Agriculture is an inherently risky business. It is subject to a number of random price, climatic, biological, and geological shocks that require coping strategies and financial management instruments to deal with the implications. Traditional risk management strategies and ex post government provided emergency relief have often not proven to be sufficiently effective and robust in preventing serious economic loss or permitting a speedy recovery. This paper focuses on production risk management, explaining key concepts, understanding why crop insurance markets have been slow to develop, and making recommendations about how to build sustainable markets in developing country contexts.
For the most part, producers in developing countries are quite exposed to weather vagaries and have little access to formal agricultural insurance products that would allow them to transfer production risk to other parties. Agricultural insurance was more widespread in Latin America and other developing regions of the world during the 1960s and 1970s. However, most of the comprehensive, multiple peril programs common then, encountered financial difficulties and were either scaled back or completely closed. At present in Latin America, the volume of agricultural insurance premiums is a miniscule share of total insurance premiums.
Nonetheless, agricultural insurance is reemerging as a topic of interest, especially in light of the need to improve agricultural competitiveness in increasingly integrated commodity markets. The challenge is how to overcome obstacles and deliver efficient and sustainable agricultural insurance products. The principal obstacles—lack of high quality information, inadequate regulatory frameworks, weak supervision, lack of actuarial expertise, lack of professional expertise in designing and monitoring agricultural insurance products, a mass of low-income, dispersed clients, who may not be willing or able to pay actuarially sound premiums for multiple peril products, and the tendency of governments to undermine market development through inappropriate use of subsidies and disaster relief funds--are highlighted and discussed. Case studies on Uruguay, the Dominican Republic, and Peru reveal how crop insurance products are evolving and/or what government-supported initiatives are under the way to expand coverage. Recommendations of how to build markets step-by-step and the importance of applying new technology to lower costs are made.
Agricultural insurance is presented as important financial risk management tool but not as a panacea for unprofitable farms, management failures, underinvestment in public infrastructure, or compensation for other poorly functioning factor markets. Different types of agricultural insurance products—single peril, multiple peril, parametric, and revenue—each have a niche but should adhere to basic principles of actuarial fairness, seek to minimize problems with adverse selection, moral hazard, and administrative costs. Governments have a vital role to play in providing the necessary information needed to measure, evaluate, and monitor risk, in maintaining an auspicious but sound regulatory and supervisory framework, in helping with reinsurance and catastrophic disaster relief, and supporting private insurance providers with technical assistance and training. Often time, the argument is made that “public subsidies for premiums” are necessary in order to make premiums more affordable for the majority of farmers. The argument presented here is that scarce pubic monies may be better spent on creating favorable market conditions for the development of the industry ( i.e. the maintenance of databases, training, and pilots) than on making transfers to private individuals. In the context of developing countries, with large rural populations (often exceeding 20%), sizeable agrarian sectors (agricultural share of GDP >10 %, agricultural exports as a share of total exports > 30%), and severe fiscal constraints, agricultural insurance systems should be cost effective and operate as part of a larger, layered risk management framework. Installing comprehensive and universal systems, as is the case for several industrialized countries, may be an inefficient use of scarce public monies for developing countries. In a layered framework, farmers should be trained how to reduce and cope effectively with some of the production risks on-farm through better management practices and diversification strategies; how to transfer some of the production risks to financial markets through efficient and sustainable instruments (insurance, savings, and credit); and how to rely on the government assistance for catastrophic events. In the latter case, rules for accessing disaster relief should be clear ex ante and not remove or undercut incentives for the adoption of better on-farm management techniques (moral hazard), the purchase of private agricultural insurance, or the accumulation of personal savings.
AGRICULTURAL YIELD INSURANCE: A PRIMER
Importance of Production Risk in Agriculture
Agriculture is a risky business. Producers face a host of different risks among them production or yield risk. While production risk cannot be totally eliminated, it can be reduced and managed. In order to address the financial implications of this type of risk, producers have historically relied on a variety of strategies and coping mechanisms that can be categorized into three general classes: risk mitigation, risk transfer, and risk retention. This paper focuses on how to effectively transfer risk. Producers often report that production and price risks as their two major concerns. Each year, unmanaged production risk contributes significantly to high economic losses throughout the developing world and helps to perpetuate poverty and income inequality.
Among the numerous sources of production or yield variability, weather is universally recognized as the dominant one. Figure 1 lists the principal sources of yield variability—quality of soil, planting date, genetic potential of the plant or animal, application of fertilizer, husbandry practices. Recent research from the Baltic states show that weather differences alone explained 35% of the variation in yield for a representative sample of farmers (See Figure 1). Of course, the relative importance of the factors may vary from place to place and with the level of technology employed. But what distinguishes climate risk from the other listed factors, however, is the degree of human control possible. The non-weather factors can be significantly reduced or mitigated with on-farm strategies, with the principal constraints being farmer knowledge and financial resources. In contrast, weather cannot be controlled and constitutes a residual risk that should be transferred, and it cannot be retained with serious financial indications. For example, the farmer can select the best seeds for planting, match plant agronomic requirements with soil characteristics, take preventive actions to minimize the risks of insect infestations or disease, and fertilize according to a schedule based on best available extension service knowledge and nutrient analysis of the soil. In practice, however, the degree of effective control is far from ideal, more so in the case of developing countries where extension services are weak, farmers have less access to information, and have less years of education, and limited access to credit. Thus, the combination of management shortcomings and weather vagaries makes agriculture more risky than most other economic enterprises in developing countries. This high degree of riskiness, especially in a sector dominated by producers with low-incomes and scant assets, has serious implications for economic growth, social equity, and poverty alleviation. Market and government-based solutions are needed.
Three major types of natural phenomena contribute to yield risks in agriculture: hydro-meteorological, geological, and biological. Hydro-meteorological risks include excessive rain, floods, droughts, high winds, tornados, hurricanes, hail, frost, abrupt temperature changes, heat waves, blizzards, prolonged cold spells, avalanches, landsides, high waves, storm surges. Geological risks include earthquakes, volcanic eruptions, and tsunamis. Biological risks include diseases and insect infestations. Each of these risks can then be categorized as either catastrophic or non-catastrophic, depending on frequency, scale, intensity, and duration. Catastrophic risk refers to natural disasters (earthquakes, hurricanes, volcanic eruptions, tsunamis, tidal waves, storm surges, etc.) that inflict large-scale damage over an extended area but are infrequent, low probability events. Non-catastrophic climatic risks (droughts, floods, landslides, mudslides, hail storms, freezes, heat waves, etc) affect localized areas (one or two provinces) or sometimes only a few farms. They tend to be more frequent, last longer, but cause less total economic damage. Biological risks such as insect infestations and disease epidemics tend to be localized but in some instances, if the disease is highly communicable, livestock may have to be slaughtered over a very wide swath surrounding the original outbreak point as a containment measure. Thus, an outbreak of brucellosis, a bacterial infection that affects ruminants, on one farm is a non-catastrophic risk whereas an outbreak of mad cow disease (BSP) would be a catastrophe for the entire cattle industry in a particular country.
Over the last decade the occurrence of natural disasters has been trending upward. The set of countries experiencing the least growth in disasters are the Central European States and the former Soviet Republics. By the year 2050, the United Nations estimates that natural disasters will cost 300,000 lives and approximately $250 billion in economic losses per year worldwide, if more measures are not taken to mitigate risks and reduce global warming (UNISDR, 2002). For the period 1994-2003, the countries that are most vulnerable to natural disasters are developing countries, particularly those in Sub-Saharan Africa and East Asia. Latin America and the Caribbean are most exposed to windstorms, floods, and droughts, and windstorms (See Table 1). Within the region, however, there are marked differences. For example, while windstorms are the most frequent threat for the Caribbean, whereas floods are virtually six times more frequent than windstorms for South America. The seven Central American states, in turn, differ markedly from the other two sub-regions. They are disproportionately exposed to a larger array of threats—windstorms, floods, droughts, earthquake, volcanic eruptions, and epidemics-- and thus accounted for 35% of all the region’s disasters. South America, with a larger number of states and with each state being larger in landmass than any of the Central American states, accounts for 47% of the natural disasters.
According to the World Bank (2001), between 1988 and 1997 natural disasters in the developing countries claimed an estimated 50,000 lives and caused direct damage valued at more than US$60 billion a year. In the Latin American and Caribbean region, as can be seen in the figures that follow (Figures 3a, 3b, and 3c), the economic losses for the Latin American and Caribbean region was US$31.9 billion for the period 1990-99 (CRED International Disaster Database, 2005). In the period 2000-05, the losses sum to US$17.5 billion (CRED International Database, 2005), the Caribbean having experienced double the losses than the entire previous decade as a result of hurricane damage in Haiti, Dominican Republic, and Grenada in 2004. These costs create dislocations at many levels— farm household, local regional economies, and national economies.
