Hmm, OK, I think I see. So a new quote: “a 1% change in manufactures (dY/Y) causes a 1/Z (>1) percent change in employment.”

So if dividing by Y can be a reasonably large number (1/100), that means dY is not a differential. If it was, the result would be as close to zero as makes no odds. But that at least explains what it is: a change proportional to its current value. Still seems very odd to me.

Does that sound right? It seems mathematicall unsound: to know that it's a 1% change, you need to make the change at t+1 and compare to what Y was at t. I guess this is convenient shorthand, but is this common notation?