insurance contract relates just to a given weather event.
Insurance of crops and forests involves insurance of an expected future value. This sets this type of insurance apart from other property covers (e.g. motor vehicle, buildings, machinery) when the value (frequently maximum value) exists at the commencement of the insurance.
One of the factors which can determine whether or not a particular crop/peril combination is suitable for insurance is the ease and economy by which losses can be satisfactorily assessed. This will be touched on below, with some of the more general loss assessment issues discussed in greater detail under the section, Loss Assessment.
Wheat, maize, rice, soybeans, sorghums, cotton, beans etc. are all insured in various parts of the world. As annuals, any loss or damage is just to one seasons crop - unlike for perennial crops and forests. This simplifies loss assessment, in contrast with the situation of Perennial Crops, taken up below.
As a general rule, the more commercial the nature of the crop, the greater will be both the potential demand for insurance, and the likelihood of a cost-effective role for crop insurance in risk management. Crops of the high value input - high value output variety are often financed with the assistance of banks, and lenders increasingly insist on insurance coverage, when this is available.
Another important issue in commercial crop production is the marketing chain. With crops such as sugar cane, coffee, tea and cotton, virtually all of the harvested production enters the commercial market, and requires processing. This means that there is control over quantities produced, year after year, together with an opportunity for establishing a strong database of producers and of details of production enterprises. Information management of this sort is vital to creating the climate of confidence necessary for efficient and economical insurance transactions.
It will be evident from the above that food crops, especially those for which there is an active, unrecorded local market, are difficult to trace after harvest. This means that insurance assessments are similarly difficult for this type of crop.
Perennial crops pose a special problem. In the event of a loss event, should the loss be calculated solely on the basis of the current seasons expected production, or should reduced production levels for the next season(s) be included? The difficulty of making accurate assessments for future years will be evident, and crop insurers in Chile and Cyprus, for example, include only the current seasons lost production.
On the other hand, when a peril such as windstorm causes serious damage to tree crops such as oil palm, coconut, rubber, and mango or to temperate fruit crops such as pip and stone fruit, growers naturally expect the longer term loss to be indemnified. Technically, when losses are severe, it is possible to make assessments. These could even include the costs of replanting and/or regrafting. Paradoxically, the problem is greater when the damage to the wooded parts of the plants is less severe, but still sufficiently serious to mean a diminution in the following seasons crop. In such cases the approach taken by Chile and Cyprus appear to be appropriate. An alternative is to formulate wording such that fruit and trees are separate parts of the same policy. This is done in the British Columbia Ministry of Agriculture crop insurance programme.
Crops grown under glass, plastic or other coverings generally fall into the high value input - high value output category. As such, risk management planning is very important, since loss of the crop and/or the structures can mean a heavy financial blow. In fact in those countries where glasshouse and plastic house cultivation is important, insurance is usually an integral part of the production financial plan, and the potential liability for insurers is very substantial. Sometimes insurers offer policies which cover the structure together with the growing crop. Generally these also specify minimum standards for construction and the materials used in the structure.
The economic role of forests is undergoing a partial change. This change affects risk management, and also insurance as part of risk management. The transition of national economies from a commodity to a service orientation, and stream of products, also affects forestry. This is because a forest today is not just a source of timber, for paper, for building and for furniture, but is also a provider of environmental services. Increasingly it is becoming possible for forest owners to generate income from the sale of carbon credits. This opens up forestry to a new, more commercially oriented class of investor, and this change will affect developing and developed countries similarly.
A further change is the move towards the certification of forests as environmentally sound entities, under some sort of recognised certification system. The implications for forest managers are twofold. Firstly, such certification opens up access to markets which will only accept timber from forests certified as being sustainably managed. Secondly, when insurance is involved, such certification, since it is based on the achievement of a high standard of management, including risk management, could lead to substantial reductions in insurance premiums.
