2009-08-012024-05-18https://scholars.lib.ntu.edu.tw/handle/123456789/713062摘要：健康(health status)是一種人力資本(human capital)，有如教育一般，其大小與多 寡常常是個人決策行為中的一重要變數。然而，健康卻不如教育，有一較為客 觀的衡量一教育年限(years of schooling)，最常被使用的健康變數是一具有主觀 意識衡量的 “自我評估健康指數＂(self-report health)。使用“自我評估健康指 數＂當做健康的衡量有許多公認的僭在問題。這一些潛在性的問題，將會造成 使用自我評估健康指數作為健康狀況，來估計勞動市場行為，在估計上產生偏 誤。不少文獻試圖著建立一較為客觀之健康指數之衡量。但是這一些研究都只 限制在單一指數模型，本研究主要想要提出一個合併自我評估、多個客觀的身 體功能以及多個勞動市場指數建立一個包含工作能力(work capacity)以及健康 狀況的健康指數。 本研究之結構方程式立基於Grossman (1972) 的 “demand for health＂ 的 經濟理論模型，資料將採用2005 年American Community Survey 調查資料。本 研究之研究方法分成二部份，第一部份利用Lee (1982)所提出的a simultaneous equation model with multiple discrete indicators 的估計方法，建構一多面向健康 指標。第二部份，為檢驗此一健康指標是否較自我評估之健康指標為優，我們 將利用第一部份所建構之多面向健康指標估計式，模擬出每個人的多面向健康 指標，進而以此指標以及自我評估之健康指標分別來估計每人之退休行為，並 比較其解釋能力，以評估我們的健康指數是否優於文獻中所使用的自我評估指 數。 <br> Abstract: Health, like education, is a kind of human capital, which is one of the major determinants of human behaviors, such as labor supply decision, retirement decision, occupational choice, and application of disability benefit. The measurement of health, however, is often not objective. The literature has commonly used global questions such as, “Does health limit the amount or kind of work you can perform?” or “How would you rate your health? Is it excellent, very good, good, fair, or poor?” There are a number of potential problems with such self-reported health measures, which leads to different kinds of biases (Bound 1991). A variety of authors have used what they hope to be more objective indicators of health—responses to questions regarding specific functional limitations. An alternative is to create a health index based on the more objective measures and the global measure. Previous authors who have used a latent variable framework to construct disability indexes have all worked with single indicators. The contribution of our project is to use multiple indicators. The advantages of using multiple indicators are the efficiency gains and the relaxation of identification restrictions. Data for this research come from the 2005 American Community Survey (ACS). The estimation method consists of two folds. First, we adopt a three-stage estimations procedure proposed by Lee (1982) to construct a multiple-dimension health indicator. The structural model is based on the “demand for health” economic model by Grossman (1972). We treat health status as an unobservable phenomenon which is both causally related to a number of exogenous characteristics of an individual and correlated with a number of observed indicators of health, impairment and qualifications for employment. Second, we use the ACS data and the estimator of our health indicator to calculate the health status for each individual. Then, we use a simple model of retirement decision to compare the prediction of using the constructed health indicator with self-reported health measure on understanding the determinants of retirement behavior and choice.多面向健康指數自我評估指數健康之需求multiple-dimension health indicatorself-reported health indicatordemand for health健康狀態之衡量 : 客觀相對自我評估