The Central Problem of Criminal Justice Reform
It has to do with data science and incentives
The US is an extraordinary country in many ways, but one place where it clearly is head and shoulders above the rest of the world is in the size of its prison population. The US imprisons more than 2 million people across local, state, and federal institutions.
Indeed, the US imprisons far more people than the typical authoritarian country (defined using the Economist’s Democracy Index). Its incarceration rate is closer to illiberal countries like Rwanda and Russia than to democratic counterparts like France and the UK.
It is no secret that this mass incarceration has a disproportionate impact on minorities and the poor. Over half of Black men in their 30s/40s in Milwaukee for example have been incarcerated. This has resounding implications on their livelihood and fuels the costs of racism.
Obviously, reform is needed and progress is being made. Recently, the US passed the First Step Act, which allows people in Federal prison to secure early release with good behavior among other changes. While laudable this act doesn’t tackle the core issue: most of the prison population is in state prisons and local jails.
Indeed, research has shown that the main driver of the explosion in the prison population is from prosecutors increasing the rate at which they file charges. The main target in criminal justice reform therefore lies in local politics and its incentive structure, but this is where the fundamental problem lies.
Prosecutors face a tradeoff when making a decision to prosecute an individual as there are two types of errors they can make. Prosecutors can accidentally prosecute innocent people (false positive) or choose to not prosecute a guilty person (false negative). Trying to reduce one type of error will lead to increases in the other.
Because prosecutors are judged by local voters mostly on their ability to keep crime low, they have an incentive to file charges and lock people up, even if they are innocent or have only committed a misdemeanor. This increases the risk of them locking up innocent people.
There is also pressure to expand the kinds of things that cause people to go to prison. The easiest way for a prosecutor to build public support is to appear tough on crime.
The same tradeoff exists for other local officials when it comes to criminal justice reform. When thinking about passing laws that could reduce the local prison/jail population, government officials could make two kinds of mistakes: releasing someone who will commit another crime (false negative) or keeping someone in jail who will not commit another crime (false positive).
Just like with prosecutors there is an asymmetric cost to these mistakes. Appearing tough on crime is popular with voters while releasing someone who commits another crime provides stories that the opposition can use to paint elected officials as causing a crime wave (just look at the influence Trump’s stories of violent immigrants have had).
Therefore, local officials often have an incentive to not pass criminal justice reform because doing so and letting out even one person who commits another crime can make their administration look ineffective at keeping the public safe. And even when they do pass reforms there is pressure to revoke or amend them because of fears of increasing crime (which is exactly what is happening with bail reform in New York right now).
This all occurs despite research showing that rising prison populations don’t really explain the drop in crime rates, especially in the 2000s and that keeping people in prison longer can actually increase the likelihood of repeat offenders since prison can teach people how to commit crimes and connect them to criminal networks.
We know that people often believe violent crime is on the rise, despite the fact that it has been consistently falling, suggesting that voters are unable to understand the impact of different criminal justice policies.
The central problem behind criminal justice reform is this incentive structure that voters create for local politicians. They implicitly overweight the cost of not prosecuting guilty people and releasing future criminals much more than the reverse. This signals to local politicians to not care about the cost of expanding prison populations in order to create the appearance of increasing safety.
What we need for true criminal justice reform is for people to better understand these tradeoffs and accept that we can never completely remove crime. Reforming existing policy will lead to some mistakes, but it is clear that on net society will be better off. The current system imposes too great a cost on families and has ruined thousands of lives.