The major risks to forests, namely fire and windstorm, will affect virtually all species of timber trees, although some are more at risk than others. For example, in recent years there have been extensive commercial plantings, in many parts of the word, of various types of Eucalyptus species. This tree type is popular because it is very fast growing and has considerable drought resistance. However, it also has a high content of oily, volatile sap, meaning that it burns readily.
When forests are insured against fire risk then considerable attention is given to management procedures to reduce the possibilities of loss in the event of a fire outbreak. Some of these points have been made above, under the heading, Fire.
In summary, the changes to the forestry scene, worldwide, mean greater commercialisation of tree cultivation, and therefore greater opportunities for introducing insurance as a risk management device.
Loss assessment is a key element of standard insurance. With crop and forestry insurance it is essential that loss assessment procedures can be designed for the crop and the perils involved. This is not always the case. A common problem is when a loss occurs which could have been caused by more than one peril. When the policy is not all-risks but rather named-perils then any loss assessment process should be able to ascertain as to whether the loss was caused by an insured peril. Unless this is possible then the crop/peril combination may be impossible to insure.
In any insurance contract it is vital that the process of loss assessment is made clear, so that in the event of a loss, the assessment process can start in a manner which has the prior agreement of both insurer and insured. The first element is to check that the loss falls within the scope of the policy. This is not always a straightforward issue, since some losses have more than one cause, and some of these might be covered by the policy, others not.
The loss must then be measured, and the indemnity to be paid determined. The whole process of assessing the loss, determining the indemnity and paying it is known as loss adjustment.
Unlike other types of property insurance, when a loss can be assessed without the biological factor, crops and trees have the capability of compensating for damage. This is covered in section 3.6.2 below, which is followed by a brief note on the closely associated issue of salvage.
Compensatory growth is a plants response to damage. Some examples will illustrate how this can impact on insurance and on the assessment of losses. Hail can do devastating damage to grapevines. If the hail event is in the spring, fruiting parts can be knocked off. However, the plant will normally grow new fruiting parts from existing buds, and a crop will result. The loss in this case is likely to be a reduction in the quantity and also in the quality of the fruit, but there will be something to harvest.
On the other hand, late summer hail damages the grape bunches themselves, and can cause an almost complete loss of the seasons production. It is too late for compensatory growth, so an insurer, working with the grower, will assess whether or not any salvage can be undertaken. Table grape market values are heavily hit by partial hail damage to the shoulders of the grape bunches. In such cases, even though the bulk of the fruit in the bunch may be undamaged, the prominence of hail damage on the shoulders of the bunch means that the grapes may not find a market. This may lead a loss assessor to declare a constructive total loss (see under Salvage, below).
Cotton and maize are among other economically important field crops which will compensate for damage by a peril like hail. But the new growth will not completely replace the lost parts, so some diminution of yield can be expected, depending largely on the stage of growth at the time the damage occurs.
Compensatory growth is something that a crop loss assessor will take into account, drawing on the considerable research which has been done on the more important field crops, which gives an indication of the extent to which some of the loss is made up by natural processes.
Salvage is the human equivalent of compensatory growth. Sometimes salvage of a damaged crop will involve sale into a different market. For example, apples grown for whole fruit sale, which are damaged by hail, might find a market at a juicing factory. Similarly, a forest damaged by fire or windstorm may yield timber which is marketable, although the extraction cost might be high.
Again, salvage possibilities are taken into account by assessors in calculating the indemnity. In the event of a constructive total loss the insurer pays a full indemnity to the insured grower, and then may try to salvage some value from the damaged crop or forest.
The management of insurance, as a business, has several stages. These are: market identification; product development, marketing, setting indemnity and premium levels, collecting premiums, handling claims. The over-riding aim in the design of administrative structures and procedures is to lay a foundation for minimising costs. Since the potential clientele comprises small and often widely dispersed growers, costs can easily escalate to the point of non-viability of the business, unless special care is taken. In this connection, the new insurance products, mentioned earlier, offer much scope for drastically lowering the costs of administering a financial risk management mechanism.
