1. THE RISK OF OFFENDING: WHEN DO EX-OFFENDERS BECOME LIKE NON-OFFENDERS? Keith Soothill Emeritus Professor of Social Research Brian Francis Professor of Social Statistics

2. RESEARCH QUESTION When can ex-offenders with no further convictions be considered as exhibiting the same risk of reconviction as non-offenders?

3. ITS IMPORTANCE? The issue is relevant for the retention and disclosure of early criminal records, and is a controversial issue.

4. THE CONTEXT OF OUR STUDY Replicating American work by Kurlychek and her colleagues (2007), this study focuses on England and Wales.

5. PREVIOUS WORK Kurlychek, M. C., Brame, R., & Bushway, S. D. (2006). Scarlet letters and recidivism: Does an old criminal record predict future offending? Criminology and Public Policy, 5, 483-504. Kurlycheck, M. C., Brame, R. and Bushway, S. D. (2007) Enduring risk: old criminal records and predictions of future criminal involvement, Crime and Delinquency, 53, 1, January, 64-83.

6. THE CONTEXT OF OUR STUDY The controversy relates to the tension between being a more policed society and retaining our freedoms. Greater surveillance v. Human rights

7. EXAMPLE – IN ENGLAND AND WALES In November 2007 the Information Commissioners Office (ICO) ordered four police forces to delete old criminal convictions from the Police National Computer (PNC) for four cases. The cases were now aged over 40, where their convictions were for a single minor offence committed when the cases were under 20 and where there had been no further convictions since (Information Commissioners Office, 2007).

8. THE RATIONALE? The concern of the ICO was that this old conviction information is held contrary to the principles of the Data Protection Act because the information is said to be no longer relevant and is excessive for policing purposes. SHOULD HUMAN RIGHTS OVERRIDE THE POTENTIAL VALUE OF THE INFORMATION FOR SURVEILLANCE?

9. OUR INTERPRETATION OF THE ISSUE Essentially the issue is when are past convictions of a particular age of little or no value in the prediction of future criminality. In other words, at what stage is the likelihood of an offender committing a future offence the same as a non-offender?

10. PART OF THE DESISTANCE DEBATE? However, in all the desistance research, the notion of non-offenders – the group to which offenders might aspire - is rarely addressed. The contribution of Kurlychek and her colleagues is crucial in this respect. In brief, all so-called non-offenders have a chance of committing a crime and being convicted. But what are these chances? Age and sex are usually good indicators. If you are female, you are much less likely to be convicted and being older helps too.

11. BACK TO PREVIOUS LITERATURE Kurlychek, Brame and Bushway (2007) used police contact data from the US 1942 Racine birth cohort to determine whether individuals whose last contact occurred many years ago exhibit a higher risk of acquiring future police contact than do individuals with no contact at all.

12. KURLYCHEK, BRAME AND BUSHWAY (2007) They consider 670 males born in Racine, Wisconsin, in 1942 who are followed until age 32. 349 individuals had at least one police contact before the age of 18 (called the baseline offenders) while the other 321 individuals did not (called the baseline nonoffenders). However, the crucial feature is the apparent convergence between the groups by age 23.

13. POLICE CONTACT HAZARD RATES FOR THOSE WITH AND WITHOUT A POLICE CONTACT BEFORE AGE 17. With police contact Without police contact

14. POLICE CONTACT HAZARD RATES FOR THOSE WITH AND WITHOUT A POLICE CONTACT AT AGES 18, 19 OR 20 (YOUNG ADULTS). With police contact Without police contact

15. SOME LIMITATIONS OF THE KURLYCHEK ET AL. STUDY Use of police contacts means the inclusion in the baseline offenders group of many who would not be involved in court proceedings. Problem of locally-based criminal statistics in the U.S. While there is scope for both the baseline offenders and the baseline nonoffenders to move away from Racine and lost to follow up, it would seem more likely to be an option exercised by those with something to hide.

16. SOLUTION USING A HOME OFFICE DATASET The Home Office Offenders Index provides a comprehensive data source for convictions, being a court-based database of all standard list criminal convictions in England and Wales from 1963 to the present day. We used the cohort data. This is a subset of the Index consisting of six birth cohorts – a sample of all offenders born in four specified weeks (one in each of March, June, September and December) in 1953, 1958, 1963, 1968, 1973 and 1978, with conviction histories recorded until the end of 1999. Here we focus on the 1953 and 1958 cohorts.

17. ADVANTAGES OF OUR STUDY 1. It is a replication in a different cultural context. 2. The two birth cohorts used are more recent than the Racine cohort. 3. The use of two cohorts enables the consideration of possible changes over time within the same location. 4. The numbers are much larger than the Racine study, for this is a national cohort rather than the inhabitants of one small part of the United States. 5. Our study is concerned with court conviction as a measure of recidivism rather than a police contact and thus requires the element of proof in a court of law. Arrest to arrest studies are not what is needed.

