Shifting to a cost/benefit approach to spending on crime

When talking about crime and alternatives to incarceration, we often end up talking about particular instances of crime and what to do with that particular person. This framing causes us to miss systemic effects and solutions. Instead of talking about what we do with this incident, like a particular theft by a particular person, we should be thinking about what we want to do about reducing crime in this community over time. This is a way to lessen the number of future incidents and specific people harmed by them. There is a lot of implicit resistance to this approach, perhaps because looking at a particular incident causes us to think about retribution and “just desserts,” rather than improving safety over time for everyone. If we shift our perspective to systemic interventions, we are in the realm of cost/benefit analyses, which take us to a very different place than the current American consensus on punishment. 

Hypothetical: what if I told you that for every dollar spent on jailing a person for stealing, the number of future theft victims would increase by X%, as compared to the lesser number of thefts if we spend that dollar on housing and other services. Would you spend the dollar on jailing the thief? In other words, does punishment have its own value for you distinct from reducing crime/harm in the future, or would you support punishment because you think it works to reduce crime? 

We have a lot of research on types of interventions that reduce crime in the aggregate, and they don’t include incarceration. Here are a few studies on interventions shown to reduce crime: keep foster kids on welfare benefits after they turn 18; increase the number of community organizations; increase access to health care; providing a guaranteed income; invest in community-based mental health services; provide grants to residents for home repairs. Basically the interventions that successfully reduce crime boil down to giving people money, housing, and health care. And yet, our budgets strongly prioritize policing and jails, and the response to crime has been to increase the police/jail/prison budgets at direct cost to social services, housing, and health. For example, check out this Austin budget allocation: 

Our social policy has been to spend the dollar arresting and incarcerating people, even if that is worse for reducing crime than alternatives. People aren’t often given this information in the form of a zero sum choice, which is what it it actually is. Instead, they’re asked whether they want to increase police budgets. Sure, why not, many voters say. But if we have a choice between marginally increasing the law enforcement/jail/prison budget  by 10%, or spending that same amount of money to give a 39% increase to the social welfare budget (for housing and health) (using the budget numbers above), the latter is the better choice, for three reasons: (1) adding marginal police dollars leads to more misdemeanor arrests (which destabilize a lot of lives and fill our jails), not lower crime; (2) dollar for dollar, social welfare investments have been shown to produce a much better return on investment for reducing crime than police investments, while also producing positive social benefits and avoiding the social costs of increased policing; and (3) independently of the first two points, we get more bang for the buck by increasing social welfare funding, since the baseline budgets are low that we are nowhere near hitting marginal returns.

Where does this leave us? If you think that punishing a person for a crime has a high value independent of actually reducing crimes in the future (aka, retribution is important in itself, separate from deterrence), then you might remain in favor of investing vast resources into jailing and imprisoning people for breaking laws, notwithstanding the social costs of punishment and the reduction in resources for health and housing. But if your goal is to reduce harm in the aggregate — fewer thefts/assaults in the future, for example — then you should pay attention to research on what actually works to reduce these things in the aggregate. It’s not about whether a person ‘deserves’ a punishment (so the work isn’t about separating people into deserving and undeserving categories), but about how to reduce crime and harm.

One reason I think this tradeoff hasn’t been discussed more is that people implicitly assume that the punishment serves the purposes of retribution and deterrence at the same time. But if those two goals are contradictory - the thing that achieves retribution will actually increase crime, then we have to work through these tradeoffs.