晕倒，怎么发出来会这样！
在这将问题重发一下


proc model data=ipo.fna1;
parms b0-b29；
retnfstd=b0+b1*age+b2*lnamnt+b3*shrcntr+b4*bj+b5*gg+b6*indcd1+b7*indcd2+b8*indcd3
+b9*indcd4+b10*indcd5+b11*yr90+b12*yr91+b13*yr92+b14*yr93+b15*yr94+b16*yr95+b17*yr96
+b18*yr97+b19*yr98+b20*yr99+b21*yr00+b22*yr02+b23*yr02+b24*yr03+b25*yr04+b26*yr05+b27*yr06
+b28*yr07+b29*yr08;
fit retnfstd/gmm kernel=(bart,1,0);
run;

这是关于回归的一段程序，indcd1-indcd5是行业虚拟变量，yr90-yr99 及yr00-yr08是年份虚拟变量。

但log 显示如下错误：
ERROR: The number of parameters (31) is greater than the number of unique instruments (30) times
       the number of equations (1).
请问是哪里出问题了？

U have 30 equations, but there are 31 parameters. It means u can find countless solution and all of them are best.
i.e.  x + y + z= 10,  x + z = 20
now, we have 2 equations and 3 variables, the answer u find would be the fundesmantal set of solution.

不是这样的吧。我用“parms"定义的变量数与回归式中的变量数（包括截距项在内）是一致的呀！
能否提供修改建议呢？

请问解决了吗，遇到了同样的问题