本人用计量方法测算居民外汇存款余额与汇率等关系：
用最小二乘法回归，结果如下：
| Dependent Variable: JMCK |
| Method: Least Squares |
| Date: 12/02/06 Time: 14:16 |
| Sample: 2004:01 2006:10 |
| Included observations: 34 |
| Variable | Coefficient | Std. Error | t-Statistic | Prob. |
| C | 117378.4 | 9833.642 | 11.93641 | 0.0000 |
| SR | 5.119204 | 2.975313 | 1.720560 | 0.0956 |
| JHE | -6.878060 | 1.292402 | -5.321920 | 0.0000 |
| HL | -848133.8 | 78726.11 | -10.77322 | 0.0000 |
| R-squared | 0.814597 | Mean dependent var | 12139.68 |
| Adjusted R-squared | 0.796056 | S.D. dependent var | 1755.721 |
| S.E. of regression | 792.8861 | Akaike info criterion | 16.29937 |
| Sum squared resid | 18860053 | Schwarz criterion | 16.47894 |
| Log likelihood | -273.0892 | F-statistic | 43.93643 |
| Durbin-Watson stat | 0.624097 | Prob(F-statistic) | 0.000000 |
由于存在一阶正自相关，使用EViews消除一阶自相关，得下图：
| Dependent Variable: JMCK |
| Method: Least Squares |
| Date: 12/02/06 Time: 23:32 |
| Sample(adjusted): 2004:02 2006:10 |
| Included observations: 33 after adjusting endpoints |
| Convergence achieved after 47 iterations |
| Variable | Coefficient | Std. Error | t-Statistic | Prob. |
| C | 4150.736 | 11245.75 | 0.369094 | 0.7148 |
| SR | 0.698423 | 0.491725 | 1.420354 | 0.1665 |
| JHE | -0.469269 | 0.412157 | -1.138568 | 0.2645 |
| HL | 29950.48 | 79798.35 | 0.375327 | 0.7102 |
| AR(1) | 0.965565 | 0.018899 | 51.09129 | 0.0000 |
| R-squared | 0.990584 | Mean dependent var | 12046.36 |
| Adjusted R-squared | 0.989238 | S.D. dependent var | 1695.166 |
| S.E. of regression | 175.8534 | Akaike info criterion | 13.31591 |
| Sum squared resid | 865883.4 | Schwarz criterion | 13.54265 |
| Log likelihood | -214.7124 | F-statistic | 736.3839 |
| Durbin-Watson stat | 2.299418 | Prob(F-statistic) | 0.000000 |
| Inverted AR Roots | .97 |
虽然消除了自相关,但Variable:HL由负变成正，本人认为与当前的外汇存款实际相矛盾：即人民币升值态势下，居民外汇存款不断下降。
请问：
1、 该回归分析结果对吗？可用于解释相关的经济现象吗？
2、 HL由负变成正说明什么什么问题？
3、各变量如SR、HL对应的Prob值的用处是什么？
谢谢！
以后贴子的标题要写清楚，否则扣分！
[此贴子已经被eblog于2006-12-3 2:02:39编辑过]
扫码加好友，拉您进群
全部版块
我的主页
收藏
