Blog
Risk, Poverty, and Safety Nets

by Stefan Dercon

Poor people in developing countries must cope with droughts, floods, illness, recession, and political instability. Much of their energy goes into coping with these shocks and into day-to-day survival, leaving little for efforts to improve their lives for the longer term. In contrast, insurance and credit markets, combined with widespread social security, provide an important cushion against poverty in rich countries. 

The poor in developing countries do have some informal mechanisms to cope with risk and misfortune. These are based on self-insurance by means of savings together with family or community-based mutual assistance. Nevertheless, their scope remains limited. They are better at dealing with individual specific idiosyncratic risks, such as household illness, than covariate or systemic risks associated with drought, flood, or recession. These events devastate whole communities, thereby limiting the opportunities for risk pooling within the community or via local markets. 

We now know much more about vulnerability to risk, and how poor people cope. Building on the insights of micro-level studies, the WIDER project on ‘Insurance against Poverty’ draws lessons for policy, including the design of more effective insurance mechanisms and safety nets.

Measuring Vulnerability and Risk

The relationship between risk, vulnerability and poverty is one theme of this research. These three dimensions are closely connected, but the exact links are not that well understood. Measurement is a key issue: are vulnerability and risk completely captured by simply using available poverty measures? 

In answering this question, it helps to consider two aspects of vulnerability and risk. Ex ante, poor people face a high level of risk that a shock will occur. Ex post, poor people have a limited range of options to cope with a bad shock once it has occurred. 

Information on poverty does not fully capture these two aspects of risk and vulnerability. This is despite much progress in modelling transient and potential poverty, using outcome measures (such as consumption and health) as well as the asset base that the poor can draw upon to cope with bad outcomes. 

Relatedly, recent work on India, Tanzania and elsewhere suggests that risk causes poverty—to the extent that it induces households to take up activities that have high security, but offer a lower mean income in the long-term. Moreover, risk can cause poverty traps; states of deprivation following large shocks to assets and income that are almost impossible to recover from. For example, evidence from China suggests that shocks are followed by only very slow recovery by poor households. We must therefore reduce vulnerability in order to maximise the growth in living standards of the poor—not just to protect them from the immediate effects of shocks.

Designing Insurance

But where markets fail, the design of policies is not self-evident. Informational problems (such as moral hazard and adverse selection) as well as the problem of reinsurance (to protect the insurer) make insurance notoriously difficult to provide in poor countries.

Nevertheless, there are ways forward. A good starting point is to understand how group, community, and household-based support systems operate. As with micro-credit, these can be an ideal starting point in delivering microinsurance. Private and informal systems can provide some protection, but offer only limited risk sharing. Even in closely-knit communities, moral hazard and enforcement-problems limit the extent of support offered. Moreover, the support they do offer is typically for idiosyncratic risk, not covariate or systemic risk. We must therefore avoid expecting too much of mutual support within poor communities

But we should also avoid using scarce resources to substitute for effective mutual support when it already exists. Public funds are better used in assisting network or community-based systems, and filling the gaps left by informal insurance. Such groups can provide effective monitoring and self-control, allowing an outside agent to reduce the otherwise prohibitive transactions costs that limit financial intermediation. Indeed, this may well prove the most cost-effective way to enter into social security provision; by providing insurance capital and reinsurance possibilities to groups, many of the standard crowding out problems may well be avoided.

New Initiatives​

However, problems still remain. The poor are less likely to be part of community-level networks of mutual support, even for idiosyncratic risks such as ill health. Indeed, this exclusion may be one reason that they are poor. So, even for idiosyncratic risks, more public help is needed. Basic support systems with universal coverage offer possibilities. But the resources of very poor countries may not permit this, resulting in rationing and new mechanisms of exclusion. Since means testing is difficult and costly, a basic self-targeted system—with limited stigma and maximum access—may prove the best starting point. In addition, since the poor face diverse risks, linking micro-credit provision to basic health and other insurance policies can be cost-effective. Such interlinked contracts exploit informational and group-liability externalities.

But what do we do about covariate or systemic risk? This is the most dangerous risk for communities, since they are least able to deal with it. Insuring for large shocks, for example crop insurance, has often been unsustainable. The administrative costs of monitoring and assessing crop damage, and the costs of reinsurance, frequently undermined traditional crop insurance

Still, there are some promising new alternatives such as rain-indexed insurance contracts. These are effectively rainfall lotteries where anyone can buy a ticket and payout occurs when a priori established poor rainfall outcomes occur. This product does not require extensive monitoring and is administratively simple.

One crucial advantage is that it is self-selecting: anyone can buy it, whether they are directly or indirectly affected by rainfall, or even just gamblers. Obviously, by its very nature, the issuer of these rainfall lotteries is exposed to potentially very high claims. Governments and their agents could protect their holdings by reinsuring these issues on international markets, tapping into the rapidly developing market of new instruments to deal with catastrophic events.

There may even be a strong macro-economic case for subsidising these schemes (including through aid). Shocks such as drought or large commodity price changes have large externalities for the economy as a whole. Insulating the producers directly is to be preferred. But since the private benefits of insurance are smaller than the social benefits (including the externalities), subsidies are warranted to avoid underinsurance, even though moral hazard puts limits on these subsidies.

Finally, we should not forget that households often use savings to self-insure. In India, for example, self-insurance via savings is far more important than mutual support for smoothing consumption over time. The key is to have good-quality savings instruments available to all.

At present, the poor have a limited choice of savings instruments and these have major shortcomings: inflation destroys monetary savings, storage losses degrade savings in kind, and covariate price cycles reduce the value of livestock. They need access to better financial products, including those provided by microsavings institutions as well as better means to pool the price risks associated with asset and goods markets. This dimension has long been neglected but provides another important starting point in helping the poor cope more effectively with risk.
 

Stefan Dercon is Fellow and Tutor in Economics, Jesus College, Oxford University, and director of the WIDER project on ‘​Insurance against Poverty’.