请教一下，GMM两步估计时报错，而一步却可以。


. xtdpdsys yyy l(0/1).x1 l(0/2).(x2 x3 x4 x5 x6 x10) ye*, lags(2) twostep vce(robust)
note: year1 dropped from div() because of collinearity
note: year2 dropped from div() because of collinearity
note: year13 dropped from div() because of collinearity
note: year1 dropped because of collinearity
note: year2 dropped because of collinearity
note: year8 dropped because of collinearity
variance-covariance matrix of the two-step estimatoris not full rank
    Two-step estimator is not available.  One-step estimator is available and variance-covariance matrix
    provides correct coverage.