News · Opinion
When a Crown Prosecutor talks.
MG v R [2007] NSWCCA 57 is the case where the NSW Court of Criminal Appeal stayed a trial until a Crown Prosecutor other than Margaret Cunneen was appointed to run it. Nearly twenty years on, with prosecutorial conduct back in the headlines, it is worth a fresh read. This is a brainstorm, not a brief.
By Aaron Kernaghan · Published 8 July 2026
Read this first. This is brainstorming. It is strictly my opinion and my reading of a published judgment. It is not legal advice, it is not a statement of fact about any person, and it is not a comment on any matter now before a court. The findings described here are the Court's own findings in the reported case. If you have a live matter, get advice on your own facts.
What did the Court decide in MG v R [2007] NSWCCA 57?
On 5 March 2007 the NSW Court of Criminal Appeal (McClellan CJ at CL, Bell J and Hoeben J) upheld an appeal under section 5F of the Criminal Appeal Act 1912 (NSW) and made a short but striking order. It stayed the applicant's trial until a Crown Prosecutor other than Ms Margaret Cunneen was appointed to prosecute it. The Court did not quash a conviction and did not throw the charge out. It said, in effect, that this trial could go ahead, but not with this prosecutor.
The reason lay outside the courtroom. The prosecutor had spoken publicly about the case, and about the criminal justice system, in a way the Court found displayed a lack of the detachment a prosecutor is required to keep. That, combined with the publicity, meant a fair-minded observer could conclude the trial would not be seen to be fair. The result is now one of the leading NSW authorities on when a court will step in and move a prosecutor off a case.
What did the prosecutor actually say?
Two things, on two occasions. After the first trial of the co-accused, the prosecutor was recorded on radio and television saying she commended "the quality of the police investigation and the fortitude of the victim." An earlier bench had already said, plainly, that a Crown Prosecutor should never comment to the media about a trial they have run.
Then, on 10 March 2005, she gave the Sir Ninian Stephen Lecture at the University of Newcastle. The address walked through the facts of the Bankstown matter, named the evidentiary rulings, and argued that the appeal system lets guilty people reopen their cases on minor points. The Daily Telegraph and The Australian reported it the next day. On 12 March 2005 the Weekend Australian ran long extracts under her by-line, headed "Just say not guilty, bro." The Court read the address as telling the audience that, in the prosecutor's view, the applicant was guilty and his successful appeal had turned on technicalities. That is the problem. A prosecutor is allowed to hold that view privately. Saying it in public, before the retrial, is what crossed the line.
What is the test for removing a prosecutor?
The Court asked whether a fair-minded, reasonably informed member of the public would conclude that the proper administration of justice requires the prosecutor to be prevented from acting. If yes, the court can act to protect the integrity of the process and the appearance of justice. This is not a finding that the prosecutor is actually biased. It is the same logic the law uses for apprehended bias in a judge: justice must not only be done, it must be seen to be done.
The Court built the test from a line of cases. Grimwade v Meagher (1995) 1 VR 446 restrained a former prosecutor from later appearing for a party in related civil litigation. Kallinicos v Hunt (2005) 64 NSWLR 561 set out the inherent power of a superior court to control its own officers, and stressed that the jurisdiction is exceptional and used with caution. R v Khazaal [2006] NSWSC 1353 applied the same principle to defence counsel, and Szabo (2000) 112 A Crim R 215 quashed a conviction over an undisclosed relationship between defence counsel and the prosecutor. The apprehended-bias roots come from Ebner v Official Trustee (2001) 205 CLR 337, Livesey v NSW Bar Association (1983) 151 CLR 288 and Webb v The Queen (1994) 181 CLR 41. The stay itself sits in the abuse-of-process jurisdiction from Jago v District Court (NSW) (1989) 168 CLR 23 and Subramaniam v R (2004) 211 ALR 1.
The connections, as I read them.
The question I was asked is whether there are connections worth drawing. There are several. This is where the opinion label matters most, so treat all of this as a brainstorm.
