論「精確線型理性預期模型計量方法」應用之限制
Limitations of Econometric Analysis in Exact Linear Rational Expectations Models
國立中正大學經濟系(Department of Economics , National Chung Cheng University)
- 卷期
- 26卷3期
- 出版
- 1998年09月
- 頁碼
- 243–257
- 編號
- NTUEJ00007035
摘要
在理性預期的假設之下,經濟模型往往需要將變數前瞻求解以達成體系之收斂,這當中包括部份經濟變數的實現值與其他經濟變數的預期值之間,存在著無誤差項動態關係的模型。 Hansen and Sargent(1981) 廣為應用之精確線型理性預期模型計量分析,說明了如何針對前瞻解模型導出跨式限制以進行假說檢定。本文指出應用 Hansen and Sargent 的這套計量分析方法,不能在檢定模型引申推論的同時,也將理論上變數之自發或外生性質加諸在 VAR 模型之因果關係中; 內生的前瞻解變數必須在統計上有助於外生變數之預測,才可以保証時間序列模型不會出現與定態性質相矛盾的參數估計結果。
Abstract
Many rational expectations models have forward-looking solution. This includes models in which there is an exact linear relationship between the expected future values of exogenous variables and the current and past values of endogenous ones. Hansen and Sargent (1981) called such kind of models exact linear rational expectations models (ELRE). They derived the time series model's cross-equation restrictions and proposed a method to specify and estimate ELRE. However, the ELRE econometric analysis cannot be applied to models which assume that exogenous variables follow an AR process. More specifically, when using the ELRE econometric analysis, one must have the Granger causality from endogenous variables to variables which are exogenous in a structural model to get the correct eigenvalues in the corresponding restricted time series model.