FORUM on Corrections Research

Reoffending rates for parolees and non-parolees: A five-year comparison

In Ontario, parole decisions about provincial offenders (those with sentences of less than two years) are the responsibility of the Ontario Board of Parole (OBP). One of the questions most frequently asked of the OBP is, "Does parole work?" This question has been asked by high school and university students preparing papers, by journalists writing articles and by standing committees that must report back to the legislature. It is an important question, and it raises the issue of whether parole is a valid correctional program for promoting offender rehabilitation and public safety in Ontario.

Over the years, the OBP has used various types of data to measure the success of its decision making. Statistics on parole revocations have been published regularly as an indicator of success; the lower the revocation rate, the larger the portion of paroles that are completed "successfully." The difficulty with this measure is that revocation rates may have as much to do with parole supervisors' tolerance of various forms of violation or the OBP's willingness to revoke parole as with the actual behaviour of offenders on parole.

It may be argued that a more objective measure of success is whether there is further offending. In 1994­1995, the OBP began to monitor parole cases that had been suspended for serious reoffending. The OBP also has information on reoffending associated with revocations, regardless of the seriousness of the offence, beginning in 1995­1996. These measures of parole success have two problems:

1. parolees are monitored only for the parole period; for Ontario provincial parole, this ranges from around 3 to 16 months, with 6 to 8 months as the average monitoring period; and
2. there is no comparison with offenders who are not released on parole.

In the current study, a sample of parolees and a sample of offenders discharged at the end of sentence were monitored for further reoffending during the two-year period after their release into the community. For this study, "end of sentence" is defined as the two-thirds point in the sentence, the point at which most provincial offenders who have not been granted parole are released without supervision or further reporting requirements and are deemed to have satisfied their sentence. Samples were selected from five consecutive fiscal years to permit examination of relationships between recidivism levels and parole grant rates.

Methodology

Samples of parolees and non-parolees (defined as any offender not released on parole, either because parole was denied or because the offender was never considered for parole by the OBP) were selected from each of the five consecutive fiscal years beginning in 1991­1992, from the Offender Management System database maintained by the Ontario Ministry of the Solicitor General and Correctional Services. The following criteria were used:

• Offenders with aggregate sentences of less than 91 days or more than 729 days were excluded, thereby excluding intermittent sentences and federal cases.
• Offenders with probation to follow were excluded, thereby excluding non-parolees with community supervision on release and parolees with any form of community supervision other than parole.
• Offenders who had been released on parole and had their parole revoked were excluded from the non-parolee group.

The follow-up period for both parolees and non-parolees was defined as two years after release.

"Further offending" was defined as any admission to the Ontario correctional system within the two-year period after release on a warrant of committal, a remand warrant, a probation order or a fine warrant. Some admissions during the follow-up may have been from offences that occurred before the period under study, but it was assumed that this portion would be similar for the two groups.

Results

In each of the five consecutive fiscal years under study, a substantially smaller proportion of parolees was readmitted during the two years after release for further offending than were non-parolees (see Table 1). In fact, for the 1995­1996 samples, parolees reoffended at about 40% of the rate of offenders who served their sentence in custody and were released at their normal discharge date (23.2% vs. 57.4%).

Table 1

Readmissions for Further Offending
	Parolees			Non-parolees
	Number of readmissions	Parolee sample	Percentage readmissions	Number of readmissions	Non-parolee sample	Percentage readmissions
1991-1992	405	1 301	31.1	2 030	3 372	60.2
1992-1993	473	1 495	31.6	1 971	3 383	58.3
1993-1994	520	1 701	30.6	1 802	3 194	56.4
1994-1995	338	1 291	26.2	1 838	3 237	56.8
1995-1996	208	897	23.2	1 852	3 226	57.4

Over the five-year period, the reoffending rate for the parolee group dropped by 8% (31.1% to 23.2%), while the reoffending rate for non-parolees dropped by 3% (60.2% to 57.4%). More substantial differences between the two groups are evident in the last three fiscal years: the reoffending rate of parolees dropped by more than 7% (30.6% to 23.2%), while the rate for non-parolees increased by 1% (56.4% to 57.4%).

To determine whether a relationship exists between reoffending rates and parole grant rates, OBP grant rates were examined for each of the five fiscal years from which samples were chosen (see Table 2). The percentage of parole grant decisions fluctuated over the five-year period, with an increase in 1993­1994 followed by substantial declines in 1994­1995 and 1995­1996. Another notable trend is that the number of offenders considered for parole declined over the five-year period; fewer with short-term sentences were applying for parole and more with long-term sentences were waiving their right to a parole hearing. As a result, 6,599 offenders were considered for parole by the OBP in 1991­1992, compared with only 4,404 in 1995­1996. The decline in numbers of candidates and the decline in grant rates combined to produce a 45% net reduction in parole grants between 1991­1992 and 1995­1996 (from 3,400 to 1,868).

Table 2

Parole Grants
	Paroles granted	Parole applicants	Granting rate
1991-1992	3 400	6 599	51.5
1992-1993	3 427	6 506	52.7
1993-1994	3 833	6 477	59.2
1994-1995	2 748	5 602	49.1
1995-1996	1 868	4 404	42.4

When the figures in Table 2 are compared with the figures in Table 1 showing the rates of readmission for reoffending, a couple of observations can be made. The increase in the grant rate from 52.7% in 1992­1993 to 59.2% in 1993­1994 does not appear to be associated with a change in the parolee reoffending rate (31.6% in 1992­1993 and 30.6% in 1993­1994). However, in 1994­1995, when the grant rate dropped to 49.1%, the parolee reoffending rate fell from 30.6% to 26.2% and, in 1995­1996, when the grant rate dropped to 42.4%, the parolee reoffending rate fell again to 23.2%. Although the non-parolee reoffending rates fluctuate slightly, they do not appear to be associated as clearly with changes in parole grant rates.

Conclusions

In the current analysis, statistics obtained from a provincial correctional database show that, over a two-year period after release, paroled offenders are far less likely to commit new offences than offenders who are released after serving their sentence in custody. This observation holds true for samples of offenders selected from five consecutive fiscal years.

What this study does not reveal are the reasons why parolees and non-parolees differ in this way. An offender file-based study is required to determine:

• how well the OBP distinguishes offenders who are likely to reoffend from those who are not; and
• the effect of community supervision on reoffending.

If we are to learn to reduce the long-term risk to public safety posed by offenders, we need to understand these factors and their relationship to the differences in reoffending rates between parolees and non-parolees.

This analysis compared grant rates and rates of reoffending, expecting to find either that a lower grant rate might lead to the release of lower-risk offenders who are less likely to reoffend, or that a higher grant rate would be associated with higher reoffending rates -- that is, taking chances with public safety.

The current analysis showed that, when grant rates increased in 1993­1994, the parolee reoffending rate did not appear to increase; however, when grant rates fell noticeably in the last two years under study, the parolee reoffending rate also dropped. As follow-up data on reoffending become available for subsequent years, it will be interesting to see whether the relationship continues between grant rates and reoffending rates, given that, in 1996­1997 (for instance), the grant rate dropped to 35.1%.

Getting back to our original question, "Does parole work?", the current analysis suggests that it does, although we do not yet know why. The results of this analysis also suggest that we should be equally concerned about the risk to public safety posed by offenders who are not paroled. These offenders are a greater risk to the community, in that their reoffending rates are so much higher than those of parolees.

---

1. 2195 Yonge Street, Suite 201, Toronto, Ontario.