Unfortunately, there are enough of those stories to keep the pipeline full. However, selection bias is the problem. Your neighbor probably doesn't realize or even encounter the (most likely) far more numerous cases of "person out on bail / released early goes back to normal life and things are fine."

Same reason I don't watch prime time news unless I know a particular story will be covered (hapy local event things, basically). Sensational sells, happy normal everyday life far outnumbers all those stories but who watches the nightly news to hear "man woke up, went to get his newspaper, and saw the first bloom of the year - much happier today"?

Point of order: If I was going to tell someone a story, I'd probably tell the dramatic one ("guy out on bail murders bus of kittens by driving it into a cancer ward") vs boring ones "a man out on bail went to court and was found not guilty".

Data points are much more interesting when talking about how to create functional systems and allocate resources than some emotional tale. If my town spent millions of dollars buffing up their homicide department when we only see common assaults in 99.999% of the time, I'd be pissed as a taxpayer.

These stories stick in the mind. You won't hear the stories about how something didn't happen, like someone with an expired license was let go and allowed to get home in time to keep their job and therefore didn't end up stealing because they lost their job and had no other way to earn money.

Going to jail can have a huge ripple effect on an individual and their community that isn't accounted for.

The parable of the broken window (https://en.wikipedia.org/wiki/Parable_of_the_broken_window) seems apt here.

Perhaps your neighbor could tell the stories of every person who didn't do that as a counterpoint? It's the only way to be fair.