Insurance.

Part I.

Insurance, basics:

Consider a group of 10 people such that one person falls sick every year. The sick person cannot work and looses his/her yearly income. This sickness randomly chooses one person from the group. At the beginning of the year no one knows for sure whether he/she will fall sick. In such a case the group has an incentive to get together and agree that whoever falls sick will be provided for by the other members of the group. THIS IS INSURANCE IN THE PURE SENSE. What the non-sick persons end up paying is called the premium.

Now consider the same group but a sickness which hits eight people every year. In such a situation two incomes may not be adequately able to provide for all the eight sick people. As such the group may not be able to provide insurance to its sick members.

The first kind of sickness represents what is called idiosyncratic shocks and the second represents aggregate shocks. Insurance lessens the pain caused by idiosyncratic shocks and is ineffective against aggregate shocks. Note that the nature of the shock is defined contingent on the group. For example the second kind of shock (eight people falling sick) may be considered idiosyncratic in a group whose size is 100. By induction therefore, for example, a global recession cannot be insured against but regional recessions can!

Of course there is the time factor involved. Again go back to the earlier example where 8 out of 10 people fall sick but only after every 5 years. Here individuals can SELF INSURE (inter-temporally) through personal savings! In the parlance of the commercial world, self insurance is not called "insurance". So let us leave this issue aside for the time being.

It is not the case that rural farmers do not have ANY insurance in India. Family, caste, kinship, biradiris are networks which provide informal insurance in India. However, due to migration, these networks are becomming less efficient.

Importantly, what is an aggregate shock for a local biradiri can be made into an idiosyncratic shock if a "group" can be formed across biradiris. This is where the formal anonymous market comes in. The name of the game is to covert local aggregate shocks into idiosyncratic shock for a suitably chosen or created group. Doing so is not easy. And it is here that things like monitoring, rating, geographic spread, market accessibility etc. become important.

Part II:

Problems with basic insurance, adverse selection and moral hazard.

In the example provided in Part I, we assumed that when a person falls sick and is not able to work, the other members of the group can verify this occurance. This is an important assumption. If such verification were not possible then people would claim that they are sick, not work and get paid by the rest of the group. Note that this is a problem of asymmetric information. The claimant knows whether or not she is sick but the others do not. If this were to be the case then insurance would fail. To understand this, note that every healthy person has an incentive to lie, not work and claim insurance. At the limit none of the 10 people would work. Ex-ante the group would figure this out and not agree to provide insurance at all. As a result the insurance "market" would be missing!

Of couse, in small groups or clubs, like the village biradiri, monitoring is easy. You would readily find out whether your neighbour is genuinely sick or not. But it may not be so easy across bigger groups. More importantly, it would be difficult (costly) in formal markets for an insurance company to figure out the true health of one villager located in a far off village. Thus, the problem highlighted above is more likely to hit formal markets than informal markets. This problem is known in the literature as the problem of adverse selection.

There is a second problem associated with symmetric information, it is the problem of moral hazard. Consider the examples in Part I again. Suppose it is the case that individuals can privately take care of their health such that the incidence of sickness is reduced. Also assume that the kind of sickness is not life threatning or does not have long term effects. Now note that if a person has full insurance (i.e. his disposable income is the same whether or not he is sick given the insurance). Then the person has an incentive to not take care of his health. If such care is unverifiable (i.e. there is asymmetric information between the group and the person) then insurance cannot be made contingent on "care". And this would lead to a higher risk of sickness in the entire group. To reverse this perverse incentive, the group may decide that the sick person would be given an amount less than what her yearly income would be. This could mitigate the perverse incentives. The amount by which the insurance is less than her yearly income is called the "deductible". Note that if the deductible is very high it is akin to no insurance!

Part III.

When informal insurance drives out formal insurance.

Consider the logic of moral hazard and deductibles as stated above. Now suppose an insurance company decides to enter the market and insure against thousands of groups, biradiris, similar to the ones in the example in Part I. All insurance is now provided by this company. It calculates the risk associated with moral hazard and charges the "appropriate" deductible.

Now suppose a person in a certain group fall sick. He is reimbursed by the insurance company minus the deductable. Of course, in the beginning of the year, the group does not know which member would pay the deductable as they would not know who would fall sick. So they have an incentive to come to an arrangement whereby healthy persons would pay the deductible for the sick person. In other words they would insure against the deductible! But then people would again loose the incentive to take care of their health. If the insurance company realizes this they would increase the deductible to nullify these perverse incentives. If the process were to repeat then in the limit there would be zero insurance provided by the formal insurance company! That is informal insurance would drive out formal insurance. Thus the feasibility of formal insurance is crucially dependent on the availibility of informal insurance.

Part IV:

The role of the competition and the legal system.

Consider the 10 member group again. Recall that we have argued that ex-ante (i.e. in the beginning of the year) the group has an incentive to come to an insurance agreement. Now suppose that one person falls sick after the agreement is signed. The healthy people may not want to pay the sick person. The reason the agreement may not be violated has to do with time. The issue is that if I, the healthy person, do not pay today, next year if I were to fall sick other members of the biradiri would not pay me. The biradiri, usually formalizes this arrangement by throwing the "cheater" out of the biradiri. This works fine, but there are two problems.

The first problem has to do with competition. In the previous paragraph, I had assumed that the person thrown out of the biradiri would find it difficult to enter other biradiris (i.e. a rigid caste-syatem). As a result he would end up with no insurance in future. But suppose the thrown out person were to MIGRATE to the "city" and procure insurance from the formal sector, then he would not care much about the punishment. Therefore he would have less incentive to pay when he is healthy! As a result the informal group insurance scheme would not come about in the first instance. The problem is the same for competing formal insurance companies. With the breakdown of caste, biradiri networks and with the increased incidence of migration in India today a large "insurance" problem is looming. The legal system of the country has to recognize this problem and come up with "punishments" which replace those that could be imposed by informal networks in the past.