Restructuring the exercise of prosecutorial discretion in England

Oxford Journal of Legal Studies

Volumn 26, Issue 4, 2006, Pages 775-803

  Rogers, Jonathan ^a  

^a UNIVERSITY COLLEGE LONDON   (United Kingdom)

Author keywords

[No Author keywords available]

Indexed keywords

EID: 33845752628     PISSN: 01436503     EISSN: 14643820     Source Type: Journal    
DOI: 10.1093/ojls/gql027     Document Type: Article

References (155)

1. This is the sense in which I use the term 'prosecutorial discretion' in this article. It necessarily includes decisions to drop or to amend charges or to recommend a caution, but not other discretionary aspects of the prosecutor's job such as trial tactics.
2. See in particular the works of Andrew Sanders, below nn 14, 82, 92, 119
3. and A. Hoyano, L. Hoyano, G. David and S. Goldie 'A Study of the Impact of the Revised Code for Crown Prosecutors' [1997] Crim LR 556.
4. See A. Ashworth and M. Blake, 'Some Ethical Issues in Prosecuting and Defending Criminal Cases' [1998] Crim LR 16
5. and A. Sanders and R. Young, 'The Ethics of Prosecution Lawyers' (2004) Legal Ethics 7 (2) 190.
6. An interesting work for comparative purposes is M. Proulx and D. Layton, Ethics and Canadian Criminal Law (Irwin Law Inc, 2001).
7. See J. Fionda, Public Prosecutors and Discretion (Oxford: Clarendon Press, 1995)
8. and J. Williams (ed.), The Role of the Prosecutor (Avebury, 1988).
9. On the earlier versions of the Code, See A. Ashworth, 'The Public Interest Element in Prosecutions' [1987] Crim LR 595
10. A. Ashworth and J. Fionda, 'The New Code for Crown Prosecutors: Prosecution, Accountability and the Public Interest' [1994] Crim LR 894
11. A. Sanders 'The Silent Code' (1994) 144 NLJ 946.
12. References to 'the Code' in this article refer to the fifth Code for Crown Prosecutors, which was published on 16 November 2004 and is accessible on the CPS website www.cps.gov.uk.
13. In cases where the state would be under a duty to deter a type of offending by punishing those who have been found to commit it, there may be a presumptive obligation to prosecute provable cases in order that the court has an opportunity to inflict punishment. See the discussion of R v DPP, ex parte Manning and Another [2000] 3 WLR 463 (QBD)
14. in Ashworth and Emmerson, Human Rights and Criminal Justice (Sweet & Maxwell, 2001) at 551-2
15. and generally J. Rogers, 'Applying the Doctrine of Positive Obligations in the European Convention of Human Rights to Domestic Substantive Criminal Law in Domestic Proceedings' [2003] Crim LR 690.
16. The possibility that the institution of a prosecution might be 'unlawful' under the Human Rights Act 1998 is briefly canvassed in R v Attorney-General, ex parte Rusbridger [2003] UKHL 38 (HL) at [40], per Lord Scott; but such cases should be few and far between.
17. See further J. Rogers, 'Prosecutors, Courts and Conduct of the Accused which Engages a Qualified Human Right' (2005) 58 CLP 101.
18. But the Serious Fraud Office both investigates and prosecutes, and it is surprising that more thought was not given towards providing for the same arrangement for the new Serious and Organised Crime Agency. See R. White, 'Investigators and Prosecutors or, Desperately seeking Scotland: Re-formulation of the Philips Principle' (2006) 69 MLR 143, who argues that the wholesale integration of investigative and prosecutorial power has worked better in Scotland.
19. Section 5.7 of the Code. All recent Codes have proudly recited the words of Lord Shawcross (then Attorney-General): 'it has never been the rule in this country-I hope it never will be_that suspected criminal offences must automatically be the subject of prosecution' (Hansard, H.C, Deb. Vol. 483, col. 681, 29 January 1951).
20. This is potentially the key gateway to a whole host of difficult ethical issues. It is all very well to say that the prosecutor must assist the defence where assistance is needed (e.g. at disclosure, and in citing relevant authorities which are favourable to the defendant if his own counsel has overlooked them, etc.), which is what is often loosely referred to as the prosecutor's role as 'Minister for Justice'. See Sanders and Young, above n 3 at 194-5.
21. Proulx and Layton note the hidden conflicts, however, noting, for example, that at some point the prosecutor might consider the effects of 'winning' cases upon his career prospects: Above n 3 at 640-7.
