The Policy Research Working Paper 7720
Solving commitment problems in disaster risk finance
D.J. Clarke and L. Wren-Lewis
The initial premise of this paper is that, in the event of a natural catastrophe, the population does not receive the aid it needs. Three causes have been indentified: the first is that households tend to be under-protected in the face of natural disasters –either when taking out insurance or when putting material resources into place for mitigating the consequences– because they are confident that, if a catastrophe occurs, they will receive assistance from governments and private institutions (
the Samaritan dilemma). Moreover, the reality of such situations shows that, often, the support arrives late and, due to political corruption and illegitimate interests of the private institutions donating the aid, the relief is misallocated and not distributed in proportion to the damage suffered, reaching the point where some people receive assistance who have not undergone any loss at all.
It is assumed that these problems do not arise –or do so to a lesser degree– in developed countries where the “reputation” of the donors –governments or private institutions– precludes failure to fulfil promises and where there is a separation of powers making it technically impossible to avoid meeting commitments.
According to the authors, the root of the three problems lies in the incapacity of the donors to put their prior commitments into practice, and they propose transferring the risk to third parties as the solution, either through insurance or by using risk transfer financial instruments, such as catastrophe bonds or swaps. The paper also makes a detailed analysis of how and when each kind of instrument should be used in order to prevent over-selling, as if it were a miracle cure against disasters.
The authors support themselves on a model based on the behaviour of
recipients (those affected by the catastrophe) and
benefactors (the donors of the relief) in order to demonstrate the following propositions:
Proposition 1. When people perceive that, in the event of a catastrophe, they will receive aid from an institution, the
recipient will not invest in self-protection, even when it would be clearly beneficial to do so. This will lead to insufficient relief, except in those cases where the
benefactor is absolutely altruistic.
Proposition 2. Benefactors are more willing to invest money in preventing the inadequate allocation of aid before a disaster strikes, for example, by establishing rigorous auditing mechanisms capable of preventing corruption at any point in the chain of distribution of the relief.
For this reason, it is more effective if donors commit to providing aid prior to a catastrophe than after it has occurred.
Proposition 3.
Benefactors take longer in disbursing aid if the number of donors is sufficiently large.
One way of avoiding this tendency would be for each
benefactor to unilaterally commit to providing the relief as soon as possible, regardless of what the rest may do.
The final conclusion reached is that a natural disaster relief system based on the generosity of individuals or institutions is inefficient in practice.
The authors of the paper have examined a number of measures that relief donors could adopt to improve the efficiency of the system:
1. Transfer the risk to third parties, either by subsidising the insuring of the recipient or by taking out insurance themselves instead of handing over post-disaster cash.
Making insurance by the recipient compulsory –the laws of some countries establish this obligation– is an incentive for recipients to take measures for mitigating losses (thereby preventing the Samaritan dilemma), since the insurance premium is established in relation to the risk. In addition, the insuring of the benefactor eliminates the possible increase in the marginal cost of seeking post-disaster financing.
Moreover, involving the insurance sector would help to solve the problem of corruption in the allocation of the relief, because the sector would be the first to be concerned with ensuring that the compensation paid is received solely by those who have suffered losses. Likewise, this would drive donors to provide truthful information on the losses.
Finally, the payments would arrive on schedule without delays due to the re-negotiation of the aid by donors after the loss has occurred.
2. Integrate the information systems on the payments made. Adopting common systems could prevent the lack of commitment on the part of some donors to collect or disclose accurate information on losses from rendering the relief system inefficient. Once again, the insurance companies would be the first to have an interest in seeing the system operate effectively, in order to maintain the confidence of their insured.
3. Use “disaster indices”, which are basically statistics built on the basis of objective data (for example, data supplied by weather satellites or seismic stations), correlating them with the loss data. This eliminates, in part, the problem of the misallocation of the relief due to the corruption of intermediaries, since an index is less vulnerable to manipulation than mere data on losses, as it is supported on a technology which is more difficult to falsify.