The various stages of standard insurance administration offer some scope for economies. The tasks involved in these stages are briefly described below, with mention of particular examples where efficient procedures have been developed in order to save costs.
The extent of involvement of the public sector varies from country to country, but it always has a role, even if this is exercised in the main through setting supportive and regulatory policies. It may be particularly important in the early stages of developing crop insurance, and in situations where financial support is considered both desirable and possible.
This is a vital stage. Buying insurance involves increasing the up-front costs for a grower. The advantages of buying cover must be clear, with careful positioning of any proposed insurance product. Firstly, this means recognising that insurance as such may not have a legitimate role in a particular industry for the major perils, as seen by the owners. Secondly, where there is believed to be a role, it means that careful attention must be paid to benefit/cost considerations for both contracting parties - the insured and the insurer. These two conditions can best be met by identifying the real points of financial risk in an enterprise type, and examining whether a financial risk-sharing mechanism can be economically applied. In general, the more commercial the operation, the more likely is it that insurance could be designed to address certain of the risks involved. This applies, in particular, to the intended market for the produce of the grower. A formal, commercial market implies the ability to collect information on quantities of production from particular growers. Time series data of this type, since they are based on transactions involving payment, is likely to be highly accurate. A market outlet may also facilitate administrative economies in arranging the cover, or even in paying premiums.
At this stage too it is important to identify the insurer. Is it to be a local insurance company, perhaps one that has no previous experience with crop insurance? This is the model which is commonly found in Latin America, and which was recently proposed for Syria, by an international team of specialists. Or is it to be handled by a special agency, as is the case in Iran, and in Mauritius. A third possibility is that a special farmers cooperative structure be formed to handle the insurance, as is done in France and in South Africa. It is not possible to give an opinion on which of these alternatives is the best. One can, however, note that if an existing company were to take on crop and forestry cover as an additional line of business, then it will start a number of advantages:
It will already have staff trained in insurance;
it will have, in place, the necessary systems to handle information concerning the sums insured, and claims;
it will have accounting systems in place;
it is likely to have existing business relationships with re-insurers;
it will have a capital base, one which may be sufficient for it to enter into a new area of business.
However, crop and forestry insurance in developing countries is not likely to be any more attractive to existing insurers than lending in these sectors is to most commercial banks. This may then call for additional inducements or alternative administrative arrangements.
The establishment of crop insurance as a new line of business, whether in an existing company, or in a new entity, can benefit from the best experience available. At the time of writing the required expertise is most likely to be found within the reinsurance industry, and with specialized consultants/researchers. There is a role here for international agencies in making the necessary contacts, and assisting with the costs.
Once the administrative business structure is in place, attention must be given to developing a product or line of products to meet the already identified demand. It is at the stage of product development that it is necessary to identify the point at which insurance could most economically impact on and contribute to growers risk management strategies.
Whereas each industry will have its own special features, problems and opportunities, one general point can be made. Product development is a highly skilled task, requiring both detailed knowledge of farming and/or forestry, coupled with a sound appreciation of the principles and operational imperatives of insurance. As such, this can be an expensive stage in the process, and one with which international agencies can often assist. This assistance might be in the form of direct partnership in product design, or training existing insurance staff to handle the new challenges. In practice it is likely to start with both approaches. What is important to note is that the design of insurance products, like the design of products for other financial services, is an ongoing task.
Implicit in any moves to start crop insurance is the assumption that there is a demand for the product. Whereas automatic insurance has many advantages, as noted earlier, it is not always possible to design this type of policy. Marketing therefore is important. Several factors are important here:
Close links with the representatives of farmers and foresters, and speedy response to new needs for insurance;
similar linkages with banks, farm product buyers and others with business connections with insured growers. The possibility of insurance being rolled into a seasonal cropping loan has already been mentioned. In this type of arrangement the marketing is automatic, at very low cost;
attention to appropriate publicity;
scrupulous fairness in loss assessment and claims handling;
speedy payment of claims.
