Keir Starmer: Will you forgive me if I spend some time explaining the current procedure, because that has informed the approach that we intend to take? I hope that it will help put it in a proper context.

The position at the moment is that any private individual can go before a magistrate with information and ask for an arrest warrant. Broadly speaking, what they have to show at that stage is that they are alleging an offence known to law and that what they allege shows a prima facie case, in that the elements are laid out—the magistrate will not see the evidence, but, on a prima facie basis, the elements will be laid out. The court has jurisdiction, which obviously affects the immunity position, that any applicable time limits have been complied with—there is a retrospective element to some of these offences—and that, where necessary, it has the authority to proceed. If it can show all that, the magistrate can then issue an arrest warrant.

Normally, the arrest warrant is to be executed by a police officer, and it is an arrest warrant not to take the person to a police station, but to bring them to court as soon as possible. Once the warrant is executed, that will very often be the same afternoon or the next morning, Sundays and bank holidays excepted. At that stage, the court proceedings begin. The court can remand the individual or bail them and so on.

In all cases within the clause in question—apart from piracy, which is an odd exception, but most of the cases have universal jurisdiction—the Attorney-General’s consent is needed. Very soon after that exercise, consideration will have to be given by the Attorney-General as to whether or not he consents to the proceedings. The timing of the consent is important because consent has to be given as soon as reasonably practicable, and always before the plea. There is some recent case law on that. If consent is not given before the plea is entered, the case becomes a nullity. That is important in terms of the approach that we have taken. It is clear from that that there is a threshold for the magistrate to issue the warrant at the moment. Things will then move very fast because, once in custody the individual has to be put before the court swiftly, as you would expect. Before he or she enters a plea, the Attorney-General has to consider consent.

Most of the cases are indictable cases, so there will be a period between arrest and entering the plea, but it can vary. The Attorney-General will act as swiftly as reasonably practicable. When he comes to consider the case, the principles that he will apply are essentially the principles taken from “The Code for Crown Prosecutors”. Is there sufficient evidence to provide a realistic prospect of conviction, and is it in the public interest? That is the test that we apply to all the cases of public prosecutors. If that test is not made out, he will not give his consent and the case falls.

In temporal terms, there is a short window between the warrant and A-G’s consent. There is another important feature: because the warrant is issued at the behest of a private prosecutor at the moment, there is no power to interview or search the individual or their premises. They are all police powers, only exercisable by police officers. Therefore, the opportunity for anyone to bridge the evidential gap between prima facie or what is needed for the warrant and the test that the Attorney will apply is pretty limited. In an ordinary case, we might expect that the interview would throw something up. You might expect that the search would throw something up, either of the person or of the money. None of that will happen, so within a reasonably short period the Attorney will apply the consent test.

Forgive that preamble, but it informs how we approach the matter. We have approached it on the basis that the offences are obviously serious international offences, and England and Wales should not be a safe haven. That is a general understanding. When there is sufficient evidence, and it is in the public interest, there should be a prosecution and the Crown Prosecution Service would prosecute. We have also approached it on the basis that we should be transparent and accountable, so we intend to publish guidelines on our approach to the exercise. Everyone will then know in advance how we shall deal with the decision whether or not to give consent. I accept that any decision about that consent would be susceptible to judicial review so, if we get it wrong, someone can go to court to challenge us. I would not quarrel with that proposition. I certainly would not advance the argument that it is a non-reviewable decision.

I should add that we have a stand-in, ad hoc committee when individuals or groups with a particular interest in such areas consult us. We meet about every six months. We had a meeting last week. Among others, we have talked to Amnesty about such matters. I shall describe our overall view and what we intend to put in the guidance. It is not detailed, and the guidance will not come out until the legislation is in place. Our much preferred route is that, if anyone wants to pursue a crime of universal jurisdiction as set out in the clause, they should engage us very early in the process. They should come to us with whatever evidence they have, and we will undertake to look at it and to advise. What often happens in such cases is that there is some evidence, but it may or may not be in admissible form. For evidence from abroad to be admissible in our courts is fraught with all sorts of difficulties, so early engagement is necessary.

If there is a case that we think ought properly to be investigated, our preference is that it should be referred by us to the police and that they should then investigate. That allows them to collate the evidence and to exercise all the powers that the private individual cannot exercise. So there is a huge advantage in the case going to the police and we would encourage that. So we say, “Come to us early. We will put people in place. This is an important issue.” We would try to direct people down that route. There will be cases in which people do not want to go down that route, or—this is the most acute problem—in which time is of the essence, and it is simply not possible to do that. There may be a need to go for a warrant for arrest. In those circumstances, we would propose applying the code test, so we would broadly apply the principles of, “Is there sufficient evidence to provide a realistic prospect of conviction?” and, “Is it in the public interest?” If those tests are satisfied, I give consent to the warrant.

So we are now in the “time is tight” category, or the individual may not be prepared to pass this over to the police to investigate. You then get the “time is very, very tight” category, in which even more difficult decisions have to be made. The approach that we propose to take there—if we are talking about a matter of hours—is to apply what in our code we call a threshold test. It is not asking ourselves the question, “Is there already sufficient evidence?”, which is what we would ask in most cases, but “Is this a case in which there is enough to satisfy us that within a reasonable period of time there will be sufficient evidence to provide a realistic prospect of conviction?”

Quite rightly, a number of groups and individuals have said to us, “We may have practically everything. We just need to change the nature of the evidence and it won’t take long. You surely wouldn’t refuse us consent on that basis?” So we have an exception that allows us to apply the threshold test—is there enough for reasonable suspicion and do we anticipate that, within a reasonable period, the evidential gap, as it were, could be plugged? There would then be sufficient evidence for a realistic prospect of conviction. That prompts the question, what is a reasonable period? It seems to us that it is probably best measured in the period between the application for arrest and the likely time that the Attorney-General will consider consent, because that is the existing window. That is the only period that can sensibly be used for that purpose.

It is a long answer, but it comes to this. We will usually apply the code test—evidential sufficiency and public interest. In very acute cases we would apply the threshold test to give a bit of room, but we would have to be persuaded that any gap in the evidence could actually be made up in quite a tight window time-wise. If we are satisfied with that, we will then give consent. As I have said, I would accept the proposition that the decision one way or the other should be susceptible to judicial review, so that if we get it wrong, we can be challenged in the High Court.