22. Subject to the power of the High Court (at the instance of the Attorney-General) to stop 'vexatious prosecutions': Supreme Court Act 1981, s 42(1)(c), and to any legislation which requires the consent of the DPP or the Attorney- General to start a prosecution for a particular offence or type of offending. On (what should be) the criteria for preventing the institution of a private prosecution (rather than taking it over), see infra pp 801-2.
23. The Director of Public Prosecutions (hereafter 'the DPP') may take over any private prosecution (Prosecution of Offences Act 1985, s 6) but he is unlikely to do so purely on account of his own unfavourable assessment of the public interest: See T. Hetherington, Prosecution and the Public Interest (London: Waterlow, 1989) at 159-61 (but see now the fact sheet at http://www.cps.gov.uk/publications/communications/ fs-privatepros.html where the CPS states that it may discontinue a private prosecution which 'falls far below the public interest test').
24. See instead K. Hawkins, Law as Last Resort: Prosecution Decision-Making in a Regulatory Agency (Oxford: OUP, 2002)
25. cf. A. Sanders, 'Class Bias in Prosecutions' [1985] Howard Journal of Criminal Justice 176 at 194-6.
26. Section 5.3 of the Code.
27. Such as the possibility of a successful plea of abuse of process, upon which developing area of law the CPS now offers very detailed guidance to its prosecutors: See http://www.cps.gov.uk/legal/section3/ chapter_e.html.
28. Section 5.1 of the Code. This has always been the rule but not surprisingly there is evidence that in some serious cases the assessment of the evidential test has been exaggerated so as to enable prosecution: see J. Baldwin, 'Understanding Judge Ordered and Directed Acquittals in the Crown Court' [1997] Crim LR 536.
29. Sections 5.8, 5.11 of the Code.
30. The converse may also be true. One imagines that the final anti-prosecution factor, the fact that 'details may be made public that could harm sources of information, international relations or national security' (s 5.10(i)), could by itself outweigh a whole host of 'pro-prosecution' factors.
31. Section 5.7 of the Code.
32. Hart famously defined punishment (of someone thought to be guilty of a recognized offence) as a consequence normally considered unpleasant and intentionally administered with the authority of the legal system (emphasis added): 'Prolegomenon to the Principles of Punishment' in Punishment and Responsibility (Oxford: OUP, 1968) at 4-5.
33. For a discussion of the importance of defining punishment as a consequence which is intended to be experienced unpleasantly by the offender, see G. Fletcher, Rethinking Criminal Law (Boston: Little Brown, 1978) at 409-14. We shall return to this because it will be necessary to appreciate that the experience of being prosecuted, by contrast, should not be (directly) intended to be an unpleasant one (in the sense that the prosecutor should not want or desire this to be the case).
34. See infra pp 782, 787-9 and 792-3.
35. A judge or magistrate 'must' have regard to the reduction of crime (including reduction by deterrence), the reform and rehabilitation of offenders, reparation and the protection of the public as aims of punishment: Criminal Justice Act 2003, s 142. Presumably he may also choose to have regard to public education and communication and other purposes of punishment too. But there is no statutory (or non-statutory) guidance on what aims of punishment should generally be considered by the prosecutor.
36. See Sanders and Young, above n 3 who argue that this is an essential background to discussing a range of ethical dilemmas faced by prosecutors, but who do not demonstrate how it might drive the application of the 'public interest' test.
37. Undoubtedly the institution of a prosecutorial system is wholly consequential. It exists in order to reduce crime. See generally H. Packer, The Limits of the Criminal Sanction (Stanford: Stanford University Press, 1968).
38. But that does not mean that all individual decisions are taken for wholly utilitarian purposes; other factors (such as humanitarian ones) matter too. We might compare this to some extent with the familiar distinction that even if the institution of punishment is justified by deterrence, its distribution in individual cases should still be informed by a variety of other factors. See Hart, above n 21 at 4-5, 231-5.
39. The prosecutor may direct that the defendant be conditionally cautioned, or request the police to administer a caution or an informal (non citable) warning. See further on conditional cautions, infra n 99.
40. See L. Griffin, 'The Prudent Prosecutor' (2001) 14 Georgetown Journal of Legal Ethics 259.
41. Sanders and Young, above n 3 at 207 discuss this work and seem to agree that individual prosecutors should not apply their own 'philosophical substantive theories of justice', although in their opinion some (uncertain) ethical scope seems to be left to 'refuse to prosecute on the basis that the criminal law. . . in question . . . [is] immoral' depending upon the extent of wider consultation upon the matter within the Service.
