Justice Committee
Oral evidence: Proposed changes to the Totality sentencing guideline, HC 1188
Tuesday 7 March 2023
Ordered by the House of Commons to be published on 7 March 2023.
Members present: Sir Robert Neill (Chair); Janet Daby; Maria Eagle; Dr Kieran Mullan; Edward Timpson.
Questions 1-39
Witnesses
I: Professor Andrew Ashworth CBE KC (Hon), Emeritus Vinerian Professor of English Law, Faculty of Law, University of Oxford; Professor Mandeep Dhami, Professor in Decision Psychology, Middlesex University London; Dr Rory Kelly, Lecturer in Criminal Evidence and Criminal Law, Faculty of Laws, University College London.
Witnesses: Professor Andrew Ashworth, Professor Mandeep Dhami and Dr Rory Kelly.
Q1 Chair: Welcome to this session of the Justice Committee and to our panel of witnesses, to whom we will turn in a moment. It might just be worth saying to those who may be watching now, or watching a recording of our proceedings, that this is the first time we have had an evidence session in relation to a Sentencing Council guideline. Normally we are statutory consultees on the guidelines and make written responses based on, in effect, a desktop exercise and our own discussions. But today we thought it would be helpful for our panel of academic witnesses to come and give evidence because we want the session to inform the work we are doing in relation to the Sentencing Council guidelines.
Under the Coroners and Justice Act 2009, the Sentencing Council must consult the Justice Committee about draft guidelines. It might be worth quoting our predecessor Committee in 2009: “Parliament sets the framework for sentencing in legislation. Sentencing guidelines are a key element to how this legislation works in practice. It is vital that Parliament, representing the public voice, contributes to sentencing guidelines as they are produced and in doing so identifies the crucial issues of public confidence and the effectiveness of sentencing.” I think that has always been recognised by the Sentencing Council itself in its successive memberships.
Today we are looking at the Council’s proposed revisions of the totality guidelines. The Council launched its consultation in October last year. That closed in January, and the Council kindly shared with us the responses it received. Each witness today has submitted a response to that consultation, which we have seen and have therefore been able to read. The Council has always been very co-operative with us in these regards. After today’s session, we will write a letter with our response to the Council, which we will publish on our website. We are grateful for its, and our witnesses’, co-operation in this.
Recently we have been looking, by way of another inquiry, into public understanding of sentencing and the guidelines. Some of the changes to the guidelines we are considering today rightly highlight the importance of the public’s understanding what can, to the layperson, be quite technical and complex matters. That is why we are adopting this novel approach.
We now have to make our declarations of interest. I am a non-practising barrister and former consultant to a law firm.
Maria Eagle: I am a non-practising solicitor.
Edward Timpson: I am a former Solicitor General. I have a current practising certificate but am not currently taking on any casework.
Chair: Will our panel of witnesses introduce themselves? Shall we start with Professor Ashworth and work along the panel?
Professor Ashworth: I am Andrew Ashworth, a retired professor of law at Oxford, where I taught sentencing among other things.
Professor Dhami: I am Mandeep Dhami. I am a professor of decision psychology at Middlesex University in London. I specialise in how people think and make decisions, primarily in the criminal justice system.
Dr Kelly: I am Dr Rory Kelly from the UCL Faculty of Laws, where I lecture in criminal law and criminal justice topics.
Q2 Chair: Professor Ashworth and Dr Kelly, you are also the authors of a quite well-known textbook on sentencing and criminal justice that was published a while ago.
Can I start by asking a question of you all? The principle of totality predates the introduction of sentencing guidelines. I remember when I was in practice—well before the sentencing guidelines—that judges would talk about totality, so of itself it is a longer-standing concept. It might be worth explaining, for people who are not lawyers or practitioners in the field, your description of the principle of totality, and how it came to use in the English and Welsh courts.
Professor Ashworth: We are dealing here with people who are sentenced on one occasion for more than one crime, so it is about multiple offending. There are two forms of dealing with that. One is concurrent sentences—sentences that are served alongside one another—and the other is consecutive sentences, where the sentences are served one after the other. The totality principle comes in because it is assumed that the judge will have an overall appreciation of the sentences; whether they are concurrent or consecutive, the judge will be interested in looking at the overall sentence.
The totality principle, which, as the Chairman said, has quite a history, is in fact a principle of mitigation, because it comes in at the end of the sentencing process—it comes in before it, too, but certainly at the end—to ask whether the sentence is fair and proportionate as a whole. That usually means, “Is it too much?” I don’t know how much detail to go into at this stage, because we are just starting off, but it is important to bear in mind that this is in effect a discount for bulk offending.
Chair: I was going to come to that.
Professor Ashworth: It is a discount, because if a person commits multiple offences and they are sentenced on one occasion for all those offences, then inevitably, some of the latter offences, if I can put it that way, will be sentenced much lower than they would be if they stood alone. The contrast is with a person being convicted one day for one offence, and then the next day for another offence. If they are all separate, that is a different matter, but we are dealing here with the sentencing of multiple offences at one sitting, and that requires the judge to take an overall view.
Q3 Chair: Dr Kelly, you wrote the book with Professor Ashworth. Isn’t part of the point that you make in your book that that is partly because the law may provide for a number of offences to cover the same set of facts—the same bit of criminality—and therefore you have to have a system that recognises that? In other instances, it may be that, as a matter of law, there is one offence where you might think of it otherwise as two or three. That is why I think you say that you cannot have a purely arithmetical approach to totalling things up, because otherwise you get something that is wildly out of proportion and the issues, which you might talk about, of inter-offence proportionality and intra-offence proportionality. Could you just explain what those are?
Dr Kelly: Yes, of course. This turns from what the principle is to why we have it. The core reasons have really stayed the same from the early case law in the ’70s until now. Some offences overlap. Take robbery and possession of a firearm. The possession of the firearm can be dealt with under the robbery guideline. The risk is that if you deal with it under the robbery guideline and then you deal with it again when you are looking at the possession offence, you have a sentence that is too high—you have counted some factors twice.
The other reason for this principle is the intuitive sense of fairness that large clusters of smaller-level offences should not receive higher total sentences than more serious ones. Take one theft for which, on its own, a fair sentence might be six months. The fact that you have 20 of those thefts after a spate of offending doesn’t mean that the total should be higher than that for a serious assault or a serious sexual offence. It doesn’t feel right.
