I would like to estimate a model in STATA, in which I have a variable called Export Intensity (Exports/Total Sales). As the measurement units are different, I prefer to use natural log of all the variables. Export intensity is in ratio (Decimal numbers) format, taking its ln gives negative numbers. Does it create problem if I use it in the estimation?

To avoid the problem of giving negative numbers, I can multiply the whole series with any multiple number (x100, x1000...), and then take ln, which will then give me positive numbers. Is it a right way to do so and use this new series in the estimation of the models? Thank you.

More Shahid Mahmood Waqar's questions See All
Similar questions and discussions