42. Sanders and Young reject suggestions that prosecutors become 'anaesthetised en masse' to the moral sensibilities and political implications of their work, and regard the size of the CPS as offering realistic opportunities to consider the applications of policies: Above n 3 at 205. This is surely correct. The CPS now runs a Policy Directorate which is split into a Prosecution Policy Division, a Criminal Justice Policy Division, and a European and International Division, all of which are concerned with promulgating policies and responses to governmental and other proposals affecting the criminal law and its practice.
43. There may be an overlap with ethical concerns here (as elsewhere), e.g. prosecutors might be concerned about playing a part in the nation's record of adding to its jails poor people convicted of minor offences, especially if various minorities are over-represented in such cases. See Sanders and Young, above n 3 at 198-9 and generally W. Heffernan and J. Kleinig (eds), From Social Justice to Criminal Justice (Oxford: OUP, 2000).
44. We shall say more of the danger of de facto decriminalization shortly, infra pp 785-7.
45. C. Cobley, T. Sanders and P.Wheeler, 'Prosecuting Cases of Suspected "Shaken Baby Syndrome"_A Review of Current Issues' [2003] Crim LR 93 at 102-5 suggest that sentences are often lenient in such cases, and point towards worrying judicial phrases such as 'not offences of throwing the child to the floor . . . [but] simply shaking a child'.
46. Criminal Justice Act 2003, s 172.
47. The Council is still a relatively new body, but the Sentencing Advisory Panel has already been influential in sentencing decisions which relate to particular offences, many of which pose particular problems in sentencing, e.g. causing death by dangerous driving (see R v Cooksley and Others [2003] 2 Cr App R 18 (CA)) and domestic burglary (see R v McInerney and Keating [2003] 1 Cr App R 36 (CA)).
48. R v Atkinson [1978] 2 All ER 460 (CA).
49. As Fionda points out, above n 4 at 41-6, since the judge remains independent and may receive advice in a presentence report, it is hard to see how he can be unduly influenced by the recommendation or argument for a particular sentence. It is unfortunate that this potentially liberating reform does not appear among the hundreds of legislative reforms made to the criminal justice system almost every year.
50. [2005] 1 WLR 2532 (CA).
51. Attorney-General's Reference (No 52 of 2003) [2003] EWCA Crim 3731 (CA),
52. v Pepper and Others [2006] 1 Cr App R (S) 20 (CA). See too s 11 of the Code.
53. Criminal Justice Act 2003, ss 22-27. For further details, see http://www.homeoffice.gov.uk/docs3/conditional_cautioning_cp.pdf for the Home Office's Code of Practice which was brought into force by statutory instrument (SI 2004/1683).
54. The Court of Appeal is never required to correct an unduly lenient sentence: Attorney-General's Reference (No 20 of 1991) (1992) 13 Cr App R (S) 557 (CA) and sometimes it declines to do so, e.g, in order to encourage work in reconciling the offender with his family since the original sentence: Attorney-General's Reference (No 80 of 2004) [2004] WL 1808935 (CA). Even when it has increased a sentence, it has almost always awarded a discount from what the appropriate sentence would have been in order to take account of the element of 'double jeopardy' suffered by the defendant who is effectively facing sentence a second time for the same offence.
55. See generally R. Henman, 'Attorney-General's References and Sentencing Policy' [1994] Crim LR 499.
56. See the above definition of punishment, Hart above n 21.
57. Moreover, since Crown Prosecutors present most cases in the Magistrates' courts in person, it will often be the same counsel who decided to prosecute_i.e. the person who knows best why the Crown preferred to proceed_ and who must yet remain neutral at the sentencing stage.
58. I am not, however, arguing that the prosecutor should be obliged to argue for a particular sentence or for a particular objective in sentencing to be adopted.
59. Section 5.9 starts with the advice 'The more serious the offence, the more likely it is that a prosecution will be needed in the public interest' before articulating a list of 17 particular factors to be taken into account.
60. This had been recommended in the Narey Report, where it was thought that prosecutors should assume that there was a good reason for prosecuting a minor offence from the very fact that the police had sent the file to the CPS, rather than taking their own decision not to proceed. See the subsequent discussion in the Glidewell Report, n 44 below at 82-3.
61. In The Review of the Crown Prosecution Service: A Report (Cm. 3960 of 1998) ('the Glidewell Report') the authors advised that 'the occasions on which the CPS should discontinue [on account of the lack of seriousness or the expectation of a nominal penalty] should be rare' and suggested that a passage in the third edition of the Code which seemed to provide for a presumption of prosecution in the public interest only 'in cases of any seriousness' should be deleted (which it was, in the fourth edition).
62. Such as the fact that the offence was premeditated or that the victim was vulnerable (ss 5.9(g) and 5.9(i) of the Code).
