Earlier quoted context omitted.
I hope one day not to have to say this in one of these debates: Anecdotal evidence is not evidence. At all. Not only is it invalid reasoning but it taints the rest of the debate because it often sounds so convincing, going hand in hand as it usually does with appeals to emotion. This applies doubly when discussing macro-level societal issues like the justice system.
However, anecdotal evidence can point to a line of inquiry that might uncover important statistical evidence. It presents categories: "Crimes not addressed" as a cause of low incarceration. Truly, to combat his assertion, you should find statistical evidence to show that his anecdotes are actually contrary to the trend. Anecdotal evidence _is_ evidence. It is just not strong statistical evidence.
It is never my responsibility to disprove any assertion, as that is fundamentally impossible. It is the responsibility of the person making the assertion to provide a sufficient quantity of relevant evidence.