Effects of Shocks and Common Traditional Production Risk Management and Coping Mechanisms
The typical effects of a hydro-meteorological, geological, or biological disaster are one or more of the following. The number of effects detectable is a function of the scale, severity, and duration of the initial shock. The more spatially correlated the shock, the more effects will be noted.
• A decrease in farm income.
• A decrease in employment for hired farm workers.
• A generalized fall in demand throughout the local or regional economy as a result of the reduced agricultural income of affected farm families and agroindustries.
• An increase in loan defaults in affected region, affecting both financial intermediaries and agricultural input suppliers who sold on credit.
• A decrease in government tax revenue and foreign earnings due to a fall in agricultural exports.
• An increase in the price of basic food items, if the affected commodities where normally marketed domestically.
Ex post, the affected rural population in response to the shock may engage in more than one of the following behaviors depending on the severity of the shock and initial economic conditions. Note that some of the behaviors have primarily micro or local effects, that is they affect mostly the household and immediate neighbors while other behaviors have macro or farther reaching external effects in that they affect not only the household in question but other parties such as municipal, state, and national governments, financial intermediaries, and urban residents.
MICRO EFFECTS:
• Drawdown accumulated savings
• Seek loans
• Reduce consumption, including food intake and withdrawing children from school nutritional threshold that would impair health.
• Liquidate assets
• Seek off-farm wage employment
• Depend on remittances.
• Depend on informal reciprocal sharing arrangements with neighbors and kin (mutual insurance)
MACRO EFFECTS:
• Seek to refinance existing loans
• Seek debt forgiveness for existing formal loans
• Petition the regional and national government authorities for emergency relief
• Exit farming permanently and migrate to an urban area
• If persistent over time, competitiveness of agricultural production unit(s) reduced.
Ex ante, farm households, in the absence of affordable, formal risk transfer instruments such as insurance, tend to rely on a series of informal risk reducing and risk retention coping strategies. The only informal risk transfer strategies are share tenancy and mutual aid:
• Use lower yielding but drought resistant varieties
• Stagger planting times to assure that a fair percentage of the crop receives sufficient rain in first stage of development
• Fragment plots to take advantage of different soils, elevations, slopes, and microclimates, going as far as to rent land that has distinct agroecologic characteristics from own land
• Intercrop two or more crops and/or tree species on a parcel
• Diversify income streams
• Conserve soil moisture by applying litter and mulch to the roots of plants or using raised beds.
• Use of integrated pest management
• Allocate relatively more labor resources of the household to other non-farm businesses or off-farm employment opportunities if crop yields are threaten by bad weather or pests during the growing season
• Adopt irrigation technology
• Sharecrop
• Engage in reciprocal lending where in a household provides a no interest loan to another distressed household in good times and in turn expects to be able to borrow in bad times.
• Engage in gift giving to build social capital in the community and to create “chits” that can be called in time of economic distress.
• Participate in informal group mutual aid , savings, and insurance schemes such as ROSCAs and ASCAs (Rotating savings and Credit Associations and Accumulating Savings and Credit Associations).
• Accumulate buffer stocks or liquid assets
• Maintain a credit reserve with a bank or agricultural supplier
• Reduce amount of purchased agricultural inputs and thereby minimize debt load or reduce the amount of target income needed to assure survival of the household
Public Policy Implications of Inadequate Traditional Risk Management and Coping
Strategies
While some of the on-farm, risk mitigation practices are time tested and highly recommendable such as crop diversification, intercropping, soil humidity management, integrated pest management, irrigation, and accumulating savings. Many of the other practices such as such as plot fragmentation, economizing on purchased inputs, and the use of low-yielding but drought resistant varieties, represent production efficiency losses (Rosenweig and Binswanger, 1993; Morduch, 1995; and Kurosaki and Fafchamps, 2002). By foregoing specialization, farmers tradeoff income variability for lost profitability and reduced future earning ability. Others such as reciprocal lending, gift giving, and participation in ROSCAs and ASCAs work for idiosyncratic risk but maybe overwhelmed and useless if the risk is covariant, that is affecting with more or less equal severity all the households in a particular community or region. Consequently, these costly risk mitigating techniques and can contribute to chronic poverty and increased vulnerability. In a setting of increasing trade liberalization and integration, the absence of absence of agricultural insurance instruments places developing country producers at a serious disadvantage vis-à-vis farmers in industrialized countries that have greater access to such instruments. The result is less trade competitiveness ceteris paribus.
In short, the farm household produces inside its production possibility curve and increases its chances of remaining below or close to the poverty line. At the level of the regional economy, a large number of households engaging in risk avoidance behaviors and producing at a suboptimal level, reduce tax revenue, limit the ability to finance social services, makes for a stagnant regional economy due to less effective demand. At the national government level, a slow growing or stagnant agricultural sector contributes to less marketed food output, less export earnings, high rural-urban migration rates, incomplete financial markets, and increases the demand for extraordinary fiscal assistance to cope with major emergencies.
The cumulative micro effects create the setting for the existence of a poverty trap and the cumulative macro effects create the setting where government intervenes inappropriately and instead of solving a market failure aggravates the situation by creating a government failure.
Lack of Modern Risk Management Instruments Increases the Vulnerability of the Rural Poor and May Contribute to the Persistence of a Poverty Trap: One of the distinguishing characteristics of the poor is their vulnerability to risk. Poor people in developing countries depend heavily on agricultural production and selling their labor to survive. Since consumption takes a greater share of income among low-income families, shocks that create a marked drop in income can easily force the household below minimal nutritional thresholds. Some can recover quickly, especially if they have enough tangible assets. Others do not fare as well and are unable to break the cycle of poverty and stagnation and remain in a poverty trap. The lack of formal, risk transfer instruments makes the poor and near poor more vulnerable and adverse to making risky and uncertain investment decisions that would put at make their income levels more variable than it is and risk their meager stock of physical assets. Thus, a low-productivity, poverty equilibrium could arise.
Recent research has shown that not only is the magnitude of poverty different between industrialized countries and developing countries (rates of poverty in developing countries surpass 30% on average while it is less than 20% for higher income countries) but also the dynamics of poverty (Naifeh, 1998). Whereas the monthly poverty exit rate hovers around 7% for the U.S., exit rates in developing countries such as Cote d’Ivorie and KwaZulu Natal state in South Africa range between .7%-1.3%, meaning that the median duration of poverty for many in developing countries can last a lifetime while the median time in poverty is 4.5 months for the U.S. (Barrett and McPeak, 2005). In developing countries, some individuals have accumulated sufficient assets that qualifies them as non poor but a transitory shock reduces their income and expenditures levels and forces them below the poverty line (stochastic poverty). Others lack a sufficient endowment of assets and can never seem to improve returns to their assets or accumulate more, so they languish below the poverty line (structural poverty). This phenomenon is illustrated in the two figures below.
Taking a static view in figure 4, a household is pictured as mired in a vicious cycle. Insufficient income and assets contribute to hunger, lethargy, and poor health, which in turn reduce income-earning ability and prevents investments in activities that would enhance upward mobility such as education.
Taking the dynamic perspective in figure 5, a poor household (II) is much more vulnerable to falling deeper into income poverty and staying poor, than say household (I), when both households experience the same adverse shock at Time period 4 resulting in an income decline, systemic economic growth processes would be sufficient for household (I) to reemerge from poverty but it would be insufficient to break the cycle of poverty for household (II). In the case of household (II), the causes of structural poverty and income inequality will have to be attacked through asset based poverty reduction strategies—asset redistribution, social safety nets, targeting, better protection of property rights, more investments in human capital. The introduction of a more modern risk management instrument such as crop insurance as one element in a larger package of interventions could theoretically help household (II) protect what few productive assets it has during a downturn and then to leverage those assets to a higher extent to grow out of poverty over time. The use of crop insurance could help to place household (II) on the dashed orange upward growth trajectory II preventing the liquidation of assets.
The Lack of Ex Ante Crop Insurance Can Retard Rural Lending, Create Fiscal Stresses for Central Governments, and Promote Rent Seeking Behavior: When crop insurance does not exist or is not used to an appreciable extent in an agrarian economy, the central government and international donors are relied upon to provide relief in the case of very severe disasters. While it can not be denied that central governments and international organizations must respond and play a role in the case of massive catastrophes, the use of ad hoc, ex post interventions sets dangerous precedents and tends to have four negative consequences if the role of government is not clear and actions are not well designed.
