I agree, yet it's a variation within so small margins that for practical purposes there is no difference. And yet once again, comparing Sep-Oct-2018 to Sep-Nov-2022 will show a decrease in stability. No conclusion over the nature of the problem can be made by cherry-picking and sticking to purism over pragmatism. If I augment your salary from 1000,0001 to a 1000,0003 , and then from 1000,0003 to 1000,0030 , and I tell you that you have experimented an augmentation rate of 10x (yourwelcome) I'm quite sure that the argument you are using now will suddenly stop having any significance to you.