Social programs are not a cost-effective way to fight crime, because most people benefitted by that sort of spending would not commit a crime in the first place

See Options for fighting crime beyond the Biden plan by Megan McArdle. Excerpts:

"I also believe that, with some targeted exceptions, such as community violence interventions or aggressive outreach on mental illness, social programs are not a cost-effective way to fight crime, because most people benefitted by that sort of spending would not commit a crime in the first place. As criminal justice policy expert Mark Kleiman pointed out in his 2009 book, “When Brute Force Fails,” transferring all government spending on criminal justice to education would increase education budgets by a fraction — perhaps a third. Yet no one seriously thinks that a 33 percent increase in the education budget would eliminate crime, or even hold it to current levels, without police on the street."

"we might be able to get to a better equilibrium using what Kleiman calls “dynamic concentration,” such as hot-spot policing: Pick one neighborhood or group of offenders and flood the zone, promising that any offense will be punished. As crime falls, the virtuous cycle frees up enforcement resources that can then be redirected to the next problem."

"As a share of gross domestic product, total prison spending here is high, but police spending is low compared to many peer nations, even though our homicide rate is dramatically higher — more than twice that of Canada, and more than five times the rate of Germany or Britain. We should have more police, train them longer and pay them better. This could help attract better candidates, for one thing. Also, government generally keeps its wage bill down by compensating civil servants with off-the-books perks like shields against liability or being fired. Pay money, instead, and demand accountability in exchange."