Newey West HAC Covariance Matrix and LDVs

[Mila]

Hi
I would really appreciate it if someone could clarify the following:
I have a multivariate regression with lagged dependent variables as regressors. LM test indicates serial correlation. Is the Newey West HAC Covariance Matrix method for standard errors still valid? I was looking at the formula and it seems that an assumption needed is E[u/x]=0, i.e. exogeneity of regressors. Are the robust standard errors valid in a large sample scenario?

Many thanks.

[startz]

If you have a lagged dependent variable and serial correlation, then the OLS estimates are biased and inconsistent.