In standard, traditional insurance, the basic issue to be addressed is whether the insurance is meant to substitute for farm income in the event of a loss event, or whether the indemnity would merely cover the cost of inputs lost, because of crop damage. The second option is certainly the easier and lower cost alternative, as the level of overall coverage would be significantly less.
With index policies the choice would be more flexible, since an insured individual could choose the level of coverage, purchasing the number of units which suits his or her needs.
In any case, it is vital that an actuarial balance is struck between premium and indemnity levels, and that this balance be continually checked in order to ensure the financial sustainability of the programme, and its ability to meet commitments to insured growers. A key issue is the level of deductible (excess) which applies. The effect is twofold. Firstly, and more obviously it impacts directly on the premium level through an inverse relationship between the quantum of deductible and the pure premium required for a given level of insurance protection. Secondly, it also impacts through economies in loss assessment and adjustment costs, a deductible means that minor losses will not prompt a claim, and therefore no loss assessment will take place.
A major area of difficulty in setting indemnity and premium levels is the lack of data linking the incidence of adverse weather events and actual losses in the field. Experience has shown that historic newspaper reports are unreliable (they usually exaggerate the losses) and that reports kept by government ministries are similarly inaccurate, since in the absence of insurance there is little incentive, or need, for precision.
In any case, insurance products in agriculture are seldom launched on the basis of all the data an actuary would wish to have in order to set premiums at the level required to meet expected indemnity liabilities. Experience must be gained during the early years of a programme. During this period adjustments can be made to the indemnity and premium levels, and also to the percentage of deductible applied.
One of the oldest crop insurance programmes in the developing world, the Mauritius Sugar Insurance Fund, has already been mentioned. It is hardly surprising, given its maturity, that this programme has developed a sophisticated and very fair system for setting indemnity and premium levels. It has been able to do this through the availability of detailed and accurate data of cropped areas, ownership, tonnages delivered as well as data generated in the course of settling claims.
The main objective here is to keep costs as low as possible, so there is a strong incentive to build linkages with existing providers of services to the farm and forestry sector. In Cyprus the Agricultural Insurance Organization utilizes the connections that processors, wholesalers and exporters already have with growers. In Mauritius the Sugar Insurance Fund taps into the linkages that its 35 000 grower clients have with the 19 sugar mills in the industry. Similarly, in the Windward Islands, all registered growers have an ongoing account with the centralised, single channel marketing system. This provides a ready means for the collection of insurance premiums.
Perhaps the most obvious linkage is between the insurer and banks serving the same clientele, with the loan included as a component of the seasonal cropping expenses. Since the premiums in such cases are paid in bulk by the banks to the insurer, costs are minimized.
Again, cost containment is very much an objective in designing procedures for the notification of claims, for assessing the losses and for paying indemnities. Clearly the big divide is between the older, traditional type of policy, in which losses need to be assessed on each farm or forest, and the newer types of policies in which a more wholesale approach is possible (see Section 2.5, New Insurance Products, above). It has already been mentioned that some perils, e.g. hail, still require in situ inspection in order to determine the loss, as the incidence of this peril is very geographically confined.
A further potent field for cost economies is through building linkages with entities already providing services to growers. These include banks, input suppliers, processors and other buyers. Sometimes, when loss assessment is done on an individual basis, the process can be made more efficient by the ready availability of detailed information. In the Windward Is. an assessor is provided by WINCROP with all details of the claim on which he is to work - including data on the cropped area and full cropping history.
Whereas, as a business, insurance belongs in a business setting, the very nature of crop and forestry insurance means that there is bound to be strong governmental involvement. Most governments have a close interest in risk management in agriculture, both for productivity reasons, and concern for the wellbeing of rural populations. This often means, in practice, that governments are active not only in an overall policy sense, but can be more intimately involved in various ways. These can range from initial investigation of the feasibility of introducing crop and/or forestry insurance products, leading to eventual promotion, and even financial participation.
