why would I get missing values in residual series?

[Fregensburg]

Dear experts,

After running a simple OLS regression, I looked at the residuals as I needed them for something else, and the series turns out to have long stretches of empty observations (as well as actual values) across the sample. Both left and right hand side variable in the regression do not have any missing values, so this shouldn't happen. Any suggestion on what the issue is?

Many thanks

[EViews Gareth]

Any observation included in the estimation will have non-NA values. Any observation not included in the estimation will have NA values.

Without seeing the workfile, hard to give more information than that.

[Fregensburg]

Here it is

[EViews Gareth]

Your independent variables has lots and lots of NAs.

[Fregensburg]

Gareth,

I don't quite understand. When I look at the rhs variables _12mf and wtispot, there is a full data set from the start of the sample, 1/02/1990, through to 7/24/2013. However, the residuals turn out to have long stretches of NA values even when there is data for the independent variables.


(data was missing between 7/24/2013 to mid 2014. I've just shortened the workfile to include only available data, values for residuals still turn up on/off NA).

[startz]

You're taking the log of a negative number.

[Fregensburg]

ah. Completely overlooked that. Generally eviews alerts you with an error message whenever one's trying to take logs of negative values.
Thanks a lot!