First, ad hoc emergency programs disrupt budget planning and administration. Funds often time have to be diverted from other ongoing and approved government programs to attend to the agricultural emergency. In the absence of a well-established emergency disaster fund with transparent rules and adequate funding, governments can easily fall into the trap of “robbing Paul to pay Peter.” If the country in question is under budget stress, it may have to engage in deficit financing and as a consequence contribute to upward pressure on interest rates in the banking system. Higher lending rates reduces the demand for loans and makes agricultural financing ever more problematic, since as a whole agriculture is a sector noted for smaller profit margins than others.
Second, the knowledge that the government is likely to “bail out” affected parties creates moral hazard conditions and depresses the market for private crop insurance. Farmers do not do all that they can do to reduce individual vulnerability to adverse climatic conditions and other biological threats. Similarly, insurance companies have little incentive to enter rural markets and offer costly insurance products since they fear that demand for their products will be weak since farmers would prefer free ex post assistance from the government as opposed to paying a premium ex ante. The negative results are represented in Figure 6. Central governments should provide disaster assistance but should set rules of eligibility so as to encourage the purchase of private insurance and/or precautionary actions to reduce vulnerability to losses.
Third, well-organized groups of farmers have a strong incentive to lobby the government for relief from a wide and varied number of adverse climatic and price effects. Thus, government can be called upon to provide relief for non-catastrophic events, which normally should be in the domain of private insurers. Many times, the farmers that are more organized and most influential tend not to be the poorest. They, tend to produce a crop that is “strategic” i.e. a domestic staple like rice or an export commodity like beef cattle, wine, cotton, or sugar. Thus, the government provides transfers to relatively non-poor farmers. To counteract and limit rent seeking opportunities, the government should be clear as to how risk management will “layered” in that private individuals will be responsible for certain types of risk and up to a certain limit, private insurance markets for another segment, and as last resort, central government will be responsible for risk losses that surpasses the limits of both private insurance and reinsurance companies.
Fourth, the ready willingness of central governments to use debt forgiveness of formal agricultural loans undermines the solvency of banks, destroys repayment culture and dampens the willingness of financial intermediaries to expand and innovate in rural areas. Financial intermediaries tend to retreat from the agricultural sector after such an event, and if they stay, they only lend to collateral rich and well-known clients. Thus, financial markets remain shallow, noncompetitive, and incomplete. Debt forgiveness, while timely and easy to implement from the perspective of politicians, also tends to be regressive in nature. It benefits larger farmers much more so than smaller one because they tend to have more access to formal credit and take out on average larger loans. Smaller farmers tend to experience greater difficult accessing formal credit, and if successful, borrow smaller amounts. Over time, the combination of intermediary weakening and regressive transfers, contributes to increased income inequality. Those with access to finance can invest in more productive technology, diversify faster, expand their scale of operations, and thus experience faster income growth, all else equal. The use of debt forgiveness should be avoided and used sparingly.
Formal Agricultural Yield Insurance: Definition, Basic Requirements, and Benefits
In previous sections, we have argued that traditional risk management and coping mechanisms are often time neither sufficiently robust nor cost effective. The amount of residual risk that remains with the household in question may induce asset liquidation and poverty. Ex post government relief actions also create incentive problems and are costly to the treasury. But what is agricultural yield insurance and how does it function?
Agricultural yield insurance is a financial contingency contract that transfers production risk from a producer to another party via the payment of a premium that reflects the true long-term cost of the insurer who is assuming the risks. The insurer pools the risks faced by a large number of individuals and covers losses incurred by any one individual in the pool. It serves to essentially protect assets, stabilize income, and smooth consumption. However, for insurance to be viable and sustainable, certain “ideal” conditions for the risk to be considered insurable and for a self-sustaining market to appear.
1. Symmetric Information: The insurer and the insured should have the same approximate knowledge of the distribution of probable losses so that proper risk classification can occur. Insurers typically do not develop premium rates on an individual basis since it would be extraordinarily expensive. Instead, insurers classify applicants into homogeneous risk pools and calculate a premium for everyone in that group. In order to estimate probable losses for different groups of risks, extensive amount of reliable and accurate information is needed on weather patterns, yields, market trends, farm conditions, farm management ability, risk attitudes, and capacity to pay for the insurance.
2. Large Number of Similar Exposed Units: The statistical Law of Large Numbers upon which the actuarial models use to calculate coverage, indemnity, and premium levels , states that the more uncorrelated risks that are added to a portfolio the lower the variance of outcomes for the entire portfolio. Thus, for the actuarial models to be accurate the size of the pool or portfolio should be large and the risks faced in a particular class or group should be similar.
3. Statistical Independence of Risks: Risk should be nearly or perfectly independent across insured individuals and spatially uncorrelated. Insurance is based on the principles of diversification so that a major consideration is the degree of correlation in financial losses caused by the risk to be insured. The more spatial correlation there is the less efficient insurance will be as a risk transfer mechanism. When losses are catastrophic, the risk–pooling advantage of insurance breaks down because the contributions of the unaffected are insufficient to cover the damages of the affected.
4. Calculable Expectancy Frequency and Magnitude of Loss: The insurance company should be able to estimate both average frequency of the random event to be insured and the average severity of loss. For low-probability risks with potentially catastrophic outcomes it is often difficult to estimate the average expected loss, because there are so few data points.
5. Actual Losses must be Determinable and Measurable: The actual loss should be clearly and causally linked to the random event insured and it should be a tangible and measurable loss. If this is not the case, claims settlements will tend to be highly contentious. Purchasers will lose faith in the process and insurer’s administrative costs will skyrocket.
6. Potential Losses Must be Significant and an Insurable Interest Must Exist: Potential buyers must perceive the probable loss as significant and beyond their own means to cover; otherwise there will be no incentive to purchase insurance. Furthermore, insurance cannot be provided to policyholders who have a vested interest in a loss occurring. For example, a property insurance policy cannot be sold to anyone other than the owners of the home and/or the owners of the furniture in the case of a renter with an unfurnished lease. If someone else could purchase such a policy, they would experience no loss if the house or furnishings were damaged or destroyed but would receive a pay-out from the insurance company. Owners and renters with “insurable interests” would have incentives to take precautions because of deductibles.
7. Limited Policyholder Control over the Insured Event: Insurance protection should not be offered if policyholders can control whether an insured event will occur. If a policyholder has sufficient control over whether a risk can occur, they can take advantage of the insurance and generate “moral hazard or suspect claims”. For example, a farmer can fail to properly care for livestock, which could induce disease causing the death of the animal, and then file a claim for loss.
8. Premiums should be Economically Affordable: In general, for an insurance policy to be attractive to potential policyholders, the annual premium cost must be substantially less than the potential benefit offered by the policy, should the insured event occur. A market for insurance may fail to appear, if the majority of clients are very poor, very isolated, and/or the chances of losses are high. A fully loaded premium could exceed the estimated cost of a one-time loss and make the product uneconomical and useless. When insurance premiums are very high, credit and savings instruments become preferable risk management instruments.
If the above conditions are met, agricultural insurance can be offered on a sustainable basis and has five main benefits.
First, agricultural insurance is often time a more efficient and potent financial instrument than either using liquid savings or credit in managing yield risk. If a household or farm enterprises is subject to a series of shocks in a short span of time, it may deplete its entire savings and not have enough to invest to improve future earnings. In many countries rural formal credit markets are very undeveloped and access is problematic. Thus, in the event of a sudden income loss, a credit-constrained household may have to rely on informal sources, friends, family, and moneylenders that may not extend sufficient volume of credit necessary to meet the crisis or at a very high interest rate. Recent empirical research from rural China, that analyzed portfolio behavior in represent to income and health risks shows that households in the lowest and highest quintiles did not appreciably reduce wealth held in liquid forms while those in the middle quintiles did to a higher extent (Jalan and Ravallion, 1998). The authors reason that the rich do not need to hold unproductive precautionary liquid wealth to deal with income losses because they had access to credit and the poor could not afford to hold precautionary savings. Thus, in the context of undeveloped savings and credit markets, making formal insurance accessible to the very poor households would permit them to transfer unmitigated residual risk to an external party and thus avoiding sinking deeper into poverty.
Second, the use of agricultural insurance can facilitate the adoption of higher yielding technologies and intensification of production by risk adverse farmers. The presence of insurance gives added comfort to innovators.
Third, agricultural insurance reduces credit default risk for financial intermediaries financing agricultural production. Crop insurance policies can serve as a substitute for physical collateral and give financial intermediaries more comfort and incentive to lend to the sector. Insurance policies can be made endorsable to a credit lender.