One prosecutor, one episode, many proceedings.
MG v R is not a standalone case. It is the tail end of a cluster of Bankstown sexual assault prosecutions arising out of a single night in August 2000. The same prosecutor ran the trial of TS, whose conviction was quashed in R v TS [2004] NSWCCA 38, and the joint trial of the four co-accused, dealt with in the anonymised appeal [2004] NSWCCA 36, where the applicant's conviction was set aside over the wrongful admission of a phone call. The same prosecutor was then reappointed for his retrial. So the "connection" is a single prosecutor threaded through multiple linked trials and appeals in one criminal episode. When the appellate history is that tangled, and the prosecutor then speaks publicly about it, the risk to the appearance of fairness is not hypothetical.
A photograph that could travel between cases.
There is a second connection buried in the lecture itself. The prosecutor noted that the same photoboard photograph used to identify the applicant in the Bankstown matter had also been used to identify him in a separate earlier trial concerning an assault at Greenacre. She observed that if the identification ruling went one way, he might try to reopen the other conviction on the strength of it. That is a genuine cross-matter connection, and it is a live theme in criminal practice: a ruling on evidence in one case can ripple into another where the same exhibit or the same method was used. It is exactly the kind of connection a defence lawyer looks for.
The doctrine borrowed from judges and civil counsel.
The most useful connection for practitioners is doctrinal. MG v R takes a test developed for judges and jurors, apprehended bias, and a test developed for restraining counsel in civil cases, and applies both to a Crown Prosecutor in a criminal trial. It ties them to the special duties a prosecutor carries under Whitehorn v The Queen (1983) 152 CLR 657 and R v Apostilides (1984) 154 CLR 563: fairness, detachment, and the job of putting the whole picture before the court. The case is the bridge that lets a defence advocate argue prosecutorial conduct as a fair-trial and appearance-of-justice issue, not merely a professional-discipline one.
The later chapter most readers remember.
The prosecutor at the centre of MG v R, Margaret Cunneen SC, later became a Deputy Senior Crown Prosecutor and the respondent in Independent Commission Against Corruption v Cunneen [2015] HCA 14, where the High Court held that ICAC had no power to investigate the conduct alleged against her because it did not affect the probity of a public official's functions. That is a connection of a different kind. It is not a legal thread from one case to the other, but the same public figure sits at the centre of two very different leading appellate decisions, one on prosecutorial detachment and one on the reach of an anti-corruption body. For a reader trying to place MG v R, that later chapter is the one they are most likely to already know.
Why it feels current again.
The reason I picked this case up now is that prosecutorial conduct in NSW is back in the news. The role of the Director of Public Prosecutions, media contact, and the confidence the public places in the office are being argued over again, as the firm's note on the parliamentary report into the DPP sets out. MG v R is the older bookend to that conversation. It is the case that says, in hard terms, that what a prosecutor says outside court can cost the Crown the carriage of a trial. The two are not legally linked, and I draw no conclusion about the current matter. The connection is thematic: public confidence in the prosecution is a fragile asset, and the courts have shown they will protect it.
What is the impact on future court proceedings?
Here is where I think MG v R earns its keep, again as opinion.
First, it confirms a remedy short of the nuclear options. A defence team does not have to choose between letting a compromised prosecution run and asking for a permanent stay of the whole charge. There is a middle order: stay the trial until a different prosecutor is briefed. That is a proportionate remedy, and the Court in MG v R noted how easily the Director could appoint one of many other competent prosecutors. A proportionate remedy is easier to persuade a court to grant.
Second, and this is the part that has aged into new relevance, the principle does not care about the medium. In 2005 the vehicle was a public lecture picked up by newspapers. Today the vehicle might be a podcast, a post on X, a LinkedIn essay, or a comment to a journalist that runs online within the hour. The reasoning in MG v R applies to all of them. A prosecutor who publicly signals a view on a defendant's guilt, or who publicly aligns with a complainant against an accused, or who publicly criticises a court ruling in a live matter, is doing the same thing the Court condemned, on a faster and wider platform. If anything, social media makes the risk sharper, because the comment is permanent, searchable, and easy to put in front of the court.