Q4 Chair: The bulk discount that Professor Ashworth referred to goes back to the Barton case in 1972, which was unreported. That even predates my starting in practice, so it must be a while ago. Part of the way this works is through a decision whether you make the sentences concurrent or consecutive when assessing what the totality principle should be. Can you give some examples of how sentencers use concurrent or consecutive sentences?
Professor Ashworth: The core of concurrent sentences is when the crimes can be seen as a single incident—or a single transaction, as some of the cases put it. We have already had one mention of that from Rory—the robbery. If you do a robbery and there is also possession of a firearm, then the sentences should be concurrent. If the two different offences are charged, the firearm should be concurrent with the robbery, because the robbery will have as an aggravating feature the carrying of a firearm. Concurrent sentences are generally single incident, but they are also presentational sometimes. In other words, on some occasions, the court wants to emphasise the seriousness of all the offences by making the sentences concurrent.
A good example of that is a case called Jenkins in 2009, where the court found itself sentencing someone for five armed robberies of security vehicles. The judge gave 25 years for that. That meant 25 years concurrent on all five charges, which of course is 25 years in all, because they are served concurrently. An interesting feature of that case is how the court came to 25. Where did that figure come from? In reviewing it, the Court of Appeal said that it thought 21 was the right figure, but no reasons are given. This is the totality principle in action. The totality principle is saying, “Overall, we think this is worth 21 years,” but there was no clue as to how the calculation had taken place.
That is concurrent sentences. With consecutive sentences, the core is unrelated incidents; whereas the core of concurrent sentences is related single incident, here the core is unrelated incidents. Also, when there is a distinct time factor, you would expect consecutive sentences. A good example of that is the case of Hartley in 2012, where a number of defendants were prosecuted for counterfeiting banknotes. They were placed on bail and while on bail they set up another counterfeiting of banknotes, so when they came to trial there were two sets of counterfeiting offences. The court dealt with that through consecutive sentences to mark the fact that they were entirely separate incidents and also, to some extent, in defiance of the courts.
Q5 Chair: In layman’s terms it is saying, “Well, you made it worse, because you did it while you were on bail for the first lot.”
Professor Ashworth: Yes.
Chair: Similarly, with the robbery, “It’s worse because you are a professional robber, because you have committed a string of offences, hence I go for the upper end of the scale.” Is that a fair way to look at it?
Professor Ashworth: It is a way to look at it, yes, but it does not get us to the difference between 21 years and 25 years.
Chair: No, exactly. That is the bit that was not explained—what the upper end of the scale should be. That is helpful.
Professor Dhami: Can I just add in terms of concurrent and consecutive that you can have concurrent sentences in situations where the offences are not in one incident? Imagine shoplifting incidents where you have different harms, different amounts stolen from different shops, and there are different victims. They do not even happen on the same occasion, yet the offender will serve their sentence concurrently rather than consecutively. It is a little bit more complex.
Q6 Chair: Do you think that leads to significantly different sentences for those with multiple offences as opposed to single offences?
Professor Dhami: I have done research looking at that. Would you like me to tell you a little bit about that?
Chair: Yes please.
Professor Dhami: I published a study in the British Journal of Criminology in 2021 in which I analysed sentencing data that was collected by the Sentencing Council through its Crown court sentencing survey. That data was collected in 2015 and released in 2018. The datasets contain details of sentences given out for groups of offences that are covered by specific guidelines. In those datasets, we know the sentence that was received for each offence, but also the seriousness of the offence, the aggravating and mitigating factors, any guilty plea reduction, previous convictions and so forth—all those sentencing-relevant factors—as well as the age and gender of the defendant.
I took the most common offence type from each dataset, so I looked at a variety of offences, and then I compared the sentences given to the multiple-offence case versus the single-offence case, controlling for the seriousness of the case and all those sentencing-relevant factors and the age and gender of the defendant. I found that there was no difference in the sentence that was meted out, in terms of either the defendant’s likelihood of getting immediate custody or some other disposal, or, if they got immediate custody, the length of custody they received. There was no difference in six out of the seven offence types that I looked at. There was no difference if you had committed one ABH versus an ABH plus other offences, and also no difference for dangerous driving, domestic burglary, fraud, robbery, and shoplifting.
There were two exceptions. One was possession with intent to supply, where the odds of immediate custody were just over two times greater if you were a multiple-offence case compared with a single-offence case. The other was robbery, where even though your odds of receiving an immediate custody was not greater than if you were a single-offence offender, if you received immediate custody and you were a multiple-offence offender, then your odds of getting a sentence of over three years were over two and a half times greater.
For the most part, we are not finding evidence of substantial difference in the sentences received for multiple-offence cases, which links back to what Andrew was saying about the bulk discount.
Q7 Dr Mullan: Thank you very much for coming. I found your written evidence really helpful. Following on from that point you just made, the extract I have says: “I found that for the large majority offence types examined an offence in a MO case received the same or a less severe penalty than its counterpart in a SO case.” Is that not saying that you did find a difference?
Professor Dhami: No. I meant there was no difference. If you were a multiple-offence case person, you did not get a longer sentence; it was the same.
Dr Mullan: Okay, but what we have written in front of us—I think it is a direct quote.
Professor Dhami: I think I meant “less severe” in the sense that one would expect a more severe penalty, so people did not get that. What we did not know in this dataset was whether those sentences were being served concurrently or consecutively.
Chair: I see. You don’t have that.
Q8 Dr Mullan: That brings me on to the question about the data. I will come back to sentences, but we are talking about data. You described really well the different scenarios in which people might be sentenced differently, but do we have an understanding of what proportion of multiple-offender sentences relate to multiple separate offences or to multiple offences within the one course of crime? Do we know how it is being used in reality?
Professor Dhami: In terms of the applicability of the guidelines, the 2015 Crown court sentencing survey main dataset shows that 47% of those cases were what we call multiple-offence cases, and that figure is comparable with the 49% that I uncovered when looking at specific offences. It is also compatible with the Sentencing Council’s own survey of their 130 sentencers who said, for the most part, that either half or more of their case load was made up of multiple-offence cases.
Dr Mullan: So it is a big proportion.
Professor Dhami: A significant proportion, and it is basically common court business.
Q9 Dr Mullan: That is really helpful, but I guess my question was about that 47%. Do we know which ones are single courses of criminal action leading to multiple charges or multiple separate offending incidents?
