Image Credit: iStock
Index-based insurance for crops and livestock could be an effective means of protecting farm families in poor countries from income losses due to natural disasters. Numerous pilot programs have shown great promise but often fail to scale up commercially. Basis risk—the chance that farmers with insurance will suffer losses but not receive any compensation—has been identified as a major culprit for this lack of uptake. In a new paper in press at the Journal of Development Economics, Erik Lichtenberg of the University of Maryland and Eva Iglesias Martínez of the Universidad Politécnica de Madrid argue that the true reason for the lack of uptake might be that farmers aren’t offered the kinds of insurance policies they really want.
Background: Natural Disasters and Developing Country Agriculture
The livelihoods of farm families in poor countries are always at risk from natural disasters. Their crops may be ruined and the livestock may die from drought, floods, diseases, insect and parasite infestations, and the like. Unlike farmers in the US and other developed countries, they have few means of defending themselves against the perils of nature. They can’t put savings into banks or other secure financial institutions because those institutions don’t exist in rural areas. They can’t borrow during bad seasons to tide them over until good seasons for the same reason. They can’t send family members to work in industry or commerce because there aren’t factories and stores in rural areas. They can’t start nonfarm businesses because they’d have no customers.
All they can do is stick to tried and true crop and livestock raising practices that long experience has taught them to know as the most resilient. That inability to protect themselves against risk has consequences for their ability to increase their incomes: They don’t plant new crops that are higher value or new crop varieties that are higher-yielding because those crops and varieties could be more at risk of failure. They don’t invest in new farming practices like using fertilizers, again because doing so is risky—if weather conditions are bad, they’ll have paid for fertilizers without getting any return on that investment, which means they’re poorer than they would have been otherwise.
Problems and Pitfalls of Insurance
Insurance could be an answer to the problem of providing protection against natural disasters financially. But insuring farmers against crop and livestock losses is hard. Insurers want to be sure that farmers do their best to protect their crops and livestock from natural disasters and to salvage as much as they can when natural disasters do strike, even though farmers have incentives to do otherwise, since they’ll get lower insurance payouts in returns for their expenditure of labor and resources (a problem known in the trade as moral hazard). And insurers want to be sure that they’re not just insuring those most at risk, which would make providing insurance very costly (a problem known in the trade as adverse selection).
Index-based insurance solves these problems by tying insurance payouts to an index like observed weather conditions (rainfall and temperatures), average crop yields in an area, or crop and pasture biomass inferred from satellite imagery, all of which are independent of an individual farmer’s actions and are thus not prone to moral hazard and adverse selection. Government-sponsored programs in the US, Canada, and Spain all offer this kind of insurance. The World Bank, the World Food Program, and development aid agencies around the world have been convinced that index-based insurance is a good means of protecting farm families in developing countries against natural disasters: They’ve funded pilot programs with more than a million contracts and insured values in the hundreds of millions of dollars. But with few exceptions, these pilot programs have failed to scale up successfully; most of those that don’t fail outright remain viable only because premiums are heavily subsidized.
Economists who study index insurance place a great deal of blame for these scale-up failures on the phenomenon known as basis risk. Index insurance is insulated from moral hazard and adverse selection because the index is not subject to manipulation by the actions of individual farmers. That independence also means that the index will not be perfectly correlated with losses actually experienced by individual farmers, so there will be occasions when farmers suffer losses but don’t receive any payout. (There will also be occasions when farmers don’t suffer losses and receive a payout anyway, but those count for less than the occasions where losses aren’t compensated for.) Studies of index insurance programs around the world indicate that the chances of suffering losses without insurance payouts tend to be in the range of 16-36%. That’s quite large!
Basis risk—especially downside basis risk—is an unavoidable feature of index-based insurance. And some economists have argued that it may be an insurmountable obstacle to successful commercialization of index insurance. Other economists have disagreed; they argue that the cost of purchasing insurance is the most important obstacle to successful commercialization.
What We Do and What We Find
Our article is theoretical, aimed at providing proof of concept. We take a closer look at the problem of basis risk in index insurance by modeling the choices that a risk-averse individual farmer would make. We compare the likelihood of purchasing insurance under two scenarios: (1) where farmers can choose the level of coverage (i.e., the share of losses to be reimbursed for) only and (2) where farmers can choose both the level of coverage and the level of the index that triggers reimbursement. We show that if farmers can choose the trigger level of the index as well as coverage, they will structure policies to protect against catastrophic risk—even if doing so means greater downside basis risk. As a result, they’re more likely to purchase insurance when they can choose the trigger level of the index as well as the coverage level. We show further that this difference in the likelihood of uptake is especially great when insurance is provided at commercially viable rates, i.e., when premiums cover the cost of administering policies as well as payouts to cover losses.
We then use a numerical simulation model to look at whether the ability to choose a trigger level of the index in addition to coverage makes a large enough difference to be worth looking at more closely in practice. The simulation model uses a range of parameter values characteristic of developing country agriculture. The simulation model shows that, when free to choose, the best choice of risk-averse farmers is a trigger level of the index substantially lower than the standard level offered. The coverage level chosen is substantially higher, as is downside basis risk. Insurance premiums are lower, substantiating our finding that the best policy involves trading off lower premiums against higher downside basis risk. And the likelihood of purchasing insurance is substantially higher when insurance is offered at rates that are commercially viable in the sense of covering the costs of administering policies.
Overall, our analysis adds weight to the argument that cost rather than basis risk is the main obstacle to index insurance uptake. Policies that they protect against catastrophic risk are cheaper because they pay out less often. Thus, structuring policies so that they provide protection against catastrophic risk involves trading off lower premium costs (which are incurred whenever insurance is purchased) against greater basis risk.
In practical terms, our analysis suggests that scale-up failures of index insurance policies may well be due to offering farmers the kinds of policies they don’t want to buy because those policies offer them too much protection at too high a cost. Field trials that offer farmers the opportunity to choose a level of the index that triggers payouts in addition to coverage levels—as is common in index insurance programs in the US, Canada, and Spain, for instance—can provide a practical test of whether that hypothesis is true. Conducting such field trials is thus a logical next step.
