日志提示错误:
WARNING: The covariance across equations (the S matrix) is singular. The error variance of the following 1 equations is near
zero: dayillqdt
我的代码如下
proc model data=usedata222;
endogenous dretwd dayillqdt;
parms k beta c M A E;
DAYILLQDT=M*EXEDINDEX2+A*DRETWD/dsmvosd+E/dsmvosd;
DRETWD=K*exedindex2+BETA*RETINDEX2+C;
fit dretwd dayillqdt ;
ods output parameterestimates=outfinal;
ods listing;
RUN;
结果,dayillqdt这个等式的系数就都回归不出来
理论上来说,SEM不是只要每个式子都有一个各自的外生变量就可以了么,我第一个式子的外生变量是exedindex2,第二个式子的外生变量是retindex2,为啥不行……