Professor Dhami: Yes, there are certain categories of offences that are more likely to be in multiple-offence cases than single-offence cases. The general categories of offences would be fraud, money laundering, driving offences, sexual offences, possession of indecent photographs of children. When we are looking at very specific types of offences, I found that 50% or more of cases involving dangerous driving, fraud, threats to kill, possession of an indecent photograph of a child, possession of drugs with intent to supply and shoplifting were multiple-offence cases. There are specific types of offences that are more likely to be part of a multiple-offence case.
Q10 Dr Mullan: That is also illuminating, but I guess my question was specifically about—there is no pressure for you to know the answer to this, because it might be that it does not exist as a dataset, which is what I might be getting at—whether you know, in terms of what you explained at the start, what proportion of these case involve a single course of events with multiple offences, so the burglary with a firearm, or three or four totally separate incidents. Do we know that?
Professor Dhami: That is not in the data, unfortunately.
Chair: I think Dr Kelly wanted to come in on the data.
Dr Kelly: It was on the earlier point about the data on concurrent as against consecutive sentence, and the effect on total sentence length. The choice should be one of form. The substantive principle is that the overall sentence should be proportionate, and there is no inflexible rule on whether a concurrent or consecutive sentencing package is used to get to that endpoint. However, there is a structural risk in the guidelines in that a choice that should be one of form becomes one of substance. That is step one under the present guideline. At step one, if the draft guideline is accepted, a judge is tasked with sentencing each offence individually. At step two, they do and will bring that work together and ask, “Is this going to be consecutive or concurrent?” The time at which they are asking the question, which ought to be one of form, sets your starting point for the sentence, and that can have a huge effect. Will it be five plus three equals eight, or will it be five to include three for five?
Q11 Dr Mullan: Thank you for raising that, because it was my other question. What flows from all your evidence is the complexity of the situation, how variable it is and how it is down to judgment. When I read the guideline, what really struck me was the sentence that says: “It is usually impossible to arrive at a just and proportionate sentence for multiple offences simply by adding together notional single sentences.” To me, that seemed a very strong blanket statement that speaks to your suggestion that people get a discount; it is almost encouraging that mindset. Is it usually impossible to arrive at a judgment by adding sentences together?
Professor Ashworth: If you believe in the principle of totality at all, you have to accept the principle that just adding all the sentences together is not going to cut the mustard, because it will inevitably get too high for inter-offence comparison; it will not be proportionate. For example, you might get 11 or 12 years for 10 burglaries, but if you think of what 11 or 12 years is, we are in the territory of armed robbery and manslaughter—really serious offences. That is why many members of the public—perhaps Mandeep will deal with this aspect of public opinion—believe that the totality principle has some mileage.
Q12 Dr Mullan: For me, it is often determined by what examples you use to illustrate the case. I will use a constituency example of Barry Bennell, the prolific sex offender who is in all likelihood responsible for hundreds and hundreds of extremely traumatising sexual offences towards young people that have ruined their lives. I imagine the public might say that the maximum limit for any one of those offences will really not cut it. That is an example of where it probably is possible to have a just and proportionate sentence while not taking into account the idea that you do not sentence consecutively.
Professor Dhami: Part of the problem—this is interrelated—is when you calculate whether a sentence is just and proportionate, and when you decide whether it should run concurrently or consecutively. At the moment, you are trying to think about how the sentence should be run after you have decided on these individual sentences, and then you are making another adjustment; you are adjusting upwards for a concurrent sentence and downwards for a consecutive sentence. There is some question about that. I don’t think the Sentencing Council has really considered in their revision of the guideline when this decision should happen. When it comes to public understanding, ultimately your constituents will want to know what time the person is serving, rather than what is notional and written on a piece of paper. You might then think, “Well, maybe that decision should come before.”
Q13 Dr Mullan: Does any of you want to touch on any strengths and weaknesses of this approach that you haven’t already mentioned?
Professor Ashworth: If we go back to the Coroners and Justice Act, which the Chair mentioned right at the beginning, we note that one of the Sentencing Council’s duties is to produce a guideline on any rule of law as to totality of sentences. My question is: have they produced a guideline? I don’t think repeating “just and proportionate” is a guideline. It’s just a phrase; there are no numbers. We need numbers if we are to have a guideline.
Of course, you get the answer coming long and loud from the judiciary that this is all a matter of experience and you cannot reduce it to numbers, but that was exactly the argument that was run at the beginning of guidelines—and look how far we have got. It is greatly to the credit of the Sentencing Council that we have such developed guidelines, but why have they not done anything in the totality area?
Q14 Chair: That context is that there was objection to guidelines at all when they first came in, wasn’t there? It was a matter of the experience and feel of the judge in the individual case, and there was a concern that there would be tramlines. That is the history, isn’t it?
Professor Ashworth: That’s exactly it.
Chair: Ms Eagle, do you want to come in on this point?
Q15 Maria Eagle: I am interested in knowing whether the application of the principle of totality affects the public’s understanding of sentencing. We have begun to touch on that. While watching the TV yesterday I was struck that, when we heard some sentencing remarks in a particular case, the judge—I am not criticising her in any way—referred to totality. I am a lawyer by trade, but I was struck by the fact that if I was a member of the public who wasn’t legally trained, I might wonder what on earth she was talking about. Does it affect the public’s understanding of sentencing? Professor Dhami, do you want to respond?
Professor Dhami: Yes, and Rory and I would then like to continue on the limitations of the current approach, because it is kind of related. I think what Andrew is getting at is that the current guideline and the current approach to applying the totality principle or even sentencing these multiple-offence cases is vague, so it is unclear how the sentence has been reached. It is unclear why a person is serving a concurrent versus a consecutive sentence, because, as has already been laid out, it is not always one incident; there could be multiple different incidents.
It is interesting that I am not aware of any research that looks at the public’s understanding of the totality principle. The Sentencing Council commissioned a survey in 2022 and a few years ago, looking at public knowledge of and confidence in the criminal justice system and sentencing, but they did not ask about multiple-offence cases and totality, so they missed an opportunity. Given that about 50% of sentences are for multiple-offence cases, that is a shameful oversight.
