Let's write a tautology : A = A
Let's rewrite it : A = A/B * B
Now, I often see people (in economics, mostly) do this little trick :
ΔA = Δ(A/B * B)
Δ * ln(A) = Δ * ln(A/B) + Δ * ln(B)
Then, they will show that, empirically, A/B has always been constant (more or less), thus, if you want to increase A, you just have to work on increasing B.
Can this kind of equation actually say something about reality ?
Can it be falsified, so it can be taken seriously to be used for predictions ?
(to add some context, I'm thinking here about Kaya identity)