Fourth, agricultural insurance would help both rural households and governments manage natural hazards better and reduce the vulnerability of the rural poor. Insurance could help a rural house avoid falling into poverty traps. It would help forestall political demands for ad hoc disaster relief. Governments normally provide monetary compensation to affected households in ex post disaster relief efforts but often distribution of the aid is not timely.
Fifth, agricultural insurance in a world marked by increasing agricultural trade liberalization and integration is a means to enhance agricultural competitiveness. In a global marketplace, producers that enjoy the benefits of crop insurance are better able to assume new investment risks without mortal fear of losing a significant share of their asset base or being forced to exit agriculture if the undertaking fails due to adverse weather. Many producers in OECD countries enjoy the benefits of crop and livestock insurance and the spread of agricultural insurance to developing regions with help to level the playing field.
Impediments to the Development of Agricultural Insurance Markets
Despite the inadequacies of informal risk management systems and problems with ex post government actions, agricultural insurance is grossly underdeveloped in middle and low-income countries. One may ask why this is so given the clear benefits. The fundamental reason is that the ideal conditions laid out in the previous section are not often met in reality and the adjustments and compromises made often prove to be inadequate so one veers between markets with a few insurers offering sustainable but limited appeal single peril products to markets heavily intervened by governments either directly or indirectly offering multiple peril products with broader appeal but which are unsustainable. Many of the crop insurance programs that appeared in the 1970s and 1980s failed miserably because the “golden rules” were not adhered too. Below is a complete list of impediments to a more stable and complete insurance market.
Lack of Statistical Independence
Asymmetric Information
High Administrative Costs
Mismatch between Farmers Preferences and Capacity to Pay
Inadequate Legal and Regulatory Frameworks
Distorted Government Incentives
Reluctance of Reinsurers to Enter the Market
Lack of Statistical Independence: Formal insurance work best when the risks to be insured are perfectly independent and spatially uncorrelated, but agricultural production risks are in between.
Agricultural production losses, deviate from the ideal and tend to fall between the two extremes of being 100% uncorrelated and 100% correlated. Agricultural yield losses tend to be characterized by some degree of positive spatial correlation. The degree of spatial correlation is often inversely related to the size of the region or country where activities will be insured. Thus, relatively small countries are likely to be characterized by more positively correlated agricultural losses than a large country. Moreover, positive spatial correlation in losses reduces the benefits that can be obtained by pooling risks from different geographical areas. When, risks are perfectly correlated, insurance fails as an instrument of risk transfer, and capital market instruments such are derivatives are more appropriate.
A good agricultural insurance risk would be an idiosyncratic or largely uncorrelated one, a risk that is unique to a household and unrelated to neighbors and possibly due to management differences. Examples would be hail or fire. Hail and fire tend to be very localized events. In the case of fire, people can take preventive measures against fire. Thus, with inspections and a large and geographically diverse pool, theses risks are insurable. On the other hand, private insurers do not like to insure against drought or hurricanes (systemic or correlated risks), which affect large areas, unless reinsurance is available.
Asymmetric Information: Problems arise when prospective farm insurance clients have more knowledge about their own distribution of probable losses than the insurer cannot correctly classify potential clients by risk type and subsequently calculate premium rates that accurately reflect the true likelihood of losses for individual farmers, or monitor them effectively once a contract has been purchased. As a result, two attendant problems emerge—adverse selection and moral hazard. In the case of adverse selection, persons with very risky profiles will purchase the insurance in greater proportion than persons with less risky profiles, generating an imbalance between indemnity payments and premium revenue. If the insurer raises the premium higher in subsequent periods, less risky clients will withdraw and the profits of the company will fall further. In order to overcome adverse selection problem, the company will have to invest more heavily in obtaining better information, especially farm level yield data for long periods, so as to permit better risk classification. The other related information problem is one of moral hazard, wherein the insured changes behavior and may become less diligent in minimizing production risks knowing that potential losses are covered. Since monitoring the behavior of the insured is costly and imperfect, this could lead to potential losses for the insurer. To overcome this problem, the insurer has to design better contract designs and rely on less costly systems of monitoring.
High Administrative Costs: Information is vital to risk measurement and evaluation yet it is tends to be costly to obtain, process, and analyze. Agricultural insurance companies have to gather significant amounts of data on climate, production conditions, yield distributions, prices; capacity to pay; develop models to determine probable losses; design appropriate contracts and set premiums and indemnity levels; establish, inspection, monitoring and claim adjustment processes; and seek reinsurance. The more disperse the client base, the more heterogeneous the farm production systems, and the smaller the insured value, the higher the administrative costs are as a percentage of premiums. Compared to other lines of insurance, agricultural underwriting and claims adjustments are generally much more costly. In the context of developing countries, where data tend to be unreliable and difficult to obtain in a timely manner, the costs escalate. In rural areas with poor roads and telecommunications systems, the cost of client monitoring and making quick claim adjustments escalates.
Mismach between Farmers Preferences and Willingness to Pay: Many farmers seem to have a limited willingness to pay a premium that covers the cost of the service provided. As a result a sustainable market does not appear. Farmers seem to prefer insurance that protects a sizeable proportion of income from multiple threats as opposed to ones that partially cover income loss from a specific threat. These types of insurance products, revenue and multiple peril, are the most costly and difficult products for private insurance companies to provide in a profitable and sustainable manner (Goodwin, 2001). The financial performance of multiple peril insurance programs has been universally disappointing (Just et al, 1999). The fact that insurance these insurances are designed to protect against losses from a multitude of perils makes the calculation of probable losses and the determination of actuarially fair premiums very difficult if not impossible. In the countries were these types of policies are offered, they normally require substantial government subsidization. The products that can be delivered profitably and at affordable premiums are specific or single peril policies—hail and fire – and parametric or indexed based products. But they have less broad based appeal. In the case of parametric products, ones that pay an indemnity when an easily observable and independently verifiable “trigger”, usually a particular temperature or rainfall level is struck, suffer from basis risk. Basis risk is when insurance is brought and an economic loss is realized but the indemnity payment is not made. Due to differences in microclimates and quality of information, an individual farmer’s crop yield distribution may not closely correspond to the distribution used for the index. In the case of developing countries, much more empirical research is needed to measure farmer’s risk attitudes and capacity to pay for crop insurance.
Cognitive Failure: Some farmers may perceive the risks they face as being smaller than they actually are. This phenomenon is called “cognitive failure” and can stem from either insufficient information or an inability to properly process and assess information. In common language, it is the feeling of invincibility: “That can’t happen to me”. Also it refers to the common feeling among farmers that “premiums paid are lost money if nothing happens”. Many farmers tend to dismiss low probability but high cost events in their decision-making processes and to just focus on developing risk management strategies for high frequency, low cost events, the “commonplace threats” (Skees, Barnett, and Hartell, 2005).
Inadequate Legal and Regulatory Frameworks: Legal and regulatory frameworks can either help promote or hinder the development of agricultural markets. The most common areas of complaint from insurers and observers concern the following. Inappropriate Reserve Requirements: Often time capital reserve requirements are adequate for life, auto, property, and casualty lines of insurance but not for agricultural insurance due to higher rates of rotation in the portfolio. Many agricultural production cycles are a few months long and if capital has to be reserved for period longer than the actual length of risk exposure, it increases the reserve load in the premium and makes the product unattractive to client. Possible solutions could entail treating agricultural insurance reserves like marine insurance reserves and use of more sophisticated calculations. Agent Licensing Requirements: How to deliver agricultural insurance to smallholders is a big obstacle and one obvious way around this obstacle would be for rural microfinance institutions and rural development NGOs, and cooperatives to serve as agents for insurance companies selling agricultural insurance products. Often the agent licensing provisions are either too strict or too lax. Some country’s insurance laws may require a long number of years, formal training, and other high qualifications, which make it difficult for young microfinance institutions to qualify. Other time, the laws may require that the agent be a natural person, thereby eliminating the possibility for a cooperative or NGO. Traditional individual agents have little incentive to sell agricultural insurance compared to auto and life. The latter two are high margin and imply less administrative costs and time. Possible solutions to help protect consumers against misselling of insurance policies but at the same time facilitate the development of agricultural insurance would be specialized training for financial intermediaries, NGOs, cooperatives etc. in the selling of agricultural insurance and the adoption of market conduct standards subject to compliance checks. Reporting requirements: Lastly, the reporting requirements can place a high burden on an insurance company that wants to specialize in a low-income, high-risk segment of the market. The impetus to the insurance company would be to specialize in more lucrative lines such as auto and life, where the high cost can be more easily borne. Regulators do need information but the practical issue of maintaining computerized databases for a dispersed, low-come clientele is a serious one for crop insurers wanting expand in developing countries. Possible solutions may involve more streamlined and effective reporting for agricultural insurers and encouragement and support for agricultural insurers invest more heavily in wireless technology, if the infrastructure of the country permits. Product Classification: Many times when an insurance company wants to introduce an weather based index there is a legal debate as to whether it is a derivative and therefore subject to the rules governing capital market securities or whether is covers an “insurable loss” and should be subject to the rules governing insurance products. If the crop insurance market is to develop, parametric or index based instruments need to be classified as insurance products and not as derivatives so that easier and more flexible delivery systems can be used to get the product to small holders. The most promising retail delivery channels for parametric products and other insurance products targeting the low-income are indirect ones. Urban-based broker and insurance office outlets will not suffice. Moreover, capital market instruments are aimed at sophisticated and knowledgeable market participants and may be subject to very little regulation or a very direct regulatory regime than insurance products. The capital market regulatory regime may not include sufficiently strict financial reserve requirements nor be subject to market conduct rules equivalent to those that international standards require should apply to the sale of insurance. Thus, small farmers with in regime where parametric products are classified as derivatives will not have the benefit of the regulatory protection that he needs. In short, the farmer client is at serious risk of abuse. If index based risk management products are not recognized as insurance products with an "insurable interest" and the requirement that an insurance policy indemnifies a loss, there is a risk that the framework will not recognize payment against an index.
