近年來由於人口死亡率的改善，長壽風險對於個人退休規劃以及保險公司經營退休商品有很大的影響。為能有效的評估長壽風險，了解未來死亡率改善趨勢以及對長期死亡率動態的預測是非常重要的。本研究將探討三個長壽風險模型建構與應用的議題，第一個年的研究主題是著重與死亡率改善之時間趨勢分析，死亡率改善具有時間之趨勢，共整合分析可以用來了解死亡率之長期均衡關係，然而，若考慮跨年齡或跨國間之死亡率資料，其為Panel 之資料型式，因此，本研究延伸文獻上的做法，以Panel 共整合的方法及檢定來探討各年齡間及男女死亡率改善之長期均衡關係，並以人口資料庫(Human Mortality Database)中之國家死亡率資料來進行分析。第二個研究主題是探討死亡率相依性（Mortality Dependence），目前文獻上探討之死亡率模型，不論是連續型態或離散型態的模型，大都假設不同年齡或性別或國家之死亡率趨勢是獨立的，而Loisel and Serant (2007)首先提出死亡率於年齡及期間之相依性，因此，本研究進一步以近年來較新的Copula 方法來補捉死亡率之相依特性，並且著重於Copula 模型之檢選與配適分析，因此，將以跨國之死亡率資料來進行分析。第三個主題則是探討新興退休商品反向抵押貸款之長壽風險，反向抵押貸款是利用房子的價值來取得退休之財源，而年金為最常見之貸款付款方式，本主題將以連生（Joint-life）型態終身年金做為主要之反向抵押貸款型態，探討其長壽風險。本研究延伸第二年之死亡率相依性的影響來計算連生之生存機率，並考慮二維之Copula 模型來建構死亡率相依性，除此之外，本研究亦考慮房價指數以及利率之動態來建構反向抵押貸款定價模型。 Human life expectancy has been increasing significantly since the start of the 20th century. Longevity risk is the tendency of individuals to live longer and longer. Due to the mortality improvement, longevity risk has been recognized as one of the significant risks for pension and annuity providers. To deal with longevity risk, how to model the dynamics of mortality rates is very critical. In this research, we attempt to deal with three research topics in the proposed three-year research projects. In year one, we deal with the “Panel Co-integration Analysis of Mortality Modeling”. Mortality improvement has showed a time trend in mortality rates. The use of co-integration techniques to test for the presence of long-run relationships among integrated variables has enjoyed growing popularity in the empirical literature. The first purpose of this research is to make use of the panel cointegration analysis to identify the effects of mortality improvement that cannot be detected in simple time series or cross-sectional data. In year two, we investigate the “Mortality Modeling under Mortality dependence: Selection of Copulas Models”. In the literature, most of the mortality models are assumed mortality independence in mortality modeling. Thus, one of the purposes of this project is to investigate the mortality dependence using copula approach. In year three, we extend the mortality modeling for “Analysis of Longevity Risk for Joint-Life Reverse Mortgage Scheme”. Due to the population becomes aging, the reverse mortgage products appear to offer important consumption benefits to the elderly. Modeling the future dynamics of mortality rates has great effects on valuing reverse mortgage products. Thus, we consider previous mortality study on mortality dependence to analyze longevity risk for reverse mortgage products. 研究期間：10008 ~ 10107
