missing value in the cointegraton test

[lexjl5]

Hi,

Now I have some history time seires data in hand. It is no surprising that there are some missing values in each series.

I found that Eviews can run the engle-granger cointegration test and give the adf t-statistic, although there are some missing values in these sequences. Could you tell me how eviews deal with these missing values while it runs the cointegration regression? Just exclude these missing values?

I am a new guy for time series analysis. Could you give me some suggestions about how to deal with missing observations in the cointegration analysis?

many thanks for your assistance

king regards

Jianan Li

[EViews Glenn]

All observations involving these missings are dropped from the cointegrating regression. Observations without valid right and left hand side regressors are then dropped from the unit root test regression involving the residuals.

[lexjl5]

All observations involving these missings are dropped from the cointegrating regression. Observations without valid right and left hand side regressors are then dropped from the unit root test regression involving the residuals.

Many thanks. :D