Distorted Incentives: When governments intervene and make ex post unconditional emergency relief payments, forgive loan contracts, and/or offer subsidized emergency loans, it removes the incentives for farmers to purchase insurance ex ante and for insurance companies to innovate and offer appropriate crop insurance products. The government intervention is often justified on the grounds that private insurance companies are unwilling or unable to supply crop insurance in an efficient manner. This dilemma of “crowding out” or market failure has raged, at least in the U.S. economic literature, for decades. The issue needs to be recast as finding an appropriate facilitation role for the government and distinguishing clearly between disguised income transfers and risk management tools.
Thin International Reinsurance Market: The market for agricultural reinsurance is limited due to the high cost of reinsurance premiums and reluctance on the part of reinsurers to develop a cadre with the necessary specialized knowledge and information systems required to properly monitor and evaluate agricultural risks. Since crop yields are highly spatially correlated, private insurance companies cannot effectively pool risk at the regional or even at a country level, especially if it is a small country. The Maximum Probable Loss and Maximum Foreseeable Loss estimates would exceed capital reserves and thus the insurer needs to cede or transfer a portion of the portfolio risk to an external party, either an international reinsurance company, a national government, or a supranational government agency. International reinsurance companies have the capacity to absorb large insured losses and for years have done so, especially for major natural catastrophes. For instance, in 1992 international reinsures paid out $23 billion to cover insured losses association with Hurricane Andrew and in 2005 will pay out in the order of $60-80 billion for Hurricane Katrina. Agricultural losses due to drought or flood are likely to be less than the cost of a major hurricane or earthquake, but the levels could be appreciable and repeated from year to year. Accordingly, only a few of the international reinsurance companies have agricultural divisions. The combination of lack of analytical capacity and expensive reinsurance premiums at the international level dampens insurers’ capacity to offer crop insurance products at the national level. At present, only four of the more than 60 reinsurers worldwide have substantial agricultural portfolios, Munich Re, Partner Re, Hanover Re and Swiss Re. Two in particular have strong expertise in analyzing weather-based indexes, Swiss RE and ACE, because they hired many former Enron employees. Enron was the pioneer the use of weather index derivatives in energy markets in the early 1990s.
Given this list of formidable problems in the path of sustainable development of agricultural insurance, the question becomes what have we learned from previous experiences with crop insurance, what trends are discernible in Latin American markets, and what should be done to promote market development?
AGRICULTURAL INSURANCE OVERVIEW: MARKET TRENDS, PRODUCT EVOLUTION,
LESSONS LEARNED, AND PROMISE OF NEW TECHNOLOGY
Market Trends
Overall, agricultural insurance is underdeveloped worldwide. In 2001, total agricultural premiums (including fishery and forestry) amounted to US$6.5 billion while the estimated total value of agricultural production worldwide was US$1.4 trillion. Thus, agricultural premiums as a share of output were a miniscule .4% (Schuetz, 2005). Moreover, the regional distribution of coverage is bimodal as can be seen in Table 4. Developed countries account for 87% of the agricultural premiums in that year as opposed for 13% for developing countries. Whereas 75% if the cultivated land in the US is insured, only five Latin American countries have more than 1% of cropped area insured, and only one, Mexico exceeds 10% (See Table 5).
Table 4:
Agricultural Insurance at a Glance
Region of the World |
Share of Agricultural Insurance Premiums 2001 |
Cumulative Share |
North America (U.S. and Canada) |
55 |
55 |
Western Europe |
29 |
84 |
Australia and New Zealand |
3 |
87 |
Latin America and the Caribbean |
4 |
91 |
Asia |
4 |
95 |
Central and Eastern Europe |
3 |
98 |
Africa |
2 |
100 |
Source: Schuetz, 2005 (FAO) |
In the developed countries, the four leading markets are found in the United States, Canada, Spain, and Japan. Within Latin America and the Caribbean, the leading agricultural insurance markets can be found in Mexico, Argentine, Brazil, and Venezuela. In Central America, the region most exposed to the widest number of natural hazards, only ten companies are active and a miniscule proportion of the cultivated area. In many countries, there is no commercial agricultural insurance available: Bolivia, Suriname, Guyana, Belize, Bahamas, Jamaica, Barbados, and Trinidad and Tobago. In many of the countries, however, active efforts are being made to promote agricultural insurance.
Products: Strengths and Weaknesses
There are four broad types of products that are offered in agricultural insurance markets: single peril, multiple peril, revenue and parametric. Historically, the first type of crop insurance to be offered was single peril for hail in Europe and North America in the 19th century. In the developing world, there were some early adopters of single peril and mutual insurance products--Uruguay (1914), Mexico (1926) and Mauritius (1945).
In the 1930s, the U.S. government started to experiment with multiple peril policies as a means to help farmers recover from the devastating effects of the Great Depression and the Dust Bowl (a prolonged drought that affected the Plain states). After WWII, the use of this product was introduced in Western Europe and Japan. Later on it spread to African, South Asian, and Latin American countries. Some of the developing country pioneers of multiple peril programs were Brazil (1954), Costa Rica (1970), Mexico (1971), India (1972), Chile (1980), Dominican Republic (1984), and Venezuela (1984).
By the late 1980s and early 1990s, however, most of the multiple peril programs in developing countries where experiencing substantial losses. With the advent of structural adjustment programs and the general reduction in public subsidies available for the agricultural sector, reforms and retrenchments occurred.
In the 1990’s, the U.S. started to experiment on a large scale with area yield, crop revenue, and income insurance products. In the developing world, India has been the leader in introducing area yield products. In most other developing countries, single and multiple peril products continue to predominate. In the late 1990s and early part of this millennium, Spain, Mexico, India, and Mongolia have either introduced parametric or indexed based products or have products under design. Canada has an area-yield product since the 1970s and India and Morocco have had such products since the 1990s.
Table 5:
Principal Characteristics of Crop Insurance Markets in Latin America and the Caribbean
Country |
Types of System |
Principal Crops Insured and Risks Covered |
Typical Contract |
Percent of Area Cultivated that is Insured |
Public Subsidies | ||
Argentina |
Private |
Grains, fruits, and livestock
Hail, fire, frost, high winds, excessive rain, pest infestations, plant diseases, replanting, livestock death due to disease or accidents |
Pay 60-90% of difference between actual and historic yield |
1% in 2002 |
No (Subsidy plan has been proposed in late 2003 and since 2004 a pilot has been operating in State of Sao Paulo) | ||
Brazil |
Private |
Most crops and livestock
Climate, pests, disease, livestock death due to sickness or accidents |
Contract 1: Cover cost of production; Contract 2: Cover the difference between expected and actual |
.22% in 2000/01 |
No (Pilots with state subsidies where executed in Rio Grande do Sul (2000) and Sao Paolo (2003)) | ||
Chile |
Mixed (Public-Private) |
Grains, pulses, vegetables, industrial crops
Drought, excessive rain, freezes, hail, snow, and high winds |
Difference between insured valor and actual yield |
2% in 2002 |
Yes | ||
Colombia |
Mixed |
Banana, cotton, potato, sugar cane
Drought, flood, excessive moisture, hail, high winds. |
Covers project investment costs, including use of borrowed funds |
Less than 1% |
Yes. Government pays between 10-45% of premium depending on crop | ||
Costa Rica |
Public Monopoly (Will liberalize by 2008-09 if CAFTA treaty is ratified) |
Crops and cattle. Note: 90% of coverage is for rice, a “strategic” good.
