Statistical Estimation and Testing for Partial Causal Measures in Multivariate Economic Time Series
| Project/Area Number |
17530160
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Economic statistics
|
|---|
| Research Institution | Meisei University (2006)
Tohoku University (2005) |
|---|
Principal Investigator |
HOSOYA Yuzo Meisei University, Department of Economics, Professor, 経済学部, 教授 (40004197)
|
|---|
| Project Period (FY) |
2005 – 2006
|
| Project Status |
Completed (Fiscal Year 2006)
|
| Budget Amount *help |
¥3,600,000 (Direct Cost: ¥3,600,000)
Fiscal Year 2006: ¥1,800,000 (Direct Cost: ¥1,800,000)
Fiscal Year 2005: ¥1,800,000 (Direct Cost: ¥1,800,000)
|
|---|
| Keywords | causal relations / causal measures / multivariate time series / economic time series / macroeconomics / Whittle estimators / Wald test / frequency domain analysis / 因果測度 / 偏因果性 / 多変量経済時系列 / 共和分モデル / スペクトル密度分解 |
|---|
| Research Abstract |
In the literature of time-series causal analysis, research workers have proposed a variety of third -series effect elimination method in the presence of such a series. In this research of the year 2006, I dealt with the problems of specification of time-series models, estimation and testing of one-way causal measures and developing feasible computational algorithm for statistical inference and application to real economic time-series data sets in order to characterize quantitatively and more precisely the dependency between a pair of time series in the presence of a third series. Those partial causal measures are estimated based on such a specific model as the cointegrated ARMA model and the research focused on numerical method of estimating and testing the cointegrated ARMA under the precise conditions of a specified rank condition and the research was able to produce a feasible procedure. The partial causal measures requires a canonical factorization method of ARMA spectral densities in the in the process of evaluating the measures. The research of this year could develop a numerical procedure for spectral factorization which is an essential step in deriving the partial measures and hence this made feasible the whole project of statistical inference on the partial measures.
|
Report
(3 results)
Research Products
(8 results)