Dr Kelly: I will come back on both questions, because I think the risk to public understanding is a possible weakness here. I will start a bit more positively than my colleagues, with a possible strength of the principle at least—if not of the guideline in its present form. It is right that we have some means to check a merely arithmetical approach that could result in ridiculous sentences. If a person is brought forward with 30 pretty low-level thefts and is told, “It’s going to be 15 years,” where is the justice? At its core, it is trying to do the right thing, but in its present form at least the guideline suffers from real issues of vagueness.
Professor Ashworth is right to refer to the 2009 Act, which in fact says that in preparing the totality guideline the Council is to be mindful of consistency. Well, if it is to promote consistency of approach, it cannot rely on broad phrases like “overall criminality”, “single incident” or “multiple incidents”, or there being “no inflexible rule” about how you structure a sentence. That is not consistency of approach; that is, perhaps, carte blanche.
How does this relate to public understanding? I share your intuition, Ms Eagle, that there is a lot here that can be misunderstood by non-experts. Some people are told, “You get two years plus three years, and you get five years total.” Others are told, “You get two years plus three years, so you get three years.” On its own terms, a concurrent sentence can be hard to get your head around. When we have concurrent sentences on the one hand and consecutives on the other, the issue can get all the messier. Professor Dhami is, of course, right; we just need the data.
Q16 Maria Eagle: Thanks. I noticed that analysis by the Attorney General’s Office of cases that it referred to the Court of Appeal in 2022 showed that in almost half—32 out of 67—it raised issues of how totality had been addressed. Does that indicate that there is an issue even among the judiciary regarding how they ought to address this in their sentencing decision making? What is your response to that figure? Does that fit in with your own work?
Professor Dhami: The timing is a little bit off, but I am currently working on a project that is looking at multiple-offence cases that have been appealed—sent to the appeal court—before this guideline was introduced, and directly after the guideline was introduced.
We are looking at the reasons for appeal, the outcome of the appeal, and the issues around totality that have been brought up in those offences. Did the guideline help, or did it muddy the waters? It is looking at that. We know there are some differences before and afterwards.
At the moment, looking at this very briefly, as this is still preliminary because we have not finished collecting the data, we have about 70 cases before and 70 cases after this guideline. In 100% of those cases after the guideline was introduced, when the sentence was considered too lenient, it was allowed, compared with 29% when it was considered too excessive.
There is no such relationship before the guideline, so something different is happening in the appeal court, and the cases coming to appeal after the guideline was introduced. But, as I said, it is quite early days and we have not yet finished that project.
Q17 Dr Mullan: It strikes me that this area perhaps suffers the most because it can least rely on precedent, which we see in single offending, and the fact that there is not a serious or consistent attempt to measure whether our justice system is actually delivering justice. I recognise that is difficult, because it is subjective, but does that mean we should not try in a systematic, ongoing way?
Some of you have done studies, and been party to studies, that happened to take place, and might show some insight, but the justice system, I noted with the annual report from the MOJ, talks nothing about justice, and whether it is actually delivering justice.
It talks about whether someone has been prosecuted or sentenced, but there is no attempt to grip the issue of whether our justice system delivers justice. Because we do not have that framework, when you come to something complex like this, there is nothing for anyone to hang their hats on.
Professor Dhami: Imagine someone said, “Yes, we want to find out, do we deliver justice?” They cannot because the data is not there. That is because the official sentencing statistics refer only to the sentence given out in what they call the principal offence. That is the offence in multiple-offence cases. That is the offence that received the highest penalty, or in the cases of ties, the offence that would carry the highest maximum penalty.
Even in the Sentencing Council’s own research, where there was the opportunity to find out what happened in those multiple-offence cases, even they did not collect data on the other offences and what sentences were given to the other offences, and even whether the sentences were served concurrently or consecutively.
For the vast majority of sentences that are meted out, we do not know what they are, we do not know what those offences are receiving. We do not know and will not know, because the data is not there, and no one seems to be collecting it. The Sentencing Council has decided it is not a priority for them for the next five years.
Q18 Dr Mullan: Do any other witnesses want to comment on how we would measure whether we are delivering justice through the totality guidelines?
Dr Kelly: I think totality lends itself to this sort of question, because it draws together sentencing for so many different offences. We are trying to make almost an instinctive judgment about whether that is fair in that case. Is it fair compared with single sentences? Perhaps the direction we are being pushed in is to say that the Sentencing Council has done good work on preparing offence-specific guidelines, its response to theft, its response to assault occasioning actual bodily harm and so on.
It might now just be time, and the totality exercise might be showing it is time, to start doing some more holistic overarching work. When you look at your theft guideline beside your assault occasioning actual bodily harm guideline, and beside every other guideline, do they promote proportionality, not individually but as a system?
Professor Ashworth: One other thing that we have not yet mentioned is that part of the scheme that is put forward is that the judge has a duty to explain how the sentence is structured. Should that include an explanation of how the sentence is calculated? “Structured” means how much is concurrent, how much is consecutive, and the decision.
My argument earlier was that we need numbers. Here, we particularly need numbers. In other words, if the judge is explaining the sentence, it should be an explanation of how the sentence is calculated. Going back to my example of the Jenkins case, 25 years for five robberies, or 21 years. That is a big difference, but how do we reach that decision? That is something that we ought to urge on the Council.
The judicial response is that this is a matter of experience. You can’t bottle it. It’s experience, and experience tells the judge what the difference and the numbers should be. But I think that’s inadequate. If experience is crucial, there must be something about the features of the cases that are in common. You cannot just say every case is different and it is all fact specific, to use the current term. If experience is crucial, all cases cannot be unique.
Professor Dhami: Following up on what Andrew is saying, if we want to tell the public or the victims, or even the offender, how a sentence was arrived at, the guideline cannot help with that explanation because of the lack of clarity, which is one of its main weaknesses. So we do not know what is just and proportionate. Proportionate to what? What does that mean? The guideline does not tell us what factors would result in an upward or a downward adjustment for concurrent or consecutive sentences. When you think about totality in terms of factors that are personal to an offender, is it too harsh a sentence? What are those personal factors? It does not tell us how much the adjustment should be and on what basis that is made.
The other problem with the guideline is the potential for double counting. Basically, because you are applying the offence’s specific guidelines to the final sentence for the individual offence, and then you apply the totality guideline, what happens is you apply, potentially, the personal mitigating factors twice, which has a penalty-reducing effect. The question then becomes: how can we afford this double counting of personal mitigating factors twice? Obviously, you do not look at aggravation again, although there is now something in the revised guideline to look at aggravating factors, but the question is: what are these so-called personal aggravating factors? What are these personal mitigating factors? And how can we ensure that they are different from the ones that are in the offence-specific guidelines so that there is no double counting?
