It does not make, nor need, that assumption. If only a few people drove better from understanding increased risk, then that may be enough to lower bad outcome rates.
Why did you claim Tullock "assumes that everyone has the same ability to assess risk"? I can find no such reference or claim online - except yours. Have some info about it?
(1) Keep driving anyway without a license or with a suspended license.
(2) Use public transit that takes two or three hours one way. When the bus is late again one day they get fired possibly setting off a downward spiral and that's another person on public assistance.
(3) Vote for politicians who allow ridiculous policies like Arizona's lifetime driver licenses.
*Free to motorists, not the taxpayers and real estate consumers.