63. Sections 5.9 (m) and (o) of the Code.
64. Often the word is left undefined, apparently so as to allow the authorities wider powers. It is striking that there is no definition of 'seriousness' in the Serious and Organised Crime and Police Act 2005, notwithstanding that the legislation sets up the Serious and Organised Crime Agency, and it may be all too easy to slip into the mindset whereby any 'organised' crime is thought also to be 'serious' simply by virtue of the number of participants.
65. Criminal Justice Act 2003, s 152(2).
66. Lord Woolf argues that in addition to an offence which is 'so serious that it can only be marked by a significant prison sentence', a judge might also imprison those who are dangerous, or who will not comply with lesser measures, or whose crimes need to be deterred by this measure (the 'clang of the prison door sentence'). See 'Making Sense of Sentencing', the first Radzinowicz Lecture, delivered in Cambridge on 12 May 2005 and reproduced at http://www.dca.gov.uk/judicial/speeches/ lcj120505.htm.
67. The Code even discusses the selection of charge and the mode of trial without reference to resources, and one notes too its euphemistic tone in s 10.1 ('Crown Prosecutors must never accept a guilty plea just because it is convenient': Emphasis added). In this respect, the earlier versions of the Code were more forthcoming. In s 8.1 of the second edition of the Code (in 1992) it was plainly stated that the prosecutor should 'weigh the likely penalty with the likely length and cost of proceedings'. But it seems that the conflicting concerns about the quality of justice and its costs will always be with us. See too Fionda, above n 4 at 55-6 who notes both the taboo nature of costs and that the 1994 Home Affairs Committee 'were particularly concerned with allegations that the CPS had been known to reduce charges on the grounds of economy, despite the fact that this is exactly what the Royal Commission on Criminal Justice had recommended'.
68. We shall say more about the costs in bringing a prosecution (including the humanitarian costs to witnesses who will be forced to testify) in section 4.
69. Perhaps the most telling indication of the bureaucratization which underpins the Code is the fact that the word 'discretion' itself is nowhere to be found (as noted in O. Quick, 'Prosecuting Gross Negligence Manslaughter: Manslaughter, Discretion and the Crown Prosecution Service' 33 Journal of Law and Society 421 at 433. The author notes too that it had appeared in earlier versions).
70. [2002] 1 AC 800 (HL).
71. See too on the inability of the prosecutor to 'suspend' the law in favour of individuals Yip Chiu-Cheung v R [1994] 3 WLR 514 (PC)
72. and cf. the Canadian case R v Catagas (1977) 38 CCC (2d) 296 (Man.CA).
73. See the instructive judgment of Laws LJ in the Divisional Court in R v DPP, ex parte Kebilene and ors [2000] 2 AC 326 at 351. This part of the judgment was not doubted subsequently by the House of Lords, which instead over-ruled the Divisional Court on the scope of judicial review of a decision to prosecute.
74. Per Laws LJ in Kebilene. It was said in Ex parte Pretty that the DPP could promulgate policies as to how he would consider prosecutions for certain offences, but these policies could not go so far as to be equivalent to rewriting the criminal law: Op.cit at [65], [121].
75. See too D. Calvert-Smith and S. O'Doherty, 'Legislative Technique and Human Rights: A Response' [2003] Crim LR 384.
76. Section 2.5 of the Code.
77. [1996] 2 Cr App R 241 (CA).
78. In fact, the Court of Appeal decided that the conduct was not criminal after all, but in this case one could hardly blame the prosecutor for having misapplied the evidential test. The decision of the Court surprised the commentators too, who at first had difficulty in distinguishing it from R v Brown and ors [1993] 2 All ER 75 (HL).
79. Alan Wilson had been sentenced to a conditional discharge for 12 months.
80. The Crown Prosecution Service stated that 'the prosecution was carefully considered and reviewed in accordance with the Code for Crown Prosecutors and our understanding of the current law and in the light of medical reports received': Daily Telegraph, 1 March 1996.
81. The police now refer possible cautions for indictable offences to the CPS, which takes the view that the 'particular circumstances of the offence or the offender' which might justify this are 'likely to be unusual': http://www.cps.gov.uk/legal/section3/chapter_k.html. Again, Mr Wilson's case might be expected to have slipped through this net because it is not obvious what 'the particular circumstances of the offence or the offender' might be. The case for issuing a caution should rather be that no worthwhile aim of punishment would be served by prosecution, but (consistently, at least) police and prosecutors are not encouraged to think in such terms about cautions any more than they are about prosecutions.