The lack of clarity does not help with allowing sentencers to provide an explanation for how they reached a sentence.
Q19 Chair: Can I play devil’s advocate and say isn’t this all making it altogether too mathematical—almost like an algorithm? Sentencing is a human thing. It is about human instinct and the person that saw the person during the trial and saw the individual in front of him. Are we not trying to make it too scientific?
Professor Ashworth: Well, I might disagree with “too scientific”, Chairman. We should start from the starting point of guidelines. You can operate guidelines by having numbers that are interpreted by the judge. That is the system that we have. We do not have a grid where the sentencer is pushed into an area that he or she cannot get out of. We have a grid—you can depict it in that way—but there is the discretion of the judge, which is a hallmark of the English system. I think we should say, “Discretion of the judge subject to his or her explanation of what they are doing.” Once we go down the line of judicial experience, that is good, but we have to have reasons.
Chair: So it is the explanation of reasons. Dr Kelly?
Dr Kelly: I don’t think the choice is between an arithmetical approach and an instinctive one. We can have a structured approach in the middle. Even if you do not want numbers in a totality guideline, you could follow the broad structure of relevant factors in offence-specific guidelines. When we want to know how serious an assault was, we look at the harm and the culpability. We could follow a similar approach here.
Chair: Professor Dhami?
Professor Dhami: I agree. What I would call a quasi-rational approach might be the way forward because it also lends itself to the notion that a human is judging a human.
Chair: Thank you. Mr Timpson.
Q20 Edward Timpson: I am going to ask about some of the proposed revisions to the totality guideline. Before I do, if you will indulge me, we just had a very interesting discussion. Clearly, none of us wants to end up in a situation where we have an algorithm deciding sentencing. One issue that totality is trying to resolve is the tension between how prescriptive the guidelines on sentencing are and the level of judicial discretion within those guidelines. Do you think we have that balance right?
Professor Dhami: There is a lot of discretion allowed in the offence-specific guidelines. Even though those guidelines provide a list of factors that could be used to calculate harm, culpability and offence seriousness, and they also look at specific aggravating and mitigating factors, what they do not tell you is how those factors need to be weighted and integrated, or how they should be weighted relative to each other.
Actually, we are nowhere close to anything remotely like an algorithmic approach in the offence-specific guidelines. There is a lot of room for interpretation there on the judge’s part. The problem with the totality guideline is that it doesn’t help; it just gives you more room for manoeuvre. What we are trying to do is have a consistency of approach. It will ultimately also help the sentencer, because then they will know. They will feel more confident and comfortable with what they are doing as well.
Dr Kelly: I agree that overall in the guidelines, there is going to be the question of how much certainty you want and how much discretion you need. The problem with the totality guideline is that so many of the key terms in it are undefined. I will take perhaps the key term, “a just and proportionate sentence”. We are told repeatedly in the present guideline—and if the new guideline is accepted, we will be told still more—that we need to impose a just and proportionate sentence.
To be fair to it, the guideline does say how you can amend a sentence to make it proportionate—for example, how you can make sentences shorter when they are concurrent—but it doesn’t tell you how to measure that you have arrived at an appropriate sentence. In my mind, that is too much discretion. It is not enough guidance for a judge who is trying earnestly to reach a fair and proportionate figure. Perhaps that is why we are seeing appeals in 32 out of 67 cases, as has been mentioned. How are you to know when you have got it right?
Professor Dhami: In the Sentencing Council’s own survey of sentencers, about 47% of them said that they found it really difficult to apply the totality guideline in certain types of case. So they themselves are admitting it is not very helpful.
Q21 Edward Timpson: Moving on to some of the proposed revisions, can I start by asking what led to the latest consultation? What research did the Sentencing Council undertake to inform what it then proposed? Would you like to share anything in relation to the research exercise that led to the consultation, and how robust it is in giving the fullest picture of the current situation? Professor Dhami, I know this is a particular interest of yours.
Professor Dhami: The Council conducted an online survey of 130 out of the 550 sentencers in its research pool. The majority of the respondents were based in the magistrates courts, not the Crown courts, so one could question the representativeness of that sample. The survey was followed up with 30-minute interviews with 10 out of the 80 sentencers who volunteered to be interviewed.
The Council’s approach to this sort of research is based on small, unrepresentative samples, and relies primarily on self-report data. The analysis of the data itself is quite basic and descriptive, just providing some proportions of what sentencers said. It does not even look into different groups of sentencers or say, for example, that maybe one opinion reflects one type of sentencer more than another type. The basic approach to even asking the questions in the survey results in responses such as “somewhat useful”—what does “somewhat useful” mean?—or results in 10% or 15% of respondents saying, “I neither agree nor disagree with the statement,” so there is no opinion on the issue there.
Overall, the Sentencing Council has been relying on methods that we would consider not particularly rigorous. They are also open to bias. That means that its findings lack reliability, validity and generalisability. It could use other approaches. That also means that the Sentencing Council cannot draw any conclusions about the relationship, causal or otherwise, between policy and practice. As I said before, the Council has unfortunately said that it will not prioritise analysis of sentencing in multiple-offence cases in its next five-year plan. We will not know anything more than we already do.
Edward Timpson: Does anyone want to add anything?
Dr Kelly: I think there is an opportunity to go beyond the judicial research pool and even further than just other members of the judiciary. There have been sustained criticisms of this totality guideline since it came into force and it is unlikely at this late stage of public consultation that a large structural change will happen to a guideline, because then it would not have been consulted on. It would have been one or two people responding and saying, “You can make this massive change”. The time for considering big structural changes was in early-stage consultation. They should have been consulting more widely with experts and putting forward provisional and alternative models.
A comparison would perhaps be the Law Commission style of consultation. I should say that I have previously worked with the Law Commission, but I do not think that it affects the point. When they engage in developing a consultation paper, they will talk to the public and to experts to get the initial draft right so that it can be made as good as possible in the late-stage consultation.
Professor Dhami: And if they are going to make drastic changes, there are opportunities to use what we call randomised controlled trials to look at the effect of those changes. There is some data out there—the Council’s own data. I analysed their data on multiple-offence cases; they didn’t analyse it. It is unfortunate that the Council has missed opportunities to collect some useful and relevant data.
