co-integration estimation

[lnp3]

1) do I need continuous observations for co-integration analysis? in other words, if some years are missing for some variables, can i still use co-integration technique?


2) how can I check robustness in co-integrating regressions estimated by FMOLS, DOLS, etc on Eviews 8? There is no LM test for serial correlation (there is only Correlagram-Q statistic). there is no hetoreskedasticity test either;

i am using Eviews 8. My data is single country/time-series.

I appreciate any feedback.

best,
MD