82. Though in both cases the harms are of course tolerated or authorized in law.
83. Or even his physical safety. If prosecuted for (say) an indecent act against a child, he may fear vigilante action against himself or against his family from those who will not wait for the trial.
84. There is a whole multitude of related difficulties which might arise while the defendant awaits trial. E.g. he may be less likely to receive a mortgage loan and thus be unable to pursue a plan to buy a house, or may be denied insurance, etc.
85. The judge might make some allowance for the harms of prosecution in sentencing_indeed, he must take into account time during which the defendant was imprisoned on remand when determining his sentence, unless it would be unjust to do so: Criminal Justice Act 2003, ss 240-43. But that measure can only occur when the accused is convicted. It appears that there is no duty of care owed by the prosecutor to the accused (even the innocent accused) to avoid any of these harms. Thus there is no claim in damages where an innocent accused has been detained pending trial, and where the prosecutor has unaccountably delayed in realising that he is innocent and thus procuring his release: Elguzouli-Daf v Commissioner of the Police for the Metropolis [1995] QB 335 (CA). This might now be subject to challenge via s 6(1) of the Human Rights Act on account of incompatibility with Art 5.1(c) of the ECHR; but even if it is, it is significant that there should have been no remedy in domestic law.
86. But it is readily conceded that research shows that prosecuted persons feel as though they are thereby being 'punished'. See M. Feeley, The Process is the Punishment (Russell Sage Foundation, 1979).
87. See the Bail Act 1976, s 4 and Sch 1; Criminal Justice and Public Order Act 1994, ss 25-30. The fact that conditions may be attached to bail and that a court may reconsider a decision to grant bail if new information about the accused comes to light should encourage the prosecutor to oppose bail only where there are particularly serious concerns about the likely behaviour of the accused.
88. Furthermore, even when bail is denied, the state tries to minimize the harms of imprisonment by granting certain concessions to prisoners on remand, though very probably it could try a lot harder. See e.g. N. Padfield, Text and Materials on the Criminal Justice Process (3rd edn, Butterworths, 2003) at ch. 6. In principle, the state ought always to seek to find imaginative ways in which to minimize the suffering of prosecuted persons. The problem seems to be that such developments (including the granting of 'extended bail' at police stations, which allows suspects time to apply for legal aid) are sporadic and not part of a wider strategy in the criminal justice system to reduce the harms of prosecutions. So all we can hope for is that appropriate recommendations will be made when each individual agency is externally reviewed. For example, see the recommendation in the Glidewell Report (at paras 21-22, 26 of the Conclusions and Recommendations) that prosecutors should be involved in the listing of cases (which do not follow the fast-track) in a Magistrates' Court and given a greater say in listings for the Crown Court, so that cases where delay causes particular hardship to the defendant might be expedited.
89. See s 8.9 of the Code.
90. But see A. Ashworth, 'Is Restorative Justice the Way Forward for Criminal Justice?' (2001) 54 CLP 345, who writes at 371 that 'the primary claim made for restorative justice must be that it can bring other desired improvements, particularly in victim satisfaction'.
91. The prosecutor might also agree to drop a charge if the local authority proposes to apply for an Anti-Social Behaviour Order against the defendant in respect of the alleged offence. See paras 8.1 8.6 of the Home Office guidance at www.homeoffce.gov.uk/docs/asbo2.html.
92. See s 5.10(g) of the Code (discussed in more depth later in the text). A. Ashworth and M. Redmayne, The Criminal Process (3rd edn, OUP, 2005) at 185 write that there is 'presumably some notion of mercy or equity' which explains this part of the Code.
93. In this respect, the Prosecution Code in Scotland (accessible via www.copfs.gov.uk) is superior. Section 6(ix) points out that 'In some cases prosecution may have the potential to affect the accused in a way or to an extent which is wholly disproportionate to the gravity of the alleged offence'. In relation to less serious offences, this may influence the prosecutor's decision as to the appropriate action. It will be noted, however that this Code also puts the seriousness (or 'gravity') of the offence at the centre of the decision-making process, and as such can be criticized for the reasons given in section 2.
94. We should recall that the prosecutor, as a 'Minister of Justice', must act fairly when presenting his case, and this might compensate for any disadvantage suffered by the defendant who has no choice but to accept a less than fully competent counsel on legal aid.
95. See s 6.5(e) of the Code.
96. A court should only stop a case where the delay prejudices the defendant's ability to defend himself at trial: Attorney-General's Reference (No 2 of 2001) [2003] UKHL 68 (HL). But it might be proper for the prosecutor to take account of the defendant's worries about being able to defend himself properly even if the court itself is likely to be persuaded that they are unfounded.
