为什么用EViews每次对时间序列做的ADF检验结果都不一样？

襄阳城

<p>为什么用EViews每次对时间序列做的ADF检验结果都不一样？</p>
<p><img src="http://img.baidu.com/img/iknow/icn_point.gif"> 悬赏分：0 -</p>
<p>解决时间：2010-7-19 09:31</p>
<p>第一次结果：</p>
<p>Null Hypothesis: RESID_B has a unit root</p>
<p>Exogenous: None</p>
<p>Lag Length: 2 (Automatic based on SIC， MAXLAG=7)</p>
<p>t-Statistic   Prob.*</p>
<p>Augmented Dickey-Fuller test statistic   -0.565107  0.4633</p>
<p>Test critical values: 1% level  -2.650145</p>
<p>5% level  -1.953381</p>
<p>10% level  -1.609798</p>
<p>*MacKinnon (1996) one-sided p-values.</p>
<p>Augmented Dickey-Fuller Test Equation</p>
<p>Dependent Variable: D(RESID_B)</p>
<p>Method: Least Squares</p>
<p>Date: 07/10/10   Time: 18:08</p>
<p>Sample (adjusted): 1981 2008</p>
<p>Included observations: 28 after adjustments</p>
<p>Variable Coefficient Std. Error t-Statistic Prob.</p>
<p>RESID_B(-1) -0.068773 0.121699 -0.565107 0.5770</p>
<p>D(RESID_B(-1)) -0.547439 0.168982 -3.239632 0.0034</p>
<p>D(RESID_B(-2)) -0.641096 0.157287 -4.075969 0.0004</p>
<p>R-squared 0.513139     Mean dependent var  333.9607</p>
<p>Adjusted R-squared 0.474190     S.D. dependent var  19767.99</p>
<p>S.E. of regression 14334.31     Akaike info criterion  22.07966</p>
<p>Sum squared resid 5.14E+09     Schwarz criterion  22.22239</p>
<p>Log likelihood -306.1152     Durbin-Watson stat  1.913124</p>
<p>第二次结果：</p>
<p>Null Hypothesis: RESID_B has a unit root</p>
<p>Exogenous: None</p>
<p>Lag Length: 2 (Automatic based on SIC， MAXLAG=7)</p>
<p>t-Statistic   Prob.*</p>
<p>Augmented Dickey-Fuller test statistic   -0.793684  0.3632</p>
<p>Test critical values: 1% level  -2.650145</p>
<p>5% level  -1.953381</p>
<p>10% level  -1.609798</p>
<p>*MacKinnon (1996) one-sided p-values.</p>
<p>Augmented Dickey-Fuller Test Equation</p>
<p>Dependent Variable: D(RESID_B)</p>
<p>Method: Least Squares</p>
<p>Date: 07/10/10   Time: 18:11</p>
<p>Sample (adjusted): 1981 2008</p>
<p>Included observations: 28 after adjustments</p>
<p>Variable Coefficient Std. Error t-Statistic Prob.</p>
<p>RESID_B(-1) -0.143976 0.181402 -0.793684 0.4348</p>
<p>D(RESID_B(-1)) -0.726501 0.200830 -3.617490 0.0013</p>
<p>D(RESID_B(-2)) -0.552279 0.174968 -3.156461 0.0041</p>
<p>R-squared 0.535522     Mean dependent var  747.2271</p>
<p>Adjusted R-squared 0.498364     S.D. dependent var  27211.96</p>
<p>S.E. of regression 19273.21     Akaike info criterion  22.67178</p>
<p>Sum squared resid 9.29E+09     Schwarz criterion  22.81451</p>
<p>Log likelihood -314.4049     Durbin-Watson stat  1.890944</p>
<p>提问者： shando3 - 二级</p>
<p>最佳答案</p>
<p>太奇怪了……是不是你的软件不够稳定呀？或者做的时候变动了数据你也不知道？</p>
<p>0</p>
<p>回答者：</p>
<p>梦里兰香hl - 二级   2010-7-14 10:13</p>
<p>我来评论>></p>
<p>提问者对于答案的评价：</p>
<p>谢谢，是数据变动了。</p>