You're reading it wrong. Those would count as vacant.

> In addition, a vacant unit may be one which is entirely occupied by persons who have a usual residence elsewhere.

The sentence says “may”. And the previous sentence says “ A housing unit is vacant if no one is living in it at the time of the interview, unless its occupants are only temporarily absent.”

I guess we would have to read further on how they determine temporary. But given the sampling method, I would guess people doing a door to door survey and noting it as temporary absent if it’s well kept?

The FED graph uses census vacancy data, which is what I’m assuming the original article does as well. So the metric is apples to apples.