97. This point is recognized by the Code. In ss 5.10(e) and (g) it is said that the principle that a prosecution is less likely to be needed on account of the bad effects of delay or upon the mental health of the accused does not apply where 'the offence is serious', cf. too the Scottish Code discussed at n 73.
98. Again, the emphasis should be on finding ways to minimize the effects of these harms on the offender's family. It is possible, for example, for a court to order that the identity of the accused not be revealed if the need to respect the privacy of his or her children outweighs the interest in the freedom of the press: See Re W (Children) (Identification: Restrictions on Publication), [2005] Fam Law 868 (Fam).
99. It might also seem undesirable for pragmatic reasons if it were thought that a person who acquires a large family might generally become safer from the threat of prosecution.
100. Such as the possibility that 'the offence was committed as a result of a genuine mistake or misunderstanding': S 5.10(c) of the Code.
101. Above n 21.
102. See A. Sanders, 'Prosecution Decisions and the Attorney-General's Guidelines' [1985] Crim LR 4 at 10-11 where the author discusses inter alia a case where the police insisted that a prosecution of a theft from an employer should continue, despite uncertainty about the co-operation of the employer, because the prosecution in itself would have a general deterrent effect.
103. The ingredient of 'malice' is established only where the desire to bring the accused to justice (in court) played no part at all in the decision to prosecute: Glinski v McIver [1962] AC 726 at 766 (HL).
104. This (besides other elements of the tort) will normally be hard to prove. See too Proulx v Quebec (Attorney General) (2001), 159 CCC (3d) 225, discussed by D. Layton 'The Prosecutorial Charging Decision' 46 CLQ (2002) 447 at 455-6.
105. On the availability of judicial review of the decision whether to prosecute, see infra pp 798-800.
106. Section 5.10 of the Code; see pp 791-2.
107. One might wish to preface the three stage test with some of the introductory parts of the present Code, e.g. s 2.2: 'Crown Prosecutors must be fair, independent and objective. They must not let personal views about ethnic or national origin, sex, religious beliefs, political views or the sexual orientation of the suspect, victim or witness influence their decisions'.
108. The Supreme Court of Canada has held that the prosecutor himself need not be convinced of the guilt of the accused: Proulx v Quebec, above n 83. Sanders and Young seem to agree (after a wider discussion of the issues) but point out that there might be a difference between the prosecutor who is not convinced of guilt and the prosecutor who is quite sure of the accused's innocence: Above n 3 at 200-202.
109. See R. Daw, 'Prosecution, Accountability and the Public Interest: A Response' [1994] Crim LR 904 at 906: 'If a [biased] tribunal were to exist, its failure to discharge its duty correctly should be addressed elsewhere; it is not for the CPS to acquiesce by wrong decision-making in recognition of fault in another part of the system'.
110. See too the discussion of Ashworth and Redmayne, above n 72 at 180-82.
111. In fact, there seems to be near unanimity about this viewpoint: See the sources given by Layton, above n 83 at 469, n 92.
112. The first Code for Prosecutors in Northern Ireland, published by the new Public Prosecution Service in 2005 (see www.ppsni.gov.uk) aims to be more comprehensive than the English Code in such matters. It expressly instructs prosecutors to assume an impartial tribunal (s 4.2.3) and attaches a further five pages on ethics in s 7. However, like the Scottish Code, it only aims to improve on the English Code in marginal ways: It shares the same structural deficiencies in outlining the exercise of prosecutorial discretion.
113. See the sceptical account in J. Darley and P. Robinson, 'Does Criminal Law Deter? A Behavioural Science Investigation' (2004) 24 OJLS 173.
114. Section 5.10(c) of the Code.
115. The point is widely recognized, e.g. see L. Mills, 'Killing her Softly: Intimate Abuse and the Violence of State Intervention' (1999) 113 Harv LR 550.
116. It has recently been suggested that the majority of witnesses (emphasis added) are potentially vulnerable and that the CPS, largely reliant on the identification of such witnesses by the police, struggles to uphold their interests: M. Burton, R. Evans and A. Sanders, 'Implementing Special Measures for Vulnerable and Intimidated Witnesses: The Problem of Identification' [2006] Crim LR 229.
117. See s 5.10(f) of the Code. The CPS has now developed the 'No Witness: No Justice' programme whereby it is hoped that having a single point of contact for victims and witnesses to receive information and support will maintain their co-operation: See the CPS Annual Report 2003/2004 (The Stationery Office, 2004).