Q22 Edward Timpson: A number of parties are involved in the sentencing exercise, of course, the sentencer being just one of them. Are you aware of whether in their research they spoke to or had responses from those who were victims of those who were being sentenced, those who were being sentenced, members of the public who had an interest in the sentence, or anyone else involved, including the advocates, to understand their views on how totality fits into the sentencing guidelines?
Professor Dhami: In the survey on totality, they looked only at sentencers, but in the research they commissioned on the public understanding of the criminal justice system and sentencing, they asked whether respondents had been victims of crime and of what sort of crime. As I said, in that survey they did not look at totality or multiple-offence cases.
Q23 Edward Timpson: Shall we get into the weeds a little on what the consultation has come up with and some of the revised guidelines? You have the general principles and then it goes on to the general approach. If I can just talk to you about the general principles—we have touched on some of this a little already in this session—the consultation proposes only a small change to the general principles section of the guideline. Towards the end, there are two elements. In the second, in the reference to “offending behaviour”, the term “overall harm and culpability” is added and, before the word “factors”, we now have “aggravating and mitigating”. “As a whole” has dropped off the end, after the word “offender”. What difference do you think the inclusion of these references to overall harm and culpability and aggravating and mitigating factors will make to how the guidelines are interpreted and used in sentencing?
Professor Dhami: Obviously, there is an opportunity for that double counting I mentioned. You already have harm and culpability factors when calculating the sentence for the individual offence and now, without specifying what these overall harm and culpability factors are, what will the sentencers rely on? Given that totality is about not a discount but potentially adjusting downwards to have this ordinal proportionality, you might end up double counting the personal mitigating factors, in particular. Again, you would see that multiple-offence cases would receive less penalty.
Edward Timpson: Professor Ashworth?
Professor Ashworth: Are we on page 8? Which section is this?
Edward Timpson: I very diligently wrote it down in my own hand, so I shall have to find it for you.
Professor Ashworth: Is it the one headed “General Approach”?
Q24 Edward Timpson: No, we are on the one before, on general principles. On page 7 of the consultation, you can see that they have added in the words I referred to as well as struck out “as a whole” from the end of the second paragraph—and, in fact, the word “components” after “current and consecutive”.
Professor Ashworth: That is the one that starts, “It is usually impossible,” which is very strong language. I do not know whether this will effect any change at all. One has to bear in mind that over the years, as far as I know—Mandeep’s research may prove this wrong—looking at Court of Appeal cases that have been reported, most of them do not refer to the totality principle at all. They just do not refer to this document—well, the 2012 document.
Q25 Edward Timpson: Are you talking about the Court of Appeal judgment?
Professor Ashworth: Yes, and also, in so far as I can look at it through a Court of Appeal judgment, I include trial judges. This is a principle that judges think they know, so they do not tend to spell it out. When a report on the Sentencing Council was conducted by Professor Sir Anthony Bottoms in 2018, this was one of the factors he mentioned too.
Dr Kelly: I think the reference to harm and culpability is in one sense promising and in another disappointing. I think I need to read out the whole sentence, if I may, first: “It is necessary to address the offending behaviour with reference to overall harm and culpability, together with the aggravating and mitigating factors personal to the offender.”
It is necessary, then, for the judge—if they are going to follow the guideline—to reference harm and culpability. Why is that promising? These are tools that judges are used to when they structure a sentence for most offences. This is an area of uncertainty in sentencing. “There are the conceptual tools; go and use those and make your sentencing exercise as clear as you can for all involved”—that is the promise.
The disappointment, though, is that when we turn to the general purposes later, these concepts seem to slip away. On the one hand we are being told, “Judge, you have to make reference to harm and culpability”; on the other, the judge is told, “But here is your structure for setting a sentence without reference to harm and culpability.”
Q26 Edward Timpson: Based on the research carried out for the consultation, and the consultation document itself, can you understand the differentiation between the principles and the general approach in that regard?
Dr Kelly: If I were to hazard a guess, it would be because the Council did not want to undertake substantial changes to the guideline, perhaps due to the absence of relevant data here. Maybe they see an outlier risk of significant effects across sentencing. It is one thing to say that it is a general principle—“These are good tools”—but it is another to go further and say, “In every sentencing exercise you should be centring them,” particularly if they have not done the empirical work to show that that would not have any effects on sentencing levels across the board.
Q27 Edward Timpson: What other changes, if any, would you suggest—sticking with the general principles, before we go on to the general approach—to help clarify some of the aspects you have highlighted today, and help improve and make clearer the way that sentencing is constructed or, indeed, calculated?
Professor Dhami: I think we have already laid out that we need to identify what the overall harm and culpability factors are and how they are different from the ones that would have been considered for the individual offences, so that there is not that double counting there.
Like Rory, I also noticed a disconnect between what is said in the general principles section and later on, in the section, “Reaching a just and proportionate sentence”. It does not connect back to the general principles; it sort of stands on its own.
The other thing in the general principles section, perhaps, is the conversation about the conditions under which sentences should be served concurrently versus consecutively. The Council has given examples, but I think they could provide more, especially in relation to the sorts of offences that commonly make up multiple-offence cases and those sorts of cases that sentencers have told the Council in their own survey that they find particular difficulty with. Those examples could be more tailored.
Q28 Edward Timpson: Anyone else?
Dr Kelly: The general principles section states that “There is no inflexible rule” relating to concurrent or consecutive sentences. I think the Council have done some good work later, when we turn to the general approach, in thickening when a concurrent sentence is more appropriate and when a consecutive sentence is more appropriate. It is not clear, to my mind, whether there is any need for continued reference to “no inflexible rule”, particularly when it seems to grate somewhat with a body that is meant to promote consistency of approach in sentencing. Perhaps that last paragraph could just be scrapped now.
Q29 Edward Timpson: Does anyone agree with that view—that, rather than that element being more prescriptive, it is probably more helpful if it isn’t there at all?
Professor Dhami: Well, they are actually giving some examples, so we are saying, “Give some more useful examples.” But that sentence in itself is contradictory. It is a bit confusing, because they are saying there is no rule and then they are giving examples of where whatever it is would apply.
Q30 Chair: That is helpful. Can you think of any more useful examples that you might put in, Professor?