At the same time, and as stated above, there are strong reasons for the business operations in insurance to be handled by a commercial concern, for reasons of efficiency and convenience in terms of insurance operations complementing other commercially-run services to farming.
This dual parentage of crop insurance can lead to tensions. The most crucial areas of concern lie in the areas of premium setting and claims handling. In these areas experience has shown that undue and inappropriate political influence on an insurer can be very damaging.
Accordingly, much attention is given during the design of crop insurance programmes to avoiding these tensions to the extent possible. Such avoidance is aimed at optimising the role of the public sector, while harnessing the drive and efficiency of the private industry sector.
Several steps are involved. One listing might suggest the following as important:
Ensure that any existing company or new entity has a sound legal basis on which to offer insurance products, with the required level of business competence.
Clarify the governments objective in promoting crop insurance. Is it purely an additional risk management mechanism, or is it also an avenue of subsidy to the farming sector? If the latter is the case, then the avenue for financial support has to be ring-fenced from day-to-day political interference. This is not easily done, yet it is essential if there is to be the required continuity of financial conditions in order to build efficiency and fairness into the system.
Establish strong linkages, at an early stage, with international re-insurers. These companies can assist not only with technical advice, but can also be instrumental in ensuring the necessary adherence to correct application of premium setting procedures, and settlement of claims. Although the opportunity for profit may be some years away, such companies are often prepared to become involved in a new geographical field of business. They operate with long term time horizons, and this can work very much to the benefit of a nascent crop insurer - whether this is a new company or a new section within an established company.
The financial base for the insurer must be adequate. This must be sufficient to survive initial years in which weather conditions might be such that underwriting profits are sharply negative. On top of this loss, administrative expenses have to be met. In many developing countries there may have to be public sector participation in ensuring a sound financial base.
Work closely with representatives of the farming and/or forestry sectors. This will help ensure that the service and products are popular and therefore in demand.
 The policy calls for a
premium of 2.5 percent of the insured value, with a deductible of 15 percent,
which rises to 30 percent in the event of two successive years of claims from a
 Skees et al. (2001) suggest that a rainfall index insurance product for Morocco would overcome the high costs of the present experimental yield assurance programme, where losses are adjusted individually.
 One of the major objections to index insurance for drought is the practical difficulty of ensuring that rainfall readings are genuinely accurate. This requires rain gauges that are tamper-proof. Various methods for tackling the problem of tampering are being developed. Hazell (2001) reports on the availability of a system of multiple rain gauges, which can be sited on the top of telegraph poles, and which measure rainfall by means of tiny buckets which trip a measuring device. These have the advantage that they do not store the rainfall, merely record the volume falling at that point. The data can be downloaded to a computer. Remote access to the data could be readily arranged, and a pattern of rainfall readings built up form the multiple measuring points. Satellite imagery over the insured period could also be utilised to confirm an insurable weather event.
 A tropical cyclone originates over tropical or subtropical waters, with organized deep convection and a closed surface wind circulation about a well-defined centre. Once formed, a tropical cyclone is maintained by the extraction of heat energy from the ocean at high temperature and heat export at the low temperatures of the upper troposphere. Tropical cyclones are known as hurricanes in the Atlantic/Western hemisphere, and as typhoons in the Asia/Pacific region. They circulate counterclockwise in the Northern Hemisphere, and clockwise in the Southern Hemisphere.
 Total damage from Hurricane Andrew was estimated at US$26.5 billion, of which all but US$1 billion occurred in Florida. The greater part of this was wind damage, and the value quoted includes all damage- not just crop losses.
