Most importantly, in the model the equality holds only over the infinite time-horizon. So I don’t see how it can be confirmed or disconfirmed empirically.

Less importantly, my point was that 250% of GDP in one decade can be much less in real terms than 50% of GDP in a later decade , given growth. So that situation is consistent with the value of the rhs of the equation rising over that period.

In any case, the equation doesn’t entail that the rhs has to always move in one direction, only that it converges to 0.