Sunday, October 19, 2014

It must be true...


I read it on the internet: How to calculate the growth rate for a series of numbers, like a FRED series.

I can do it with my VBA code. But that only works in Excel, and that's not always convenient. So I've been doing it by guessing a rate and tweaking it, and watching the numbers change in Excel, and getting close that way. It's easy enough. But recently I noticed an error in that. A discrepancy. So I wanted to find the right way to do it and write it down somewhere.

Looking for that, I found How to Calculate Growth Rate at wikihow. It figures a couple different scenarios, including the one I needed:

Part 2 of 2: Calculating Average Growth Rate Over Regular Time Intervals

It's got a sample dataset:


It explains everything, and then it's got this great graphic:


It shows the formula a couple different ways, and it even gives an example based on the sample dataset. What more could you want?

Well, you could want it to be right, I guess.



In their graphic, where they use "n" as an exponent, they should be using "n-1". And where they use "1/n" as an exponent they should be using "1/(n-1)".

Now maybe I can remember the formula.

1 comment:

Jazzbumpa said...

Thanks for this.

And, yes, I think you're right about n-1.

Cheers!
JzB