118. If the witness is unavailable to give oral evidence through fear, the trial judge may give leave to admit written (or other non-oral) evidence 'in the interests of justice': Criminal Justice Act 2003, s 116(2)e, (3), (4). Early case law suggests that the courts will be concerned to admit such evidence where it is at all possible to give the defendant a fair trial: R v Sellick [2005] 1 WLR 3257 (CA). But if the accused cannot receive a fair trial, then the witness's interests may effectively force the prosecution to abandon the case on public interest grounds: See s 5.10(i) of the Code.
119. cf. M. Burton, 'Prosecution Decisions in Cases of Domestic Violence involving Children' (2000) 22(2) Journal of Social Welfare and Family Law 175.
120. Indeed, A. Hoyano et al., above n 2 at 558, found that prosecutors routinely consulted Policy Manuals first, and tended to sideline the Code for Crown Prosecutors itself, the 'general principles' of which they thought were already 'absorbed into their bloodstreams'.
121. Section 5.9(q) includes the new 'pro-prosecution' factor that 'a prosecution would have a significant positive impact on maintaining community confidence'; but, if left unchecked, there is clearly scope for this to be used as a way of bypassing some policies which ought to apply everywhere.
122. There may be a strong argument for not revealing policies of 'selective prosecution' to the public, however.
123. They might also be mentioned at the second stage of the test as an alternative way of achieving a purpose of punishment, though it is not yet clear whether a conditional caution will necessarily be more cost-effective than (say) a fast-track hearing at a Magistrates' Court. The truer benefit in conditional cautions should probably be the avoidance of the harms of prosecution; but this is just the sort of debate which is evaded in the present Code which typically conflates the issues by saying only that the conditional caution might serve 'the interests of the suspect, victim and community' (s 8.4).
124. The CPS started to discharge its obligations in 1986. Until then, most prosecutions were started by the police, who were effectively acting as 'citizens in uniform' in respect of starting prosecutions (i.e. they might have been paid for their work but had no extra powers in respect of it).
125. But in R v DPP, ex parte Hogg, The Times, 14 April 1994 (CA), the Court of Appeal indicated that disciplinary action in respect of a proper exercise of his discretion would be judicially reviewable, so his independence is some-what more secure than that of other civil servants.
126. Prosecution of Offences Act 1985, s 3(1).
127. See Hetherington, above n 13 at 37-44.
128. cf. the survey of statements given by successive Attorney-Generals in the Glidewell Report at ch. 13, paras 7-13.
129. See generally its Annual Reports, available on www.hmcpsi.gov.uk.
130. Arguably the prosecutor should be responsible in his capacity as a lawyer to his professional governing body, at least in respect of those functions where he employs purely legal functions, such as assessing the 'evidential test' in good faith, or presenting his case fairly in court. The Bar Council's Code of Conduct (8th edn, 2004, accessible via www.barcouncil.org.uk) does not address the exercise of prosecutorial discretion as opposed to the preparation or presentation of a case, but cf. the responsibility of prosecutors in Canada to their governing bodies for the exercise of discretion in bad faith or for an improper purpose: Krieger v Law Society of Alberta [2002] SCJ No. 45 (QL).
131. R v DPP, ex parte C [1995] Cr App R 136 (QBD):
132. see M. Burton, 'Reviewing Crown Prosecution Service Decisions Not to Prosecute' [2001] Crim LR 374.
133. R v DPP, ex parte Kebilene and ors [2000] 2 AC 326. It looks as though there may be an 'exceptional circumstance' which permits of review (or a plea of abuse of process at the start of the criminal trial) where there is a policy of selective prosecution, and where the decision to prosecute the defendant seems to depart from that policy:
134. v Inland Revenue Commissioners, ex parte Mead [1993] 1 All ER 772 (QBD),
135. R v Adaway [2004] WL 2582629 (CA).
136. Similarly, where a juvenile is prosecuted in apparent contravention of policy: R v Chief Constable of Kent, ex parte L (1991) 93 Cr App R 416 (QBD).
137. "See too R (D) v Central Criminal Court [2004] 1 Cr App R 41 (DC), where an "exceptional circumstance"was held to arise when necessary disclosure of material to a co-defendant would endanger the life of another prosecuted person".
138. Ex parte Kebilene, at 369-70, 393-4. This should in principle apply equally to cases which are to be heard in the Magistrates' Courts: Cf. R (on the application of Pepushi) v Crown Prosecution Service [2004] EWHC 798 (QBD).
139. Above n 4 at 46.
140. The idea that if the status of prosecutors is 'quasi-judicial' then it might be 'partly executive too' is briefly canvassed in Ashworth, 'Testing Fidelity to Legal Values: Official Involvement and Criminal Justice' (2000) 63 MLR 633.