Uncontrollable climatic risk and death of cattle caused by accident or a few specified diseases |
Cover total or partial losses produced by the risks stated in the policy |
2% in 2000 |
Yes | ||
Dominican Republic |
Public |
Rice All climatic risks and plant infestations and diseases |
Covers losses |
Less than 1% |
Yes | ||
Ecuador |
Private |
Banana, cotton, potato, sugar cane, cattle, horses Hail, drought, freezes, excessive moisture, high winds, pests, diseases, livestock death due to sickness, accident, or forced sacrifice |
Cover losses realized. |
Less than 1% |
No | ||
Mexico |
Climate, pests, diseases, livestock death |
Most crops and most types of livestock |
Many varieties of contracts offered (some cover production cost, other yield loss, revenue loss) |
15% in 2002 |
Yes | ||
Panama |
Public |
Rice, corn, vegetables, cattle, horses, Swine, and farm machinery/buildings Climate, pests, disease, livestock disease |
Covers production costs or market value of equipment. |
Less than 1% |
Yes (administrative costs of state provider | ||
Paraguay |
Private |
Wheat, soybean, corn
Hail and fire |
Covers actual value of actual losses |
.1% en 2001 |
No | ||
Venezuela |
Private |
Covers the difference between average and realized yields. |
4% |
Yes | |||
Source: ENESA, 2004
As can be seen from Table 5, the most common type of product offered in Latin America is a multiple peril product that covers natural and biological hazards. In evaluating insuring products, five variables are normally used: commodity coverage (number of products that can be insured); penetration ratio (amount of acreage insured as a share of total cultivated area); participation rate (number of farmers purchasing insurance); loss ratio (indemnities/premiums); and long-term viability (Indemnity Payments +Administrative Cost + Reinsurance Cost+ Capital Reserve Load +Profit)/ Premium <1). The financial performance of single peril products (loss ratio) has been impressive but coverage and farmer participation has limited. In comparison, the financial performance (loss ratio and viability) of multiple peril products has been universally unsatisfactory despite massive public sector subsidies for premiums, operational expenses, and reinsurance. Parametric or index based products, on the other hand, despite noteworthy theoretical features has not been widely attempted and in the few places that it had introduced, acceptance has been limited. It is still a novel product that is seen as plagued with basis risk. The biggest practical success has been the program run by ICICI Lombard-BASIX in India. The program is still very young, three years old, and no rigorous and independent evaluation has taken place. In other areas of the developing world, several pilots are in various stages of development.
Examples of Parametric or Index Insurance
Index products use any independent random variable measurement that is readily observable, protected from tampering, and is highly correlated with agricultural or livestock losses. Four examples are:
1. Weather based index uses a specific amount of rainfall or a certain number of days with temperatures in a particular range as a trigger. If the trigger is struck a payment is made. In use in Morocco, Mexico, and India.
2. Area yield Index. Uses the average crop yield in a country or particular jurisdiction as a trigger. If an individual farmer has a yield less than the reference average, an indemnity payment is made as a function of the degree of deviation from the norm. In use in the US., India, Brazil, and Quebec, Canada.
3. Satellite Vegetative Index: Satellite images are used to calculate the health of a pasture based on “previous” normal years and payment is made to the rancher based on degree of deviation. In use in Alberta, Canada and Spain.
4. Mortality Rates for Livestock: A yearly census of livestock is used as a reference point to estimate “annual average death rates” from yearly censuses comparing end of year to mid year points. The trigger will be a certain pre-fixed percentage of average mortality. When death rates surpass the “trigger” payments will be made. Under design in Mongolia.
POLICY RECOMMENDATIONS FOR SUSTAINABLE MARKET DEVELOPMENT
The development of agricultural insurance markets as can be seen in the previous exposition is saddled by inertia, lack of actuarial and agriculture specific knowledge, lack of information, and in some cases weak legal and regulatory frameworks. One of the key debates in the promotion of agricultural insurance revolves around the question of the role of public subsidies. This section discusses the pros and cons of subsidies and sets theoretical and practical guidelines for the use of subsidies and delineates the short and medium steps that national donors, donor organizations, insurance companies and farmers should follow to develop a sustainable, broad-based and competitive agricultural insurance markets.
The Role and Allocation of Public Subsidies
Many industry representatives vehemently claim that agricultural insurance can only be developed with government subsidies for premiums and operational expenses. The traditional argument for government subsidies, especially for premium payments is based on three elements. First, farmers have repeatedly expressed more of an interest in purchasing multiple peril and revenue/income insurance, than single peril or parametric instruments, which are still a novelty. Thus, if a large market is to be developed, as measured by insured area cultivated, value of policies, and number of participants, then the desired types of insurance products must be offered. Second, these two types of products are so costly to deliver to the majority of farmers that government subsidies are required otherwise the charged premiums would be unaffordable and no market or a very small market comprised only of the largest farmers will appear. Third, agrifood sector is too politically sensitive to consign to this quandary of suboptimal risk management due to “market imperfection”. Therefore, on political economy and social benefit grounds, the intervention of the state can be justified to enhance national welfare, otherwise society will suffer the negative effects of fluctuating agricultural supplies and prices, the economic and social dislocations associated with the distressed family families and communities, as well as the probable loss of international competitiveness when foreign producers have access to insurance. According to the proponents of this argument, the three roles that the state can play a role is (1) providing a subsidy directly to farmers to help play for the premium; (2) providing a subsidy to help private insurance companies obtain reinsurance, and (3) providing direct subsidies to private insurance companies in order to help defray the high costs of serving numerous, dispersed, small-scale rural clients.
This paper argues that insurance market development does indeed require government assistance and investment but that public subsidies should be more orientated to providing “public goods” that will help lay the foundation for private risk taking and less on providing “income transfers” as occurs when the emphasis is on providing subsidies in the premium in order to make the product more affordable to a larger set of farmers. In the context of fiscally constrained treasuries, public expenditures should be rationalized and allocated to the ends that would generate the highest economic return. Moreover, developing countries have a much larger farm population than developed countries, making the implementation of a comprehensive and broad-based program extremely expensive and failing universal reach, prone to politicization. For example, the farm population (owners/operators) of the U.S. is 2 million out of a total population of 295 million (.6%) whereas the farm population (owners/operators) of Bolivia is 660,000 out of total population of 8 million (8%). The US spends approximately $3 billion on crop insurance subsidies per year or approximately $1500 per each farmer to get 72% of cultivated area covered. Using extrapolation, Bolivia would have to spend $990 million to get significant coverage. Bolivia, however, is a much smaller economy than the U.S. $22.33 billion (PPP est. 2004), vs. 11.75 trillion (PPP est. 2004), meaning that Bolivia would be spending approximately 4% of its GDP on crop insurance subsidies annually, whereas the US is spending .0002% (CIA World Factbook). The question becomes one of cost effectiveness. Given all of the pressing investments needs in such a low-income country as Bolivia, would spending scarce public monies on crop insurance yield a higher rate than on other activities such as rural education, rural roads, potable water, electrification, agricultural research, extension and marketing services?
Favored Uses of Public Subsidies to Promote Agricultural Insurance:
Public subsidies and monies can and should be used to the following activities:
1. Market Development
a. creation and maintenance of information databases including purchase of data and images
b. promotion of innovative products and their pilot testing
c. training of staff in insurance companies (actuarial sciences, risk modeling, claim adjustments, analysis and familiarization with agricultural commodity markets)
d. measurement of farmer’s risk attitudes so that coefficients of risk aversion can be used to setting pricing
e. education of farmers on insurance products and contracts
f. education of government policy makers on how to create favorable settings for the introduction of insurance products.
g. purchase, installation, and maintenance of weather stations
h. payment on a permanent or declining basis the cost of broadband internet and satellite connections the entity consolidating the information and transferring it to private users.
i. development of sophisticated models of weather phenomena
j. acquisition of computers with greater and faster processing capacity.
k. acquisition of data storage devices.
l. modeling impacts of climate change on precipitation patterns and extreme weather events
2. Legal and Regulatory Reform
a. Hiring legal and economic consultants to determine if biases against agricultural insurance products exist in the regulatory framework and to determine the most logical way to remove them without jeopardizing the soundness and solvency of the entire insurance industry.
