Is it possible to regress a log-log (or log-linear) equation where the data values can be negative (such as growth rates)?
Question for any econometricians.. Watch
- Thread Starter
- 19-12-2010 02:19
- 19-12-2010 13:15
i have never seen log(growth) only log(gdp) which is used to see growth on a graph.
i dont know for sure but just mathematically i dont think that even works.
- 19-12-2010 13:22
You can't take a log of a negative number, mathematically, so as far I can understand I don't think that would work.