Professor Dhami: As I said, if you look at the Council’s own survey, they list the sorts of offences that sentencers have difficulties with. Those are examples where there are different types of offences, different victims and so forth.
Chair: So, take more from the survey.
Professor Dhami: Another one is historical sexual offences.
Q31 Maria Eagle: If I may, I will move on to talk about the “General Approach” section of the guideline. We have already touched on some of this in earlier replies. The Council are proposing a number of changes to the “General Approach” section, which uses examples to illustrate when concurrent or consecutive sentences should be appropriate, and includes a direction to the sentencer to explain how the sentence is structured. We talked about this a little earlier—particularly Professor Ashworth. Is that going to make the application of totality more easily understandable? I realise you have already made reference to increasing clarity by explaining how sentences are calculated, but is what they propose now going to help? I realise you have said it doesn’t go far enough.
Professor Ashworth: I am afraid that is my position: I don’t see how it does. If you look at the three principles at the bottom of the “General Approach” page, the first is: “Consider the sentence for each individual offence, referring to the relevant…guidelines.” That is fine. The second is: “Determine whether the case calls for concurrent or consecutive sentences.” That is what we were just dealing with above. Like Professor Dhami, I think that should be spelled out more fully and not disappearing into some sort of concept that is not useful. I think it is useful to have certain core cases of concurrent sentences and core cases of consecutive sentences, even if there are other cases that don’t fall within that. The third principle is: “Test the overall sentence against the requirement that the total sentence is just and proportionate to the offending as a whole.” Well, that really is the nub of it. I don’t think they go far enough.
Q32 Maria Eagle: Is there anything anybody else would like to add, because we have gone around the racecourse a little bit on that?
Professor Dhami: The word “test” is a strong word. What is the test? Is it an objective test? Is it the same test that everyone applies or that you apply in every case? It seemed to me that it was asking the sentencer to use their own judgment to see whether their own judgment in the case led to a just and proportionate sentence.
Q33 Maria Eagle: Let’s move on. Are there any particular offences that should be addressed in the guideline? At the moment, it is all very general. Are there particular offences that should be addressed by it?
Professor Dhami: I think this goes back to the fact that we know there are certain types of offences that are more likely to make up multiple-offence cases and so it would make the guideline more applicable or more useful if the examples that were provided in the guideline were representative of or related to those sorts of cases and, again, to any kinds of cases that the sentencers in the Sentencing Council’s own survey pointed out were difficult: historical sexual offences and offences involving different victims, different types of offences, and offences that might carry different types of disposals or sentences.
Q34 Chair: One of the things that the Council has done is take some of the things that were in the general principles and create a new section called “Reaching a just and proportionate sentence” on pages 11 and 12 of the consultation document. It says, “The Council considered that the information in this section is key and by giving it a separate section it will give it more prominence” than if it appeared in the general principles. Do you agree with that and has that been achieved?
Dr Kelly: I confess that I found it a bit odd to see this section after a general purposes section. The general purposes section makes reference midway through its process, at three out of four, to reaching a just and proportionate sentence. The risk then is that the judge follows the general process and feels they are done, and then turns the page to see, actually, at three out of four, there is still more detail to come.
On the substance of what is included in the new additional section, I think more could be done to set out this test of how you know it is just and proportionate. I think to do that, you need to have a means of saying how serious the offending was overall. We are not going to know what is proportionate or what a proportionate sentence is until we know how serious the overall offending is.
Chair: Thank you. Any other observations on that point and whether it should be separated out?
Professor Ashworth: I agree entirely and I do not see that this particular section is helpful, because if we take the just and proportionate sentence principle seriously, it should come in at every stage. As Dr Kelly said, just to bring it in at the end is not good enough. The judge would have it in mind throughout as to what the overall sentence should be and how to structure it.
Chair: Fair enough.
Professor Dhami: The Council in this section is trying to distinguish between different types of multiple-offence cases, so it talks about those cases that are similar with similar offences, or different offences with similar levels of seriousness. In those cases, you would be adjusting downwards and they would be served consecutively. That is versus offences—it does not say here whether those offences are similar or different—with different levels of seriousness. Some of them would have no separate penalty and would run concurrently.
The Council is now beginning to introduce the idea that there are two different types of multiple-offence cases—actually, more than two. I do not think it has captured all the different types. Then, it tries to give that so-called rule as to what would run concurrently and what would run consecutively. But I think they have missed something; it fits in a two-by-two box, but they have missed some of the squares in these boxes out. Sometimes they are emphasising the similarity of the level of seriousness and sometimes they are emphasising the similarity of the offence type itself. It muddies the waters and, again, we do not know what “just and proportionate” refers to.
Q35 Janet Daby: The Council says that its proposals “are unlikely to lead to substantive changes to current sentencing practice” and “are therefore unlikely to have any substantial impact on prison or probation resources”. Do you agree?
Professor Dhami: I disagree. The Council admits that it does not have the data to make reliable and valid resource or impact assessments. It states that it does not intend its change in guideline to affect the average sentence. It does not say what it means by average sentence. Does it mean the average proportion or the proportion of cases that receive immediate custody versus any disposals, or does it mean in terms of those cases that receive immediate custody the average length of that custody? It is not quite clear what they mean by sentence severity. They have left that undefined.
Anyway, on the basis of their own intention of having no effect, they have estimated that therefore there will be “no resource impact on prison, probation or youth justice services.” It does not say anything about the Court of Appeal, by the way. It is important to consider that. The Council says that there is some “uncertainty” around its estimate—clearly, if there is no data—and acknowledges that given the wide applicability of the guideline, if a change did occur, it would have a substantial resource impact. Again, at the same time, the Council does not say how it is going to monitor this. In fact, it will not necessarily monitor the impact at all.
The question is: is there potential for change to happen? One can think about the fact that the Council is now making explicit the obligation on sentencers to explain the sentence to the court. That in itself, given the reminder of that sense of accountability, could change someone’s judgment process, change how they reason through it. It would not even need any changes to the guideline; it could be just how you might interpret that guideline, because you are aware—or it reminds you—that you now have to explain this. Just that awareness could change your behaviour. So, yes, this is not what I would consider a proper impact or resource assessment.
Janet Daby: Is there anything that you would like to add?
Professor Ashworth: I agree with that entirely.
Dr Kelly: I agree.
Q36 Janet Daby: The consultation asks how the guideline could cause an increased disparity in sentencing. Do you have a view on that point, and what more could the guideline do to address disparities in sentencing?
