| Project/Area Number |
62530008
|
|---|
| Research Category |
Grant-in-Aid for General Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Research Field |
統計学
|
|---|
| Research Institution | TOHOKU UNIVERSITY |
|---|
Principal Investigator |
HOSOYA Yuzo Faculty of Economics, Tohoku University, 経済学部, 教授 (40004197)
|
|---|
| Project Period (FY) |
1987 – 1988
|
| Project Status |
Completed (Fiscal Year 1988)
|
| Budget Amount *help |
¥1,200,000 (Direct Cost: ¥1,200,000)
Fiscal Year 1988: ¥500,000 (Direct Cost: ¥500,000)
Fiscal Year 1987: ¥700,000 (Direct Cost: ¥700,000)
|
|---|
| Keywords | time-series analysis / non-linear models / equilibrium / macroeconomics / econometrics / causality / 統計的推測 / マクロ的需要均衡 / 失業率 / 雇用量 |
|---|
| Research Abstract |
Equilibrium models of the macroeconomy aim at explaining the observed economic activity level and the price level by the equilibrium of the aggregate demand and supply. For the statistical inference of econometric models of the equilibrium, we conducted in this reserch various model fitting to the time-series data of the Japanese economy. We observed that the explanation ability by linear time-series models is limited and in particular that the basic relationships such as the expectation of the future inflation rate and the potential output level seem to need a flexible nonlinear model since th observed period involves variety of economic phases such as the hyperinflation or the low economic growth. In this reseach we developped a nonparametric testing method of the causality between macro economic time series, Abased on the nonparametric time series analysis by P.Robinson. Essentially, the test is based on nonparametric estimation of the conditional expectation by the kernel method or the nearest neighborhood method. Since in general the relations between macroeconomic variables are functionally not specified a priori and need to be empirically determined, it is not desirable to delimit the nonlinear functional form beforehand, in the application of nonlinear time series models to macroeconomic data. Therefore we need a model of certain generality. In this respect the state dependent model proposed by Priestley seems suited to the macroecomonic time series analysis.
|