141. See A. Ashworth and M. Blake, above n 3 at 17.
142. See para 10.1 of the Bar Council's 'Written Standards for the Conduct of Professional Work' (2004): 'Prosecuting counsel should not attempt to obtain a conviction by all means at his command. He should not regard himself as appearing for a party. He should lay before the Court fairly and impartially the whole of the facts which comprise the case for the prosecution and should assist the Court on all matters of law applicable to the case'.
143. See Connelly v DPP [1964] AC 1254 (HL) at 1354 where Lord Devlin (in deciding that the prosecutor should not be allowed to proceed with a charge which infringed the spirit if not the letter of autrefois acquit) said that 'the courts cannot contemplate for a moment the transference to the executive of the responsibility for seeing that the process of law is not abused'.
144. See sections 2 and 3 above. Bureaucratization undoubtedly helps prosecutors to process their decisions more quickly, but Fionda notes the research which suggests that some prosecutors always believe that they are overburdened with cases and thus need a model of 'operational efficiency', even though the evidence that the caseload would otherwise be unmanageable may sometimes be missing: Above n 4 at 179.
145. The same is true of any other attempt by the prosecutor to pursue a purpose of punishment which might circumvent a clear policy adopted by Parliament. In R v J (2005) 1 Cr App R 19 (HL), Lord Steyn at para 38 said that 'The CPS as an independent law enforcement agency carries out duties of a public character . . . The powers of the CPS are extensive but not extensive enough to permit it to take decisions intended to evade the clear intent of Parliament. And it is plain as a pike staff that the CPS policy under challenge in the present appeal was intended to circumvent the intent of Parliament in creating a time limit for prosecutions under section 6(1)'.
146. That notwithstanding, one agrees with Ashworth that it would be healthy for the CPS to adopt (so far as is possible) the habit of giving reasons for their decisions, both on public interest as well as evidential grounds, whenever asked to do so: Above n 5 at 605-06.
147. R. Ede and A. Edwards write that it is already 'quite proper to request the prosecution to reconsider the decision to prosecute': Criminal Defence (London: Law Society Publishing, 1991) at 121. Interestingly, their example of a situation where this may be appropriate is where defence counsel is 'in possession of information personal to the defendant (illness, recent bereavement) not available to the prosecution'_that is, an occasion where the defendant wishes the harms of prosecution to be reassessed. One notes with approval that defendants in Northern Ireland are explicitly told that they may submit 'evidence and information to the Prosecution Service for consideration': S 4.3.7 of the Code for Prosecutors.
148. Often, it seems that the police do not pass on information which might affect this part of the 'public interest' test because it is irrelevant to the guilt of the suspect. See A. Sanders 'Incorporating the Public Interest in the Decision to Prosecute' in J.Williams (ed.) above n 4
149. and 'Constructing the Case for the Prosecution' (1987) 14 Journal of Law and Society 229. It will be interesting to see to what extent this situation might improve on account of 'the statutory charging scheme', introduced in section 5 and Sch 2 of the Criminal Justice Act 2003, and covered in s 6 of the latest Code.
150. I am not saying that the weighing of the harms of prosecution should only be a matter between the prosecutor and the defendant. Parliament should still retain some interest in the possibility that its citizens are being prosecuted when they cannot be expected to cope with the process, and pressure might then be brought upon the Attorney-General. The point is rather that this is the only respect in which the prosecutor's primary avenue of accountability might be reflected in the availability of judicial review.
151. See the report of the Law Commission, Consents to Prosecutions (1998), accessible at www.lawcom.gov.uk/docs/lc255.pdf.
152. The Commission proposed a consent provision where 'it is particularly likely that a Convention right might be engaged by a certain offence and where particular harm might be caused by the institution of a prosecution' (an example offered was the Child Abduction Act 1984, s 1) and where 'offences create a high risk that the right of private prosecution will be abused and the institution of proceedings will cause the defendant irreparable harm' (it offered here the example of misuse of public office, prosecution for which might cause 'irreparable' harm just before an election). (Emphasis added). The Commission also thought that offences 'which involve national security or have some international element' might be properly restricted by a consent provision. This is readily explicable because national security (unlike financial outlay) is a 'cost' of prosecution which is borne by the state rather than by the private prosecutor, and the state has a particularly keen interest in controlling it.
153. The Law Commission did not deny that it might be proper for the DPP to take over and discontinue such prosecutions; the point was rather that there need be no bar to private persons from at least instituting prosecutions in such cases.
154. See the discussion above pp 786-7.
155. This may be because of improper motivation, insufficient attention to the harms of prosecution in exceptional cases, or because of an automatic assessment of 'seriousness' in a case where there is little reason to wish to seek punishment.