更精確的風險趨避係數及效用函數變換的效果

Abstract

先前的相關文獻中對消費基礎資本資產定價模型(C-CAPM)中的Covariance項拆解成為FED以及SED的型式，再對消費水準以及效用函數進行假設，最後經由各國實證資料對於風險趨避係數做出估算，本篇論文藉由更高次方的期望相關(NthED)來對拆解後的C-CAPM進行更進一步的計算以達到某種程度上收斂的結果；效用函數假設的方面也從原本的exponential效用函數增加了使用power效用函數的結果，最近的文獻中對於CRRA也有實證上的討論，因此增加power效用函數假設應能夠對現實中人們的風險趨避的程度提供更多樣的解釋，在求出的相對風險趨避係數當中，相對於原先的假設，數值上比較接近其他文獻中小數值的狀況，提供了一定程度的可信度。

Previous papers change the covariance term of C-CAPM into FED and SED terms and assume the consumption and utility function to solve the numbers of risk aversion coefficients of each country by empirical data. This paper tries to use higher order expectation dependence to get a more precise converging number of risk aversion coefficient of each country. There is another assumption of utility function which is power utility function to compare with original exponential utility function. There is some discussion about CRRA empirically by recent studies, as a result, adding the assumption of power utility function may provide more widely interpretations to the degree of risk aversion. In the coefficients of relative risk aversion, there is a special case that the result of Italy in accordance with past study, which gives a certain degree of reliability.