Professor Ashworth: I think Professor Dhami has eloquently pointed out that consistency will not be enhanced by the changes in this consultation document, and I think, therefore, that that answers your question.
Professor Dhami: Any disparity that is already there will remain. Clearly, there is potential for disparity if factors such as race and ethnicity are associated with the multiple offence or single offence status of a case, or if it is associated with a certain kind of multiple-offence case—the kind that would receive a concurrent sentence versus a kind that would receive a consecutive sentence. There are two recent pieces of evidence that suggest that these disparities exist. The Stott et al review in 2021 of the Government’s own studies concluded that there was demonstrable, quantifiable and robust evidence, including where, basically, ethnic and racial disparities existed primarily due to policing practices. They included for offences that we know already are more likely to be multiple-offence cases.
A couple of weeks ago, the CPS published some independent evidence to show that there are racial and ethnic disparities in its charging decisions. We do not know whether that was for multiple or single-offence cases—it did not have that breakdown. But there is clearly potential there.
The other thing that we have not spoken about—perhaps I should have mentioned this earlier—is that when the Council talks about giving concurrent sentences, it says that it will ordinarily be appropriate, “where there is a series of offences of the same or similar kind, especially when committed against the same person” where you would give a concurrent sentence. I would be concerned here, because you could be introducing a bias against victims who suffer from these types of crimes; these victims are likely to be women who are subject to stalking and harassment, and domestic abuse, as well as children subjected to abuse and neglect. Those types of offences are a series of crimes of the same kind against the same individual. What you are saying there is that they will get a concurrent sentence, which may be less than if someone had committed one of those crimes against one person. The Council’s approach to dealing with any potential disparities is simply to tell the sentencers to read the “Equal Treatment Bench Book”, which is over 500 pages long. I think that that is inadequate. At the very least, the Council should point sentencers in the right direction in that book.
Janet Daby: Dr Kelly, do you want to add anything?
Dr Kelly: I would only re-emphasise that the risk of disparity needs to be taken very seriously in the context of the totality guideline. There are terms in this guideline that have a lot of discretion within them over all criminality: a “single” incident or “multiple” incidents; a “just and proportionate” sentence. On the one hand we are being told that the data is not there for analysis, but on the other we have terms that allow significant discretion in how the final sanction is reached. You can certainly see the possibility for disparity in sentences here.
Chair: Mr Timpson, do you have anything to add?
Q37 Edward Timpson: If I could, yes. In recognising the grand old age of our justice system, we also, I hope, have enough humility to be able to look elsewhere to try to understand whether there are others who have grappled with this idea of a totality guideline within their sentencing. Have any of you been able to look at or research how other jurisdictions deal with this aspect of sentencing? Or is there anything that we as a Committee should consider to see how some of the clarity, and the understanding of the role it plays, can be improved as our very old system gets even older?
Professor Dhami: I think that Andrew could probably give you more, but before he does, can I point out that you might want to refer to the Sentencing Council’s own document from September 2010? The Council did a review of how totality and multiple offences were being dealt with in international jurisdictions. That was quite a thorough review, so the information is there if you want to look at it.
Professor Ashworth: There are systems in Europe—Germany and Sweden, for example—where there is a sort of arithmetical treatment of totality. If a court is sentencing, say, four or five offences on the same occasion, the first one gets the full sentence according to the guidelines—let’s say that’s 10 years—the second one gets a third of that and the third gets a sixth of that. In other words, each incrementally adds to the overall sentence, but in a decreasing fashion. That is basically what the system is in Germany and Sweden. The multiple offender gets a higher sentence than the single offender, but it tapers off, and it is a decreasing increment.
Of course, that would be quite unacceptable in this country, because it does not suggest judicial discretion. But if you add judicial discretion into that, so that a judge should be able to say, “That’s too little,” or “That’s too much of a reduction,” that is something that we ought to at least look at. It is anathema to our judiciary, I’m afraid, so it would be a very difficult sell for the Council, but it is a system that appears to work in other European countries.
Q38 Edward Timpson: Do you know whether in Germany the level of public understanding of and confidence in sentencing are greater than in this country?
Professor Ashworth: I would hazard a guess that it is, but the whole system is different, of course. The system separates out the young judges at the early stage, and they go through a judges’ school and become judges at, say, 30, having had a formation in a school for judges, whereas we do it very differently. We expect our judges to be drawn from people who have been in practice for a substantial length of time. It is a very different system. I do not know whether Professor Dhami has any data on that.
Dr Kelly: I think it is right to look to other jurisdictions here, because the risk of severe sentences for a large number of small sentences and the risk of sentences for offences that conceptually overlap are not just national issues. You always have that question in the back of your head: has someone else got it right already? Another example of something done elsewhere is the model penal code developed in the United States. That is a more presumptive-based approach, which is perhaps a bit closer to what Professor Ashworth was suggesting. There, the strong presumption is that a total sentence will not be two times more than the most serious sentence. For example, if you have 20 years for the most serious offence, the strong presumption is that you are going to go only up to 40, or if it was two, it would be four. It can be arithmetical in some jurisdictions, perhaps, but it can also be presumption-led too.
Edward Timpson: That’s very helpful. Thank you.
Chair: That is interesting. That is a common law system, in effect.
Professor Dhami: Just to remind you, in 2010, the Sentencing Council considered these alternative approaches internationally when it was going to develop the totality guideline, and it decided that they were not appropriate.
Q39 Chair: Thank you very much. That has been extremely helpful to us, and very wide ranging. Are there any topics that we have left out?
Professor Dhami: May I tell you what I think are some of the strengths?
Chair: Yes, absolutely.
Professor Dhami: It is worth pointing out that one of the strengths is that it is applied as a limiting factor—rather than at the beginning, it is at the end. I think that is a strength. The fact that we have these individual sentences calculated for the individual offences is great, because it lends some transparency to the process. It means you are demonstrating a commitment to the punishment of the individual crimes and a commitment to victims. It also means that those individual sentences may be appealable. If there is any revision, those are some of the strengths that we would want to retain.
Chair: Anything from the other two witnesses? Nothing more to add. Thank you very much. We are very grateful to all three of you giving up your time to give evidence to us today. It is very useful, and it will be very helpful for us to feed into our letter to the Sentencing Council in pursuit of our statutory duty. We are much obliged to you.