The guy completely ignores the time and risk factors, namely being short-term-payoff-oriented and long-term-payoff-oriented. This is a subject of microeconomics (not 101, of course - it's a lot of maths, integrals and derivative functions).
Poor people do not save extra money exactly because of being short-term oriented. Drugs, entertainment and having early sex also deal with short-term payoff. Short-term payoff orientation also makes a person more risk-tolerant, which is a common attribute of criminals.
All the given examples are easily explainable with a utility function that takes into account risk perception and time discounting (long- or short-term orientation) and current posessions.
Committing a crime or not is a choice between risky and not risky opportunities. This deals with risk perception. Few posessions make the low-risk choice even less valuable. Fixing or not fixing dents on a car fit into choice between long-term and short-term payoff.
Writing this I suspected that the author hasn't worked with advanced microeconomics' maths. Indeed, Karelis is a philosopher. I haven't seen the book, but by his words ("Econ 101 is to blame") suspect that he doesn't know any more complicated economics. That's the reason he finds it wrong.