I have written several times about risk assessment of individual defendants and offenders and the role of socioeconomic factors such as employment, educational level, income, financial situation and housing. Risk assessment is used by criminal justice agencies – in the US, UK, Canada and the Netherlands, as well as other countries – to inform decisions about bail, pre-trial detention, sentencing, probation, parole, treatment, and/or supervision.
- Performance at work
- Housing stability
- Neighbourhood crime rates
- Dependence on social assistance
- High school grades
- Chances of finding work above minimum wage
One powerful argument for risk assessment, according to proponents, is that it would keep people who pose little risk to public safety (how a low risk to commit other crimes) out of prison. In the US, for example, it is said that risk assessment is a way to ending mass incarceration.
My major concern is that including socioeconomic factors in risk assessment – which comes down to treating marginality as a risk factor – results in disparities in how people experience the criminal justice system based not on the crime they have committed but on their socioeconomic status. (in addition, because ethnic/racial minorities are more likely to have a marginal socioeconomic status, it also helps produce ethnic/racial disparities in criminal justice decisions, among which prison sentences.
There seems to be little debate about the question whether this is justifiable – usually debates about risk assessment focus on how well risk assessment tools predict or sort people into different modalities and kinds of treatment.
Here are several arguments against assessing socioeconomic marginality as a risk factor.
It is unconstitutional
Sonja Starr, in an article in the Stanford Law Review, argues that it is (in the US) unconstitutional to include demographic and socioeconomic factors of individuals in their risk assessment in order to inform sentencing decisions about them:
The technocratic framing of evidence-based sentencing should not obscure an inescapable truth: sentencing based on such [risk assessment] instruments amounts to overt discrimination based on demographics [e.g. gender] and socioeconomic status. […]
in fact the [Supreme] Court has specifically condemned the notion of treating poverty as a predictor of recidivism risk. (p.806).
More generally, the Supreme Court has rejected ‘statistical discrimination’: the use of group tendencies as a proxy for individual characteristics—for example saying that a male individual deserves a longer sentence because men on average are more likely to re-offend than women, or saying that an unemployed individual deserves a longer sentence because people without jobs on average are more likely to re-offend than people with jobs.
Statistical discrimination […] violates a core value embodied by the Equal Protection Clause: people have a right to be treated as individuals (p.827).
End mass incarceration?
About the argument of advocates of risk assessment that it would help reduce incarceration rates, Starr says:
But what they do not typically emphasize is that the mass incarceration problem in the US is drastically disparate in its distribution. […] High school dropouts, for example, are 47 times as likely to be incarcerated as college graduates [Sum et al. 2009]. [Risk assessment] produces higher risk estimates, other things equal, for subgroups whose members are already disproportionally incarcerated, and so it is reasonable to predict that [risk assessment] will exacerbate these disparities (p.837).
In other words, it is not likely that risk assessment will keep those of marginal socioeconomic status out of prison, and that is exactly the group that is overrepresented in prisons. To reduce the incarceration rate, we would need strategies that keep those of marginal socioeconomic status –people who are unemployed, uneducated, homeless – out of prison.
The difference SES makes
As a side note: It is sometimes argued that socioeconomic status (SES) is among many factors in risk assessment tools, and that the weight of socioeconomic status is small, compared to for example past convictions. However, as Starr shows, it could make all the difference. Considering a risk assessment tool that is used in Missouri, US, based upon which a score is calculated for each defendant ranging from -8 to 7:
An employed high school dropout will score three points worse than an employed high school graduate—potentially making the difference between “good” and “average”, or between “average” and “poor” (p.813).
Also, referring to the same tool:
A 20-year-old high school dropout can never score higher than “1” on the scale (“average risk”), even if she has no criminal history and no other risk factors and has committed a relatively minor offense (p.841).
Image by julie faith on Flickr