The Power of Distaso (2008) Test When the Stochastic Coefficient and Errors are Correlated
| Autor: | Yi-Fang Hsieh, 謝宜芳 |
|---|---|
| Rok vydání: | 2019 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 107 Unit root test has always been an important issue in time series data analysis. Based on the ALM test of Distaso (2008), this paper makes the random coefficient and the error term have the correlation, derives the limit distribution of the ALM test under the alternative hypothesis and uses Monte Carlo simulation to verification. The study found that under the correlation between random coefficients and error terms, the ALM test needs to be divided by two-thirds of the sample number to converge to the limit distribution, so when the number of samples increases, the power of ALM test will increases. In addition, it is found that both the correlation coefficient of the random coefficient and the error term or the variance of the random coefficient increase, which also increases the power of the ALM test. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|