Say I want to change the inclination of the ISS by one degree. The ISS orbits at 7660m/s so the manoeuvre will cost me 133.7m/s of $\Delta v$:

From time to time, the station must be reboosted because of orbital decay. Let us just assume that this is done like a Hohman transfer, with the first burn being 10m/s to change the velocity of the station from 7660m/s to 7670m/s. But what if I do a slight plane change at the same time, say 0.1 degrees?

Well, that costs me an additional 6.7m/s on top of the original 10m/s. But wait! If I do this on 10 consecutive reboosts, I have changed the inclination by one degree, but only spent 67m/s of $\Delta v$!

It only gets better. With 100 steps of 0.01 degrees, the total $\Delta v$ spent is only 8.9m/s, at a 1000 0.001 degree steps, only 0.9m/s. It is approaching zero.

Is this really a free propulsive inclination change? I think a way to summarize this odd effect is something like this: