This page contains further detail on the large number of references on which this project builds.
There are several main methods used to estimate the costs of crime, of which I focus on the two most prominent: the bottom-up and the top-down approaches. Research using the bottom-up approach attempts to list all the ways in which crime can inflict costs on society; then, it estimates each of these costs and aggregates them. The other main approach goes by several different names: top-down, willingness-to-pay, or contingent valuation. Rather than aggregating up various costs, research using this method relies on surveys of individuals that ask a series of questions aimed at determining the total amount a reduction in crime would be worth.
Prison costs include costs of feeding, housing, and monitoring prisoners, healthcare costs and capital costs. The largest component of prison costs (over two-thirds) is spent on employee compensation. Twenty-six percent of costs go to prisoner living expenses, with medical care accounting for just under half of that component. Utilities, housing, and food make up the rest (with food accounting for just four percent of overall prison expenditures). Capital expenditures also account for only four percent of the prison expenditures.
Incarceration (and other penalties) can reduce crime simply by its threatened use. This effect is called general deterrence, and it has the advantage of potentially reducing crime without increasing costs at all (or even reducing them) – because crime reduction due to longer sentences is concomitant with a reduction (although increase in duration) of incarcerated individuals. The papers referenced below generally find a non-zero, but relatively small general deterrent effect.
While general deterrence operates on the general public through the threat of sanction, specific deterrence acts on the individual offender who has already received a punishment. The relationship between time served and recidivism could be positive or negative. Incarceration could have a rehabilitative effect, which would lead to a negative relationship, or a criminogenic one, which would result in a positive one. Most likely is that both effects are present and either may dominate for different individuals or different sentence lengths. Any empirical study will only detect the net effect, it cannot distinguish the positive and negative components.
While the deterrence effects discussed so far focus on how the threat or experience of incarceration modify current or future commission of crime, incapacitation measures a much more direct effect: the current reduction in crime due to removing inmates from society.
When offenders are incarcerated, they are restricted from performing work that could be beneficial both to themselves and to society. They also lose a number of other opportunities afforded to those who are free, including the right to spend time with loved ones and friends, to eat what they please, to engage in preferred activities, and to go where they would like.
Some commentators believe that inmates lost freedom should not be incorporated into any sort of cost-benefit or welfare analysis. Foregone wages may be included since they could give an estimate of societal and not just offender loss due to gains from employment. Whether to include the impact on families and friends is also a difficult question.
Incarcerating offenders will lower necessary spending on other criminal justice costs, including those for policing, courts, and probation officers. I account for these costs by assuming that a large fraction of offenders would otherwise be on probation if they were not incarcerated and that these costs would be of similar magnitude.
A large number of inmates would likely impose costs on the state through welfare, foodstamps, and other social benefits if they were free. Thus this may be seen as a net savings. At the same time, detention facilities must provide medical and mental health care to their wards. These costs are included in the costs of prisons, so it is likely that on net there is no appreciable cost-benefit effect from social benefits.
One concern often raised about incarceration is its impact on long-term employment prospects. A forward-looking offender should incorporate this into the value of freedom. Diminished job opportunities may also result in higher recidivism rates, which are reflected in a lower specific deterrence estimate.
Rehabilitation is a goal of incarceration that is sometimes put forth. While I do not attempt to quantify rehabilitation in moral dimensions, the economic effects may be quantified in terms of increased job prospects and decreased criminality. As mentioned above, any change in offending behavior due to incarceration will be incorporated into the specific deterrence estimates.
Much has been written about the long-term societal consequences of mass incarceration. These are perhaps the most difficult elements to quantify, and include such phenomena as the promotion of racial stigma, poverty, absent parents, loss of economic mobility, distorted marriage markets for Black women, detrimental effects on children, and increases in juvenile crime. These large-scale, societal harms are certainly of a magnitude that they would significantly impact a cost-benefit analysis. However , in this article I confine myself only to the consideration of policy changes with relatively short-term impacts on incarceration. Because the policy changes discussed in this article do not result in the abolition of the large-scale use of incarceration in society, I assume that these long-term effects will remain relatively unchanged.
An additional cost that has not been included thus far is the dead-weight loss of taxation. This is simply the notion that since taxes distort choices (for example, an individual will work less than she would otherwise if she must pay income tax) they impose a social cost. Since incarceration is paid for out of tax revenue, these costs should be adjusted upward to account for the deadweight loss. Estimates for appropriate adjustment factors range from 25% upward. Some cost-benefit analyses incorporate this cost, while others do not. I leave it out of the current calculation, but it can be easily added in.