b. Hiring legal and financial sector consultants to help draft laws and regulations
c. Educate and train staff responsible for the supervision of the insurance industry
d. Help improve contract enforcement and consumer protection mechanisms
3. Reinsurance
a. Provide assistance to the local insurance industry in attracting private reinsurers interested in agricultural portfolios
b. Provide subsidies to private insurance companies to purchase reinsurance or facilitate in bundling contracts
c. Act as a co-reinsurer or direct reinsurer as a last resort
4. Disaster Relief Funds
a. Design and financing of disaster preparedness and emergency relief funds aimed at dealing with the consequences of low probability but catastrophic events.
b. Use innovative index insurance and bond instruments to transfer catastrophic the risk to international markets
Unfavored Uses of Public Subsidies to Promote Crop Insurance
In practice, special interest groups mobilize and lobby for a subsidy that benefits them but rarely is it known if the impact would be broader if the scarce public monies were spent on public versus private goods or one particular industry/sector versus another. Recent work by Ramon Lopez suggests that countries that dedicated a greater share of rural public expenditures dedicated to public goods rather than private goods, scored better on variables of policy concern, i.e. growth in agricultural output, reduction in rural poverty, and increase in rural employment (Lopez, 2004). A public subsidy toward the payment of a crop insurance premium is more of a private good than an expenditure on the eradication of an animal disease or investment in a rural university with research, teaching, and extension mandates. In the case of a catastrophe, where the welfare of many has been reduced and it has serious negative externalities, transfers to private individuals are warranted.
In general public subsidies should not be used for premium because of a host of implementation problems—identification, sustainability, development effectiveness, efficient targeting, and adverse selection.
Theoretically, a case could be made to subsidize only the operational and administrative cost elements in the premium and not the pure risk element, otherwise investment decisions could be skewed to favor riskier crops than would be the case without the subsidy. It would be possible to raise a farm household out of poverty by providing subsidized insurance (Expected Utility with insurance could be concave curve greater than Expected Utility without insurance over a range). The beneficiary household could be placed on a higher income growth trajectory that would get it out of poverty and keep it out of poverty over time.
In practice, however, implementation problems loom large. Therefore, it is advisable le to apply scarce public subsidies to market development efforts rather than to private participation incentives. The implementation problems are listed below.
First, it is difficult to distinguish between what is the risk element and the other non-risk element in the premium—the identification problem. See the significant differences reported in risk premiums in section 4 using different estimation methodologies. In practice, the subsidy is applied as a percentage of the total premium cost. To be theoretically consistent, one would have to calculate the “pure risk premium” for each policy or product and try to be certain that the correct yield distribution is being used. The additional effort and costs required to correctly “identify” the acceptable subsidy would be costly and even counterproductive.
Second, even if the identification problem could be solved, the non-risk elements could easily dwarf the pure risk premium and still signify large fiscal outlays. Overtime, the central government may find it difficult to continue making these outlays and the market will tend to grow only as a function of subsidy availability (sustainability). In practice, one sees start and stop patterns in the data and in several cases it can be traced back to the availability of subsidies, i.e. BANDESA in Honduras and Banco Agrícola-AGODOSA in the Dominican Republic.
Third, if the government is interested in pro-poor growth and reducing inequality as rapidly as possible, then a cost-benefit analysis may be warranted to determine how best to allocate scarce public monies. Government authorities should empirically decide whether if subsidizing crop insurance would generate higher rates of return than would the subsidizing other productive support services or investing in infrastructure such as rural roads, rural education, health services, potable water, sanitation, electrification, vocational training, agricultural research, or agricultural extension services (development effectiveness).
Fourth, government subsidy programs tend to be captured by the higher income persons unless mean testing or targeting is used. Historically, the subsidy on insurance premiums has tended to be captured by larger farmers in both the developed and developing world (Skees, 2002 and 2005; Goodwin, 2001; Makki, 2001; and Hazell, 1986 and 1992). Differential subsidies can be paid, wherein small can benefit more but this would imply additional administrative costs to identify and verify who is a small farmer. Thus a problem of efficient targeting problem may arise (efficient targeting).
Fifth and lastly, subsidies in practice tend to be a palliative for the problem of adverse selection. Insurance companies depend on premium subsidies to grow markets and overcome adverse selection. Without the subsidy, the lower-risk candidates would opt out of the market (Goodwin, 2001; 1993; Makki; 2001; and Just et. al., 1999).
In conclusion, it would be advisable to focus scarce public monies on developing the conditions favorable for the emergence of insurance markets and the development of low cost insurance products such as weather and area-yield indexes where the premium would be affordable without the need for a subsidy.
Recommendations
Fiscally stressed, low-income countries would do best on developing an integrated and layered risk management system and not view crop insurance as a panacea. Crop insurance is but one element in an arsenal of instruments and policies that a government may need to rely on in order to spur agricultural growth, improve agricultural competitiveness, and reduce farm poverty.
Adopt an Integrated Risk Management Strategy
A “layered risk management strategy” should be followed wherein a series of coordinated and reinforcing activities are pursued. The government, international reinsurance companies, national insurance companies, insurance supervisors, and farmers have to work together.
The cornerstone of the strategy is an effective and improved agricultural extension service that helps farmers educate themselves about risk management and to take individual on-farm actions to reduce vulnerability and mitigate risks.
1. Improved Extension Services
a. Use improved seed and animal breeds
b. Site plots away from flood plains, areas susceptible to land and mudslides, and avoid planting on easily erodible soils.
c. Planting on time
d. Use soil and water conservation aimed at improving fertility, increasing soil moisture, reducing soil borne diseases, maintaining soil structure, and reducing water run-off
e. Use Integrated Pest Management whenever feasible
f. Make environmentally sound use of pesticides and herbicides
g. Apply fertilizer at the right time and the right amount
h. Make greater use of drip irrigation
i. Engage in crop diversification and rotational cropping
j. Use good animal husbandry practices
k. Use intercropping
l. Adhere to animal and plant health inspection and control protocols
m. Use of custom hire labor and equipment for timely harvest and post-harvest responsibilities
Once the farmer has adopted as many of these good management practices as possible, the farmer should try to avail him/herself to formal financial instruments and marketing contracts that are within reach in order to further reduce risk exposure.
2. Financial Market Improvements
a. Use of formal savings
b. Use of futures, options, sales contracts, guaranteed marketing schemes to control price risk
c. Use of single peril insurance for idiosyncratic risks
d. Use of area-yield and weather index insurance for non-catastrophic, covariate climatic risks
e. Use of multiple peril, income, and revenue insurance products to stabilize income
f. Use of reinsurance to lower premium cost and transfer risk outside the country to parties willing to bear it.
For risk that is not covered by these financial and market contracting instruments, such as severe earthquakes, massive volcanic eruptions, hurricanes of category 3 or greater intensity, massive flooding etc, the farmer will need to depend on government emergency assistance.
3. Improved Emergency Disaster Relief Systems.
a. Non-reimbursable cash payments for immediate survival needs
b. Temporary housing and allowances for relocation if necessary
c. Distribution of in-kind materials
d. Refinancing of existing loans
e. Emergency low-interest loans for rebuilding and farm recovery
The government, however, should condition the level of assistance on demonstrated prudence and diligence--the adoption of good management practices, avoidance of excessive risk, and the use of formal financial instruments whenever feasible prior to the event.
Develop Crop Insurance Markets Sequentially
Interested parties should focus on the following sequential activities.
First, review of the legal and regulatory framework with particular attention to the following common impediments:
(i) Barriers to entry. Foreign companies should be free to enter as long as they are solvent and have a track record. Exclusion of foreign companies and reinsurers impedes the diffusion of technical know-how as well as price competition.
(ii) Reserve policies: The amounts that have to set aside for current risks should be consistent with the duration of the agricultural insurance policy and the loadings should not be for a catastrophic level.
(iii) Agents: Few restrictions as to who can be an insurance agent capable of selling insurance should apply. i.e. only an individual with specific training. Greater flexibility should be allowed for legal entities such as NGOs, cooperatives, farmer and community associations, and credit granting institutions to affiliate with a recognized insurance and serve as an indirect delivery platform. In the developing rural areas, indirect delivery mechanisms may be preferable to direct mechanisms in order to lower the fixed costs of establishing an extensive branching system.
(iv) Recognition of Parametric Instruments as Insurance Products and not as Derivatives: If parametric instruments are governed by capital market security frameworks, parametric instruments may never be widely adopted by farmers. It would be an instrument used primarily by corporations.
After a thorough review, steps should be taken to remove biases and to educate insurance market supervisors about recent developments in agricultural insurance. Because agricultural insurance is such a tiny fraction of the total policies, many supervisors are not knowledgeable about this specialty market.
