我最近在做四川省FDI与工业废气排放(1996-2005年的数据)因果关系检验时,先对两个时间序列做了单位根检验,得出以下结果:
| ADF Test Statistic | -5.456580 | 1% Critical Value* | -5.4776 | |
| | | 5% Critical Value | -4.0815 | |
| | | 10% Critical Value | -3.4901 | |
| *MacKinnon critical values for rejection of hypothesis of a unit root. | ||||
| Augmented Dickey-Fuller Test Equation | ||||
| Dependent Variable: D(LNFDI) | ||||
| Method: Least Squares | ||||
| Date: 05/03/07 Time: 02:53 | ||||
| Sample(adjusted): 1997 2005 | ||||
| Included observations: 9 after adjusting endpoints | ||||
| Variable | Coefficient | Std. Error | t-Statistic | Prob. |
| LNFDI(-1) | -1.332239 | 0.244153 | -5.456580 | 0.0028 |
| D(LNFDI(-1)) | 0.069504 | 0.174867 | 0.397467 | 0.7074 |
| C | 13.35371 | 2.492890 | 5.356718 | 0.0030 |
| @TREND(1995) | 0.171684 | 0.028116 | 6.106346 | 0.0017 |
| R-squared | 0.886927 | Mean dependent var | 0.077652 | |
| Adjusted R-squared | 0.819083 | S.D. dependent var | 0.298082 | |
| S.E. of regression | 0.126787 | Akaike info criterion | -0.991510 | |
| Sum squared resid | 0.080375 | Schwarz criterion | -0.903855 | |
| Log likelihood | 8.461795 | F-statistic | 13.07305 | |
| Durbin-Watson stat | 1.878508 | Prob(F-statistic) | 0.008397 | |
| ADF Test Statistic | -5.003936 | 1% Critical Value* | -5.4776 |
| |
| | | 5% Critical Value | -4.0815 |
| |
| | | 10% Critical Value | -3.4901 |
| |
| *MacKinnon critical values for rejection of hypothesis of a unit root. | |||||
| Augmented Dickey-Fuller Test Equation | |||||
| Dependent Variable: D(LNGAS) | |||||
| Method: Least Squares | |||||
| Date: 05/03/07 Time: 03:53 | |||||
| Sample(adjusted): 1997 2005 | |||||
| Included observations: 9 after adjusting endpoints | |||||
| Variable | Coefficient | Std. Error | t-Statistic | Prob. |
|
| LNGAS(-1) | -0.468097 | 0.093546 | -5.003936 | 0.0041 |
|
| D(LNGAS(-1)) | 0.226034 | 0.165866 | 1.362749 | 0.2311 |
|
| C | 3.804087 | 0.797578 | 4.769550 | 0.0050 |
|
| @TREND(1995) | 0.043921 | 0.007130 | 6.159916 | 0.0016 |
|
| R-squared | 0.938944 | Mean dependent var | 0.025874 |
| |
| Adjusted R-squared | 0.902310 | S.D. dependent var | 0.119857 |
| |
| S.E. of regression | 0.037462 | Akaike info criterion | -3.429890 |
| |
| Sum squared resid | 0.007017 | Schwarz criterion | -3.342234 |
| |
| Log likelihood | 19.43450 | F-statistic | 25.63064 |
| |
| Durbin-Watson stat | 3.180534 | Prob(F-statistic) | 0.001835 |
| |
两个都是对原序列采用截距+趋势的模型,滞后一期得到的ADF检验数据,是不是说明两个原序列都是平稳的呢?如果是平稳的,那么我应该可以不用做协整检验,直接对二者进行格兰杰因果检验了吧?
还有,为什么要选择滞后长度呢?滞后长度有何意义?选择滞后长度,经济意义是否发生了改变呢?
我们的计量经济学教材没有涉及这方面的内容,所以恳请大家指点!!


雷达卡



京公网安备 11010802022788号







