EViews.com

There is sort of a kludge to do this. After estimating an equation do View/Coefficient diagnostics/Wald test.

Then enter the nonlinear formula you're interested in, like c(1)/c(2) = 0. You get the value of the formula, the standard error, and the t-statistic.

Some of the dates look like 201001. Is there a good way to handle that?

If it helps, I get the same error you do. But if I use File/Open it appears to work. No idea why.

If you could post that last row in the text file there's better chance someone might catch what's wrong.

Yes. Suggesting R is doing something not quite standard. If it's easy to catch, would be nice if EViews handled it.

Not clear this is exactly an EViews bug...but...

If you read in an Excel file using write_xlsx() EViews will not pick up the variable names in the header row. Quite possibly something not quite right at the R end.

Thanks.

Yes. That is the issue.

The point is that the command defaults to setting the save type as wf1. It probably should not. Much as I am personally fond of the wf1 file format, it is rather dated.

The command creates a workfile of type wf1 even when the default is set to wf2.

IV would make sense.

Autocorrelation and heteroskedasticity have nothing to do with endogeneity.

You almost certainly cannot estimate both equations by OLS.

Maybe a better error message, but I was surprised I couldn't replace the object with a new one.

works fine, but gives an error message

MTOS cannot convert a series/group to a matrix/vector

When I try, I get the correct NAs. What version are you using? Can you post a short example?