'growth' rate is likely to be a geometric constant, not an arithmetic one.

I agree. I find it unlikely that Piketty's thesis rests on such an elementary mistake as interpreting arithmetic means as geometric ones. Academic economists are basically applied mathematicians. (In the book, Piketty actually bemoans the fact that economists are preoccupied with proving mathematical theorems at the expense of engaging with the real world.)