18. CONVICTION HAZARD RATES FOR FOUR GROUPS USING THE 1953 COHORT DATA

19. CONVICTION HAZARD RATES FOR THE 1958 COHORT DATA

20. THREE EMERGING PRINCIPLES WHICH STEM FROM THE RESULTS All non-offenders have a risk of being convicted within the next year – from around 1 in 100 at the age of 21 years, around 1 in 200 at the age of 25 years, around 1 in 300 at the age of 30 years and around 1 in 700 at the age of 35 years. This very real but declining hazard was similar for both the 1953 and 1958 cohorts. The three groups with convictions between the ages of 10 to 20 years have differential likelihoods of a further conviction in the first ten to twelve years after their 20th birthday, but then they seem to converge. While they get close, the three convicted groups do not finally converge with the non- offending group.

21. BACK TO THE REAL WORLD The four individuals whose cases provoked the order by the Information Commissioners Office (ICO) to four police forces to delete certain old criminal convictions from the Police National Computer (PNC) had sizeable periods of conviction-free years, but presumably they still feel stigmatised and that their life chances are diminished – why else would they bring their complaints to the ICO? In other words, their criminal past, meagre though it apparently is, still causes them difficulties. In short, is it reasonable to expect that ex-offenders will, indeed, match the risk levels of non-offenders who do not have to battle against such stigmatisation?

22. A PUZZLE AND A POSSIBLE EXPLANATION The puzzle is why the Kurlychek figures seem to produce a convergence between their baseline offenders and their baseline nonoffenders while ours – despite our series seeming to do remarkably well – do not. Methodological? It is likely that more of their baseline offenders than their baseline nonoffenders move out of the town of Racine and thus avoid the follow- up, while we believe that there are fewer such leakages in our two national cohorts.

23. BUT THE MESSAGE IS CLEAR – OR, AT LEAST TO US! Ultimately it is a value judgment whether or when ex-offenders should be treated exactly like non-offenders. However, the evidence might help. The present study presents clear evidence that, if persons remain crime-free for a period of, say, ten years after the age of 20 years, then those with an offence record in their youth and/or early adulthood have a very similar likelihood of a further conviction compared with the non-offending population of their age.

24. OUR BELIEF We believe that it is time to wipe the slate clean for most offenders with convictions after, say, ten years if ex-offenders have managed such a significant crime-free period. There will be some offences, of course, which should never be expunged, but provisions, such as the Sex Offender Registration Scheme, should take care of these.

25. OUR BELIEF(2) We need to remember that the greatest protection of the public is when ex-offenders become like non-offenders and we also perhaps need to be more welcoming when that transition happens.

26. WHAT HAPPENED TO THE ICO JUDGEMENT? The four police authorities appealed against the judgement to the Information Tribunal. The case was heard in April 2008. The Information Tribunal supported the view of the Information Commissioners that the early police records had no predictive value for later offending. The police were told to provide a new policy on data retention. The police appealed again to the court of appeal. This case was heard in July 2009.

27. THE ICO JUDGEMENT AND THE COURT OF APPEAL The judgement of the court of appeal was issued in October 2009. The Appeal was allowed. LORD WALLER: The police are the first judge of their operational needs and the primary decision makers; the Information Commissioner 's role is a reviewing or supervisory one. If the police say rationally and reasonably that convictions, however old or minor, have a value in the work they do that should, in effect, be the end of the matter. The examination of statistics relevant only to the question as to the risk of reoffending was not to the point.

28. T HE FUTURE ? L AST W EDNESDAY IN THE C OMMONS Mr. Amess: To ask the Secretary of State for the Home Department what representations he has received following the Court of Appeal judgment on 19 October 2009 on the retention of individual records on police computers.. Meg Hillier: My right hon. Friend the Home Secretary has asked the newly appointed Independent Advisor for Criminality Information Management, Sunita Mason, to review how criminal records on the PNC are retained and disclosed. She will work closely with the police to develop proposals for a fair and proportionate approach.

29. REFERENCES Soothill, K. and Francis, B. (2009). When do ex- offenders become like non-offenders? Howard Journal of Criminal Justice, 48, 4, 373 – 387. England and Wales Court of Appeal (Civil Division) Decisions. Chief Constable of Humberside Police & Others v The Information Commissioner & Another [2009] EWCA Civ 1079 (19 October 2009) URL: http://www.bailii.org/ew/cases/EWCA/Civ/20 09/1079.html http://www.bailii.org/ew/cases/EWCA/Civ/20 09/1079.html