Error using hac (in order to get robust standard errors)
8 views (last 30 days)
Show older comments
I am reducing a (robust) linear regression using backwards elimination. In order to get robust, i.e. heteroscedasticity and autocorrelation consistent, standard errors of the remaining variables (after removal of an insignificant variable), I use the hac option (<http://www.mathworks.se/help/econ/hac.html#btsg5pc-1>) on the reduced model. So far I have eliminated 4 explanatory variables but when I am trying to calculate the robust standard errors of the remaining 7 variables, I get the error message below (first line is the code):
[EstCov,se,coeff] = hac(mdl);
Warning: Matrix is singular to working precision.
> In arima.estimate>arx0 at 1695
In arima.estimate at 732
In hac>getBW at 845
In hac at 465
Error using hac (line 467)
Input to ROOTS must not contain NaN or Inf.
What does this error message mean? And is there anything I can do to get the robust standard errors?
0 Comments
Answers (0)
See Also
Categories
Find more on Linear Regression in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!