ارزش اطلاعات ژنتیکی در بازار بیمه عمر
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|24072||2000||18 صفحه PDF||سفارش دهید||7990 کلمه|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Journal of Public Economics, Volume 78, Issue 3, November 2000, Pages 235–252
This paper analyzes the effects of additional information in a life insurance market under adverse selection. Individuals have an incentive to acquire information about their risk type if their informational status cannot be observed by insurers. In aggregate, however, the existence of a testing opportunity has an effect on the equilibrium premium. We describe the conditions under which, from an ex ante standpoint, all individuals gain, all lose or in which some gain and some lose from the existence of the test.
The purpose of this paper is to determine whether information such as that obtained from genetic screening has positive or negative value in a life insurance market which displays adverse selection. In many countries existing or proposed legislation prohibits the use of genetic tests for ratemaking purposes and so assessing the impact of such information in the context of a model in which information about risk type is both private and increasing is an important exercise. Also, other types of health tests generate what is effectively private information for insureds. The question of the private and social value of additional information in insurance markets has been analyzed by Doherty and Thistle (1996) in the context of the usual insurance model of adverse selection. As we do, they analyze the case where insurers cannot observe whether consumers have obtained a test. They show that, in this context, acquiring information usually has a positive private value since by taking the test the market possibilities (especially the price of insurance) do not change for the individual but better information allows consumers to make a more informed choice; that is, to adjust the amount of insurance they buy to what is optimal for their risk type. However, the social value of the testing opportunity is negative. If there were no asymmetric information before the test, all individuals would insure for a medium premium. If insurers could observe test results then, depending on the outcome of the test, some (those with good news) could buy insurance cheaper and for those with bad news, the premium would increase. By the martingale property of conditional expectation, the expected premium would be the same as the medium premium before, but since the premium is a random variable and individuals are risk averse, all are worse off from an ex ante perspective. Asymmetric information after the test aggravates this problem, since low risks (those with good news) cannot buy full insurance for the low premium but must signal their type by purchasing only partial coverage. Our model departs from Doherty and Thistle’s in that we analyze the same problem in a life insurance model. The fundamental difference between life insurance and other insurance policies is, from an institutional point of view, that individuals can buy life insurance from as many companies as they want and therefore price–quantity contracts are not a feasible means against adverse selection; insurance companies can only quote a uniform price for all life insurance contracts.1 A second important difference between life insurance and other insurance is that there is no natural choice for the size of loss. In most models of insurance there is a fixed insurable loss l and this loss is independent of its probability of occurring. Thus, a risk averse individual when faced with an actuarially fair premium will purchase full coverage insurance regardless of her probability of loss. Increasing symmetric information about risk type will therefore not have any potential for increased consumption efficiency in terms of the amount of insurance desired. In the context of life insurance, however, this is not the case as a change in the probability of death can, and as we show generally does, lead to changes in the amount of insurance demanded even when these probabilities are symmetrically observed. Thus, information of the type we study here has an added possible dimension for providing positive social value by allowing for better informed consumption choices. There are several reasons for concerning ourselves specifically with the impact of genetic information in a pure adverse selection model. First, the impending completion of the Human Genome Project, which is a massive international effort to map and sequence the entire human genome, will accelerate recent successes in the discovery of disease genes and the development of associated genetic screening tests. This will lead to screening tests becoming available at much lower costs which will thus be used more frequently. Second, many countries ban the use of information from genetic screening tests for ratemaking purposes 2; also, other practical considerations make it natural to assume increasing but private information. Third, most known genetic diseases are not amenable to effective medical treatments which justifies the use of a pure adverse selection model. 3 Consistent with Doherty and Thistle, we find that the private value of becoming informed is positive in our model. More surprisingly, the social (ex ante) value of information may be either negative (as in the normal insurance model) or positive and we describe the scenarios and construct examples for both possible sets of cases. The intuition why additional information might lead to a Pareto improvement is as follows. Suppose there are three groups, high risks, low risks and initially uninformed individuals who may be high or low risks. Suppose that in the reference situation only high risks buy life insurance (this is possible since the premium might be unacceptably high for low risks and uninformed individuals); hence the premium will be based on the high risks’ probability of death. Now suppose a test for uninformed consumers becomes available. If consumers were not to adapt their insurance purchases to their new information, their ex ante expected utility would be unchanged. However, although uninformed consumers who test negative (i.e. learn that they are low risks) will still not buy life insurance, those testing positive (the high risks) will. Therefore, the ex ante expected utility for uninformed consumers is increased by the new testing opportunity. Besides there being a possible positive value of insurance for those who take the test, there is also the possibility of positive price spillover effects for those who don’t take the test. To see how this can occur, suppose there is a range of risk types so that an uninformed individual who takes a test may be determined to be one of a number of higher or lower risk types. Suppose such a person initially does not buy insurance and upon testing discovers she is of a high risk type but that she carries a risk level less than the average of those who initially purchase insurance. If she now purchases insurance the result will be a lower equilibrium price of insurance which will be to the advantage of the original pool of insurance buyers. Negative spillover effects from increased information, which are generally stressed for standard insurance models such as that of Doherty and Thistle (1996), are also possible in our model. Such a situation arises in our model when the initially uninformed individuals discover themselves to be a risk type higher than the average risk of those originally purchasing insurance. The model for the life insurance market we use is similar to that developed in Villeneuve (1996) to explain the effect of adverse selection on the markets for life insurance and annuities. Our model is developed in such a way as to stress the life insurance purchasing decision of individuals for the purpose of replacing lost income due to premature death. We proceed as follows. In Section 2 we present a simple model of life insurance demand and some results concerning the comparative statics of this model. In Section 3 we analyze, for different test scenarios, the ex ante welfare implications of the testing opportunity, taking account of the fact that the equilibrium premium for life insurance will change if individuals can gather information about their risk type before buying life insurance. Depending on the scenario, it is possible (i) that all individuals in this market lose (in comparison to the situation where the test is not available), (ii) that some gain and some lose or (iii) that all individuals gain.
نتیجه گیری انگلیسی
Additional information about life expectancy, especially through genetic tests, will become more and more important as the state of science progresses. Individuals have an economic incentive to acquire such information if we assume either that insurers cannot observe whether an individual was tested or that legislation prevents insurers from using such information, both realistic possibilities. In a market where ex ante information is symmetrically distributed, the availability of the test decreases ex ante welfare; of course, in a more general model, this welfare loss would have to be weighed against possible gains due to the test since more information makes better medical treatments available. In a market with initial informational asymmetry, the welfare effects of a new test could go either way; we constructed examples for a Pareto improvement, a Pareto worsening and for a situation in which those who are tested gain and those who are not lose. At the moment, we believe that the last case is the most realistic one for genetic testing in which those who test positive for a certain gene causing a fatal illness receive very bad information, and are then probably worse risks than the average life insurance buyer, so the equilibrium premium will go up. Since only few people are tested currently, however, the price effect will be small. Hence those who are tested gain since they have the possibility to adjust their life insurance demand to their real risk type for an only moderately higher price. However, as genetic testing becomes available more widely and for less serious illnesses or also for certain other tests, the other scenarios we have investigated, in which testing can lead to either Pareto improvements or worsenings, are relevant.