0

I have to perform a Wald-Test on the estimated coefficients of a multivariate linear equation model, estimated with systemfit. The command waldtest.systemfit does not work, R does not know it (even though i have installed the package). I tried to solve it with linearHypothesis() but R gives me following message:

Error in solve(restrict.matrix %% vcov %% t(restrict.matrix)) :
fejl under evaluering af argument 'a' under valg af metode for funktion 'solve': Error in restrict.matrix %*% vcov : non-conformable arguments

My equation system has the following form:

CDSSurReg = systemfit(list(Y~CDSSur[,3] + Dummy1 + Dummy2, ...  Y~CDSSur[,20] + Dummy1 + Dummy2,  Y~CDSSur[,21] + Dummy1 + Dummy2), method = "SUR")

With 19 equations in total.

I want to test if the average Dummy1 - Coefficient is significantly different from zero.

Hack-R
  • 22,422
  • 14
  • 75
  • 131
EvilKnivelCat
  • 1
  • Instead of posting a question with an ellipsis in your formula and no data, you should find a dataset in the package, post a library call to the package first, and then post a model specification that will result in an object that supports a proper demonstration. – IRTFM Jun 25 '16 at 15:24
  • http://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example It's not required but it will help you get answers. – Hack-R Jun 26 '16 at 03:43

0 Answers0