Second, construct an information depositary that is easily accessible to insurance companies. The depositary would capture, transform, and clean data relevant to the design and monitoring of insurance products and policies. Information is vital to the measurement and evaluation of risk and is the bedrock of insurance underwriting. To design and price an agricultural insurance product a slew of different type of information is needed.
Information systems that capture most, if not all, the above information and organize, clean, standardized, and made it readily available to the general public would greatly reduce the cost that each insurance company must bear at present if it wants to enter into agricultural insurance. By having all the information consolidated in one place and to be confidently cleaned, and transformed, and standardized, would represent a tremendous savings to insurance companies and other interested parties.
Starting in the 1990s, public information systems began to appear in Latin America largely financed by international technical cooperation with the express purpose of facilitating better land use planning and land title regularization. These entities capture data from public as well as private sources of information. Once the information is cleaned and transformed it can be placed as layers in geo-referenced data facilitating spatial analysis of the data. More importantly, a wide number of different actors can use the same information for different ends. For example, banks can use the database to check title registrations and existence of liens for loan applicants. A hotel resort developer can search for the ideal location taking into account road access, location of major population centers, weather patterns, and scenic attractions. An insurance company can use the topological, crop use, and weather data and agronomic model to determine exposure to risks and calculate probable losses, i.e. a risk map.
In Central America, there are four public information systems. The most advanced being Sistema Nacional de Información Terretorial (SINIT) in Honduras which by law receives data from a number of public agencies and then complements it with information from private sources such as Tela Rail Company and Standard Fruit to maintain a GIS database with over 3,000 layers. Via the web anyone can access the information and download it. The principal users are other government agencies. Other countries in Latin America should emulate this example.
Third, finance pilot experiments and support training for staff in the insurance industry. Agricultural insurance requires specialization and since it is an underdeveloped market segment, there are few people in developing countries who dominate all the intricacies. The pilot experiments should strive to develop products that are low-cost to administer, actuarially fair, easy to understand, and attractive to clients. Client participation in the design should be high.
Fourth, monitor and evaluate the pilots rigorous and make necessary changes.
Fourth, scale-up the experiences that are documented to be successful.
Fifth, educate farmers and policy makers about the limits and benefits of insurance.
Sixth, invest in and maintain the infrastructure and recurrent activities necessary to support information flows, i.e. weather stations, household surveys, marketing information (volume, price, grade transacted in different places), agricultural censuses, broadband internet infrastructure, aerial photo surveys, obtaining satellite images documenting land use and monitoring in real time of weather systems, improved internet connectivity and affordability, and construction of relational geographic databases.
The systems are in Honduras, El Salvador, Nicaragua, and Guatemala and all grew of international cooperation projects to improve land use planning and title registration. Honduras is fully operational and open to the public. El Salvador and Guatemala should be on–line and fully functional within a few months.
Bad Practices to Avoid
Mandatory Insurance: Governments interested in promoting agricultural insurance policies should shy away from making agricultural insurance mandatory to access credit or guarantee loan programs. The insurance product, accordingly, converts into being “bank insurance” and do not force lenders and insurance companies to adequately measure and evaluate risks. Invariably, the private insurance companies demand backing from government to participate in such schemes or only publicly owned insurance companies would participate. At the end of the day, the government is using a roundabout scheme to guarantee a loan. The challenge is to reduce asymmetric information problems, reduce transaction costs, and increase profitability at the farm level.
Premiums Need to Adjust Frequently as New Information Becomes Available: Government run insurance programs tend to be reluctant to adjust premiums from year to year based on loss ratios and new available information out of political sensitivity. Policyholders tend to complain vocally about high premiums. The failure to adjust premium rates undermines program viability.
Set Insurance Sale Dates Well in Advance of Harvest Time and Honor Them: Selling insurance policies after closing dates, invites opportunistic behavior. More reliable and accurate information on weather becomes available the closer one gets to planting time. Farmers will use this information to their advantage. They will not buy if the weather forecasts are good leading to low income for the insurance company or they will pay buy if the forecasts are bad, leading to heavy losses for the insurance company.
Extending Insurance Coverage to All Regions and Most Crops: For political reasons, many government backed comprehensive agricultural insurance programs seek to extent “affordable” insurance to high-risk areas and to accept all clients regardless of management skills, character, and risk profile. This violates one of the Golden Rules—all risks are not insurable. Either the premiums should be set sufficiently high, active client screening engaged in, or the insurance company should withdraw.
Using Third Party Claim Adjusters: The monitoring and claim inspections cannot be delegated to third parties or strategic allies. The staff of the insurance company underwriting the policy should make the physical inspection to verify and make claim adjustments. To use third parties creates the risk of collusion and fraudulent manipulation between the policyholder and the third party.
CONCLUSION
Agricultural insurance is a complex and difficult product to deliver in a sustainable manner. In the region, the agricultural insurance market is nascent but there are encouraging signs. More and more policymakers and farmers recognize the need for more modern risk management systems in order to stabilize incomes, prevent asset depletion, and to enhance competitiveness. Traditional risk management systems sometimes are not sufficiently robust to deal with the vagaries of weather and disease and as a result these uncontrollable events cause significant economic losses that negatively affect households, communities, and government themselves.
Nonetheless, yield insurance must be kept in perspective. It should not be seen as a substitute for unprofitable farms, failures of farm management, changes in technology innovations, market access, disaster aid, or government policies that suddenly affect the rate of return. Neither should the provision of insurance be seen as a sufficient condition in order to improve agricultural competitiveness. If other necessary investments are not made in rural infrastructure, market information systems, and production support services, competitiveness will not improve.
Insurance, however, can be beneficial in improving access to credit by serving as a guarantee against involuntary default. On the other hand, insurance policies should not be made a mandatory condition to access credit because invariably such a dictum undermines both the bank’s and insurance company’s capacity to evaluate creditworthiness, measure risk, and assess farmer management capacity. Some farmers may have adequate on-farm risk management strategies and will be forced to bear additional financial costs in order to access credit. Many others will have no incentives to engage in on-farm risk management activities and will increase the loss probabilities of the insurance company. Markets that evolve spontaneously and are based on solid fundamentals tend to be deeper and more efficient in the long-run, i.e. Mexico since 1990 vs. state banks in Honduras and the Dominican Republic that depend on mandatory purchase of insurance in order to access credit.
Last but not least, developing countries should not convert crop insurance into an entitlement or disguised income transfer tool. Many do not have the economic wherewithal and it would be more advisable to keep insurance as a risk management tool.
In developing agricultural insurance markets, the role of governments is crucial. An action agenda was laid out—adjusting legal and regulatory frameworks, if necessary; developing public information depositaries easily accessible by insurance companies and others; training staff; educating farmers, policymakers, and superintendents; conducting pilot experiments; scaling-up activities; designing catastrophic disaster relief programs that do not undermine incentives to undertake on-farm risk management activities and/or to purchase formal insurance---that could serve as a model for operations. Moreover, it was argued that in this model, all public money should be spend on creating public goods and sustaining favorable conditions and not necessarily on subsidizing the insurance premium. The principal reasons for this allocation are based on efficiency and sustainability. Several different types of insurance products were reviewed—single peril, multiple peril, parametric, and revenue--- and their respective strengths and weaknesses were noted. Regardless of the product, the guiding criteria for design and implementation of products should be based on achieving the lowest administrative cost possible, pricing for actuarial soundness, fostering transparency, and maintaining affordability. A tradeoff, however, does seem to exist between actuarially fair crop insurance schemes and the limited financial means of farmers in developing countries. Farmers prefer individual, multiple risk coverage but an actuarially fair premium would be unaffordable for most. Parametric products (indexes based on area yield averages or weather triggers) are less costly but imply basis risk and would be attractive to less risk adverse farmers. The historical record for writing multiple peril products is generally unsatisfactory and great caution should be exercised in expanding these products unless historical data exist that would permit reliable loss estimations and actuarially sound premiums are charged (UNCTAD, 1995; Hazell, 1992, Just, et. al, 1999). Greater emphasis and government support should be given immediately to developing information systems, modeling yield losses, quantifying degrees of risk aversion, determining better fits between individual losses and aggregate triggers so that less costly insurance schemes can be introduced that are attractive and of interest to low-income farmers. In the short- to medium-term, more attention should be paid to promoting better on-farm risk reducing and risk coping strategies through better extension services and the use of single peril and parametric products. Much work remains to be done to further develop agricultural insurance in Latin America and the Caribbean. Ripe areas for research and pilots include eliciting farmer risk attitudes, blending crop insurance with other financial products; using modern information technology to reduce costs; better modeling and understanding of weather phenomena and the impact of climate change; and improving reinsurance capabilities.