 At the high-cost end of physical frost damage prevention, kiwifruit and other fruit growers in New Zealand occasionally hire helicopters at crucial times - usually in the early hours of the morning in the late spring, when exceptional weather conditions are indicated by forecasts, and in-orchard temperature recording instruments, linked to alarms, warn of an imminent danger. The helicopters hover over an orchard, selecting the warmest layer of air, then use their rotor blades to blow this relatively warm air downwards. The cost is high, but the procedure is usually effective. In other countries, e.g. France and Syria, propellers set on towers are used to circulate air.
 A notable exception is the worst hailstorm of modern times in Western Europe. This struck Southern Germany, including the area of the city of Munich, on the 12th July 1984. Several swathes of hail struck across an area some 250 km long by 15 km wide. Total damage was estimated at US$3 billion, with the insured portion of this being US$750 million. Several hailstones of 300g in mass were registered. Agricultural damage was severe, as was damage to vehicles, aircraft and glasshouses/plastic houses.
 Hail nets are used by some apple growers in the Granite Belt zone of Queensland, Australia. The cost of the structures is high, but those who use this form of management against hail risk do so because they have marketing contracts to fulfil, and need an annual apple crop, and not just an indemnity payment, in order to maintain a valued outlet for their fruit. In many countries cherries are protected with netting, to protect against hail and birds. Hail netting is also in widespread use in France and Italy.
 Among the provisions under the Finnish Forest Insect and Fungal Damage Prevention Act is the requirement that in areas of high blow down, snow breakage etc., trees must be harvested so that the risk of insect damage to living trees is minimized.
 An example comes from Mauritius. The Mauritius Sugar Insurance Fund (MSIF) was established to provide automatic windstorm (cyclone) damage insurance for all the islands sugar growers. After 27 years of experience losses from fire, and from excessive rain, were included in the policies. After a further 10 years of operations (i.e. in 1984) the MSIF recognised that yellow-spot disease could not be controlled by normal farm management practices in conditions of excessive rainfall. Therefore the MSIF agreed to indemnify losses from yellow-spot when excessive rain also occurs.
 In Central Italy, in the peak fire risk (summer and autumn) seasons, it is a requirement that owners of pasture land keep a ploughed margin around the perimeter of their land, to act as a fire break. In many countries large commercial forests are watched over by fire wardens sited in high towers, linked by radio to a control centre which has command over equipment which can be sent to the scene. This may mean the use of aircraft to drop water, and/or bulldozers to fell trees in the path of the fire, thus creating an additional firebreak.
 As an example, in France, the cooperative farmers insurer, Groupama, publishes a manual setting out construction standards for glass and plastic houses in which crops are grown. These standards must be adhered to for Groupama insurance contracts to be valid. The standards and the manual are frequently updated.
 Cottle, Phil (2001)
 Also known as CERUs - carbon emission reduction units.
 Loss assessment is not a part of the operations of an index policy. This pays out a predetermined indemnity, on the basis of weather or other records. The state of the crop of the insured is not a factor affecting the eventual payout. Because loss assessment in the field is a very expensive process, this is a major advantage of index policies.
 From the beginnings of the frost and hail insurance on the New Zealand kiwifruit crop, the cover was arranged by the export authority which handled all export kiwifruit. This meant that cover was automatic for all growers, and only one premium payment was made. The administrative cost economies were very significant.
 It has already been noted above that insurance is one area of governmental assistance to the farming sector which is permitted under WTO regulations.
 In India the premium levels for the major state-run crop insurance programme were set by government, on political rather than actuarial grounds. In Cyprus the premium rates were written into the Act establishing the parastatal insurer. In both cases the inflexibility has resulted in problems for the insurance entities involved.
 See also Section 2.3 above.
 Some years ago the then Government of The Philippines provided a substantial fund to establish the Philippines Crop Insurance Corporation (PCIC). A substantial part of the fund was invested in order to provide a regular adjunct to the premium income of normal insurance operations. Establishing this endowment fund was a long-sighted move on the part of the then government. It has provided the PCIC with a valuable degree of autonomy and responsibility.