This is only a fair comparison if you also measure how much people in the office are actually working and not just being present at their desk. The results would probably be similar.