Not independent observations!

I am trying to estimate a model and I have there observations but 2 of them are obtained from the third, would that be a problem? I have data on output, but I only have data on tax revenues as a % of gdp and informal sector as a % of gdp, so I multiply gdp by these %s to get series for Tax Revenues and Informal output.
I ask because I get the following warnings:
1- “Warning: Matrix is singular, close to singular or badly scaled.
Results may be inaccurate. RCOND = NaN.”

2- “Warning: Matrix is singular to working precision.”

And the code does not produce the smoothed shocks. Any help would appreciated.

As long as your observation equations are correctly specified, that is not a problem, because these variables still contain information separate from GDP.
The error messages you describe are most likely due to something else.