Backing out the level variable from a first-differenced log

[Mike Kraft]

Hi
I have a variable that is only available in my dataset as first-differenced log variable (dtfp, which represents first differenced log total factor productivity).
I need to have this variable as in level form (or log level).
How can I do this in Stata?

Thanks
Mike

[Nick Cox]

You reversing differencing with a cumulative sum and logging with exponentiation, do you not?

So, with natural logarithms implied exp(sum(diff)) would imply an initial logarithm of 0 and so an initial value of 1. Some conventional prefactor such as 100 may also seem convenient.

[Mike Kraft]

Thanks a lot Nick
I did now

 
gen level_tfp=exp(sum(dtfp))

The level variable looks reasonable.