Attend Spring Meetings on development topics from Apr 18-23. Comment and engage with experts. Calendar of Events

Syndicate content

Behind low rates of participation in micro-insurance: a misunderstanding of the insurance concept?

Jed Friedman's picture
Micro-insurance pilot programs begin with grand hopes that the target population will enroll and obtain program benefits, but many are disappointed that after much planning and effort so few actually take up the program. Apparently take-up rates greater than 30% are rare and often do not exceed 15%. Furthermore, only a fraction of beneficiaries choose to renew their participation after the initial enrollment period. There are various reasons for low take-up, but a recent study highlights one aspect that hasn’t received enough attention in the economic literature: the popular misunderstanding of what insurance does.

The appeal of micro-insurance to development policy wonks is clear. In the health arena, a high proportion of health spending comes from out-of-pocket expenditures (OOP) and many low-income countries are unable to provide universal health coverage (UHC). In response to this situation, an intermediary form of insurance, community-based health insurance (or CBHI), has emerged. It is the hope of many that CBHI can serve as a bridge towards eventual UHC. But there is the stumbling block of low take-up mentioned in the first paragraph.

Jean Platteau and Darwin Ontiveros study a CBHI pilot program in Maharashtra, India, that notably did not only cover hospital (inpatient) care but also reduced outpatient costs in order to try and significantly reduce OOP payments. However this program ran into an extreme case of low take-up: only 1.5% of the catchment population enrolled in the CBHI and two-thirds of those beneficiaries declined to continue after the first year.

The authors originally intended to evaluate the program, but this was precluded by the low rates of take-up. Instead the authors take advantage of the situation to look at the determinants of low take-up and continuation. There are many potential reasons on both the supply side (such as low quality of services) as well as the demand side, most notably the lack of beneficiary knowledge of program features – an information failure – for low take-up. Information failures are certainly a factor in the case of the Maharashtra program: some enrollees failed to use the insurance even though they suffered an adverse health event. Multivariate analysis shows that enrollees are 50% more likely to use the scheme if they are well informed, conditional on actual illness.

But another demand side failure that the authors focus on is the misunderstanding of how insurance operates in principle, even when the beneficiary has full information on the particular program conditions. Inquiry into the perceived meaning of insurance is not a common topic for economists, however a previous paper by Platteau suggests that misunderstanding may be a fundamental source of low take-up and dissatisfaction.

Platteau argues, based on work in Senegal, that traditional and pervasive notions of “balanced reciprocity” run counter to the contemporary functioning of insurance. Balanced reciprocity belief implies that an insurance premium, which the payee views as going to a household in need, should be “repaid” in a reasonable amount of time. In essence, beneficiaries view insurance more as a form of credit and if the premium is not “repaid” then the payee feels empowered to leave the scheme. Corollary to this is the view that if a participant did not receive a payout over the previous year then premium payments should be suspended for coming years. Of course these beliefs run counter to the risk-pooling function of insurance where the lucky in any given period compensate the unlucky.

For the Maharashtra study the researchers devise a measure of insurance understanding based on the yes/no answers to the three following questions:
  1. If the discounts (for care) turn out to be smaller than the premium paid, should the insurer reimburse the premium?
  2. Is it unfair that everybody pays the same premium whether falling sick or not?
  3. Is it shocking that other people benefit from the premium that you have paid because they have been sick?
Less than 8% of study respondents answered no to all three of these questions, conveying that the vast majority of the study population exhibits some degree of insurance misunderstanding.

When it comes to the decision to renew, it appears that both information and understanding are important influences on whether the household opts-out of the scheme in the second year. Conditional on a host of observable information included in multivariate analysis, households with a high level of information are 30-40% more likely to continue participation while those with a high level of understanding are 10-20% more likely to continue. (These same factors are also highly predictive of program satisfaction). Households that experience a negative premium payout are significantly less likely to continue unless they exhibit a high degree of insurance understanding, indicating the importance of enhancing understanding for the sustainability of micro-insurance programs.

So it appears the lack of the particular program information as well as low understanding of the general principles of insurance play key roles in the low levels of take-up and continuation. On the other hand it doesn’t appear that perceptions around the availability or quality of care play a large role in the failure to renew participation – almost no respondent mentioned dissatisfaction with the availability or quality of care as a factor.

Clearly there is scope for the program providers to improve beneficiary information about the program. This in turn should increase take-up and continued participation. But the findings here suggest that additional effort, either from the provider or government more broadly, targeted to increasing the understanding the principles of insurance and how they are distinguished from other credit or transfer schemes should help to improve population coverage still further.


Submitted by Rob Yates on

Time to stop flogging the CBHI dead horse

These dismally low coverage rates for community based health insurance in Maharastra are not at all unusual. In fact they are typical of CBHI schemes across Africa. Over the last 30 years there have been hundreds of schemes like this and many have tried to up their coverage rates with intensive efforts to "sensitise the community" But the impact has been minimal.

Therefore wouldn't it be better if we all stopped flogging the dead horse of CBHI because it suffers the fatal flaw of all voluntary insurance schemes - adverse selection. The only way to force the healthy wealthy to cross-subsidise the sick and the poor is through compulsory publicly-governed health finanicng systems. In other words tax finaning or social health insurance or a combination of the two. As even the world's largest economy has shown you can't reach UHC using private insurance as your primary mechanism. Public health finanicng is the key to UHC.

I entirely agree with the conclusion drawn by Jed Friedman and in the last version of my paper with Darwin Ugarte it is actually featured almost word by word! Where I partly disagree, however, is with the statement that CBHI suffers the fatal flow of adverse selection. In fact, our paper shows that this is NOT a problem in the program that we have studied. The main, hard-to-remedy problem is the understanding failure. Here is another, entirely different reason why voluntary insurance schemes may fail.

Hello Jean and Rob, thanks very much for the insightful comments and discussion. I featured the working paper by Jean and Darwin not to make a specific statement about micro-insurance but becuase the attempt to understand how potential program beneficiaries understand the mechansims and the meaning of the program is a very useful undertaking. An undertaking that economists often don't attempt and, as we see here, these features may be quite important for our understanding as program evaluators. Many thanks to you both!

Add new comment