Third, it makes a non-frivolous complaint about the prosecutor part of the picture. The Court compared the position to a judge who has been the subject of a serious complaint by a litigant and could not then sit on that litigant's case. It was careful to say that spurious complaints made to manufacture a conflict will not work, and that each case turns on its facts. But a genuine complaint, especially one a regulator has found could amount to unsatisfactory professional conduct, is a fact the court can weigh.
Fourth, it is a caution as much as a weapon. The jurisdiction is exceptional and exercised with caution. Courts give real weight to a party keeping the counsel of their choice, and the Crown has a legitimate interest in continuity, including for a vulnerable complainant. So the case is not a licence to try to knock out any prosecutor a defendant dislikes. It is a remedy for the rare case where the appearance of fairness has genuinely been damaged.
How we would use it.
If a prosecutor in one of our matters had spoken publicly about the case, the first step is evidence. We would gather the exact words, the date, the platform, and the reach. Screenshots, transcripts, and republication all matter. The second step is to test the words against Bar Rule 59 and the DPP media guidelines, and to ask whether they express a view on guilt, on the complainant, or on the correctness of a ruling. The third step is to frame the application around the appearance of fairness, not personal criticism of the prosecutor, because that is what the test is about. The order we would seek is the MG v R order: a stay until a different prosecutor is appointed.
None of this is a substitute for advice on a real file. It is a brainstorm about a case I find genuinely interesting, and a reminder that the appearance of justice is a live issue, not a slogan.
Frequently asked.
Can a Crown Prosecutor be removed from a criminal trial in NSW?
Yes, in exceptional cases. In MG v R [2007] NSWCCA 57 the Court of Criminal Appeal stayed a trial until a Crown Prosecutor other than the one who had spoken publicly about the case was appointed. The test is whether a fair-minded, reasonably informed member of the public would conclude that the proper administration of justice requires the prosecutor to be prevented from acting. The jurisdiction is exceptional and used with caution.
What is the test for restraining a lawyer from acting?
It comes from cases such as Grimwade v Meagher (1995) 1 VR 446 and Kallinicos v Hunt (2005) 64 NSWLR 561: whether a fair-minded, reasonably informed member of the public would conclude that the proper administration of justice requires that the lawyer be prevented from acting, to protect the integrity of the judicial process and the appearance of justice. MG v R applied that test to a Crown Prosecutor.
Can a prosecutor comment to the media or post about a case?
A prosecutor is bound by Bar Rule 59 and the DPP media guidelines, which prohibit publishing material about current proceedings and commenting on the correctness of a court's decision. MG v R treats a breach as capable of undermining the appearance of a fair trial, whether the comment is made to a journalist, in a lecture, or, on the same logic, on social media.
Does a complaint against a prosecutor mean they must step aside?
Not automatically. MG v R says spurious complaints made for an ulterior purpose are always a possibility and each case turns on its own facts. But a complaint that is not frivolous or vexatious, especially one a regulator has found could amount to unsatisfactory professional conduct, can support a reasonable concern that the prosecutor may not keep the required detachment.
Is MG v R [2007] NSWCCA 57 still good law?
It remains a leading NSW authority on removing or restraining a Crown Prosecutor and on staying a trial to protect the appearance of fairness. This article is opinion and general information, not legal advice. Whether it helps a particular case depends entirely on the facts of that case.
To speak to a criminal defence lawyer, call Aaron Kernaghan on 0421 717 019.
This article is written by Aaron Kernaghan, criminal defence lawyer at Good Legal Lawyers, as at 8 July 2026. It is commentary and reflects the personal opinion of the author only. It is not legal advice, it is not a substitute for advice on your own matter, and it does not purport to be accurate or complete. Any matter referred to may be before the court and any person charged is presumed innocent. For advice on a specific matter, contact Aaron Kernaghan on 0421 717 019 or contact the firm.
More criminal-law commentary by Aaron Kernaghan.