A 20 ton bus starts out 160,000 times worse than a 1 ton car.
If you get 13.3 people off the road, and people weigh 150 pounds, then you have taken off 13.3 units of car wear, but added 34,000 units of bus-wear (going from 20-to-the-4th to 21-to-the-4th).
If the argument is "reduce road wear" and damage is proportional to weight to the 4th power, you want lots and lots of smaller units over bigger units.
EDIT: I did some more math. Assuming a very light 5-ton bus, adding a 90-pound person to it is the same road wear as adding a 2.18 ton vehicle to the road. O(n^4) is nasty.
(Someone once evoked it like this: Would you rather have 300 1lb kittens walk on you, or one 300lb human?)
But the wear is per tire, not per vehicle. So you have to divide the bus weight by the ratio of number of wheels compared to a car.
It looks like the conclusion was read exactly right. A car with 2000lbs per axle will put sixteen times more wear on the road than one with 1000lbs.
And the wear increases again at very high speeds, for unrelated reasons.