27% ROW, minus the parts for pedestrians and the parts designed to protect pedestrians from cars, plus all the off-street parking (including driveways)? 33% sounds pretty darned plausible to me.
But you're right, this is one of those academic stats I should be able to cite properly. Since I can't find it (I swear it was on Seattle Public Utilities website once upon a time), perhaps this will be an afternoon task to validate. Should be doable by carving up the city into grid cells of some appropriate unit and then using orthophotos to aid blocking out car-accessible hardscape in a random sampling of grid cells.