序列都是I(1), 请问Johansen cointegration test 检验出有一个equation之后, 该如何操作才能得到cointegration regression? 是不是用PROC里的MAKE EQUATION，然后选择LS? 根据以下结果，协整回归是否应为 log_reer = - 0.1361log_open +0.045180log_NFA +0.0986637 log_NEER +0.214544 log_GEXPGDP +0.060965 ? 但是T检验显示大部分自变量都不显著，我又检验了残差在10%的水平上是稳定的，而且这个方程的R2值挺高，那这个协整回归方程是对是错？

Dependent Variable: LOG_REER
Method: Least Squares
Date: 05/16/07 Time: 00:43
Sample (adjusted): 3 28
Included observations: 26 after adjustments

Variable Coefficient Std.Error t-Statistic Prob.

LOG_OPEN -0.136159 0.132320 -1.029012 0.3152
LOG_NFA 0.045180 0.029116 1.551724 0.1357
LOG_NEER 0.986637 0.119720 8.241200 0.0000
LOG_G_EXP_GDP 0.214544 0.142855 1.501824 0.1480
C - 0.060965 0.199158 -0.306113 0.7625

R-squared 0.937607 Mean dependent var 2.053714
Adjusted R-squared 0.925723 S.D. dependent var 0.162583
S.E. of regression 0.044310 Akaike info criterion -3.224167
Sum squared resid 0.041231 Schwarz criterion -2.982226
Log likelihood 46.91418 F-statistic 78.89396
Durbin-Watson stat 0.556831 Prob(F-statistic) 0.000000