On Lee-Carter model for catastrophic illnesses health insurance

• Main Authors: Yen-Ting Chou, 周彥廷
• Other Authors: Chih-Hua Chiao
• Format: Others
• Language: zh-TW
• Published: 2010
• Online Access: http://ndltd.ncl.edu.tw/handle/26125082834036479985

碩士 === 東吳大學 === 財務工程與精算數學系 === 98 === As medical technology advances, people are living longer than their elderly. The increase of life expectancy may not be all beneficial, it has some unexpected risks. The decline of mortality is also the major cause of population aging. Along with the longevity is the risk of decay health condition and the unexpected health care needs. We paid our attention to the health condition of our aging society and its impact on health insurance. It is a well known fact that the younger generation lives longer and there are models to discuss it in the literature. Among which the Lee-Carter (1992) dynamic mortality rate model gets the most attention for its clearly understood and transparency assumptions. Brouhns et al. (2002) recommends embedding in a Poisson log-bilinear model to modify the Lee-Carter model. One indicator of the population health condition is the incidence count of catastrophic illnesses, including cancer. The incidence count curve of many catastrophic illnesses has the shape of death count curve suggested in all the public health statistics. That is, the curve is relatively high for the newborns; inverse proportion to age from baby to adolescence; increasing with age between 20 and 85; and drops after age 85. Many catastrophic illnesses incidence rate has been increased over the year in all ages also supported by medical literatures as well as health insurance claims settlement statistics. That means younger generation has higher incidence rate than their elderly. Modeling incidence rate with time-varying variables has not been widely discussed. In this paper, we adopt the two models mentioned in above for mortality improvement to indicate the declination of population health condition. In this paper, we utilize the database from Taiwan’s National Health Insurance Research Database (NHIRD) from March 1, 1995 to the end of 2007 to implement the incidence rate of the catastrophic illnesses. NHIRD has extensive and complete information on the insured with catastrophic illness and disease code. As of 2007, 22.60 million of Taiwan’s 22.96 million populations were enrolled in the National Health Insurance program. Both the mortality improvement and the increase of incidence rate have a large impact on health insurers. Many applications could be made for health insurers and regulators based on our findings, in particular the projection of future cash flows in the risk management. As an example, we exam one health insurance policy, the net single premiums are compared based on different morbidity assumptions.