Mandatory Minimum Penalties: Their Effects on Crime, Sentencing Disparities, and Justice System Expenditures

5. The impact of mandatory minimum sentences (continued)

5. The impact of mandatory minimum sentences (continued)

5.5 Estimating the Effects of Hypothetical Mandatory Sentences

Brown (1998) assessed the potential impact of MMS through an analysis of the characteristics of a national sample of 613 offenders released from New Zealand prisons during 1986. These offenders had been either released halfway through their sentences on parole or at two-thirds of sentence on remission. A subsequent two and one-half year follow-up suggested that the marginal incapacitation effects of additional confinement during the post-release period would have been modest. Just five percent of the sample were convicted for a serious offence during the follow-up period. Furthermore, three-quarters of these individuals had been imprisoned for non-violent offences – offences that would not have predicted the serious offences of those who did recidivate nor have warranted additional confinement.

In Franklin County, Ohio, investigators studied the criminal records of 342 individuals arrested for violent felonies in 1973 to determine whether that offence could have been prevented by a five-year MMS for a previous juvenile or adult felony conviction (Van Dine, Dinitz, and Conrad, 1979). Over half of these offenders were found to have no previous felony conviction. According to the authors' calculations, a five-year MMS imposed for violent felonies in the five years leading up to 1973 would have reduced the 1973 felony rate in the county by just 1%-4%. A re-calculation of the figures, using more generous assumptions, yielded a 15% reduction in the felony rate (Boland, 1978). The Ohio study underscored the notion that, each year, many new offenders come to the attention of the justice system. In Michigan, Johnson (1978) further found that only about 24% of those arrested for violent offences had a previous violent felony conviction. Harsh incapacitation policies would be most effective where most crime was confined to a relatively small and specialized (e.g., exclusively violent) offender group.

In a Swedish study of all persons born in 1953 and living in the Stockholm area in 1963, Andersson (1993) estimated the crime preventive effect that would be achieved if a two-year MMS was imposed for any second criminal offence. He concluded that 28% of all sentences for crimes would be prevented, while the prison population would increase by 500%. Such a policy would bring Sweden near the top of the list of nations in terms of its per capita prison population (Mathiesen, 1998). Furthermore, Mathiesen contends that the preventive effect would be far more modest as Andersson's calculation fails to account for undetected offenders, the substitution of prevented crimes by others, the adverse effects of incarceration on post-release behaviour, and the entry into crime by new cohorts of offenders.

One major concern with MMS based exclusively on the commission of a set number of designated offences (i.e., collective incapacitation) is that they are applied to individuals across the entire risk spectrum. Hence, false positives, the false assumption that those receiving such penalties will repeat their crimes if not incarcerated, tend to be problematic. Even if we accept Andersson's very liberal estimates of crime preventive gains from blanket MMS for all second offenders, a 28% reduction in known crimes would require a five-fold increase of Sweden's prison population – an increase from 5,200 to 26,000 inmates. Many low risk offenders would be ensnared in the web of such sweeping legislation, exacting a massive economic and human toll.

The inefficiency and costs seemingly associated with collective incapacitation strategies have led investigators to project the preventive effects of sentencing strategies that would consider individual risk factors. A subsequent study by Andersson revealed that if a MMS was also imposed on the basis of criminal history and prior drug use/offence data, 44.5% of those predicted to be poor risks would actually be false positives. Such selective incapacitation strategies try to refine incapacitation policies and make them economically more feasible.

One of the best known efforts to develop an effective and parsimonious incapacitation strategy was that of Greenwood and Abrahamse (1982) at the Rand Corporation. Using a self-report study on a large inmate sample, the authors classified the inmates into high, medium, and low-rate offenders. They then developed a prediction instrument to identify high-rate offenders. Predictors included age of onset of criminality, substance abuse, and employment history. Their hypothetical scheme involved the incarceration of just the high-rate offenders. The propriety of predictors such as employment history has been questioned (Beres and Griffith, 1998). This issue aside, the authors estimated that the implementation of their prediction scheme could reduce the California robbery rate by 15%, while simultaneously reducing the number of incarcerated robbers by 5%. Their scheme correctly classified offenders as low, medium or high rate just 51% of the time, meaning that prediction errors occurred in 49% of the cases. Reanalyses of their data place the false positive rate (i.e., the rate at which individuals are erroneously predicted to be high-rate offenders) at over 50% (Visher, 1986). Auerhahn (1999) replicated the Rand study using a more current sample of 2,188 California state prison inmates. Using a very similar prediction instrument, her studied improved predictive accuracy to just 60%. Even so, just slightly over a third of those predicted to be high-rate offenders were correctly classified – i.e., were actually found to be high-rate offenders.

Another predictive instrument developed by the Institute of Law and Social Research in Washington identified 200 “chronic offenders” from a sample of 1,708 federal parolees in the US (cited in Mathiesen, 1998). Eighty-five percent of those predicted to be chronic offenders recidivated during a five-year follow-up, while 35% of those predicted to be non-chronic offenders recidivated. The results therefore improved over the Rand study in terms of false positives, but revealed a problematic number of false negatives (i.e., those predicted to be low-rate offenders who recidivated).

Estimates of the potential crime preventive effects, through incapacitation, of several hypothetical mandatory sentencing policies triggered by a violent crime indicate that the benefits would be modest relative to the substantial prison costs. This is due to low reconviction rates on further violent offences during the time these penalties would be in effect. Blanket or collective incapacitation strategies would needlessly incarcerate many individuals not at risk of reconviction. Selective incapacitation strategies aimed at high-rate offenders would be more efficient; however, they are plagued by prediction errors.