因生命價值之衡量可作為評估政策效益或分配資源時的一個指標，故生命價值之估計在成本效益分析中是相當重要。因此，本研究主要探討生命價值，利用民國91年至95年人力運用調查，以及勞保局之各行業的勞工職業災害死亡統計資料做實證研究。 本研究採用特徵工資模型為主要架構，並且以最小平方法進行迴歸分析，透過「工資─工作風險」間的抵換關係，估算民國91年至95年各年度的生命價值，以及各年齡層勞動者的生命價值，並探討生命價值與年齡間之關係。其中，本研究分別利用行業別死亡率，以及不同年齡層於各行業的死亡率資料衡量工作風險。 由本研究實證結果可得知，各年度的生命價值在以行業別死亡率為工作風險變數，所推估的生命價值皆高於以不同年齡層對應的工作風險之估計結果，表示風險變數衡量方式的不同會影響生命價值之估算。另外，風險變數衡量之差異亦會影響生命價值與年齡間的關係，而再考量年齡和地區等外在條件因素對工資補償的影響後，亦使兩者間的關係改變。故以台灣為例，本研究分析生命價值和年齡間在實證上的關係可能呈遞增、倒U型、U型或為更複雜之關係曲線，此結果與Shepard and Zeckhauser (1984)和Johansson (2002)的研究指出，兩者為模糊不清之關係相一致。Because the measurement of the value of a statistical life serves as an indicator of the estimation of the benefits of the policy and the allocation of resources, it is very important to estimate the value of a statistical life in cost-benefit analysis. Therefore, this paper explores the value of a statistical life, and uses the Taiwan Labor Force Survey data from 2002 to 2006, and uses work-related fatality statistics data at the industry-level from the Taiwan Labor Insurance Agency for the corresponding years. This paper applies the hedonic wage model, and OLS method is used to estimate the hedonic wage regression, based on wage-job mortality risk tradeoffs in the labor market to estimate the value of a statistical life from 2002 to 2006, and calculates age group-specific VSLs, and explores the relationship between the value of a statistical life and age. The study uses measures of fatality risk by industry and by both age and industry. From this paper’s empirical results indicates that the annual value of life estimates by industry are greater than those generated by fatality risks by both age and industry, and shows the difference of the measures of risk variable significantly influences the estimated of value of a statistical life. In addition, the choice of the measures of risk also influences the relationship between the value of a statistical life and age. When considering the influence from external circumstances such as age and region factors on wage compensation, also make changes the relationship between the value of a statistical life and age. Taking the case of Taiwan, this paper infers that the empirical relationship between the value of a statistical life and age may be presented on the increase, inverted U-shape, U shape or more complicated curve. This result which the relationship between VSL and age is ambiguity is consistent with Shepard and Zeckhauser (1984) and Johansson (2002).
