Development of econometric testing method of causality in dynamic macroeconomic models.

在本研究中，我开发了适用于因果关系检验的同时<X^2>，Beta和F检验，当备择假设具有嵌套结构时。其基本思想是基于任何广义似然比检验。在检验非因果关系的西姆斯检验中，给出了自回归方案的阶，但这并不是一个现实的假设。为了检验日本宏观经济中名义收入与货币收支数据之间的因果关系，本文采用了嵌套<X^2>检验，得到了比以往研究更清晰的单向因果关系模式。此外，巢式试验还用于鉴别圣路易斯F.R.B.。模型本文还提出了一种适用于嵌套度不超过13的嵌套<X^2>检验P值的算法。

In this research, I developped simultaneous <X^2> , Beta and F tests which apply to causality test when alternative hypotheses have a nesting structure. The basic idea is based on any generalized likelihood ratio test. In the Sims test for testing noncausality, the order of autoregressive scheme is given, but it is not a realistic assumption. For detecting causality between nominal income and money balance data in Japanese macroeconomy, a nested <X^2> test is applied and it gave a sharper pattern of one-way causality than the previous studies in this field. Also the nested test was applied for the purpose of identification of the St. Louis F.R.B. model. Also in the present research, an algorithm for a nested <X^2> test P-value which applies to the case where the degree of nesting is at most 13.